TuW3	HG D7.1
Workshop 3: Time Delay Systems – Stability & Control in Applications (Workshop)
09:00-17:00	TuW3.1
TIME DELAY SYSTEMS – Stability & Control in Applications*
Atay, Fatihcan	Max Planck Inst. for Mathematics in the Sciences
Breda, Dimitri	Univ. of Udine
Michiels, Wim	K.U. Leuven
Niculescu, Silviu-Iulian	UMR CNRS 8506, CNRS-SUPELEC
Ozbay, Hitay	Bilkent Univ.
Sipahi, Rifat	Northeastern Univ.

TuW4	HG D7.2
Workshop 4: Fault Diagnosis, Fault-Tolerant Control, and Cooperative Control of Manned and Unmanned Systems (Workshop)
09:00-17:00	TuW4.1
Fault Diagnosis, Fault-Tolerant Control, and Cooperative Control of Manned and Unmanned Systems*
Zhang, Youmin	Concordia Univ.
Chen, YangQuan	Univ. of California, Merced
Edwards, Christopher	Univ. of Exeter
Fulford, Cameron	Quanser Inc.
Goupil, Philippe	AIRBUS Operations S.A.S.
Hu, Qinglei	Harbin Inst. of Tech.
Liu, Hongtao	Univ. of Toronto
Marcos, Andres	Deimos Space S.L.
Noura, Hassan	United Arab Emirates Univ.
Puig, Vicenc	Univ. Pol. de Catalunya
Theilliol, Didier	Univ. of Lorraine
Tsourdos, Antonios	Cranfield Univ.

Technical Program for Wednesday July 17, 2013

WePT1	HG F30, F1, F7
Plenary 1: A Computationally Tractable Theory of Performance Analysis in Stochastic Systems (Plenary Session)
Chair: Morari, Manfred	ETH Zurich
08:30-09:30	WePT1.1
A Computationally Tractable Theory of Performance Analysis in Stochastic Systems*
Bertsimas, Dimitris	M.I.T.

WeA1	HG F1
Sparse and Low-Rank Representation Methods in Control, Estimation and System Identification (Tutorial Session)
Chair: Wahlberg, Bo	KTH
Organizer: Wahlberg, Bo	KTH
09:50-10:30	WeA1.1
Convex Optimization Algorithms for Sparse and Low-Rank Representations (I)*
Vandenberghe, Lieven	Univ. of California, Los Angeles
Convex methods for computing sparse or low-rank representations and their extensions typically require the minimization of the sum of a differentiable function (for example, a quadratic error term) and a simple nondifferentiable function (for example, a structure-promoting norm). A wide variety of algorithms has been proposed for this purpose. The lecture will give an introduction to algorithms based on dual decomposition, proximal mappings, and operator splitting. This includes proximal gradient methods, the alternating direction method of multipliers (ADMM), and other popular methods. After presenting some elements of convex analysis, we will discuss three general classes of algorithms: the proximal point algorithm, the forward-backward iteration, and the Douglas-Rachford splitting algorithm, applied to the primal, dual, or primal-dual optimality conditions. The algorithms will be illustrated with nuclear norm minimization problems arising in system identification.
10:30-11:10	WeA1.2
Sparsity-Promoting Optimal Control of Distributed Systems (I)*
Jovanovic, Mihailo	Univ. of Minnesota
This tutorial is about design of feedback gains that achieve a desirable tradeoff between quadratic performance of distributed systems and controller sparsity. Our approach consists of two steps. First, we identify sparsity patterns of the feedback gains by incorporating sparsity-promoting penalty functions into the optimal control problem, where the added terms penalize the number of communication links in the distributed controller. Second, we optimize feedback gains subject to structural constraints determined by the identified sparsity patterns. In the first step, the sparsity structure of feedback gains is identified using the alternating direction method of multipliers, a powerful algorithm well-suited to large optimization problems. This method alternates between promoting the sparsity of the controller and optimizing the closed-loop performance, which allows us to exploit the structure of the corresponding objective functions. In particular, we take advantage of the separability of the sparsity-promoting penalty functions to decompose the minimization problem into sub-problems that can be solved analytically. Even though the standard quadratic performance index is in general a nonconvex function of the feedback gain, we identify classes of convex problems that arise in the design of sparse undirected networks. In this case, the corresponding synthesis problem can be formulated as a semidefinite program, implying that the globally optimal sparse controller can be computed efficiently. Several examples are provided to demonstrate the effectiveness of the developed approach and the accompanying software LQRSP (which is available at www.ece.umn.edu/users/mihailo/software/lqrsp/)
11:10-11:50	WeA1.3
Sparse and Low Rank Representation Methods in System Identification (I)*
Rojas, Cristian R.	KTH Royal Inst. of Tech.
Wahlberg, Bo	KTH Royal Inst. of Tech.
Since the 90's, sparsity has been playing an important role in several aspects of statistics, machine learning and signal processing, among other fields. In this talk, we will focus on some specific connections of sparsity to important problems in system identification. In particular, the problem of sparse estimation will be seen in the context of model structure selection, where $l_1$ regularization becomes a handy tool for overcoming the curse of dimensionality in model selection (when the number of candidate model structures is large) and for imposing prior knowledge, thus delivering parsimonious models. Then, the so-called nuclear norm relaxation, a convex surrogate of the rank function, corresponding to the $l_1$ norm of the singular values of a matrix, will be used for subspace-like identification and for the estimation of structured systems. Finally, we will discuss the use of $l_1$ regularization for change point detection. These techniques unfortunately are not fool-proof, so we will discuss some conditions under which these sparsity-inducing methods might fail and when they can be more safely used.

WeA2	HG F3
Model Predictive Control I (Regular Session)
Chair: Richards, Arthur	Univ. of Bristol
Co-Chair: Findeisen, Rolf	Univ. of Magdeburg
09:50-10:10	WeA2.1
Fast Model Predictive Control with Soft Constraints, pp. 1-6
Richards, Arthur	Univ. of Bristol
This paper describes a fast optimization algorithm for Model Predictive Control (MPC) with soft constraints. The method relies on the Kreisselmeier-Steinhauser function to provide a smooth approximation of the penalty function for a soft constraint. By introducing this approximation directly into the objective of an interior point optimization, there is no need for additional slack variables to capture constraint violation. Simulation results show significant speed-up compared to using slacks.
10:10-10:30	WeA2.2
Integrating the RTO in the MPC: An Adaptive Gradient-Based Approach, pp. 7-12
Limon, Daniel	Univ. de Sevilla
Alamo, Teodoro	Univ. de Sevilla
Pereira Martin, Mario	Univ. de Sevilla
Ferramosca, Antonio	CONICET - UNL
González, Alejandro H.	CONICET
Odloak, Darci	Department of Chemical Engineering, Univ. of Sao Paulo
Model Predictive Control (MPC) is the most used advanced control technique in process industries, since it ensures stability, constraints satisfaction and convergence to the setpoint. The optimal setpoint is calculated by the Real Time Optimizer (RTO), minimizing the economic objective taking into account the operational limits of the plant. Since RTO employs complex stationary nonlinear models to perform the optimization and a larger sampling time than the controller, the economic setpoints calculated by the RTO may be inconsistent for the MPC layer and the economic performance of the overall controller may be worse than expected. The aim of this work is to propose an MPC controller that explicitly integrates the RTO into the MPC control layer. The proposed strategy is based on the MPC for tracking; the optimization problem to be solved only requires one evaluation of the gradient of the economic cost function at each sampling time. Based on this gradient, a second order approximation of the economic function is obtained and used in the MPC optimization problem resulting in a convex optimization problem. Recursive feasibility and convergence to the optimal equilibrium point is ensured.
10:30-10:50	WeA2.3
Multivariable MPC Design Based on a Frequency Response Approximation Approach, pp. 13-18
Shah, Gaurang	Tech. Univ. Dortmund
Engell, Sebastian	TU Dortmund
This paper presents an improvement to the MPC tuning approach presented in [1]. The drawbacks of the former approach are discussed and a new tuning approach is proposed to overcome them. The primary aim is to treat linear MPC as a classical controller design problem and to use the tools of linear control theory for determining the tuning parameters. MPC tuning is performed for a desired open-loop frequency response which results from an achievable closed-loop performance determined using the Youla parameterization technique. The tuning approach is performed in two steps and both resulting minimization problems are convex optimization problems. The approach is tested on a challenging example.
10:50-11:10	WeA2.4
A Barrier Function Based Continuous-Time Algorithm for Linear Model Predictive Control, pp. 19-26
Feller, Christian	Univ. of Stuttgart
Ebenbauer, Christian	Univ. of Stuttgart
In this paper, we present a novel linear model predictive control (MPC) scheme that relies on a continuous-time, barrier function based algorithm which asymptotically tracks the solution of a time-varying open-loop optimal control problem. In particular, the control input is obtained as the sampled output of a continuous-time dynamical system and no iterative optimization algorithm is needed in the on-line implementation. In addition, we present a new approach towards stabilizing MPC based on gradient recentered logarithmic barrier functions that allows enlargement of the employed terminal set.
11:10-11:30	WeA2.5
On Efficient Predictive Control of Linear Systems Subject to Quadratic Constraints Using Condensed, Structure-Exploiting Interior Point Methods, pp. 27-34
Koegel, Markus	OVG Univ. Magdeburg
Findeisen, Rolf	Univ. of Magdeburg
We describe a primal-dual interior point method tailored for predictive control of linear systems subject to convex quadratic costs and constraints. In particular, we consider only the inputs as decision variables to reduce the memory demand. Moreover, we propose to exploit the problem structure, which decreases the computational burden for long horizons. For the implementation we consider tailored solution methods of the arising Riccati recursion and the line-search. Furthermore, we discuss the memory and computational demand and problem specific methods to further save computations. Finally, we outline with an example the applicability.
11:30-11:50	WeA2.6
Accelerating Online MPC with Partial Explicit Information and Linear Storage Complexity in the Number of Constraints, pp. 35-40
Jost, Michael	Ruhr-Univ. Bochum
Mönnigmann, Martin	Ruhr-Univ. Bochum
Model predictive control (MPC, or online MPC for clarity here) is a widely acclaimed method for the control of constrained systems. Since a constrained optimization problem has to be solved in every time step, the online computational effort is high for MPC, however. Explicit MPC, on the other hand, provides an analytical solution to the same optimization problem, but explicit MPC is only useful for small systems, since the storage requirements for the explicit control law grow exponentially in the number of constraints of the optimization problem. We show that online MPC can be accelerated with information on the structure of the control law, where this structural information is calculated ofﬂine with techniques from explicit MPC. Our two main contributions are as follows: We demonstrate that online MPC can be sped up signiﬁcantly if only q state space regions, the regions of activity, are stored, where q is the number of constraints. Note that this linear growth in q is obviously very different from the exponential growth in q of the number of polytopes that need to be stored in explicit MPC. Secondly, we claim that the proposed method is a variant of a family of methods, which comprises online MPC and explicit MPC as two limiting cases.

WeA3	HG F5
Distributed Parameter Systems I (Regular Session)
Chair: Bonnans, Frederic	INRIA-Saclay and CMAP
Co-Chair: Witrant, Emmanuel	Univ. Joseph Fourrier
09:50-10:10	WeA3.1
Iterative Solution of Operator Lyapunov Equations Arising in Heat Transfer (I), pp. 41-46
Reis, Timo	Univ. Hamburg
Wollner, Winnifried	Univ. Hamburg
We consider an iterative method for the numerical solution of Lyapunov equations of infinite-dimensional control systems governed by an heat equation. Inspired by the 'alternating direction implicit (ADI)' iteration, which has been successfully applied to the solution of matrix Lyapunov equations, we present a method to determine approximations of the Gramian operator corresponding to the heat equation system. This method provides approximations of finite rank and is shown to be convergent in the nuclear norm.
10:10-10:30	WeA3.2
Combined Distributed Parameters and Source Estimation in Tokamak Plasmas Heat Transport, pp. 47-52
Mechhoud, Sarah	Univ. Grenoble 1
Witrant, Emmanuel	Univ. Joseph Fourrier
Dugard, Luc	CNRS-INPG
Moreau, Didier	CEA
Joint estimation of distributed time-space parameters and inputs in the non-homogeneous linear reaction-diffusion equation which models the heat transport in tokamak, is investigated, in the finite dimensional framework. Once the equation's discretization in space using the cubic b-splines finite elements method, the application of the parameter projection method results in a linear time-varying state-space model with unknown parameters and inputs. The DAISYS method proves the structural identifiability of the problem and the EKF-UI-WDF estimates simultaneously the states, parameters and inputs. These steps are applied on the tokamak plasma heat transport in order to reconstruct simultaneously its coefficients and its source term. Computer simulations on both mock-up and real data show the performance of the proposed technique.
10:30-10:50	WeA3.3
Computation of Empirical Eigenfunctions of Parabolic PDEs with Non-Trivial Time-Varying Domain, pp. 53-58
Izadi, Mojtaba	Univ. of Alberta
Dubljevic, Stevan	Univ. of Alberta
In this article, a methodology to compute the empirical eigenfunctions for the order-reduction of parabolic partial differential equation (PDE) systems with time-varying domain is explored. In this method, a mapping functional is obtained, which relates the time-evolution of the solution of parabolic PDE with the time-varying domain to a fixed reference domain, while preserving space invariant properties of the raw solution ensemble. Subsequently, the Karhunen-Loeve decomposition is applied to the solution ensemble with fixed spatial domain resulting in a set of optimal eigenfunctions that capture the most energy of data. Further, the low dimensional set of empirical eigenfunctions is mapped (``pushed-back'') on the original time-varying domain by an appropriate mapping resulting in the basis for the construction of the reduced-order model of the parabolic PDE system with the time-varying domain. Finally, this methodology is used for the order-reduction of the Czochralski crystal growth process model which is a two dimensional parabolic PDE system on the time-varying domain with non-trivial geometry. The transformations which relate the raw data on the time-varying and time-invariant domains are designed to preserve dynamic features of the scalar physical property and comparisons among reduced and high order fidelity models are provided.
10:50-11:10	WeA3.4
On the Structure of Robust Controllers for Infinite-Dimensional Systems, pp. 59-62
Hamalainen, Timo	Tampere Univ. of Tech.
Pohjolainen, Seppo	Tampere Univ. of Tech.
Starting from a very general formulation of the Internal Model Principle it is shown that a robust controller tracking/rejecting signals generated by an exosystem can be decomposed into a servocompensator and a stabilizing controller. The servocompensator contains an internal model of the exosystem generating the reference and disturbance signals and the stabilizing controller stabilizes the infinite-dimensional closed-loop system.
11:10-11:30	WeA3.5
Receding Horizon Observer and Control for Linear 2x2 Hyperbolic Systems of Conservation Laws, pp. 63-68
Pham, Van Thang	Univ. of Grenoble - Gipsa-Lab.
Georges, Didier	Grenoble Inst. of Tech.
Besancon, Gildas	Ense3 - Grenoble INP
This paper presents an infinite-dimensional Receding Horizon Observer for linear 2x2 hyperbolic systems with boundary measurements. The initial state is estimated as the optimal solution of an optimization problem which minimizes the distance between the measurements and the observer output. A constructive method is used to derive the existence and uniqueness of the solution. A composite strategy combining Receding Horizon Optimal Control and Receding Horizon Observer is also presented. Its effectiveness in guaranteeing closed-loop stability is also demonstrated. For the implementation, the calculus of variation approach is used to derive the adjoint state which will be discretized and solved with the observer state to obtain the optimal solution. Finally, a simulation with a linearized model of an open-channel system is carried out to validate the here-proposed approach.
11:30-11:50	WeA3.6
Singular Arcs in the Optimal Control of a Parabolic Equation, pp. 69-74
Bonnans, Frederic	INRIA-Saclay and CMAP
We present a theory of singular arc, and the corresponding second order necessary and sufficient conditions, for the optimal control of a semilinear parabolic equation with scalar control applied on the r.h.s. We obtain in particular an extension of Kelley's condition, and the characterization of a quadratic growth property for a weak norm.

WeA4	HG F7
Stochastic Systems I (Regular Session)
Chair: Bitmead, Robert	Univ. of California San Diego
Co-Chair: Barmish, B. Ross	Univ. of Wisconsin
09:50-10:10	WeA4.1
Optimal Decentralized Control of a Stochastically Switched System with Local Parameter Knowledge, pp. 75-80
Mishra, Anshuman	Univ. of Illinois at Urbana-Champaign
Langbort, Cedric	UIUC
Dullerud, Geir E.	Univ. of Illinois at Urbana-Champaign
This paper considers an optimal decentralized control problem for a linear system with stochastically switched input/output matrices depending on local parameters. These stochastic parameters are assumed independent in time and available instantaneously to the local controller but with a one time step delay to the other. We first solve this problem for the case of a single time step and show that the optimal policy can be reached through iterating the best responses of each player. For the optimal multiple step problem, a dynamic programming approach is employed while using the result of the one step control at each step.
10:10-10:30	WeA4.2
On a Time-Varying Stochastic Small Gain Theorem, pp. 81-86
Aberkane, Samir	UHP, NANCY 1
Dragan, Vasile	Romanian Acad.
In this note, a small gain result is presented for a class of discrete-time time-varying Markovian jump systems with state-dependent noise. Using the stabilizing solutions of some generalized Riccati equations, the result is obtained based on a Lyapunov type argument.
10:30-10:50	WeA4.3
A Drawdown Formula for Stock Trading Via Linear Feedback in a Market Governed by Brownian Motion, pp. 87-92
Malekpour, Shirzad	Univ. of Wisconsin-Madison
Barmish, B. Ross	Univ. of Wisconsin
Control of drawdown is one of the greatest concerns to both stock traders and portfolio managers. That is, one typically monitors "drops in wealth" over time from highs to subsequent lows, and investors often shy away from funds with a past history of large drawdowns. With this motivation in mind, this paper addresses the analysis of drawdown when a feedback control is employed in stock trading in an idealized market with prices governed by Geometric Brownian Motion. We begin with a result in the applied probability literature which is applicable to cases involving buy-and-hold. Subsequently, after modifying this result via an Ito correction to account for geometric compounding of the daily stock price, we consider the effect on drawdown when a simple pure-gain feedback control is used to vary the investment I(t) over time. That is, letting V(t) denoting the trader's account value at time t > 0, when a feedback control I(t) = KV(t) is used to modify the amount invested, the buy-and-hold result no longer applies. Our first result is a formula for the expected value for the maximum drawdown in logarithmic wealth log(V(t)). This formula is given in terms of the feedback gain K, the price drift mu, the price volatility sigma and terminal time T. Subsequently, using a fundamental relationship between logarithmic and percentage drawdowns, we obtain an estimate for the expected value of the maximum percentage drawdown of V(t). This paper also includes an analysis of the asymptotic behavior of this drawdown estimate as T tends to infinity and Monte Carlo simulations aimed at validation of our estimates.
10:50-11:10	WeA4.4
Analysis of Controlled Biological Switches Via Stochastic Motion Planning, pp. 93-98
Mohajerin Esfahani, Peyman	Automatic Control Lab. Swiss Federal Inst. ofTechnolo
Milias, Andreas	ETH Zurich
Chatterjee, Debasish	Indian Inst. of Tech. Bombay
We consider the problem of regulating a genetic toggle switch by means of an in silico feedback control loop, which has been proposed as a complement to synthetic feedback control schemes implemented inside cells. Control of naturally occurring bistable systems can provide useful insights into their functional organization, as well as enable the generation of completely different phenotypes, by driving the cells away from the commonly observed states. To achieve this, we first introduce two basic notions of motion planning, and then establish a connection to a class of stochastic optimal control problems concerned with sequential stopping-times. To characterize the desired set of initial conditions, in the context of controlled diffusion processes, we propose a sequence of partial differential equations for which the first one has a known boundary condition, while the boundary conditions of the subsequent ones are determined by the solutions to the preceding steps. We then formulate the control of a bistable system as stochastic motion planning problem, and derive the closed-loop control law that maintains the system inside a prespecified region of its state space.
11:10-11:30	WeA4.5
Escape Time Formulation of State Estimation and Stabilization with Intermittent Communication, pp. 99-104
Huang, Chun-Chia	Univ. of California, San Diego
Bitmead, Robert	Univ. of California San Diego
The problems of state estimation and feedback stabilization of an unstable system including a communications channel are quantified as escape or survival times, which yield stochastic processes describing the time of first exit of the state estimate error or of the state itself from a specific domain. The system complications introduced by communications -- intermittency, channel noise, quantization, delay, etc -- are evaluated using a Markov hitting time formulation. This is compared to and contrasted with earlier analyses which considered: the behavior of Kalman filters with intermittent data, the evaluation of the minimal number of bits required for mean square stabilization, and the Large Deviations analysis of probability theory. The main result shows the the escape time is characterized by a Markov chain which is amenable to explicit analysis in the linear gaussian case. This is then used to develop an approach to bitrate assignment in such communications-based systems.
11:30-11:50	WeA4.6
H-Infinity Control for Singular Stochastic Systems, pp. 105-110
Barbata, Asma	CRAN
Zasadzinski, Michel	CRAN
Souley Ali, Harouna	Cran
Messaoud, Hassani	Ec. Nationale d'Ingénieurs de Monastir
In this paper, we deal with the bounded real lemma for stochastic singular systems with multiplicative noises. Based on the adaptation of Ito calculus, the admissibility for this class of systems is defined and the bounded real lemma is derived using the mean square exponential stability. This lemma is then used to synthetize a H-infinity output feedback controller for the considered class of systems that achieves a given level of disturbance attenuation. The design is based on the solution of linear matrix inequalities coupled with an algebraic constraint.

WeA5	HG E1.1
Discrete Event Systems (Regular Session)
Chair: Mahulea, Cristian	Univ. of Zaragoza
Co-Chair: Basile, Francesco	Univ. Degli Studi Di Salerno
09:50-10:10	WeA5.1
Model Matching Control for Composite Asynchronous Sequential Machines, pp. 111-116
Yang, Jung-Min	Catholic Univ. of Daegu
Park, Seong-Jin	Ajou Univ.
Kwak, Seong Woo	Keimyung Univ.
Model matching for an asynchronous sequential machine is to design a feedback controller that compensates the closed-loop system so that it can match the behavior of a reference model. The objective of this paper is to address the model matching problem for a composite asynchronous sequential machine. Two input/state asynchronous machines in cascade connection form a composite machine, where the output of the front machine is transmitted as the input to the rear machine. Given a desired model, we present the existence condition and design procedure for an appropriate corrective controller that achieves model matching. An illustrative example is provided for verifying the applicability of the proposed scheme.
10:10-10:30	WeA5.2
Discrete Control of Computing Systems Administration: A Programming Language Supported Approach, pp. 117-124
Delaval, Gwenaël	Univ. Grenoble
De Palma, Noel	UJF, LIG / INRIA Grenoble
Gueye, Soguy Mak karé	UJF, INRIA-Grenoble / LIG
Marchand, Herve	INRIA, centre Rennes Bretagne-Atlantique
Rutten, Eric	LIG / INRIA Grenoble
We address the problem of using Discrete Controller Synthesis for the administration of Computing Systems, following an approach supported by a programming language. We present a mixed imperative/declarative programming language, where declarative contracts are enforced upon imperatively described behaviors. Its compilation is based on the notion of supervisory control of discrete event systems. It targets the programming of closed-loop reconfiguration controllers in computing systems. We apply our method to the problem of coordinating several administration loops in a data center (number of servers, repair, and local processor frequencies) : we formulate it as an invariance controller synthesis problem.
10:30-10:50	WeA5.3
Open Loop Controllers to Solve the Disturbance Decoupling Problem for Max-Plus Linear Systems, pp. 125-130
Shang, Ying	Southern Illinois Unviersity Edwardsville
Hardouin, Laurent	Univ. of Angers
Lhommeau, Mehdi	Univ. d'Angers
Maia, Carlos Andrey	Univ. Federal de Minas Gerais
This paper presents the new investigations on the disturbance decoupling problem (DDP) in the geometric control of max-plus linear systems. Due to the monotone nondecreasing property of max-plus linear systems, a modified disturbance decoupling problem (MDDP) is defined and an optimal controller to solve the MDDP is obtained using a prefilter mapping by taking the disturbances into account. Moreover, this paper presents a necessary and sufficient condition such that the same optimal controller for the MDDP can be used to solve the DDP for max-plus linear systems. When the necessary and sufficient condition is not satisfied, a new disturbance mapping is developed in order to guarantee the solvability of the DDP by the same optimal controller for the MDDP. The main results of this paper are illustrated by using a timed event graph model of a manufacturing network.
10:50-11:10	WeA5.4
Controller Synthesis for a Class of Hybrid Systems, pp. 131-136
Batis, Sonia	Gipsa-Lab. Grenoble INP
Alla, Hassane	Grenoble Uiversity
A timed control synthesis approach is proposed for hybrid systems modeled with rectangular hybrid automata. The control goal is to constrain the reachable state spaces by the addition of desired control specifications to the automaton transition guards, and to determine in a maximal permissive way the transition dates that respect those constraints. The approach is illustrated via a production system.
11:10-11:30	WeA5.5
Minimum-Time Flow Control of Timed Continuous Choice-Free Nets, pp. 137-142
Wang, Liewei	Univ. of Zaragoza
Mahulea, Cristian	Univ. of Zaragoza
Silva, Manuel	Univ. De Zaragoza
This paper addresses the problem of reaching (and then maintaining) an (optimal) flow of timed continuous Choice-Free Petri nets in minimum-time. First, we compute the optimal flow by solving a linear programming problem. Then we focus on driving the system to a steady-state (marking) corresponding to the optimal flow. The main challenge of solving this problem is the fact that, in general, the steady-state marking corresponding to a given optimal flow is not unique. We propose a heuristic algorithm, in which at each time step we estimate a ``best'' firing count vector that drives the system to the convex region where the optimal flow is obtained; then an ON/OFF strategy is applied. Later, we show that some additional firings can further decrease the time spent to obtain the optimal flow.
11:30-11:50	WeA5.6
Compact Supervisors for General Constraint Enforcement in Petri Net Models with Uncontrollable Transitions, pp. 143-148
Basile, Francesco	Univ. Degli Studi Di Salerno
Cordone, Roberto	Univ. degli Studi di Milano
Piroddi, Luigi	Pol. di Milano
Petri net (PN) models of production processes are subject to a number of heterogeneous constraints, both static, (e.g., limitation and sharing of resources, job limitation) and behavioral (e.g., liveness, reversibility). All of these constraints can be indirectly formulated as generalized mutual exclusion constraints (GMECs), which are conveniently implemented as monitor places suitably connected to the transitions of the open-loop PN plant model. The design procedure is typically sequential, dealing separately with each control objective, possibly resulting in a redundant supervisor. The process is further complicated in the presence of uncontrollable transitions. An integrated modeling approach is here proposed to solve the redundancy problem by accounting for all the specifications in a single design step that optimizes the number of required GMECs and the permissivity of the resulting supervisor. The supervisor can be then implemented as a monitor-based supervisor or, in some cases, as a logical predicate.

WeA6	HG E1.2
Nonlinear System Theory I (Regular Session)
Chair: Scherpen, Jacquelien M.A.	Univ. of Groningen
Co-Chair: Spurgeon, Sarah K.	Univ. of Kent
09:50-10:10	WeA6.1
Nonlinear Control with Approximately Linear Tracking Error, pp. 149-154
Röbenack, Klaus	TU Dresden
Paschke, Fabian	TU Dresden
Knoll, Carsten	TU Dresden
Exact feedback linearization is a well-established method in nonlinear control, where the system is transformed into a linear system by a nonlinear change of coordinates in connection with a nonlinear feedback law. For a controllable single-input system, exact linearizability is equivalent to flatness. Unfortunately, the existence conditions are quite restrictive. Even if the existence conditions are met, the explicit computation of the flat output can be very difficult. In this contribution we suggest tracking controllers based on the tangent system, where the flat output is not required explicitly. The controller can even be calculated for controllable non-flat systems.
10:10-10:30	WeA6.2
Adaptive Output Tracking Control Design of a Gun Launched Micro Aerial Vehicle Based on Approximate Feedback Linearization, pp. 155-160
Drouot, Adrien	CRAN - Nancy Univ.
Richard, Edouard	Nancy Univ.
Boutayeb, M.	Lorraine Univ.
Zasadzinski, Michel	CRAN
Zemouche, Ali	Nancy-Univ.
This paper considers the problem of controlling the position and the orientation of a Gun Launched Micro Aerial Vehicle - GLMAV - despite unknown aerodynamic efforts. The proposed approach overcomes the problem of gyroscopic coupling by taking advantage from the structure of the thrust mechanism, which is made of two counter rotating propellers. An adaptive linearizing controller is designed, allowing the trajectory tracking and the stabilization of the vehicle's position and orientation while the unknown aerodynamic efforts are estimated by means of an identifier. The overall process is shown to be stable for constant, or slowly time varying, aerodynamic efforts. However numerical simulations demonstrate the satisfying controller's performance even with non constant aerodynamic efforts.
10:30-10:50	WeA6.3
Differential Flatness of a Class of N-DOF Planar Manipulators Driven by an Arbitrary Number of Actuators, pp. 161-166
Franch, Jaume	Univ. Pol. de Catalunya
Reyes, Alex	Univ. Pol. de Catalunya-BarcelonaTech
Agrawal, Sunil K.	Univ. of Delaware
This paper deals with the problem of feedback linearization, either by dynamic or static feedback, of a planar robot manipulator with n links and m inputs. The robot is provided with a particular inertia distribution in order to reduce nonlinearities, and with n-m springs to keep controllability. Under these design conditions, the paper answers the question of where to place the actuators such that the system is static feedback linearizable. For those systems that there are not static feedback linearizable, the paper shows that are linearizable by prolongations and, therefore, dynamic feedback linearizable.
10:50-11:10	WeA6.4
On the Computation of Flat Outputs for Nonlinear Control Systems, pp. 167-172
Franke, Matthias	Fraunhofer IIS
Röbenack, Klaus	TU Dresden
A simple procedure for the systematic computation of flat outputs for nonlinear control systems is presented. After symbolic linearization of the system equations, a so-called tangent flat output is determined. For this purpose, a new algorithm which is based on the calculation of nullspaces and generalized inverses is given. If some integrability conditions are satisfied, the flat output of the nonlinear system can be determined by integration. The procedure is illustrated in detail by computational examples.
11:10-11:30	WeA6.5
Tuning of Dynamic Feedback Control for Nonlinear Mechanical Systems, pp. 173-178
Dirksz, Daniel A.	Eindhoven Univ. of Tech.
Scherpen, Jacquelien M.A.	Univ. of Groningen
Mechanical systems are often only equipped with position measurement encoders and obtain the velocity signal by differentiation. However, differentiation largely amplifies noise. In this paper we look at dynamic feedback control of Euler-Lagrange mechanical systems. Dynamic feedback is often used to avoid velocity measurements in the control feedback. In the Euler-Lagrange literature it is shown that the dynamic extension realizes an approximate differentiator, justifying its application. We show in this paper that the dynamic extension used in the Euler-Lagrange literature actually realizes a lead-compensator, proving that a lead-compensator can also globally asymptotically stabilize a nonlinear mechanical systems. Furthermore, based on the lead-compensator structure it is then possible to offer a frequency approach to tune the controller gains.
11:30-11:50	WeA6.6
Delay Independent Static Output Feedback Variable Structure Control for Affine Nonlinear Systems, pp. 179-184
Yan, Xing-Gang	Univ. of Kent
Spurgeon, Sarah K.	Univ. of Kent
Zhu, Quanmin	Univ. of the West of England
Zhang, Qingling	Northeastern Univ.
In this paper, a stabilisation problem for a class of nonlinear control systems with uncertainties involving time varying delay, is considered. It does not require that the nominal system is linearisable or partial linearisable. The bounds on both matched and mismatched uncertainties are nonlinear and time delayed. A static output feedback variable structure control is synthesized to stabilise the system uniformly asymptotically. A control strategy to enforce exponential stability is also derived. The Lyapunov-Razuminkhin approach is used to deal with the time delay, and the bounds on the uncertainties are used in the control design to enhance the robustness. The designed control does not depend on the time delay and thus it is not required that the time delay is known. Finally, the obtained results are tested on a simple mechanical system through simulation.

WeA7	HG E3
Optimal Control (Regular Session)
Chair: Summers, Tyler H.	ETH Zurich
Co-Chair: Ober-Blöbaum, Sina	Univ. of Paderborn
09:50-10:10	WeA7.1
Performance Bounds for Constrained Linear Min-Max Control, pp. 185-190
Summers, Tyler H.	ETH Zurich
Goulart, Paul J.	ETH Zurich
This paper proposes a method to compute lower performance bounds for discrete-time infinite-horizon min-max control problems with input constraints and bounded disturbances. Such bounds can be used as a performance metric for control policies synthesized via suboptimal design techniques. Our approach is motivated by recent work on performance bounds for stochastic constrained optimal control problems using relaxations of the Bellman equation. The central idea of the paper is to find an unconstrained min-max control problem, with negatively weighted disturbances as in H-infinity control, that provides the tightest possible lower performance bound on the original problem of interest and whose value function is easily computed. The new method is demonstrated via a numerical example for a system with box constrained input.
10:10-10:30	WeA7.2
Approximate Dynamic Programming Via Sum of Squares Programming, pp. 191-197
Summers, Tyler H.	ETH Zurich
Kunz, Konstantin	ETH Zurich
Kariotoglou, Nikolaos	ETH Zurich
Kamgarpour, Maryam	ETH Zurich
Summers, Sean	ETH Zurich
Lygeros, John	ETH Zurich
We describe an approximate dynamic programming method for stochastic control problems on infinite state and input spaces. The optimal value function is approximated by a linear combination of basis functions with coefficients as decision variables. By relaxing the Bellman equation to an inequality, one obtains a linear program in the basis coefficients with an infinite set of constraints. We show that a recently introduced method, which obtains convex quadratic value function approximations, can be extended to higher order polynomial approximations via sum of squares programming techniques. An approximate value function can then be computed offline by solving a semidefinite program, without having to sample the infinite constraint. The policy is evaluated online by solving a polynomial optimization problem, which also turns out to be convex in some cases. We experimentally validate the method on an autonomous helicopter testbed using a 10-dimensional helicopter model.
10:30-10:50	WeA7.3
Time-Changed Linear Quadratic Regulators, pp. 198-203
Lamperski, Andrew	Univ. of Cambridge
Cowan, Noah	Johns Hopkins Univ.
Many control methods implicitly depend on the assumption that time is accurately known. For example, the finite-horizon linear quadratic regulator is a linear policy with time-varying gains. Such policies may be infeasible for controllers without accurate clocks, such as the motor systems in humans and other animals, since gains would be applied at incorrect times. Little appears to be known, however, about control with imperfect timing. This paper gives a solution to the linear quadratic regulator problem in which the state is perfectly known, but the controller's measure of time is a stochastic process derived from a strictly increasing Levy process. The optimal controller is linear and can be computed from a generalization of the classical Riccati differential equation.
10:50-11:10	WeA7.4
A Multiobjective Optimization Approach for Optimal Control Problems of Mechanical Systems with Uncertainties, pp. 204-209
Ober-Blöbaum, Sina	Univ. of Paderborn
Seifried, Albert	Univ. of Paderborn
This work deals with the optimal control of mechanical systems with uncertainties which are associated to e.g. unknown parameters such as friction coefficients or geometry parameters in the model. While in the deterministic setting we characterize the maximal performance of the control system by fulfilling a predefined task exactly, in presence of uncertainty a performance measure is introduced whose mean has to be minimized to guarantee maximal performance of the control system. Taking this notion of system performance as additional objective into account leads to a multiobjective optimal control problem. Based on techniques from multiobjective optimization we present a methodology to numerically approximate the set of Pareto optimal solutions. The approach is verified by means of a robot arm maneuver with uncertain model parameters.
11:10-11:30	WeA7.5
On Kinematic Control Extremals, pp. 210-214
Aleksandrov, Vladimir	Moscow State Univ.
Budninskiy, Maxim	Moscow State Univ.
This paper suggests the possible means for optimal kinematic control problems classification. The classification is partially illustrated by two examples with state-constraints taken from the mechanics of controlled systems. The first of these is connected with the application of kinematic control problems to the inference of the general mathematical models of mechanical systems while the second is a problem of the time-optimal motion of a point on a flat surface that represents the movement of a vehicle on a multilane highway.
11:30-11:50	WeA7.6
Convergence of the Shooting Algorithm for Singular Optimal Control Problems, pp. 215-220
Aronna, Maria Soledad	Imperial Coll. of London
In this article we propose a shooting algorithm for optimal control problems governed by systems that are affine in one part of the control variable. Finitely many equality constraints on the initial and final state are considered. We recall a second order sufficient condition for weak optimality, and show that it guarantees the local quadratic convergence of the algorithm. We show an example and solve it numerically.

WeA8	HG E5
Consensus Control I (Regular Session)
Chair: Dimarogonas, Dimos V.	Royal Inst. of Tech. (KTH)
Co-Chair: Kibangou, Alain Yuwa	GIPSA-Lab. Joseph Fourier/CNRS
09:50-10:10	WeA8.1
Ergodicity and Class-Ergodicity of Balanced Asymmetric Stochastic Chains, pp. 221-226
Bolouki, Sadegh	GERAD, Ec. Pol. de Montreal
Malhame, Roland P.	Ec. Pol. de Montreal
Unconditional consensus is the property of a consensus algorithm for multiple agents, to produce consensus irrespective of the particular time or state at which the agent states are initialized. Under a weak condition, so-called balanced asymmetry, on the sequence (A_n) of stochastic matrices in the agents states update algorithm, it is shown that (i) the set of accumulation points of states as n grows large is finite, (ii) the asymptotic unconditional occurrence of single consensus or multiple consensuses is directly related to the property of absolute infinite flow of this sequence, as introduced by Touri and Nedi'c. The latter condition must be satisfied on each of the islands of the so-called unbounded interactions graph induced by (A_n), defined by Hendricks et al. The property of balanced asymmetry is satisfied by many of the well known discrete time consensus models studied in the literature.
10:10-10:30	WeA8.2
Consensus-Based Distributed Estimation of Laplacian Eigenvalues of Undirected Graphs, pp. 227-232
Tran, Thi Minh Dung	Gipsa-Lab.
Kibangou, Alain Yuwa	GIPSA-Lab. Joseph Fourier/CNRS
In this paper, we present a novel algorithm for estimating eigenvalues of the Laplacian matrix associated with the graph describing the network topology of a multi-agent system or a wireless sensor network. As recently shown, the average consensus matrix can be written as a product of Laplacian based consensus matrices whose stepsizes are given by the inverse of the nonzero Laplacian eigenvalues. Therefore, by solving the factorization of the average consensus matrix, we can infer the Laplacian eigenvalues. We show how solving such a matrix factorization problem in a distributed way. In particular, we formulate the problem as a constrained consensus problem. The proposed algorithm does not require great resources in both computation and storage. This algorithm can also be viewed as a way for decentralizing the design of finite-time average consensus protocol recently proposed in the literature. Eventually, the performance of the proposed algorithm is evaluated by means of simulation results.
10:30-10:50	WeA8.3
On the Weight Controllability of Consensus Algorithms, pp. 233-238
Goldin, Darina	Tech. Univ. Berlin
Raisch, Joerg	Tech. Univ. Berlin
In this paper we consider controllability of leader-follower networks running a consensus algorithm. We allow the communication links between the agents to be weighted and give the necessary and sufficient conditions for a the system to be controllable generically, i.e. for almost all choices of weights. We call such a network weight controllable, a new notion that is introduced in this paper. A new descriptor formulation for the leader-follower consensus algorithm is derived that allows modelling communication weights as free parameters without changing the system structure. This leads us to obtaining necessary and sufficient conditions for weight controllability of leader-follower consensus algorithms.
10:50-11:10	WeA8.4
Average Consensus in Symmetric Nonlinearly Coupled Delayed Networks, pp. 239-243
Proskurnikov, Anton	St.-Petersburg State Univ.
The paper addresses consensus under nonlinear couplings and bounded delays for multi-agent systems, where the agents have the single-integrator dynamics. The network topology is undirected and may alter as time progresses. The couplings are uncertain and satisfy a conventional sector condition with known sector slopes. The delays are uncertain, timevarying and obey known upper bounds. The network satisfies a symmetry condition that resembles the Newton's Third Law. Explicit analytical conditions for the robust consensus are offered that employ only the known upper bounds for the delays and the sector slopes.
11:10-11:30	WeA8.5
Global Consensus in Homogeneous Networks of Discrete-Time Agents Subject to Actuator Saturation, pp. 244-249
Yang, Tao	Royal Inst. of Tech.
Meng, Ziyang	Royal Inst. of Tech.
Dimarogonas, Dimos V.	Royal Inst. of Tech. (KTH)
Johansson, Karl Henrik	Royal Inst. of Tech.
In this paper, we give necessary conditions for achieving global consensus in homogeneous networks of discrete-time linear time-invariant agents with input saturation constraints under fixed undirected topologies. For two special cases, where the agent model is either neutrally stable or a double integrator, these necessary conditions together with a gain condition are also sufficient. We show this by designing a linear protocol based on the combination of state differences between the agent and its neighbors. In particular, for the neutrally stable case, we show that any linear protocol of a particular form which solves the consensus problem for the case without input saturation constraints also solves the global consensus problem for the case with input saturation constraints. For the double integrator case, we show that a subset of linear protocols which solve the consensus problem for the case without saturation constraints also solve the global consensus problem in the presence of input saturation. The results are illustrated by numerical simulations.
11:30-11:50	WeA8.6
Random Convex Programs for Distributed Multi-Agent Consensus, pp. 250-255
Calafiore, Giuseppe	Pol. di Torino
Lyons, Daniel	Karlsruhe Inst. of Tech.
We consider convex optimization problems with N randomly drawn convex constraints. Previous work has shown that the tails of the distribution of the probability that the optimal solution subject to these constraints will violate the next random constraint, can be bounded by a binomial distribution. In this paper we extend these results to the violation probability of convex combinations of optimal solutions of optimization problems with random constraints and different cost objectives. This extension has interesting applications to distributed multiagent consensus algorithms in which the decision vectors of the agents are subject to random constraints and the agents' goal is to achieve consensus on a common value of the decision vector that satisfies the constraints. We give explicit bounds on the tails of the probability that the agents' decision vectors at an arbitrary iteration of the consensus protocol violate further constraint realizations. In a numerical experiment we apply these results to a model predictive control problem in which the agents aim to achieve consensus on a control sequence subject to random terminal constraints.

WeA9	HG E7
Mechatronic Systems I (Regular Session)
Chair: Wagner, John	Clemson Univ.
Co-Chair: Widmer, Tobias	ETH Zurich
09:50-10:10	WeA9.1
An Electromagnet Excited Mass–Pendulum System Laboratory Experiment — Theory and Test, pp. 256-261
Austin, Kelly	Clemson Univ.
Wagner, John	Clemson Univ.
An electromagnet excited mass–pendulum system with attached spring and damper elements is introduced as a senior and graduate level engineering laboratory experiment. This laboratory offers mechanical, electrical, and control engineering challenges to the students. The derivation of the coupled equations of motion is developed using both Newtonian and Lagrangian approaches. The system is pendulum actuated by a powerful electromagnet for which the magnetic force is modeled by a magnetostatic forcing function. Representative numerical and experimental results are presented which validate the mathematical model. Further, the bench top experiment offers hands-on opportunities for the students. The numerical results agree within 2% to 20% of the experiments.
10:10-10:30	WeA9.2
Robust Control of Magnetic Guidance Lightweight AGVs Path Tracking Using Randomization Methods, pp. 262-267
Madaschi, Matteo Giacomo	Univ. degli Studi di Bergamo
Gryazina, Elena	Inst. for Control Sciences RAS
Cologni, Alberto	Dipartimento di Ingegneria dell'Informazione e Metodi Matematici
Spelta, Cristiano	Univ. degli Studi di Bergamo
Previdi, Fabio	Univ. degli Studi di Bergamo
Savaresi, Sergio M.	Pol. Di Milano
Pesenti, Ivan	R&D Business Unit Scaglia-Indeva
In this paper, a state space dynamical model for the differential drive path tracking system of an AGV (Automated Guided Vehicle) is developed. In order to guarantee desirable control specifications for a wide range of possible masses (50 < M < 1000kg), the payload mass is considered as uncertainty in the model. Robust stability domain in the controller parameter space via randomization technique is analysed. Two sets of controller parameters are selected and experimentally tested on a commercial vehicle, showing performances in agreement with the analysis
10:30-10:50	WeA9.3
The Cubli: A Reaction Wheel Based 3D Inverted Pendulum, pp. 268-274
Gajamohan, Mohanarajah	ETH Zurich
Muehlebach, Michael	ETH Zurich
Widmer, Tobias	ETH Zurich
D'Andrea, Raffaello	ETH Zurich
The Cubli is a 15 x 15 x 15 cm cube with reaction wheels mounted on three of its faces. By applying controlled torques to the reaction wheels the Cubli is able to balance on its corner or edge. This paper presents the development of the Cubli. First, the mechatronic design of the Cubli is presented. Then the multi-body system dynamics are derived. The parameters of the non linear system are identified using a frequency domain based approach while the Cubli is balancing on its edge with a nominal controller. Finally, the corner balancing using a linear feedback controller is presented along with experimental results.
10:50-11:10	WeA9.4
Pedaling Torque Reconstruction for Half Pedaling Sensor, pp. 275-280
Spagnol, Pierfrancesco	Pol. di Milano
Corno, Matteo	Pol. di Milano
Savaresi, Sergio M.	Pol. Di Milano
Electrically Power Assisted Cycles (EPACs) have been gaining increasing attention worldwide during the past few years. Market analysis show that cyclists prefer torque assisted EPACs over velocity-based ones. A torque assisted cycle requires an estimate or measurement of the cyclist torque, which increases the cost and complexity of the set-up. Among the viable solutions, sensors mounted on the pedal axis have several advantages in terms of packaging and cost; unfortunately, they are capable of measuring the torque exerted only on one pedal. This paper deals with the development, implementation and validation of two algorithms to estimate the entire revolution torque using a half pedaling sensor. The algorithms are based on the iso-power or iso-torque assumptions (checked on a high-end training bicycle). The algorithms are validated using road measurements, showing that a sufficient level of accuracy for EPAC is reached.
11:10-11:30	WeA9.5
Modeling and Analysis of an Atmospheric Driven Atmos Clock with Mechanical Escapement Control, pp. 281-287
Patel, Shreyas	Clemson Univ.
Moline, David	Clemson Univ.
Wagner, John	Clemson Univ.
The classic Atmos mantle clock operates on atmospheric differences to power the mechanical bellows and wind the mainspring. A delicate gear train and verge turn a torsional pendulum with minimal frictional losses throughout the system. A pair of hands display the time on a dial in twelve hour increments. In this paper, an Atmos 540 clock has been investigated as it offers vivid insight into a mechanical escapement controller with integrated system components, and a “green” temperature and/or pressure driven power source. The clock design, operation, and mathematical model representing the energy flow through the system will be presented. Representative experimental and analytical results will be introduced and discussed to demonstrate the clock's functionality. Finally, the question of whether this clock meets the definition of a perpetual motion machine will be examined.
11:30-11:50	WeA9.6
Stabilization of the Ball and Beam System by Dynamic Output Feedback Using Incremental Measurements, pp. 288-294
Rapp, Philipp	Univ. of Stuttgart
Sawodny, Oliver	Univ. of Stuttgart
Tarin, Cristina	Univ. of Stuttgart
The present contribution deals with the development of a stabilizing dynamic output feedback controller for the ball and beam system. The measured states are composed of the relative rotation angle and the absolute ball position. The torque applied to the beam serves as the control input. Significant practical relevance of this contribution is due to the relative rotation angle measurement, provided by most commercially available encoders. The controller is designed by employing a published inverse Lyapunov stabilizing approach together with a novel nonlinear observer based on the I&I methodology. Simulation experiments illustrate the theory and show the effectiveness of the proposed design.

WeA10	HG D1.1
Linear Systems I (Regular Session)
Chair: Kucera, Vladimir	Czech Tech. Univ. in Prague
Co-Chair: Chesi, Graziano	Univ. of Hong Kong
09:50-10:10	WeA10.1
Partial Interconnection Dead-Beat Control of Two-Dimensional Behaviors, pp. 295-300
Bisiacco, Mauro	Univ. of Padova
Valcher, Maria Elena	Univ. di Padova
In this paper the dead-beat control problem, by partial interconnection, of two-dimensional (2D) discrete behaviors, defined on the grid Z+ x Z and having the time as (first) independent variable, is investigated. It turns out that the possibility of driving to zero (in a finite number of "steps") the "relevant variables" by means of a partial interconnection controller, making use of the "measurable variables", is equivalent to the reconstructibility of the relevant variables, that are not accessible for control. On the other hand, if we search for "admissible" dead-beat controllers, the only ones that provide meaningful results in practice, we have to introduce the zero-time-controllability assumption on the relevant variables.
10:10-10:30	WeA10.2
A Remark on Conditioned Invariance in the Behavioral Approach, pp. 301-305
Pereira, Ricardo	Univ. of Aveiro
Rocha, Paula	Univ. of Porto
In this paper a definition for the property of behavioral invariance is proposed with the purpose of generalizing the state space geometric approach to the behavioral setting. Based on this notion together with the well-known notion of behavioral observer, a definition of conditioned invariance is also presented. The results obtained for the characterization of the defined properties put into evidence some problems that, in our opinion, should deserve attention. This could serve as a starting point for a discussion on the foundations of an analogue of the geometric theory within the behavioral setting.
10:30-10:50	WeA10.3
Optimal Decoupling Controllers for Singular Systems, pp. 306-311
Kucera, Vladimir	Czech Tech. Univ. in Prague
The problem of decoupling a singular, linear, and time-invariant system by dynamic compensation into multi-input multi-output subsystems is studied. The set of all controllers that decouple the system and render it regular, proper and stable is determined in parametric form. Optimal and suboptimal decoupling controllers are then obtained by an appropriate selection of the parameters.
10:50-11:10	WeA10.4
On the Definition of Negative Imaginary System for Not Necessarily Rational Transfer Functions, pp. 312-316
Ferrante, Augusto	Univ. di Padova
Ntogramatzidis, Lorenzo	Curtin Univ.
In this paper we provide a definition and characterisation of negative imaginary systems for not necessarily rational but symmetric transfer functions along the same lines of the classic definition of positive real systems. We then derive a necessary and sufficient condition that characterises symmetric negative imaginary transfer functions in terms of a matrix sign condition on the imaginary axis.
11:10-11:30	WeA10.5
On the Structure of Linear Behaviors Over Quaternions, pp. 317-322
Vettori, Paolo	Univ. of Aveiro
Some results on the time-domain structure of linear, time-invariant systems over quaternions are presented, both in the continuous and in the discrete-time case. Within the behavioral approach, a system is defined as a set of trajectories (functions or sequences). In this paper, the trajectories are solutions of linear differential or difference equation with constant coefficients which belong to the skew-field of quaternions. As in the real and complex case, the equations may be represented by polynomials whose roots are related to the solutions. However, the properties of the roots and the structure of the corresponding solutions, which are analyzed in the paper, differ in many aspects from the standard commutative case.
11:30-11:50	WeA10.6
A Necessary and Sufficient LMI Condition for Stability of 2D Mixed Continuous-Discrete-Time Systems, pp. 323-328
Chesi, Graziano	Univ. of Hong Kong
Middleton, Rick	The Univ. of Newcastle
This paper addresses the problem of establishing stability of 2D mixed continuous-discrete-time systems. Traditional stability analysis for 2D systems gives a sufficient condition based on 2D version of a Lyapunov equation. Here, a linear matrix inequality (LMI) condition is proposed that extends these results by introducing complex Lyapunov functions depending polynomially on a parameter and by exploiting the Gram matrix method. It is shown that this condition is sufficient for 2D exponential stability for any chosen degree of the Lyapunov function candidate, and it is also shown that this condition is also necessary for a sufficiently large degree. Moreover, an a priori bound on the degree required for achieving necessity is given. Some numerical examples illustrate the proposed methodology.

WeA11	HG D1.2
Control in Aerospace I (Regular Session)
Chair: Tsiotras, Panagiotis	Georgia Inst. of Tech.
Co-Chair: Kyriakopoulos, Kostas J.	National Tech. Univ. of Athens
09:50-10:10	WeA11.1
Rigid Body Motion Tracking without Linear and Angular Velocity Feedback Using Dual Quaternions, pp. 329-334
Filipe, Nuno	Georgia Inst. of Tech.
Tsiotras, Panagiotis	Georgia Inst. of Tech.
This paper takes advantage of a new, recently proposed representation of the combined translational and rotational dynamic equations of motion of a rigid body in terms of dual quaternions. We show that combined position and attitude tracking controllers based on dual quaternions can be developed with relatively low effort from existing attitude-only tracking controllers based on quaternions. We show this by developing an almost globally asymptotically stable nonlinear controller capable of simultaneously following time-varying position and attitude profiles without linear and angular velocity feedback based on an existing attitude-only tracking controller without angular velocity feedback.
10:10-10:30	WeA11.2
Nonlinear Control of the Longitudinal Rotational Dynamics of a Flexible Aircraft, pp. 335-340
Duraffourg, Elodie	ONERA
Burlion, Laurent	ONERA
Ahmed-Ali, Tarek	ENSIETA
Lamnabhi-Lagarrigue, Francoise	CNRS-EECI
This paper studies the longitudinal inner loop control design of a flexible aircraft over a large flight domain. More precisely, the objective is to construct a stabilizing flight control law to steer the angle of attack to a desired trim value while suppressing the oscillation of the bending modes. A new Lyapunov-based nonlinear controller is designed to enhance the transient of the flexible modes before reaching the zero dynamics submanifold. Moreover, the local behavior of the trajectories is finally modified by blending our controller with a LQ controller in order to achieve the additional and practical control objective of locally damping out the oscillations of the bending modes.
10:30-10:50	WeA11.3
A Simplified Goddard Problem in the Presence of a Nonlinear Media Resistance and a Bounded Thrust, pp. 341-346
Samylovskiy, Ivan	Lomonosov Moscow State Univ.
We consider a problem of maximization of the distance traveled by a material point in the presence of a nonlinear friction under a bounded thrust and fuel expenditure. Using the maximum principle we obtain the form of optimal control and establish conditions under which it contains a singular subarc. This problem seems to be the simplest one having a mechanical sense in which singular subarcs appear in a nontrivial way.
10:50-11:10	WeA11.4
Robust Attitude Control for an Unmanned Helicopter in Near-Hover Flights, pp. 347-352
Marantos, Panos	National Tech. Univ. of Athens
Dritsas, Leonidas	Univ. of Patras
Kyriakopoulos, Kostas J.	National Tech. Univ. of Athens
In this paper, a systematic procedure for designing robust attitude controllers for unmanned helicopters, based on mixed H2/Hinf methodologies, is presented. Firstly, a family of linearized models describing the near-hover flight dynamics is derived which can be compactly formulated as a nominal plant perturbed by norm bound uncertainties on the system, control and wind matrices. It is then shown that a single robust controller can be designed guaranteeing stability, robustness and gust disturbance rejection for the whole near-hover flight envelope. Performance analysis and simulation results show that the proposed attitude control strategy can also satisfy the handling qualities defined in ADS-33E specification requirements. Finally, the attitude controller is used as a module in a total control scheme offering position tracking capabilities which is implemented in a real embedded system. The efficacy of the total control structure is proved by Hardware-In the-Loop simulations on an accurate nonlinear helicopter model.
11:10-11:30	WeA11.5
Dynamics Modeling and Comparative Robust Stability Analysis of a Space Launcher with Constrained Inputs, pp. 353-358
Chaudenson, Julien	SUPELEC Sciences des Systemes (E3S)
Beauvois, Dominique	Ec. Superieure D'Electricite
Bennani, Samir	ESA/ESTEC (TEC-Ec.
Ganet-Schoeller, Martine	ASTRIUM Space Transportation
Sandou, Guillaume	Ec. Superieure d Electricite
This paper presents a comparative robust stability analysis using IQC-based and LMI-based tools. An LFR, resulting from the factorization of the equations of motion, allows assessing the robust stability of a space launcher in the face of inertia uncertainties and saturated inputs.
11:30-11:50	WeA11.6
New Algorithmic Approach Based on Integral Quadratic Constraints for Stability Analysis of High Order Models, pp. 359-364
Demourant, Fabrice	ONERA-CERT
To analyze a non linear closed loop which represents a high order aeroelastic model of a large civil aircraft interconnected with non-linearities, an Integral Quadratic Constraints (IQC) approach has been involved. This approach is particularly interesting for two reasons. The first one is that it is possible with the same stability criterion to analyze a large class of stability problems. And the second reason is that the stability criterion is based on a frequency domain inequalities (FDI). Usually the Kalman-Yakubovich-Popov (KYP) lemma is involved to transform this infinite set of inequalities into one linear matrix inequality (LMI). But this kind of approach leads to a strong increase in the number of optimization variables. Consequently a new FDI based algorithmic approach has been developed. Usually the number of FDI to satisfy is infinite. To tackle this problem a specific technique has been developed. It consists in computing a frequency domain where the solution is valid. By an iterative approach this domain is extended until it covers $[0,+infty[$. In this way the solution obtained from the FDI is necessarily valid on the frequency domain continuum and the number of optimization variables remains limited which makes tractable the IQC approach for high order models.

WeA12	HG D3.2
Adaptive Control I (Regular Session)
Chair: Fagiano, Lorenzo	ETH Zurich
Co-Chair: Fabri, Simon G.	Univ. of Malta
09:50-10:10	WeA12.1
Closed-Loop Reference Models for Output-Feedback Adaptive Systems, pp. 365-370
Gibson, Travis	Massachusetts Inst. of Tech.
Annaswamy, Anuradha	Massachusetts Inst. of Tech.
Lavretsky, Eugene	The Boeing Co.
Closed-loop reference models have recently been proposed for states accessible adaptive systems. They have been shown to have improved transient response over their open loop counter parts. The results in the states accessible case are extended to single input single output plants of arbitrary relative degree.
10:10-10:30	WeA12.2
Mirror Decent Algorithm for a Multi-Armed Bandit Governed by a Stationary Finite State Markov Chain, pp. 371-375
Miller, Boris	Monash Univ. Clayton Campus
Nazin, Alexander V.	Inst. of Control Sciences, RAS
This article further develops an adaptive approach to the control of observable Markov chains with a finite number of states. We apply the Mirror Descent Randomized Control Algorithm (MDRCA) to a class of homogeneous finite Markov chains governed by the multi-armed bandit with unknown mean losses. The article develops the approach represented in [18]. As opposed to the partially observable Markov decision process an adaptive approach does not presuppose the knowledge of probabilistic characteristics of random perturbations and permits to obtain the control strategy with known rate of convergence to the optimal solution. We propose the concrete MDRCA and prove the explicit, non-asymptotic upper bound for the mean losses at a given (finite) time horizon. Numerical example illustrates theoretical results.
10:30-10:50	WeA12.3
Multiple Stationary Solutions to the Extremum Seeking Control Problem, pp. 376-381
Trollberg, Olle	Royal Inst. of Tech. KTH
Jacobsen, Elling W.	Royal Inst. of Tech. - KTH
Extremum seeking control was originally proposed for adaptive optimization of static systems and later extended to Hammerstein and Wiener systems. More recently, stability and convergence results were presented also for general type dynamic systems with a focus on the local behavior around the optimum and under assumptions of relatively slow gradient estimation and control. In this paper we derive properties characterizing any stationary solution of the extremum seeking control scheme, i.e., we do not restrict ourselves to solutions close to optimum and allow for any frequency in the sinusoidal perturbation based gradient estimation scheme. By considering the linear properties around a stationary solution of the system, we show that stationary solutions are characterized by either a zero gradient or a phase lag condition. The former condition is satisfied at the optimum only for systems in which the zero gradient at the optimum is due to a static nonlinearity. The phase lag condition is shown to be satisfied close to the optimum for low frequency excitations, but can also be satisfied at solutions arbitrarily far from the optimum. The results imply that the extremum seeking control scheme applied to general type dynamic systems can have multiple stable stationary solutions of which some are sub-optimal and potentially far removed from the optimum. For illustration we consider extremum seeking control of a tubular bioreactor, displaying a maximum yield, and show that the closed-loop has two saddle-node bifurcations resulting in a total of three possible stationary solutions for some perturbation frequencies. A stable sub-optimal solution, with a yield less than 10% of the optimal yield, exists even with relatively slow gradient estimation.
10:50-11:10	WeA12.4
Adaptive Model Predictive Control for Constrained Linear Systems, pp. 382-387
Tanaskovic, Marko	ETH Zurich
Fagiano, Lorenzo	ETH Zurich
Smith, Roy S.	ETH Zurich
Goulart, Paul J.	ETH Zurich
Morari, Manfred	ETH Zurich
A novel adaptive output feedback control technique for uncertain linear systems is proposed, able to cope with input and output constraints and measurement noise. At each time step, the collected input-output data are exploited to refine the set of models that are consistent with the available information on the system. Then, the control input is computed according to a receding horizon strategy, which guarantees recursive constraint satisfaction for all the admissible models, hence also for the actual plant. The technique relies only on the solution of linear and quadratic programs. The effectiveness of the approach is illustrated in a numerical example.
11:10-11:30	WeA12.5
Adaptive Symbolic Feedback for Discrete-Time Uncertain Systems, pp. 388-391
Hayakawa, Tomohisa	Tokyo Inst. of Tech.

11:30-11:50	WeA12.6
Unscented Transform-Based Dual Adaptive Control of Nonlinear MIMO Systems, pp. 392-397
Fabri, Simon G.	Univ. of Malta
Bugeja, Marvin K.	Univ. of Malta
The paper proposes a multilayer perceptron neural network controller for dual adaptive control of a class of stochastic MIMO nonlinear systems subject to functional uncertainty. The neural network parameters are adjusted in real-time using the Unscented Kalman filter algorithm and no pre-operational training phase is required. Dual adaptive control aims to strike a compromise between the two control characteristics of caution and probing, leading to an improved overall performance. The system is evaluated through numerical simulations and Monte Carlo analysis. The resulting performance of the dual adaptive controller is not only consistently superior to non-dual adaptive control schemes, but also surpasses the performance of similar controllers that are based on Extended Kalman filter estimators. This reflects the enhanced accuracy of the Unscented Kalman filter estimator, despite being a local estimation method. In addition, unlike use of other estimators, the proposed approach neither requires the computation of complex Jacobian matrices as part of the design, nor the evaluation of such matrices in real-time. This renders the proposed controller inherently amenable and practical for real-time implementation.

WeA13	HG D5.2
Fault Detection and Identification I (Regular Session)
Chair: Parisini, Thomas	Imperial Coll. & Univ. of Trieste
Co-Chair: Puig, Vicenc	Univ. Pol. de Catalunya
09:50-10:10	WeA13.1
Cyber Attack Detection and Faults Diagnosis in Power Networks by Using State Fault Diagnosis Matrix, pp. 398-403
Fujita, Yuki	Keio Univ.
Namerikawa, Toru	Keio Univ.
Uchida, Kenko	Waseda Univ.
This paper discusses cyber attacks and faults diagnosis method for networked electrical power systems. In this paper, a power system is divided into some areas, and then, some parameters are estimated by Kalman filter in each area. Then we propose a diagnosis method by using state and output fault diagnosis matrices which are composed of the estimated values, output matrix and so on. These diagnosis matrices can immediately judge which nodes of which area is attacked. Finally, some simulation results are shown to validate the proposed approach.
10:10-10:30	WeA13.2
Extended-Horizon Analysis of Pressure Sensitivities for Leak Detection in Water Distribution Networks: Application to the Barcelona Network, pp. 404-409
Casillas, Myrna Violeta	Inst. Tecnológico y de Estudios Superiores de Monterrey
Garza-Castañón, Luis	Tecnológico de Monterrey, Campus Monterrey
Puig, Vicenc	Univ. Pol. de Catalunya
In this paper, a model-based leak detection and isolation approach for water distribution networks (WDN), which considers an extended time horizon analysis of pressure sensitivities is proposed. It differs from previous works based on pressure sensitivities analysis since the existing approaches were considering time instant evaluation. This fact makes those approaches very sensitive to demand changes and noise in measurements. A fault isolation approach based on new criterion, known as the angle method, is introduced. This criterion is based on evaluating the angle between the residual vector and the columns of the leak sensitivity matrix. The performance of the proposed approach is compared with two well established methods in the literature (the least square optimization and the correlation methods) when they are applied to the Barcelona WDN.
10:30-10:50	WeA13.3
Model-Based Leakage Localization in Drinking Water Distribution Networks Using Structured Residuals, pp. 410-415
Rosich, Albert	Univ. of Luxembourg
Puig, Vicenc	Univ. Pol. de Catalunya
In this paper, a new model based approach to leakage localization in drinking water networks is proposed based on generating a set of structured residuals. The residual evaluation is based on a numerical method based on an enhanced Newton-Raphson algorithm. The proposed method is suitable for water network systems because the non-linearities of the model make impossible to derive analytical residuals. Furthermore, the computed residuals are designed so that leaks are decoupled, which improves the localization of leaks with respect to other existing methods. Finally, the Hanoi water network benchmark is used to illustrate the results of the proposed approach.
10:50-11:10	WeA13.4
Process Monitoring Based on Kullback Leibler Divergence, pp. 416-421
Zeng, Jiusun	China Jiliang Univ.
Xie, Lei	National Key Lab. of Industrial Control Tech.
Kruger, Uwe	Queens Univ. of Belfast
Yu, Jie	McMaster Univ.
This article proposes to monitor industrial process faults using Kullback Leibler (KL) divergence. The main idea is to measure the difference between the distributions of normal and faulty data. Sensitivity analysis on the KL divergence under Gaussian distribution assumption is performed, which shows that the sensitivity of KL divergence increases with the number of samples. For non-Gaussian data, a recently proposed kernel method for density ratio estimation is used to estimate the KL divergence. The density ratio estimation method does not involve direct estimation of probability density functions, hence is fast and efficient. For monitoring of non-Gaussian data, the confidence limits are obtained through a window based strategy. Application studies involving a simulation example and an industrial melter process show that the performance of the proposed monitoring strategy is better than the principal component analysis (PCA) based statistical local approach.
11:10-11:30	WeA13.5
A Distributed Fault Detection Filtering Approach for a Class of Interconnected Input-Output Nonlinear Systems, pp. 422-427
Keliris, Christodoulos	Univ. of Cyprus
Polycarpou, Marios M.	Univ. of Cyprus
Parisini, Thomas	Imperial Coll. & Univ. of Trieste
This paper develops a filtering approach for distributed fault detection of a class of interconnected input-output nonlinear systems with modeling uncertainties, disturbances and measurement noise. A distributed fault detection scheme and the corresponding adaptive thresholds are designed based on filtering certain signals so that the effect of the measurement noise and disturbances is attenuated, which facilitates less conservative thresholds and enhanced robustness. Further analysis leads to a quantitative characterization of the class of detectable faults and simulation results are used to illustrate the proposed distributed fault diagnosis filtering approach.
11:30-11:50	WeA13.6
Contribution Plots for Statistical Process Control: Analysis of the Smearing-Out Effect, pp. 428-433
Van den Kerkhof, Pieter	KU Leuven
Vanlaer, Jef	Chemical & Biochemical Process Tech. & Control (BioTeC), De
Gins, Geert	Chemical & Biochemical Process Tech. & Control (BioTeC), De
Van Impe, Jan F.M.	Katholieke Univ. Leuven
Since the generation of contribution plots requires no a priori information about the detected disturbance (e.g., historical faulty data), it is a popular fault isolation technique in Statistical Process Control (SPC). However, Westerhuis et al. (2000) reported that contribution plots suffer from fault smearing, i.e., the influence of faulty variables on the contributions of non-faulty variables, which complicates the fault isolation task as variables unaffected by the fault may be highlighted and faulty variables obscured. This paper presents an analysis of the smearing effect for three general contribution computation methods: Complete Decomposition, Partial Decomposition and Reconstruction-Based contributions. The analysis shows that (i) smearing is present in all three methods, (ii) smearing depends on the chosen number of principal components of the underlying latent variable model and (iii) the extent of smearing increases for variables correlated in the training data for a well-chosen model order. The effect of smearing on the isolation performance of single and multiple sensor faults of various magnitudes is illustrated using a simulation case study. The results indicate that correct isolation with contribution plots is not guaranteed for multiple sensor faults. Furthermore, contribution plots only outperform univariate fault isolation for single sensor faults with small magnitudes. For multiple sensor faults, univariate fault isolation exhibits a significantly larger correct fault isolation rate. Based on the smearing analysis and the results for sensor faults, the authors advise to use contributions only if a sound physical interpretation of the principal components is available.

WeA14	HG D7.1
Automotive Control (Regular Session)
Chair: Brun, Xavier	Insa de Lyon
Co-Chair: Glielmo, Luigi	Univ. of Sannio
09:50-10:10	WeA14.1
Torque Control of a Diesel Engine by an Eigenpressure Based Approach, pp. 434-439
Moser, Dominik	Johannes Kepler Univ. Linz
Hahn, Sebastian	Johannes Kepler Univ. Linz
Waschl, Harald	Johannes Kepler Univ. Linz
Del Re, Luigi	Johannes Kepler Univ. Linz
Although the application of cylinder pressure sensors to obtain insight into the combustion process is not a novel topic itself, the recent availability of inexpensive in- cylinder pressure sensors has prompt an upcoming interest for the utilization of the cylinder pressure signal within engine control and monitoring. By the use of techniques, like principle component analysis, it is possible to reduce the high amount of data in the pressure signal during one cycle whilst preserving as much as possible of the fundamental information. Up to now this extracted information, the so called features, were mainly used for modeling tasks and virtual sensors. In this work a converse approach is proposed, namely to directly control these features to desired values by controlling the injection parameters. To this end, the relation between engine torque and features was identified and in addition also models for the relation between features and injection parameters, like the angle and amount of the main injection, were obtained. These models were then used for the active control of the features. The method was implemented on a 2L common rail Diesel engine at the testbench of the JKU Linz and led to initial results in a torque control application.
10:10-10:30	WeA14.2
Robust Vehicle Stability Control with an Uncertain Driver Model, pp. 440-445
Carvalho, Ashwin	Univ. of California, Berkeley
Palmieri, Giovanni	Univ. del Sannio
Tseng, Hongtei Eric	FORD
Glielmo, Luigi	Univ. of Sannio
Borrelli, Francesco	Univ. of California, Berkeley
We present the design of a robust lateral stability controller to track yaw rate and lateral velocity reference signals while avoiding front and rear tire force saturation. The controller takes into account the driver's intent at the design stage by treating it as a measured disturbance. The uncertainty in the driver's input is modeled as a set--valued function of the vehicle states. The control design is based on a hybrid piecewise affine bicycle model with input--dependent and state--dependent uncertainties. The performance of the controller and the importance of driver behavior modeling are demonstrated through experimental tests on ice with aggressive driver maneuvers.
10:30-10:50	WeA14.3
Comparison of Feedback Linearization and Flatness Control for Anti-Slip Regulation (ASR) of an Hybrid Vehicle : From Theory to Experimental Results, pp. 446-451
Chapuis, Cédric	Sagem DS
Bideaux, Eric	INSA Lyon, Lab. Ampere
Brun, Xavier	Insa de Lyon
Minoiu Enache, Nicoleta	Renault SAS
This paper compares 3 Anti-Slip Regulation control laws, a PI controller, a linearizing feedback and a flatness based control. First, the controllers are designed based on a two state equations vehicle model. Then, the controllers are experimentally validated on an hybrid vehicle (VELROUE) with rear electric wheels and front axle Internal Combustion Engine traction. Finally, the nonlinear control laws are compared to the PI controller that is classically used in the automotive industry.
10:50-11:10	WeA14.4
Optimal Energy Management for a Mechanical-Hybrid Vehicle with Cold Start Conditions, pp. 452-457
van Berkel, Koos	Tech. Univ. Eindhoven
Klemm, Wouter	Tech. Univ. Eindhoven
Hofman, Theo	TU/e
Vroemen, Bas	Drivetrain Innovations
Steinbuch, Maarten	Eindhoven Univ. of Tech.
This paper presents the design of an optimal Energy Management Strategy (EMS) for a hybrid vehicle that starts with a cold powertrain. The cold start negatively affects the combustion and transmission efficiency of the powertrain, caused by the higher frictional losses due to increased hydrodynamic viscosity effects. The excess fuel consumption of the engine and the excess power loss of the transmission are modeled by static relations as a function of the lubrication oil temperature. The thermodynamics in the powertrain during the heating period of the powertrain is approximated by a first-order dynamic model. The main design criterion for the optimal EMS is the minimization of the overall fuel consumption over a pre-defined driving cycle. Dynamic programming is used to find the globally optimal solution for six representative driving cycles. The results show that the cold start has a significant impact on the fuel consumption of the hybrid vehicle, yet its influence on the optimal EMS is negligible.
11:10-11:30	WeA14.5
Model Predictive Parking Control for Nonholonomic Vehicles Using Time-State Control Form, pp. 458-465
Oyama, Kentaro	Tokyo City Univ.
Nonaka, Kenichiro	Tokyo City Univ.
In this study, a model predictive control (MPC) for nonholonomic vehicles with steering and travel range constraints is proposed. We transformed the nonlinear dynamics into two linear subsystems using time-state control form (TSCF) where we suppress the computational effort of MPC with reduced order dynamics. Both steering and travel range constraints can be considered using 1st order approximation with enough accuracy. The parking problem for nonholonomic vehicles needs a switching movement where the switching point is automatically optimized. The performance of the proposed method is verified through actual vehicle experiment on the parking problem in narrow garage.

WeA15	HG D7.2
Wind Turbines (Regular Session)
Chair: Alexandridis, Antonios	Univ. of Patras
Co-Chair: Martinez, John-Jairo	Gipsa-Lab. INP-Grenoble
09:50-10:10	WeA15.1
Optimal Flux Magnitude Tracking with Application to Fault-Tolerant Control of Wind Turbine Generators, pp. 466-471
Lesic, Vinko	Univ. of Zagreb, Faculty of Electrical Engineering and Comp
Vašak, Mario	Univ. of Zagreb
Perić, Nedjeljko	Univ. of Zagreb
Joksimović, Gojko	Univ. of Montenegro, Faculty of Electrical Engineering
Wolbank, Thomas	Vienna Univ. of Tech. Faculty of Electrical Engineeri
This paper concerns with wind turbine generator fault-tolerant control that avoids system shut-down and enables safe operation with less than nominal power production. We focus on generator stator isolation inter-turn fault that can be characterized before triggering the safety device and applied for any type of generator used in wind turbines. A low complexity model predictive tracking controller is proposed to achieve very accurate flux modulation that prevents the fault propagation while power delivery under fault is deteriorated as less as possible compared to healthy machine conditions. Presented fault-tolerant control strategy is developed taking into account its modular implementation and installation in available control systems of existing wind turbines to extend their life cycle and energy production. Simulation results for the case of a 700 kW wind turbine are presented.
10:10-10:30	WeA15.2
Experimental Validation of Wind Turbine Higher Harmonic Control Using Shaft Loads Measurements, pp. 472-477
Petrović, Vlaho	Univ. of Zagreb
Campagnolo, Filippo	Pol. di Milano
In order to achieve reductions in wind turbine structural loads and to enable further increase of wind turbine rated power and dimensions, individual pitch control algorithms received significant portion of attention in the scientific community in the recent years. Individual pitch control typically reduces only the first loads harmonic, but as the wind turbines grow in size, it is expected that higher harmonics will have to be reduced as well. Instead of using blade loads, which is typical approach in the literature, this paper analyses the possibilities of reduction of higher harmonics from the wind turbine loads using the shaft measurements. To this aim, loads transformations capable of extracting information about higher loads harmonics from shaft measurements are derived. The controller for load reduction based on such transformations is implemented on a scaled wind turbine model and experimentally validated in the wind tunnel at Politecnico di Milano.
10:30-10:50	WeA15.3
Observer-Based Maximum Power Tracking in Wind Turbines with Only Generator Speed Measurement, pp. 478-483
Sandoval-Moreno, John	GIPSA Lab. Grenoble-INP
Besancon, Gildas	Ense3 - Grenoble INP
Martinez, John-Jairo	Gipsa-Lab. INP-Grenoble
In this paper, it is proposed a technique to optimize the power generated by a small scale wind turbine, by having no wind speed measurement and uncertain knowledge of the power coefficient curve characteristic. First, a Kalman-like observer is used for estimating the power coefficient characteristic, which is then used in a nonlinear control strategy for the electrical generator that secures power maximization. Then, the uncertain power coefficient function is adjusted with a Recursive Least-Square algorithm that uses the estimate of the power coefficient as reliable information. Finally, from the adjusted power coefficient polynomial is computed the optimal electrical generator speed reference that leads the system to operate near the theoretical optimal power. This methodology is illustrated on a simulated example and compared with a case when the wind speed is measured.
10:50-11:10	WeA15.4
Nonlinear Stability Analysis of DFIG Wind Generators in Voltage Oriented Control Operation, pp. 484-489
Bourdoulis, Michael	Univ. of Patras
Alexandridis, Antonios	Univ. of Patras
In this paper the damping and stability properties of Doubly-Fed Induction Generator (DFIG) wind systems are studied. First the complete 5th order nonlinear system of a DFIG in the synchronously rotating dq reference frame is introduced. Assuming operation under the grid voltage oriented control mode, it is possible to control reactive and active power separately through the d and q controlled voltage inputs, respectively. After that, a simple cascaded controller design is adopted for each input, comprising of proportional inner-loop and proportional-integral outer-loop controllers. A passivity analysis is conducted for both the open- and closed-loop systems with the dynamics of the proportional inner-loop controllers taken into account. The analysis shows that the damping of the system can be modulated by the inner-loop controllers while it provides an appropriate storage function for the nonlinear stability analysis. This constitutes the main contribution of the paper since it is the first time that stability and convergence to the equilibrium is proven for the complete machine model in the face of unknown torque inputs, without needing any estimation or adaptive mechanism. Finally, the analysis and the performance of the closed-loop DFIG wind system are verified through extensive simulation results.
11:10-11:30	WeA15.5
An MPC Approach to Individual Pitch Control of Wind Turbines Using Uncertain LIDAR Measurements, pp. 490-495
Mirzaei, Mahmood	Tech. Univ. of Denmark
Soltani, Mohsen	Aalborg Univ.
Poulsen, Niels Kjølstad	Tech. Univ. of Denmark
Niemann, Henrik	Tech. Univ. of Denmark
Spatial distribution of the wind field exerts unbalanced loads on wind turbine structures and it is shown these loads could be mitigated by controlling each blade's angle individually (individual pitch control). In this work the problem of individual pitch control of a variable-speed variable pitch wind turbine in the full load region is considered. Model predictive control (MPC) is used to solve the problem. A new approach is proposed to simplify the optimization problem of MPC. We linearize the obtained nonlinear model for different operating points which are determined by the effective wind speed on the rotor disc and take the wind speed as a scheduling variable. The wind speed is measurable ahead of the turbine using LIDARs, therefore the scheduling variable is known for the entire prediction horizon. We consider uncertainty in the wind propagation, which is the traveling time of wind from the LIDAR measurement point to the rotor. An algorithm based on wind speed estimation and measurements from the LIDAR is devised to find an estimate of the delay and compensate for it before it is used in the controller. Comparisons between the MPC with error compensation, without error compensation and a benchmark cyclic pitch PI controller are given. The results show that with appropriate signal processing techniques, LIDAR measurements improve the performance of the wind turbine controller.
11:30-11:50	WeA15.6
A Novel Online Controller Redesign Approach to Fault Accommodation in Wind Turbine Systems, pp. 496-501
Jain, Tushar	Aalto Univ.
Yamé, Joseph-Julien	Univ. Henri Poincaré
Sauter, Dominique	Lorraine Univ.
Aberkane, Samir	UHP, NANCY 1
This paper presents a real-time mechanism to tolerate faults occurring in a Wind Turbine (WT) system. This system is a FAST coded simulator designed by the U.S. National Renewable Energy Laboratory. The demonstrated mechanism lies under the taxonomy of Active Fault-tolerant Control (FTC) systems, namely the online redesign based approach. In the proposed approach, we do not use any a priori information about the model of the turbine in real-time. In fact, we use online measurements generated by the WT. Based on the given control specifications, and the observed measurement an occurred fault is accommodated by redesigning the controller online such that the WT generates rated power even under faulty conditions. Secondly, no explicit fault diagnosis (FD) module is used in this approach. As a result, issues of model uncertainty, false alarms, etc. associated with an integrated FD and controller reconfiguration approach to FTC systems are not experienced here.

WeST1	HG F1, F3
Semi-Plenary 1: Power Grid Stability Despite Renewable Instability (Semi-Plenary Session)
Chair: Kowalewski, Stefan	RWTH Aachen Univ.
12:00-12:45	WeST1.1
Power Grid Stability Despite Renewable Instability*
Hermanns, Holger	Saarland Univ.

WeST2	HG F7, F5
Semi-Plenary 2: The Scenario Approach to Stochastic Optimization (Semi-Plenary Session)
Chair: Goulart, Paul J.	ETH Zurich
12:00-12:45	WeST2.1
The Scenario Approach to Stochastic Optimization*
Campi, M. C.	Univ. di Brescia

WeB1	HG F1
Theory and Practice of Explicit MPC (Tutorial Session)
Chair: Jones, Colin N	EPFL, Lausanne
Co-Chair: Kvasnica, Michal	Slovak Univ. of Tech. in Bratislava
Organizer: Jones, Colin N	EPFL, Lausanne
Organizer: Kvasnica, Michal	Slovak Univ. of Tech. in Bratislava
Organizer: Herceg, Martin	Swiss Federal Inst. of Tech. - ETH Zurich
14:00-14:20	WeB1.1
Explicit Model Predictive Control: Basics, Fast Implementations, Advantages and Limitations (I)*
Bemporad, Alberto	IMT Inst. for Advanced Studies Lucca
Model Predictive Control (MPC) has been used in the process industries for more than thirty years, due to its ability to control multivariable systems under restrictions on input and output variables in an optimized way. MPC requires the solution of a Quadratic Program (QP) on line to compute the control action, so that MPC has been labelled traditionally as a technology for slow processes. Significant advances were done in the last decade to push MPC to fast-sampling applications and/or under limited CPU and memory resources. In particular, explicit model predictive control completely removes the need for on-line solvers by precomputing the control law off line as an explicit function of the state and reference vectors, so that on-line operations reduce to a simple function evaluation. Such a function is piecewise affine in most cases, so that the MPC controller maps into a lookup table of linear gains, a form that is extremely easy to code, requires only basic arithmetic operations, and requires a number of iterations that can be exactly bounded a priori. In this talk, I will introduce the main concepts, techniques, and properties of explicit MPC. I will also show some recent advances in fast implementation of linear MPC controllers based on the use of canonical piecewise affine (PWA) function approximations, that allow the implementation of MPC control laws on chip with sampling times in the order of tens of nanoseconds. I will conclude my talk with some practical considerations about the advantages and limitations of explicit MPC compared to latest results in MPC based on embedded QP solvers.
14:20-14:40	WeB1.2
General Purpose Computational Tool for Explicit MPC: MPT3.0 (I), pp. 502-510
Herceg, Martin	Swiss Federal Inst. of Tech. - ETH Zurich
Kvasnica, Michal	Slovak Univ. of Tech. in Bratislava
Jones, Colin N	EPFL, Lausanne
Morari, Manfred	ETH Zurich
The Multiparametric Toolbox (MPT) is a collection of algorithms for the modeling, control, analysis and deployment of constrained optimization-based controllers developed under Matlab. Since its' release in 2003, it has been widely used for the design and deployment of explicit MPC controller both in academic and industrial projects. MPT features a powerful geometric library that extends the application of the toolbox beyond optimal control to various problems arising in computational geometry. This talk coincides with the release of version 3.0 of the toolbox, which is a complete rewrite of the original with a more flexible structure that offers faster integration of new algorithms. The numerical side of the toolbox has been improved by adding interfaces to state of the art solvers and by basing all computational geometry primitives on a new parametric linear complementarity solver. The toolbox provides end-to-end algorithms for the design, simulation and deployment of high-speed explicit and optimization-based MPC controllers for high-speed applications. The talk will provide an overview of the new features of the toolbox, as well as a number of examples demonstrating its efficacy.
14:40-15:00	WeB1.3
Complexity Reduction in Explicit MPC: From Theory to Implementation (I)*
Kvasnica, Michal	Slovak Univ. of Tech. in Bratislava
Holaza, Juraj	Slovak Univ. of Tech. in Bratislava
Takács, Bálint	Slovak Univ. of Tech. in Bratislava
Although explicit MPC allows the implementation of optimization-based controllers on embedded hardware, its applicability is limited to small- to moderate-sized problems due to the high complexity of explicit solutions. This talk reviews several approaches to the reduction of complexity and shows how to easily implement them using the Multi Parametric Toolbox 3.0. Discussed algorithms are divided into two groups: The first class of methods approximates the explicit MPC feedback law by a simpler feedback function while sacrificing optimality for a significant reduction of implementation cost. The second group of approaches then reduces the on-line implementation effort while maintaining all properties of a complex explicit MPC feedback. The main advantages as well as limitations of each method will be discussed and the efficacy will be demonstrated on motivating real-world examples.
15:00-15:20	WeB1.4
Approximate Nonlinear Explicit MPC Based on Reachability Analysis (I)*
Raimondo, Davide Martino	Univ. of Pavia
Schulze Darup, Moritz	Ruhr-Univ. Bochum
Mönnigmann, Martin	Ruhr-Univ. Bochum
We summarize recent progress in the computation of approximate explicit solutions to constrained optimal control problems for discrete-time nonlinear systems. We focus on methods that are not tailored for a particular subclass of nonlinear system (such as input affine or bilinear systems), but apply to a broad class of sufficiently smooth finite-dimensional systems. Moreover, we only address methods that are rigorous in the sense that they provide guarantees in terms of constraint satisfaction and stability. The approaches treated here can be briey described as follows: The state space around a target set is partitioned, and a control law is constructed on each region of the partition. Partitions and inputs are chosen in order to maximize the size of the capture basin (region of attraction). The obtained regions can be further refined to better approximate the optimal solution of the control problem. The resulting control law is built on a grid hierarchy that is fast to evaluate in real-time systems. Feasibility and stability of the system in closed-loop with the approximate control law are verified using set-based techniques such as interval or affine arithmetic, DC programming, or similar methods from the field of reliable computing. We discuss some recent developments with a focus on the trade-offs between computational complexity of the offline construction of the explicit control law, its level of suboptimality, and the computational complexity of its representation and online evaluation.
15:20-15:40	WeB1.5
High-Speed High-Performance Model Predictive Control of Power Electronics Systems (I)*
Mariethoz, Sebastien	ETH Zurich
This tutorial will describe the typical problem classes that arise in the control of switched electronics systems, which are characterised by the hybrid nature of their dynamics. It describes the associated model predictive control classes and discusses how to formulate an optimisation problem tractable in real-time. The tractability issue is particularly important due to the switching and sampling frequencies that typically range from a few hundred Hertz to a few megahertz. The tutorial will also develop problem formulations that optimise the system switched behaviour in order to tackle objectives such as minimising the harmonic distortion or system losses and it will be illustrated by some selected applications.
15:40-16:00	WeB1.6
Explicit MPC Applications in the Mechatronics Industry: Technology Transfer Potential and Current Limitations (I)*
Di Cairano, Stefano	Mitsubishi Electric Res. Lab.
In this talk we discuss the impact, the potential, and the limitations of explicit model predictive controllers in industrial applications, with particular emphasis on automotive, aerospace, and precision manufacturing. We discuss several real world applications, including vehicle stability, HEV energy management, power train control, spacecraft rendezvous, and precision actuators, where different types of explicit MPC (linear, switched, hybrid) are implemented. We show how explicit solutions made possible rapid prototyping and testing the MPC controllers, with minimum effort despite stringent platform limitations. Motivated by these examples, we highlight the features of explicit MPC that make it viable and appealing for real-world implementation, such as the simple structure, the basic operations executed in real time, the bounds on the computations, the simple verification procedures. Finally, we discuss the explicit MPC limitations in these domains highlighting the research problems that still needs to be tackled in order to achieve widespread usage of these controllers, such as numerical robustness, adaptation, memory efficiency, and we discuss some of the strategies that we are pursuing to address these.

WeB2	HG F3
Model Predictive Control II (Regular Session)
Chair: Maciejowski, Jan M.	Univ. of Cambridge
Co-Chair: Falugi, Paola	Imperial Coll.
14:00-14:20	WeB2.1
Stabilizing Linear Model Predictive Control: On the Enlargement of the Terminal Set, pp. 511-517
Brunner, Florian David	Univ. of Stuttgart
Lazar, Mircea	Eidnhoven Univ. of Tech.
Allgower, Frank	Univ. of Stuttgart
It is well known that a large terminal set leads to a large region where the MPC problem is feasible without the need for a long prediction horizon. This paper proposes a new method for the enlargement of the terminal set. Contrary to existing approaches, the method uses the convex hull of a trajectory as the basis for the construction. This trajectory may be any feasible trajectory of the system terminating in an invariant set around the origin and is not restricted to consist only of equilibrium points. The method is applied in an MPC scheme where the terminal set is calculated at initialization.
14:20-14:40	WeB2.2
Model Predictive Control for Tracking Random References, pp. 518-523
Falugi, Paola	Imperial Coll.
Mayne, David Q.	Imperial Coll. London
This paper proposes a simple model predictive control scheme for linear systems, tracking a random reference and analysis its performance. In such situations it is usual to assume that the reference eventually converges to a constant in which case convergence to zero of the tracking error can be established. In this note we characterize the set to which the tracking error converges and the associated region of convergence when the reference does not converge to a constant.
14:40-15:00	WeB2.3
Stabilising Terminal Cost and Terminal Controller for $ell_{asso}$-MPC: Enhanced Optimality and Region of Attraction, pp. 524-529
Gallieri, Marco	Univ. of Cambridge
Maciejowski, Jan M.	Univ. of Cambridge
In recent literature, l1-regularised MPC, or lasso-MPC, has been recommended for control tasks involving complex requirements on the control signals, for instance, the simultaneous solution of regulation and sharp control allocation for redundantly-actuated systems. This is due to the implicit thresholding ability of LASSO regression. In this paper, a stabilising terminal cost featuring a mixed l1/l2 penalty is presented. Then, a candidate terminal controller is computed, with the aim of enlarging the region of attraction.
15:00-15:20	WeB2.4
Evaluation of Constrained Multivariable EPSAC Predictive Control Methodologies, pp. 530-535
Medina Sanchez, Robinson	Univ. de Ibague
Hernandez, Andres	Ghent Univ.
Ionescu, Clara	Ghent Univ.
De Keyser, Robin M.C.	Univ. of Gent
This paper investigates the performance of two Multivariable Model Predictive Control (MPC) strategies: selfish and solidary. These strategies are based on the main ideas developed in the EPSAC (Extended Prediction Self-Adaptive Control) approach to MPC. A two degree of freedom (2DOF) helicopter simulation has been chosen to illustrate these concepts, as it represents a complicated and challenging problem due to strong intercoupling effects, non-linear dynamics and uncertainties in the system model. The performance obtained with a Linear Quadratic Regulator (LQR) is also included as a reference to the performance of the multivariable MPC strategies. In this contribution, the performance of these multivariable MPC strategies is analyzed, providing more insight about the behavior of these controllers.
15:20-15:40	WeB2.5
Explicit MPC of Higher-Order Linear Processes Via Combinatorial Multi-Parametric Quadratic Programming, pp. 536-541
Feller, Christian	Univ. of Stuttgart
Johansen, Tor Arne	Norweigian Univ. of Sci. & Tech.
Parameter-dependent constrained optimization problems like they occur in the context of model predictive control (MPC) can be solved explicitly by means of multi-parametric quadratic programming (mpQP) techniques. We present a complexity analysis for a recently proposed combinatorial mpQP algorithm and discuss its advantages over existing geometric approaches concerning off-line explicit MPC computations for higher-order linear systems. The results are accompanied by numerical benchmark results for two suitable example problems from the area of process control.
15:40-16:00	WeB2.6
Robust Model Predictive Control of Uncertain Linear Systems with Persistent Disturbances and Input Constraints, pp. 542-547
Yang, Weilin	City Univ. of Hong Kong
Feng, Gang	City Univ. of Hong Kong
Zhang, TieJun	Masdar Inst. of Science and Tech.
This paper presents computationally attractive robust model predictive control approaches for the control of discrete-time linear systems with input constraints, structured parameter uncertainties and persistent disturbances. In order to ensure robust stability of constrained uncertain systems, constructive methods are proposed to compute robust positively invariant sets for stabilizing predictive controller. The proposed robust predictive control (RMPC) systems satisfy both recursive feasibility and input-to-state stability. In the controller design, the 0-step predictive controller with a simple structure is proposed. In order to deal with the RMPC problem with a fixed terminal set, the result is extended to the N-step predictive controller. Simulations results have demonstrated the efficacy of the proposed predictive control approaches.

WeB3	HG F5
Distributed Parameter Systems II (Regular Session)
Chair: Aschemann, Harald	Univ. of Rostock
Co-Chair: Schöberl, Markus	Johannes Kepler Univ. Linz
14:00-14:20	WeB3.1
Analysis and Comparison of Port-Hamiltonian Formulations for Field Theories - Demonstrated by Means of the Mindlin Plate, pp. 548-553
Schöberl, Markus	Johannes Kepler Univ. Linz
Siuka, Andreas	Johannes Kepler Univ. Linz
This paper focuses on the port-Hamiltonian formulation of systems described by partial differential equations. Based on a variational principle we derive the equations of motion as well as the boundary conditions in the well-known Lagrangian framework. Then it is of interest to reformulate the equations of motion in a port-Hamiltonian setting, where we compare the approach based on Stokes-Dirac structures to a Hamiltonian setting that makes use of the involved bundle structure similar to the one on which the variational approach is based. We will use the Mindlin plate, a distributed parameter system with spatial domain of dimension two, as a running example.
14:20-14:40	WeB3.2
Stability Analysis for Structured Feedback Interconnections of Distributed-Parameter Systems and Time-Varying Uncertainties, pp. 554-559
Kao, Chung-Yao	National Sun Yat-Sen Univ.
Cantoni, Michael	Univ. of Melbourne
This paper considers stability analysis of systems with constantly-proper Callier-Desoer distributed-parameter transfer functions, interconnected via feedback channels with time-varying parameters and bandwidth-limited time-varying delays. A point-wise in frequency condition that implies stability is established by applying a robust stability result based on integral-quadratic-constraints and a linear time-varying generalization of the $nu$-gap metric. The stability condition is applied in an example to illustrate the proposed analysis method.
14:40-15:00	WeB3.3
Solvability of the Output Regulation Problem with a Feedforward Controller, pp. 560-565
Laakkonen, Petteri	Tampere Univ. of Tech.
Pohjolainen, Seppo	Tampere Univ. of Tech.
In this paper the solvability of the output regulation problem with an infinite-dimensional exosystem by using a linear feedforward controller is considered. New sum conditions that are necessary and sufficient for the solvability are found. In addition, the required smoothness properties of the reference signals are discussed in detail.
15:00-15:20	WeB3.4
Boundary Port Hamiltonian Control of a Class of Nanotweezers, pp. 566-571
Ramirez, Hector	Ec. Nationale Supérieure de Mécanique et de Microtechniques (E
Le Gorrec, Yann	FEMTO-ST
Boundary controlled-port Hamiltonian systems have proven to be of great use for the analysis and control of a large class of systems described by partial differential equations. The use of semi-group theory, combined with the underlying physics of Hamiltonian systems permits to prove existence, well-possessedness and stability of solutions using constructive techniques. On other hand, the differential geometric representation of these systems has lead to finite dimension approximation methods that conserves physical properties such as the interconnection structure and the energy. These results are applied to the modelling and control of a class of nanotweezers used for DNA-manipulation. The Nanotweezer may be modelled as a flexible beam interconnected with a finite dimensional dynamical system representing the manipulated object. A boundary controlled-port Hamiltonian model for the ensemble and an exponentially stabilizing controller are proposed. A geometric approximation scheme is used to reduce the infinite dimensional system and numerical simulations of the closed-loop system presented.
15:20-15:40	WeB3.5
The Method of Integro-Diffferential Relations for Control of Spatially Two-Dimensional Heat Transfer Processes, pp. 572-577
Rauh, Andreas	Univ. of Rostock
Dittrich, Christina	Univ. of Rostock
Aschemann, Harald	Univ. of Rostock
The design of control strategies for distributed parameter systems is an important field of current research. To design control laws and state estimation procedures for this class of systems, it is essential to find approximations that represent the system dynamics with good accuracy and simultaneously allow for an evaluation in real time. Moreover, possibilities for the quantification of approximation errors are useful to determine reliable finite-dimensional models which are sufficiently accurate for the control task at hand. Approximation errors result from the replacement of the original system model that is given in terms of partial differential equations by a finite-dimensional system representation. In this paper, the method of integro-differential relations is used for the derivation of a finite-dimensional approximation of a spatially two-dimensional heat transfer process. Simulation results for control and state estimation employing the before-mentioned modeling approach are presented for a test rig that is available at the Chair of Mechatronics at the University of Rostock.
15:40-16:00	WeB3.6
Robustness Properties of Controllers with Reduced Order Internal Models, pp. 578-583
Paunonen, Lassi	Tampere Univ. of Tech.
Pohjolainen, Seppo	Tampere Univ. of Tech.
In this paper we study robust output regulation for distributed parameter systems. In particular we are interested in the internal model principle, which can be used in characterizing controllers that achieve robust output tracking and disturbance rejection for a linear system. We show that if we do not require robustness with respect to arbitrary perturbations in the parameters of the systems, then there may exist robust controllers that do not contain a full-sized internal model of the exosystem's dynamics. Our approach also establishes a convenient way of testing the robustness of a controller with respect to given perturbations. The theoretic results are applied to analyzing the robustness properties of controllers for a system of two shock absorber models, and for a one-dimensional heat equation.

WeB4	HG F7
Stochastic Systems II (Regular Session)
Chair: Olshevsky, Alexander	Univ. of Illinois at Urbana-Champaign
Co-Chair: Balakrishnan, Hamsa	Massachusetts Inst. of Tech.
14:00-14:20	WeB4.1
Approximate Dynamic Programming for Stochastic Reachability, pp. 584-589
Kariotoglou, Nikolaos	ETH Zurich
Summers, Sean	ETH Zurich
Summers, Tyler H.	ETH Zurich
Kamgarpour, Maryam	ETH Zurich
Lygeros, John	ETH Zurich
In this work we illustrate how approximate dynamic programing can be utilized to address problems of stochastic reachability in infinite state and control spaces. In particular we focus on the reach-avoid problem and approximate the value function on a linear combination of radial basis functions. In this way we get significant computational advantages with which we obtain tractable solutions to problems that cannot be solved via generic space gridding due to the curse of dimensionality. Numerical simulations indicate that control policies coming as a result of approximating the value function of stochastic reachability problems achieve close to optimal performance.
14:20-14:40	WeB4.2
Optimal H2 Mode-Independent Filter for Generalized Bernoulli Jump Systems, pp. 590-595
Gonçalves, Alim P. C.	UNICAMP
Fioravanti, Andre Ricardo	INRIA Roquencourt
Geromel, Jose C.	UNICAMP
In this paper we work with the filtering problem for a special class of discrete-time Markov Jump Linear Systems (MJLS) whose transition probability matrix has identical rows. For that class of systems we design, with the help of new necessary and sufficient LMI conditions, H2 optimal mode-independent filters with the same order of the plant. For cluster availability of the mode, we also show it is possible to design optimal filters. We complete the results with a numerical example.
14:40-15:00	WeB4.3
Combinatorial Bounds and Scaling Laws for Noise Amplification in Networks, pp. 596-601
Jadbabaie, Ali	Univ. of Pennsylvania
Olshevsky, Alexander	Univ. of Illinois at Urbana-Champaign
Motivated by the problem of designing and analyzing clock synchronization protocols for large-scale networks, we provide combinatorial bounds on the mean square disagreement of synchronization dynamics subject to additive noise. While previous work has used eigenvalues to analyze the mean square disagreement when the synchronization dynamics is governed by a normal matrix, in the non-normal case eigenvalues cease to be an adequate measure. We first show that the so-called 2-norm coefficient of ergodicity used in the study of inhomogeneous Markov chains can be used to bound mean square disagreement even when the underlying dynamics is governed by a non-normal matrix. Our main result then demonstrates that the 2-norm coefficient of ergodicity has a natural combinatorial interpretation as a combined measure of matrix non-normality and graph connectivity. We apply this result to yield new performance guarantees for distributed clock synchronization: we show that in a simple (possibly irregular) random graph model, the natural clock synchronization scheme wherein neighboring nodes average their clocks is order-optimal: it drives the expected clock offsets between any two nodes to at most a constant, independently of the total number of nodes.
15:00-15:20	WeB4.4
Stochastic Localization of Sources with Convergence Guarantees, pp. 602-607
Huck, Stephan Marc	ETH Zurich
Lygeros, John	ETH Zurich
We establish convergence guarantees for a recently proposed Markov Chain Monte Carlo (MCMC) method to locate source(s) of a certain concentration field. Our method utilizes a Markovian controller to control the motion of autonomous vehicles on a compact search domain. The distribution of the resulting discrete-time Markov chain is used to estimate the locations of the sources. To guarantee the correctness of the localization, we prove that the existing invariant measure for the Markov chain is unique. The chain is shown to be uniform ergodic and will converge to its stationary distribution. The theoretically derived convergence rate is compared to results from numerical simulations.
15:20-15:40	WeB4.5
Optimal Control of Airport Operations with Gate Capacity Constraints, pp. 608-613
Khadilkar, Harshad	Massachusetts Inst. of Tech.
Balakrishnan, Hamsa	Massachusetts Inst. of Tech.
The mitigation of airport surface congestion is an important step towards increasing the efficiency of the air transportation system, and decreasing flight delays. This paper proposes a strategy to control the release of departing flights from their gates with the specific objective of reducing their taxi times and fuel consumption, while limiting the impact on airport throughput. The proposed strategy also explicitly accounts for the practical constraints that arise due to limited gate resources at the airport. A stochastic network abstraction of the airport surface is used to model aircraft movement, and the optimal release time for each aircraft is calculated using dynamic programming. Simulations of operations at Boston's Logan International Airport in the US are used to illustrate the effects of the resultant policies.
15:40-16:00	WeB4.6
Stability of Dynamic Traveling Repairman Problem under Polling-Sequencing Policies, pp. 614-619
Huang, Jiangchuan	Univ. of California, Berkeley
Sengupta, Raja	Univ. of California at Berkeley
We establish a necessary and sufficient condition for stability in the dynamic traveling repairman problem (DTRP) [3] under the class of polling-sequencing (P-S) policies satisfying unlimited-polling and economy of scale. The P-S class includes some of the policies proven to be optimal for the expectation of system time under light and heavy loads in the DTRP literature. The number of tasks inside each polling partition is shown to be a Markov chain. Policies such as first come first serve, traveling salesman policy, nearest neighbor and Daganzo's algorithm are shown to have economy of scale.

WeB5	HG E1.1
Event Based Detection and Control (Regular Session)
Chair: Dormido, Sebastián	UNED
Co-Chair: Leva, Alberto	Pol. di Milano
14:00-14:20	WeB5.1
Feedback-Based Memory Management with Active Swap-In, pp. 620-625
Terraneo, Federico	Pol. di Milano
Leva, Alberto	Pol. di Milano
Memory and swap management is an important issue in operating systems, owing to the large difference between RAM and disk access times. This paper presents a memory management layer, designed along a fully control-theoretical approach, that decouples the swap-out and swap-in activity from the events triggered by the applications' memory access patterns. A system endowed with this layer can manage memory on a per-process basis, while at the same time avoiding the presence of swapped-out pages if RAM memory is available. Such a tunable active swap-in mechanism thus inherently avoids temporary RAM over-utilisations to slow down the system for a period significantly longer than their duration. Simulation results prove the effectiveness of the proposal, as well as the viability of its integration with existing memory managers.
14:20-14:40	WeB5.2
Experimental Study of Two Event-Based PI Controllers in a Solar Distributed Collector Field, pp. 626-631
Chacón Sombría, Jesús	UNED
Sánchez Moreno, José	UNED
Yebra, Luis José	CIEMAT-Plataforma Solar de Almería
Visioli, Antonio	Univ. of Brescia
Dormido, Sebastián	UNED
In this work we focus on the study of limit cycles that appear in a control scheme which is based on the use of a PI controller with an event-based send-on-delta sampling (SOD), which can be used to sample either the process variable or the control variable. We use an algorithm to calculate the limit cycles properties and then we compare the results obtained in simulations with experiments performed on a real plant, a distributed solar collector field at the Solar Platform of Almería (PSA, Spain). The process is identified as a first order plus time delay (FOPTD). Practical issues, such as the implementation of an anti-windup mechanism and a bumpless transfer between the manual and automatic mode are also addressed.
14:40-15:00	WeB5.3
Non-Parametric Analysis of Eye-Tracking Data by Anomaly Detection, pp. 632-637
Jansson, Daniel	Uppsala Univ.
Rosén, Olov	Uppsala Univ.
Medvedev, Alexander V.	Uppsala Univ.
A non-parametric approach for distinguishing between individuals by means of recorded eye movements is suggested. The method is based on the principles of stochastic anomaly detection and relies on measured data for probability distribution estimation and evaluation. For visual stimuli that excite the essential nonlinear dynamics of the oculomotor system, mean gaze trajectories and characterizations of their uncertainty are approximated per individual using eye-tracking data. With this information, eye-tracking profiles are established, against which independently acquired data sets are statistically tested to evaluate the probability that they belong to said profiles. Both Gaussian function fitting and kernel density estimation (KDE) techniques are used for distribution estimation and novel means for testing observations against general distributions are suggested. It is shown that the presented method yields promising results in terms of individual classification based on eye movements and that using the KDE method for trajectory distribution estimation is favorable compared to normal distribution fitting.
15:00-15:20	WeB5.4
Self-Triggered Model Predictive Control for Nonholonomic Systems, pp. 638-643
Eqtami, Alina	National Tech. Univ. of Athens
Heshmati-alamdari, Shahab	National Tech. Univ. of Athens
Dimarogonas, Dimos V.	Royal Inst. of Tech. (KTH)
Kyriakopoulos, Kostas J.	National Tech. Univ. of Athens
This paper proposes a Model Predictive Control (MPC) framework combined with a self-triggering mechanism for constrained uncertain systems. Under the proposed scheme, the control input as well as the next control update time are provided at each triggering instant. Between two consecutive triggering instants, the control trajectory given by the MPC is applied to the plant in an open-loop fashion. This results to less frequent computations while preserving stability and convergence of the closed-loop system. A scenario for the stabilization of a nonholonomic robot subject to constraints and disturbances is considered, with the aim of reaching a specific triggering mechanism. The robot under the proposed control framework is driven to a compact set where it is ultimately bounded. The efficiency of the proposed approach is illustrated through a simulated example.
15:20-15:40	WeB5.5
A Stopping Rule for Simultaneous Perturbation Stochastic Approximation, pp. 644-649
Wada, Takayuki	OsakaUniversity
Fujisaki, Yasumasa	Osaka Univ.
A stopping rule is developed for simultaneous perturbation stochastic approximation (SPSA) which is an iterative method for minimizing an unknown objective function via its noise corrupted measurements. It is shown that, when the number of iterations reaches a constant determined by the stopping rule, SPSA for the quadratic convex problem provides us with a suboptimal solution which is close to the optimal solution with a specified probabilistic guarantee. Furthermore, the number is determined by the specified guarantee and polynomial in the problem size.
15:40-16:00	WeB5.6
An Integrated Systems Engineering Framework for Supervisor Synthesis, Verification, and Performance Evaluation, pp. 650-657
Markovski, Jasen	Eindhoven Univ. of Tech.
We propose a model-based systems engineering framework that enables supervisor synthesis of stochastic (nondeterministic) discrete-event systems, and post-synthesis validation of functional and quantitative properties of the supervised system. Supervisory control theory deals with synthesis of models of supervisory controllers that ensure safe and nonblocking behavior, based on discrete-event models of the uncontrolled system and the control requirements. Typically, neither more elaborated functional properties nor performance metrics can be guaranteed by the synthesis procedure for large systems, due to high computational complexity. Thus, the supervised system must be validated to ensure that intended behavior is present. The framework employs a single integrated model that denotes all relevant aspects of the system. We rely on state-of-the-art tools to implement the proposed framework. For supervisor synthesis we employ Supremica, which models we extend to denote stochastic behavior. For verification, we provide a consistent translation of the supervised system to the model checker UPPAAL. To evaluate quantitative supervised behavior, first we transform the denoted system model to a labeled Interactive Markov chain and couple it with the synthesized supervisor. Then, we derive the underlying labeled Markov process and feed it to the stochastic model checker MRMC. We illustrate the framework on an industrial case study of coordinating printing maintenance procedures.

WeB6	HG E1.2
Nonlinear System Theory II (Regular Session)
Chair: Henrion, Didier	Czech Tech. Univ. in Prague
Co-Chair: Kablar, Natasa A.	Louisiana State Univ.
14:00-14:20	WeB6.1
Control Polynomial Dynamical Systems on Rectangles, pp. 658-663
Ben Sassi, Mohamed Amin	Lab. Jean Kuntzmann, Univ. de Grenoble
Girard, Antoine	Univ. Joseph Fourier
In this paper we focus on a particular class of nonlinear dynamical systems given by polynomial vector fields in rectangular domains (boxes). This is a generalization of the work of Belta and Habets dealing with multi-affine dynamical systems on rectangles. The main idea is to use the blossoming principle which allows us to relate our polynomial dynamical system to a multi-affine one. This technique allows us to establish sufficient conditions for invariance of a rectangle or exit of a rectangle through a given facet. We extend these results to handle control synthesis. Finally, we show how our approach can be used to solve motion planning problem.
14:20-14:40	WeB6.2
A New Approach to the Solution of Free Rigid Body Motion for Attitude Maneuvers, pp. 664-669
Pagnozzi, Daniele	Strathclyde Univ.
Maclean, Craig	Strathclyde Univ.
Biggs, James Douglas	Univ. of Strathclyde
A Hamiltonian formulation of free rigid body motion defined on the Special Unitary Group SU(2) is used to integrate the system to obtain a convenient quaternion representation for attitude engineering applications. Novel content of this paper concerns applying a modern approach, based on geometric control theory, to obtain the kinematic solution in an elegant and compact form. Moreover, this integration leads to an attitude representation which is not Euler-angle-like, thus enhancing its applicability (e.g. to attitude motion design).
14:40-15:00	WeB6.3
On the Analysis of the Bifurcation Sets of Equilibrium Points in Parameter Space, pp. 670-675
Chesi, Graziano	Univ. of Hong Kong
Tanaka, Gouhei	The Univ. of Tokyo
Hirata, Yoshito	The Univ. of Tokyo
Aihara, Kazuyuki	The Univ. of Tokyo
This paper addresses the problems of characterizing and estimating the bifurcation sets of equilibrium points in multi-parameter space of a class of nonlinear dynamical systems. Specifically, we investigate the sets of parameters that lead to saddle-node bifurcations and Hopf bifurcations at an equilibrium point of interest. First, a characterization of these sets is provided in terms of the zeros of some functions. Second, this characterization is exploited to estimate such sets through convex programming for the case of polynomial dynamical systems. In particular, two conditions are proposed for establishing whether a sublevel set of a given polynomial does not contain parameters that lead to bifurcations. By using these conditions, the largest of such sublevel sets can be estimated by solving an eigenvalue problem. Some numerical examples illustrate the proposed results.
15:00-15:20	WeB6.4
Convex Computation of the Region of Attraction of Polynomial Control Systems, pp. 676-681
Henrion, Didier	Czech Tech. Univ. in Prague
Korda, Milan	École Pol. Fédérale de Lausanne (EPFL)
We address the long-standing problem of computing the region of attraction (ROA) of a target set (typically a neighborhood of an equilibrium point) of a controlled nonlinear system with polynomial dynamics and semialgebraic state and input constraints. We show that the ROA can be computed by solving a convex linear programming (LP) problem over the space of measures. In turn, this problem can be solved approximately via a classical converging hierarchy of convex finite-dimensional linear matrix inequalities (LMIs). Our approach is genuinely primal in the sense that convexity of the problem of computing the ROA is an outcome of optimizing directly over system trajectories. The dual LP on nonnegative continuous functions (approximated by polynomial sum-of-squares) allows us to generate a hierarchy of semialgebraic outer approximations of the ROA at the price of solving a sequence of LMI problems with asymptotically vanishing conservatism. This sharply contrasts with the existing literature which follows an exclusively dual Lyapunov approach yielding either nonconvex bilinear matrix inequalities or conservative LMI conditions. The approach is simple and readily applicable as the outer approximations are the outcome of a single semidefinite program with no additional data required besides the problem description.
15:20-15:40	WeB6.5
Embedding the Generalized Acrobot into the N-Link with an Unactuated Cyclic Variable and Its Application to Walking Design, pp. 682-689
Celikovsky, Sergej	Acad. of Sci. of Czech Republic
Anderle, Milan	Czech Tech. Univ. in Prague
Moog, Claude	CNRS
The Acrobot is the well-known and widely studied underactuated mechanical system having two links and one actuated joint between them. It may be also viewed as the simplest possible walking like mechanism without knees and the ankle-joint actuation, alternatively also referred to as the underactuated Compass gait walker. To extend techniques used to control the Acrobot to a more general underactuated n-link having an unactuated cyclic variable, this paper defines the so-called generalized Acrobot. Further, it is shown that for every set of virtual constraints there exists a generalized Acrobot that is linearly embedded into this n-link. Based on this property and results valid for the Acrobot, walking strategies for the n-link are provided. Important achievement here is that the exponentially stable tracking during the swing phase only is possible, i.e. the stabilizing effect of the impact map is not needed. Computer simulations of the 4-link case are provided.
15:40-16:00	WeB6.6
Robust Control of Singular Systems with Time Delay. Part I: Continuous Time Case, pp. 690-695
Kablar, Natasa A.	Louisiana State Univ.
Kvrgic, Vladimir	Lola Inst.
Debeljkovic, Dragutin	Univ. of Belgrade, School of mechanical Engineering
In this paper for the class of continuous time nonlinear uncertain singular time delay dynamical systems we present robust stability analyze results. Next, we consider a control problem for nonlinear continuous time uncertain singular time delay dynamical systems involving a notion of optimality with respect to an auxiliary cost which guarantees a bound on the worst-case value of a nonlinear-nonquadratic cost criterion over a prescribed uncertainty set. Further we specialize result to affine uncertain systems to obtain controllers predicated on an inverse optimal control problem. In particular, to avoid the complexity in solving the steadystate Hamilton-Jacobi-Bellman equation we parameterize a family of stabilizing controllers that minimize some derived cost functional that provides flexibility in specifying the control law. The performance integrand is shown to explicitly depend on the continuous time nonlinear singular time delay system dynamics, the Lyapunov function of the closed-loop system, and the stabilizing feedback control law wherein the coupling is introduced via the Hamilton-Jacobi-Bellman equation. By varying the parameters in the Lyapunov function and the performance integrand, the proposed framework can be used to characterize a class of globally stabilizing controllers that can meet the closed-loop system response constraints. Obtained results for nonlinear case are further specialized to continuous time linear singular time delay dynamical systems.

WeB7	HG E3
Game Theoretic Methods (Regular Session)
Chair: De Schutter, Bart	Delft Univ. of Tech.
Co-Chair: Dragan, Vasile	Romanian Acad.
14:00-14:20	WeB7.1
Optimal Leader Functions for the Reverse Stackelberg Game: Splines and Basis Functions, pp. 696-701
Groot, Noortje	Delft Univ. of Tech.
De Schutter, Bart	Delft Univ. of Tech.
Hellendoorn, Hans	Delft Univ. of Tech.
In order to deal with the control of large-scale infrastructures, a multi-level approach may be required in which several groups of decision makers have different objectives. A game formulation can help to structure such a control task. The reverse Stackelberg game has a hierarchical structure in which the follower player acts subsequent to the leader's disclosure of her leader function, which maps the follower decision space into the leader decision space. The problem of finding a leader function such that the leader's objective function is optimized, given an optimal response with respect to the follower's objective function, is in general a difficult problem. So far, the set of optimal affine leader functions has been delineated. However, for the more general class of nonlinear leader functions, no structured solution approach exists yet. In this paper, we consider several nonlinear structures for a leader function based on basis functions as well as based on interpolating splines and we show how these approaches can be adopted to find an optimal leader function.
14:20-14:40	WeB7.2
Games of Network Disruption and Idempotent Algorithms, pp. 702-709
McEneaney, William	Univ. of California, San Diego
Desir, Antoine	Columbia Univ.
We consider a game on the space of network disruptions. An application in command and control is used as a guide for the development of the model. The outcome of any set of physical actions depends on the information available to the controller. We suppose that information flows along a network of humans and machines. The opposing player may act to intermittently block information flow along the network. This may be combined with physical actions such that our controller might be making decisions based on information that is not as current as would be possible without the induced delays. We find that the model of information delay dynamics is best captured as a controlled min-plus linear system. We also find that the minimax value function may be represented as a min-plus convex functional over the space of delay vectors. This is a max-min linear space. Backward dynamic programming propagation of the value function leads to (max-min) sum and product compositions. This yields a particularly nice solution algorithm. Computational and solution-representation complexity are examined.
14:40-15:00	WeB7.3
Approximate Solutions to a Class of Nonlinear Differential Games Using a Shared Dynamic Extension, pp. 710-715
Mylvaganam, Thulasi	Imperial Coll. London
Sassano, Mario	Univ. of Rome, Tor Vergata
Astolfi, Alessandro	Imperial Coll. London
A class of nonzero-sum differential games is considered and a dynamic state feedback control law that approximates the solution of the differential game is proposed. The control law relies upon the solution of algebraic equations in place of partial differential equations or inequalities and makes use of dynamics shared by the players, thus relaxing the conditions of a previously presented paper. The idea is firstly illustrated by the two-player case and then extended to the N-player case. A simple numerical example completes the paper.
15:00-15:20	WeB7.4
Game-Based Inner Approximation of Maximal Output Admissible Sets under References Unknown in Advance, pp. 716-723
Kogiso, Kiminao	Nara Inst. of Sci and Tech.
Kita, Koji	mixi, Inc.
Sugimoto, Kenji	Nara Inst. of Science and Tech.
This paper proposes a game-based inner approximation method that efficiently arranges multiple ellipsoids approximating a parameterization of maximal output admissible sets in terms of a reference vector. The approximation is contributive when considering a tracking control problem for constrained systems in which a reference value is not given in advance. Our idea for the approximation is to characterize search points of the ellipsoids in a state based game and a log-linear learning algorithm by dynamics of the constrained system. Consequently, an efficient arrangement of the ellipsoids can be achieved at a lower computation cost. A numerical example is finally illustrated to make sure of the consequence, compared to our previous works.
15:20-15:40	WeB7.5
Nash Equilibria in Normal Games Via Optimization Methods, pp. 724-729
Buttler, Jens	Company: DB Systel
Akchurina, Natalia	Control Theory and Robotics, TU Darmstadt
This paper is devoted to Nash equilibria of normal form games. We give a survey on computing Nash equilibria via optimization methods. Further on we prove, that the Nash equilibria of a game coincide with the zeros of a nonlinear, almost everywhere smooth system of equations with the same dimension as the strategy space. Thus, we can apply tools for solving nonlinear systems of equations. We present an algorithm for computing Nash equilibria, which has shown very satisfactory results.
15:40-16:00	WeB7.6
Stackelberg Strategies for Singularly Perturbed Stochastic Systems, pp. 730-735
Mukaidani, Hiroaki	Hiroshima Univ.
Unno, Masaru	NTT FINANCE Corp.
Yamamoto, Toru	Hiroshima Univ.
Xu, Hua	Univ. of Tsukuba
Dragan, Vasile	Romanian Acad.
In this paper, a linear closed-loop Stackelberg strategy for a class of singularly perturbed stochastic systems (SPSS) governed by Ito differential equations is considered. Necessary conditions for the solution are established via a set of cross-coupled algebraic Lyapunov and Riccati equations (CALREs). After studying the asymptotic behavior of the solution for these stochastic equations, two new numerical algorithms based on Newton's method and semidefinite programming (SDP) for solving CALREs are given. A numerical example is solved to demonstrate the efficiency of the proposed algorithm.

WeB8	HG E5
Consensus Control II (Regular Session)
Chair: De Persis, Claudio	Univ. of Groningen
Co-Chair: Hadjicostis, Christoforos	Univ. of Cyprus
14:00-14:20	WeB8.1
Distributed Map Merging with Consensus on Common Information, pp. 736-741
Aragues, Rosario	Clermont Univ.
Cortes, Jorge	Univ. of California, San Diego
Sagues, Carlos	Univ. de zaragoza
Sensor fusion methods combine noisy measurements of common variables observed by several sensors, typically by averaging information matrices and vectors of the measurements. Some sensors may have also observed exclusive variables on their own. Examples include robots exploring different areas or cameras observing different parts of the scene in map merging or multi-target tracking scenarios. Iteratively averaging exclusive information is not efficient, since only one sensor provides the data, and the remaining ones echo this information. This paper proposes a method to average the information matrices and vectors associated only to the common variables. Sensors use this averaged common information to locally estimate the exclusive variables. Our estimates are equivalent to the ones obtained by averaging the complete information matrices and vectors. The proposed method preserves properties of convergence, unbiased mean, and consistency, and improves the memory, communication, and computation costs.
14:20-14:40	WeB8.2
Novel Results on Slow Coherency in Consensus and Power Networks, pp. 742-747
Romeres, Diego	Univ. of Padova
Dorfler, Florian	Univ. of California at Santa Barbara
Bullo, Francesco	Univ. of California, Santa Barbara
We revisit the classic slow coherency and area aggregation approach to model reduction in power networks. The slow coherency approach is based on identifying sparsely and densely connected areas of a network, within which all generators swing coherently. A time-scale separation and singular perturbation analysis then results in a reduced low-order system, where coherent areas are collapsed into aggregate variables. Here, we study the application of slow coherency and area aggregation to first-order consensus systems and second-order power system swing dynamics. We unify different theoretic approaches and ideas found throughout the literature, we relax some technical assumptions, and we extend existing results. In particular, we provide a complete analysis of the second-order swing dynamics – without restrictive assumptions on the system damping. Moreover, we identify the reduced aggregate models as generalized first or second-order Laplacian flows with multiple time constants, aggregate damping and inertia matrices, and possibly adverse interactions.
14:40-15:00	WeB8.3
Balancing Time-Varying Demand-Supply in Distribution Networks: An Internal Model Approach, pp. 748-753
De Persis, Claudio	Univ. of Groningen
The problem of load balancing in a distribution network under unknown time-varying demand and supply is studied. A set of distributed controllers which regulate the amount of flow through the edges is designed to guarantee convergence of the solution to the steady state solution. The results are then extended to a class of nonlinear systems and compared with existing results. Incremental passivity and internal model are the main analytical tools.
15:00-15:20	WeB8.4
Asynchronous Distributed Calibration of Camera Networks, pp. 754-759
Borra, Domenica	Pol. di Torino
Fagnani, Fabio	Pol. Di Torino
This work focuses on the problem of calibrating planar networks of cameras in a distributed fashion. The camera network is modeled by a graph, and along each edge a noisy relative angular measurement is available. The goal is to achieve the absolute orientation of each camera with respect to a fixed external reference frame, in order to be able to perform monitoring and patrolling tasks. The idea is to exploit the cycles in the graph, along which all relative measurements sum to zero, in order to eliminate the noise. We design a distributed algorithm for the cameras to autonomously calibrate and we adopt an asynchronous gossip-like communication protocol. The proposed algorithm is proved to converge, almost surely and in the mean square sense, to the set of angles with zero cycle error. Finally, numerical experiments are presented to compare the performance of the algorithm on different graph topologies.
15:20-15:40	WeB8.5
Privacy-Preserving Asymptotic Average Consensus, pp. 760-765
Manitara, Nicolaos	Univ. of Cyprus
Hadjicostis, Christoforos	Univ. of Cyprus
In this paper, we develop and analyze a distributed privacy-preserving average consensus algorithm that enables all of the components of a distributed system, each with some initial value, to asymptotically reach average consensus on their initial values, without having to reveal the specific value they contribute to the average calculation. We consider a set of components (nodes) that interact via directional communication links (edges) that form a generally directed communication topology (digraph). The proposed protocol can be followed by each node that does not want to reveal its initial value and, under certain conditions on the communication topology that we characterize precisely, all nodes can calculate the average of their initial values while maintaining privacy (i.e., the initial values contributed to the average by the nodes that follow the protocol are not exposed to malicious nodes). We assume that malicious nodes try to identify the initial values of other nodes but do not interfere in the computation in any other way; malicious nodes are assumed to know the predefined linear strategy and topology of the network (but not the actual values used by the nodes that want to preserve their privacy).
15:40-16:00	WeB8.6
A Decentralized Algorithm for the Preferred Assignment Problem in Multi-Agent Systems, pp. 766-771
Khan, Usman A.	Tufts Univ.
Kar, Soummya	Carnegie Mellon Univ.
In this paper, we consider a task-allocation problem where~$N$ agents are to be assigned~$N$ distinct tasks without a central coordinator. The agents communicate over a connected network and we assume that each agent has a preference for a distinct unique task, i.e., without loss of generality, agent~$i$ prefers task~$i$ and is emph{unhappy} with any other task~$(neq i)$. With the assumption that each agent starts by picking an arbitrary random task, we can divide the emph{preferred} task-allocation problem into two phases: (i) to reach the set of~$N!$ emph{unique assignments} from the set of~$N^N$ total configurations--presented elsewhere; and (ii) to reach the single emph{preferred assignment} from the set of~$N!$ unique configurations. In this paper, we consider the phase (ii) of this emph{preferred assignment} problem, i.e., we assume that the network is already in one of the~$N!$ unique configurations. In particular, we assume that each agent has a unique task that may not be preferred and each agent does not know its preference until it sees the task and evaluates its cost. Under mild conditions on the network connectivity, we propose a emph{swap-stick} algorithm and show that this algorithm reaches the preferred assignment in finite-time (a.s.) using Markov-chain arguments. The analysis in this paper is further extendable to show the convergence of the phase (i) assignment problem.

WeB9	HG E7
Mechatronic Systems II (Regular Session)
Chair: Alleyne, Andrew G.	Univ. of Illinois at Urbana-Champaign
Co-Chair: Nikolakopoulos, George	Luleå Univ. of Tech. Sweden
14:00-14:20	WeB9.1
Adaptive Internal Model Control Scheme for a Pneumatic Artificial Muscle, pp. 772-777
Andrikopoulos, George	Univ. of Patras, Greece
Nikolakopoulos, George	Luleå Univ. of Tech. Sweden
Manesis, Stamatis	Univ. of Patras
In this article, the model identification and control problem of a Pneumatic Artificial Muscle (PAM) is being considered. The identification of the PAM's model parameters is being carried out by a Recursive Least Square (RLS) based algorithm, while an Internal Model Control (IMC) structure is being synthesized. Experimental studies are being utilized to prove the overall efficiency of the suggested control scheme, regarding: a) set-point tracking performance through selected positioning scenarios, b) adaptability through hysteresis shift compensation and c) robustness through disturbance cancellation.
14:20-14:40	WeB9.2
Soft Sensor Based Dynamic Flow Rate Estimation in Low Speed Radial Pumps, pp. 778-783
Leonow, Sebastian	Ruhr-Univ. Bochum
Mönnigmann, Martin	Ruhr-Univ. Bochum
Today, poorly controlled low power centrifugal pumps are responsible for a great fraction of the overall pump related energy consumption in the European Union [1]. It is expected that the energy efficiency can be improved considerably with appropriate automatic flow rate control. Automatic flow rate controllers are not popular among practitioners, however, since their installation and tuning is an additional effort, and since they require flow rate sensors, which result in an increased unit cost. We present a dynamic flow rate estimation method which makes automatic control without traditional sensors possible. The setup of the estimation method and all required measurements can be carried out by the manufacturer. As a result, the additional effort for the customer is kept to a minimum.
14:40-15:00	WeB9.3
Motion Control for Magnetic Micro-Scale Manipulation, pp. 784-790
Alleyne, Andrew G.	Univ. of Illinois at Urbana-Champaign
Schuerle, Simone	ETH Zurich
Meo, Alessandro	Univ. of Pisa
Nelson, Bradley	ETH Zurich
This work demonstrates performance improvement in motion control under a particular set of machine system constraints. A high performance industrial magnetic micro-manipulation system, the Minimag, is introduced and modeled with both first principles and system identification. The form of the closed loop controller is constrained by operational bounds and the system software, resulting in limits to achievable performance. A model-based motion control enhancement is developed and implemented using tools from Iterative Learning Control. The resulting performance improvements indicate the benefits of motion control even when the closed loop controller is fixed. Experimental results at two different size scales (500 microns and 4.5 microns) are given.
15:00-15:20	WeB9.4
Design of a Bench Hardware-In-The-Loop System for the Study of Chatter in Turning, pp. 791-796
Mancisidor, Iker	Dynamics and Control department, Ik4-Ideko
Barcena, Rafael	Univ. of the Basque Country UPV/EHU
Munoa, Jokin	Dynamics and Control department, Ik4-Ideko
Etxebarria, Ainhoa	Univ. of the Basque Country UPV/EHU
The violent vibrations due to regenerative effects occurring in machining processes, also known as chatter, are a classical problem that limits the productivity in machine tools. These effects motivate a reduction of the practical life of the tool and mechanical components of the machine, and prevent obtaining required surface quality. Therefore, the chatter has been a prevalent topic for academic and industrial research. On the other hand, experimental investigations on the active reduction of chatter may be difficult in a real machining environment, given the exigency of numerous machining tests and the associated problems of repeatability, due to the involvement of a large number of practical parameters. In this paper, a mechatronic Hardware-In-the-Loop (HIL) simulator for chatter in turning is presented. Such system reproduces experimentally, on a simple mechanical structure in the laboratory, certain regenerative chatter, described by stability lobes previously obtained from the real turning machine of interest. Then, by adding inertial actuators to the HIL simulator, the performance of different algorithms for the active control of such chatter may be realistically tested.
15:20-15:40	WeB9.5
Energy Saving Control in Five-Axis Machine Tools Using Contouring Control, pp. 797-802
Uchiyama, Naoki	Toyohashi Univ. of Tech.
Ogawa, Yuki	Toyohashi Univ. of Tech.
M., A. El Khalick	Toyohashi Univ. of Tech.
Sano, Shigenori	Toyohashi Univ. of Tech.
Yamazaki, Kazuo	Univ. of California, Davis
Machine tools operate around the clock in industrial applications all over the world, and hence there is a requirement not only for high-speed and high-precision operation but also for reduced energy consumption. This paper presents a novel controller design that aims to reduce energy consumption in five-axis machine tools, which typically consist of three translational and two rotary axes. The proposed design is based on a contouring controller that has been proposed for use in high-precision machining in machine tool systems. The design also takes into consideration the relative orientation error between a tool and the workpiece in five-axis machining. The synchronous controller is also used to reduce synchronous error between two parallel actuators, which is a cause of significant mechanical damage. The experimental results demonstrate the effectiveness of the proposed controller, which reduces energy consumption by about 13%
15:40-16:00	WeB9.6
Stable Region of Gravity Position of Object Grasped by Virtual Springs, pp. 803-808
Nakashima, Akira	Nagoya Univ.
Hayakawa, Yoshikazu	Nagoya Univ.
This paper considers the stable grasp in the sense of quasi-statistics, i.e., an object is grasped by a multi-fingered robot hand, where each finger is regarded as a virtual 3D linear spring. The stability analysis is to find a condition under which the equilibrium can be restored after any small perturbations of the object's position and orientation. The paper derives a necessary and sufficient condition for the grasp to be stable in terms of relations among the stiffness parameters, the contact points, and the object's COG. By using this result, a region of the object's COG can be clarified which guarantees the stable grasp.

WeB10	HG D1.1
Linear Systems II (Regular Session)
Chair: Panteley, Elena V.	CNRS
Co-Chair: Deutscher, Joachim	Univ. Erlangen-Nürnberg
14:00-14:20	WeB10.1
Spectral Conditions for Symmetric Positive Real and Negative Imaginary Systems, pp. 809-814
Bajcinca, Naim	Max Planck Insitute
Voigt, Matthias	Max Planck Inst. for Dynamics of Complex Tech. Systems
Non-Hamiltonian spectral conditions for the class of symmetric multivariable strictly positive real and strictly negative imaginary systems are derived. They represent gen- eralizations of known ones for strict positive realness to the cases with singular feedthrough matrix, and are novel in the context of strict negative imaginariness. Moreover, we propose a concept of strong negative imaginariness and establish its links to strict positive realness of symmetric systems. The proposed spectral conditions are useful in the corresponding assessment and enforcement procedures, as well as in quadratic stability analysis of uncertain and switched systems.
14:20-14:40	WeB10.2
Optimal Cost Actuator/Sensor Placement for Large Scale Linear Time-Invariant Systems: A Structured Systems Approach, pp. 815-820
Pequito, Sergio	Carnegie Mellon Univ.
Aguiar, A. Pedro	Faculty of Engineering, Univ. of Porto (FEUP)
Kar, Soummya	Carnegie Mellon Univ.
We address the problem of minimal cost allocation of actuators (sensors) placement for ensuring structural controllability (structural observability) of large scale linear time invariant (LTI) systems. In particular, for the dedicated actuator placement problem (i.e., each actuator may control only one state variable or dynamic component), we propose a design methodology, that provides the minimal placement with minimal cost (with respect to the given placement cost functional), under the requirement that the system be structurally controllable. This cost may include, for instance, the price of the physical control device and its installation cost. In addition to obtaining the global solution of the optimization problem, the methodology is shown to be implemented by an algorithm with polynomial complexity (in the number of state variables), making it suitable for large scale systems. By duality, the solution readily extends to the structural design of the corresponding sensor placement under cost constraints.
14:40-15:00	WeB10.3
On Asymptotic Equivalence of the All-To-All Kuramoto Model and Certain Linear System: Stability Analysis of Phase Locked Solutions, pp. 821-826
Conteville, Laurie	Univ. paris 11, LSS
Panteley, Elena V.	CNRS
For the classical ``all-to-all'' Kuramoto model, we construct a family of auxiliary linear models that preserves information on the natural frequencies and interconnection gains of the original Kuramoto model and depends on its phase-locked solutions. Stability properties of the family of linear systems are analyzed, we show that there is only one system in this family which is stable and for almost all initial conditions its solutions exponentially converge to a stable periodic limit cycle. Finally, we show that asymptotically, in the time limit, this linear system maps on the original Kuramoto model.
15:00-15:20	WeB10.4
Krylov Subspace Methods for Block Patterned Linear Systems, pp. 827-832
Deutscher, Joachim	Univ. Erlangen-Nürnberg
This article considers the moment matching using Krylov subspace methods for block patterned linear systems that comprise symmetrically interconnected as well as block circulant systems. Thereby, the block patterned system structure is preserved so that subsequent analysis and synthesis approaches can exploit the special structure of the reduced order system. It is shown that a specific number of moments at arbitrary points can be matched. For the order reduction numerical efficient algorithms can be applied to determine the corresponding Krylov subspaces. The proposed approximation technique is illustrated by means of a simple symmetrically interconnected system.
15:20-15:40	WeB10.5
An LMI Approach to Structured Sparse Feedback Design in Linear Control Systems, pp. 833-838
Polyak, Boris T.	Moscow Inst. of Control Sciences
Khlebnikov, Mikhail	Inst. for Control Science, RAS
Shcherbakov, Pavel	Moscow Inst. for Control Sciences, RAS
Consider the classical state feedback design in the linear system x' = Ax+Bu subject to performance specifications with an additional requirement that the control input vector u=Kx has as many zero entries as possible. The corresponding gain K is referred to as a row-sparse controller. We propose an approach to approximate solution of this kind of nonconvex problems by formulating the proper convex surrogate,---the minimization of a certain matrix norm subject to LMI constraints. The novelty of the paper is the problem formulation itself and the construction of the surrogate. The two main contributions are the design of low-dimensional output to be used in static output feedback, and suboptimal design illustrated via LQR. The results of preliminary numerical experiments are twofold. First, in many test problems, the number of controls was considerably reduced without significant loss in performance. Second, the number of nonzero entries obtained by our method is either very close to or coincide with the minimum possible amount. The approach can be further extended to handle numerous problems of optimal and robust control in sparse formulation.
15:40-16:00	WeB10.6
P·SPR·D Control Design Via LMI for Linear MIMO Systems and Its Extension to Adaptive Control, pp. 839-845
Shimizu, Kiyotaka	Keio Univ.
This paper concerns with P·SPR·D control for set-point servo problem of linear MIMO systems and its extension to adaptive control. The P·SPR·D control is a structured controller introducing a SPR (strict positive real) element instead of I element for PID control. The purpose is to design a P·SPR·D controller for aysmptotical stabilization and to adjust P, SPR, D gain matrices for the improvement of control performances. We make stability analysis of the P·SPR·D control via LMI, based on Lyapunov's direct method. We extend the P·SPR·D control to adaptive P·SPR·D control in the case where plant parameters change. Namely, we make an effort to stabilize the closed-loop system and to improve convergence speed to the desired equilibrium by adding an adaptive control mode with a time-varying gain matrix to the usual P·SPR·D controller. Consequently the closed-loop system keeps asymptotical stability against unknown plant-parameter changes. For the adaptive P·SPR·D control we propose, it is not necessary to assume the minimum phase property of the plant. The P·SPR·D control can be modified to P·SPR·D+I control for improving the steady state performance (an off-set).

WeB11	HG D1.2
Control in Aerospace II (Regular Session)
Chair: Fassois, Spilios D.	Univ. of Patras
Co-Chair: Magnis, Lionel	MINES ParisTech
14:00-14:20	WeB11.1
Gravity Gradiometer Integrated Inertial Navigation, pp. 846-851
Welker, Troy	OUSD(AT&L)
Pachter, Meir	AFIT/ENG
Huffman, Richard	AFIT/ENY
In high precision inertial navigation, gravity field modeling error becomes a limiting factor. Granted that high precision accelerometers are used, airborne gravity gradiometry can be employed in a self-contained way to accurately estimate the gravity field on the fly and eliminate the gravity field modeling error. The local acceleration of gravity will be estimated using the onboard accelerometer measurements, provided that the acceleration measurements are very accurate, as is the case in high precision INS using cold atom interferometry-based accelerometers. An autonomous free-inertial gravity gradiometer integrated aircraft navigation system is promulgated. The accurate mapping of the gravity field along the aircraft's flight path is an added benefit.
14:20-14:40	WeB11.2
Rotation Estimation for a Satellite from Sun Sensors, pp. 852-859
Magnis, Lionel	MINES ParisTech
Petit, Nicolas	MINES ParisTech
We develop a method to reconstruct the rotation motion of a satellite from Sun sensor measurements. Mathematically, the estimation objective is formulated as the problem of phase reconstruction for a 2-dimensional periodic function. A computationally straightforward solution is proposed. Theoretical developments allow to assess the robustness to model imperfections, sampling and noise. Simulation results illustrate practically obtainable performance.
14:40-15:00	WeB11.3
Self-Reconfigurable Control for Dual-Quaternion/Dual-Vector Systems, pp. 860-865
Drakunov, Sergey V.	Embry-Riddle Aeronautical Univ.
Mackunis, Will	Univ. of Florida
Price, William	Embry-Riddle Aeronatuatical Univ.
Ton, Chau	Embry-Riddle Aeronautical Univ.
In this paper we suggest self-reconfigurable control for dual-quaternion systems with unknown control direction. The technique is based on the creation of multiple equilibrium surfaces for the system in the extended state space. We describe the mathematical tools of dual quaternions and technique required to design such system. Examples are presented to illustrate the proposed method.
15:00-15:20	WeB11.4
Weighted Clustering Coefficient Maximization for Air Transportation Networks, pp. 866-871
Ponton, Julien	Ec. Pol.
Wei, Peng	Purdue Univ.
Sun, Dengfeng	Purdue Univ.
In transportation networks the robustness of a network regarding nodes and links failures is a key factor for its design. At the same time, traveling passengers usually prefer the itinerary with fewer legs. The average clustering coefficient can be used to measure the robustness of a network. A high average clustering coefficient is often synonymous with a lower average travel distance and fewer number of legs. In this paper we present the average weighted clustering coefficient maximization problem, and give several solution methods based on branch and bound algorithm, dynamic programming and quadratically constrained programs.
15:20-15:40	WeB11.5
An Adaptive Time Series Framework for Aircraft 4D Trajectory Conformance Monitoring, pp. 872-877
Kopsaftopoulos, Fotis	Univ. of Patras
Fassois, Spilios D.	Univ. of Patras
The general problem of conformance monitoring with respect to preassigned 4-dimensional (4D) trajectories equipped with corresponding 4D margins (4D contracts) is considered within an adaptive statistical time series framework. The specific issues tackled within this context are: (a) Present conformance monitoring and quality of conformance evaluation via statistical tools, which also leads to abnormal event detection; (b) future conformance monitoring, in which the conformance is predicted ahead of time, allowing for potentially corrective or other actions. The performance of the developed methods is assessed via simulations. In present conformance monitoring, an alarm is shown to be issued instantaneously, following the emergence of an abnormal event. In future conformance monitoring, the comparison with a scheme based on nominal probabilistic trajectory prediction demonstrates the benefits of the adaptive statistical time series framework.
15:40-16:00	WeB11.6
Improving Classification Performance through Kinematic Decisions, pp. 878-883
Zhang, Weijia	Univ. of Michigan
Hyun, Baro	Univ. of Michigan
Kabamba, Pierre T.	Univ. of Michigan
Girard, Anouck	Univ. of Michigan at Ann Arbor
We analyze the relationship between classification performance (i.e., the mutual information and the probability of misclassification) and sensor abilities. The analysis suggests an effective region of sensor space that can improve the classification performance when multiple measurements are to be taken sequentially, and possible sensor allocation strategies are discussed. Based on the analysis, we apply the sensing strategies to a UAV path planning problem where the sensor performance depends on the relative position (i.e., range and azimuth) of the UAV with respect to the object of interest. Specifically, we use two sliding mode controllers, each of which accounts for a particular sensing strategy, with a hybrid-system switching scheme. We validate our approach with numerical simulation results.

WeB12	HG D3.2
Adaptive Control II (Regular Session)
Chair: Plestan, Franck	Ec. Centrale De Nantes-CNRS
Co-Chair: Simandl, Miroslav	Univ. West Bohemia
14:00-14:20	WeB12.1
AdaptiveMobile Robots Formation Control Using Neural Networks, pp. 884-889
Raimúndez, José Cesáreo	Univ. de Vigo
Paz, Enrique	Univ. of Vigo
In this paper we present the tracking problem of controlling a particular formation among mobile robots, using feedback linearization techniques. Reference tracking will be made using look ahead control. Look ahead control will be obtained by linearization feedback. To cancel the modeling errors or/and external perturbations, the closed loop will incorporate an adaptive element performed by a neural network. The adaptive controller, implemented through a hidden layer feed-forward neural network, has its weights realtime updated to cope with external perturbations as well as modeling errors. The control procedures required for tracking control, are inspired in the Lyapunov stability theory.
14:20-14:40	WeB12.2
High Order Integral Sliding Mode Control with Gain Adaptation, pp. 890-895
Taleb, Mohammed	LUNAM Univ. Ec. Centrale de Nantes - IRCCyN
Plestan, Franck	Ec. Centrale De Nantes-CNRS
Bououlid, Badr	Moulay Ismail Univ. Meknes
In this paper, an adaptive integral sliding mode is proposed. The main interest of gain adaptation is the reduction of the chattering and the possibility to control uncertain nonlinear systems whose the uncertainties have unknown bounds. The proposed control approach consists in using dynamically adapted control gain that ensure the establishment, in a finite time, of a real high order sliding mode. The control is applied by simulation to an academic example to evaluate its efficiency.
14:40-15:00	WeB12.3
Implicit Dual Controller Based on Stochastic Integration Rule, pp. 896-901
Flídr, Miroslav	Univ. of West Bohemia
Simandl, Miroslav	Univ. West Bohemia
A new implicit dual control method is proposed providing a suboptimal solution of the optimal control problem for discrete linear stochastic state space model with unknown and unobservable parameters. The solution is based on Bellman optimization recursion where two stages of optimization recursion will be pursued. The resulting controller ensures both dual properties of the suboptimal control, i.e. caution and probing. In order to be able to determine the control, the stochastic integration rule is employed for approximate evaluation of expectations. The dual control is then obtained using suitable iterative numerical algorithm. The proposed implicit dual controller is compared to the explicit dual controllers which are easier to derive but require proper tuning of design parameters.
15:00-15:20	WeB12.4
Design of Performance-Adaptive PID Control System Using Model Predictive Approach, pp. 902-907
Sato, Takao	Univ. of Hyogo
Yamamoto, Toru	Hiroshima Univ.
Araki, Nozomu	Univ. of Hyogo
Konishi, Yasuo	Univ. of Hyogo
In the present paper, we discuss a new design method for a PID control system us ing model predictive approach. The PID compensator is designed based on generalized predictive control (GPC). The PID parameters are adaptively updated such that the control performance is improved because the design parameters of GPC are automatically selected to attain a user-specified control performance. In the proposed scheme, estimated plant parameters are updated only when the estimation error increases. Therefore, the control system is not updated frequently, The control system is updated only when the control performance is sufficiently improved. Numerical examples demonstrate the effectiveness of the proposed method.
15:20-15:40	WeB12.5
High-Gain Feedback Stability of a Nonlinear Drivetrain System, pp. 908-913
Loepelmann, Peter	Daimler AG
Bäker, Bernard	Tech. Univ. Dresden
The development of modern automotive drivetrain systems has seen a constant increase in complexity over the past years. It is mostly driven by the increasing effort to reduce fuel consumption of passenger cars by optimizing the drivetrain to reduce losses and achieve energy-optimal set points for engine operation. Thus, electronic control units for drivetrain systems, such as automatic transmissions, have become much more complex. This increases the need for more complex control systems that traditionally have to be calibrated in a time-consuming process before production. We propose a new method of designing controllers for the shift procedure in automatic transmissions that have the same capabilities as the traditional control systems, but require less calibration. To this end, we introduce adaptive Lambda-tracking controllers based on the high-gain feedback principle to replace the current control scheme. In order to ensure high-gain stability, we provide a comprehensive system analysis based on a nonlinear system model and investigate the relative degree and the minimum-phase property in detail.
15:40-16:00	WeB12.6
Design of Adaptive Block Backstepping Controllers for Semi-Strict Feedback Nonlinear Systems with Perturbation Estimation, pp. 914-919
Cheng, Chih-Chiang	National Sun Yat-Sen Univ.
Chien, An-Fu	National Sun Yat-Sen Univ.
Huang, Yi-Chun	National Sun Yat-Sen Univ.
Based on the Lyapunov stability theorem, a design methodology of adaptive block backstepping control with perturbation estimation scheme is proposed in this paper for a class of multi-input systems with perturbations to solve regulation problems. First of all, the virtual input controllers are designed from the first block to the (m-1)th block. Then the proposed robust controller and the perturbation estimator are designed in accordance with the last block. Adaptive mechanisms are employed in each of the virtual input as well as the robust controller so that they may automatically adapt the least upper bounds of perturbations or perturbation estimation errors which may not fulfill the strict-feedback form. Furthermore, the proposed control scheme is able to achieve asymptotic stability. A numerical example is also given for demonstrating the feasibility of the proposed control scheme.

WeB13	HG D5.2
Fault Detection and Identification II (Regular Session)
Chair: Gerencser, Laszlo	MTA SZTAKI
Co-Chair: Wahrburg, Arne	Tech. Univ. Darmstadt
14:00-14:20	WeB13.1
Integrated Fault Detection in Multiple Switched Systems, pp. 920-925
Hasan Abdo, Ali	Univ. of Duisburg-Essen
Ding, Steven X.	Univ. of Duisburg-Essen
Saijai, Jedsada	Duisburg-Essen Univ.
Damlakhi, Waseem	Duisburg-essen Univ.
This paper addresses a new strategy for constructing switched systems. This approach will be used to set an integrated fault detection filter (FDF). The basic idea behind this approach is to switch multiple sub-models with weighting factors that reflect the effect of each local sub-model. The main contributions in this paper are summarized as follows: (1) set a multiple switching technique for describing the system dynamic behavior. (2) integrate residual generation with evaluation and threshold setting. (3) enhance the fault detectability for the switched systems by utilizing available information provided by each local sub-model. (4) relax the stability condition for the switched observers in terms of switched Lyapunov functions (SLF). The proposed approach will be illustrated by lateral vehicle dynamics as an example.
14:20-14:40	WeB13.2
An Algebraic Observer for Leak Detection and Isolation in Plastic Pipelines, pp. 926-931
Navarro, Adrián	CINVESTAV
Sánchez, Juan Diego	CINVESTAV del IPN unidad Guadalajara
Begovich, Ofelia	CINVESTAV
Besancon, Gildas	Ense3 - Grenoble INP
Patiño Murillo, Julian Alberto	Univ. Nacional de Colombia
In the continuation of authors' studies on leak diagnosis in pipelines, a new model-based Leak Detection and Isolation (LDI) algorithm is designed. This system only uses measures of flow and pressure coming from sensors placed at the ends of a pipeline. The present approach is based on a finite nonlinear pipeline model, and extended with variables related to the leak. On this basis, the purpose here is to investigate the use of a so-called algebraic observer to estimate the leak position and its magnitude. The corresponding observer design is thus presented, and its performances are illustrated both with simulation results, and experimental ones, with data taken from a real pipeline prototype.
14:40-15:00	WeB13.3
Distributed Fault Detection Using Sensor Networks and Pareto Estimation, pp. 932-937
Boem, Francesca	Univ. of Trieste
Xu, Yuzhe	KTH Royal Inst. of Tech.
Fischione, Carlo	KTH
Parisini, Thomas	Imperial Coll. & Univ. of Trieste
In this paper, a preliminary novel distributed fault detection architecture for dynamic systems using sensor networks and a distributed estimation method based on Pareto optimization is proposed. The goal is to monitor large-scale or distributed systems by using a sensor network where each node acts as a local estimation agent without centralized coordination. Probabilistic detection thresholds related to a given rate of false alarms are derived in several different scenarios as far as the measurement pattern and the nominal dynamics is concerned. Preliminary simulation results show the effectiveness of the proposed fault detection methodology.
15:00-15:20	WeB13.4
Modeling Improvements for Leak Detection in Pipelines of LPG, pp. 938-942
Torres, Lizeth	UNAM
Verde, Cristina	Inst. de Ingenieria, UNAM
The fluid behavior in a pipeline of Liquefied Petroleum Gas (LPG)—according to the principles of conservation of mass and momentum—is governed by a set of Partial Differential Equations (PDE). In this work, such equations are modified with the inclusion of an unsteady friction model and approximated using the Finite Difference Method. As a result of this approximation, a model is obtained and subsequently used for the design of a high gain observer to locate leaks in a pipeline. Some results based on real data are exposed to demonstrate the repercussions of the proposed modifications in tasks of leak detection.
15:20-15:40	WeB13.5
LMI-Based Design of Robust Fault Isolation Filters for Linear Systems, pp. 943-949
Wahrburg, Arne	Tech. Univ. Darmstadt
Adamy, Juergen	Tech. Univ. Darmstadt
In this paper we employ general linear dynamic filters to robustly isolate faults in linear systems. The concept can be regarded as a generalization of observer based approaches and offers more degrees of freedom and less structural constraints than fault isolation observers (FIOs). We propose an LMI-based design approach for fault isolation filters (FIFs), where the existence of a solution for exactly known system matrices is guaranteed be the constructive design of an attainable reference model. The approach can readily be applied to systems with disturbances and/or parametric uncertainties, which is a main advantage over observer-based approaches. Its applicability is demonstrated by means of both simulations and lab experiments with a gantry crane system.
15:40-16:00	WeB13.6
Change Detection for Finite Dimensional Gaussian Linear Systems - a Bound for the Almost Sure False Alarm Rate, pp. 950-955
Gerencser, Laszlo	MTA SZTAKI
Prosdocimi, Cecilia	LUISS Guido Carli, Rome
Vago, Zsuzsanna	Peter Pazmany Catholic Univ. Budapest, Hungary
We consider the problem of change detection in the context of finite dimensional Gaussian linear systems. In particular a known initial system will be tested for eventual changes against a known alternative using a simplified version of the Page-Hinkley or CUSUM detector. We show that the detector is L-mixing, implying the existence of an almost sure false alarm rate. The derivation of an explicit upper bound for the latter will be outlined.

WeB14	HG D7.1
Vehicle Dynamics and Control (Regular Session)
Chair: Notarstefano, Giuseppe	Univ. of Lecce
Co-Chair: Abel, Dirk	RWTH Aachen Univ.
14:00-14:20	WeB14.1
Wheelie Detection for Single-Track Vehicles, pp. 956-961
Panzani, Giulio	Univ. degli Studi di Trento
Corno, Matteo	Pol. di Milano
Savaresi, Sergio M.	Pol. Di Milano
Single-track vehicles electronic control systems have been experiencing an important growth in the last years. Despite some similarities with four-wheeled vehicles the dynamics of two-wheeled vehicles have some unique features that require ad hoc solutions. One of those is the lift of the front wheel from the ground during severe accelerations, usually known as wheelie. This phenomenon is particularly important since, if not controlled, can lead to vehicle instabilities. Moreover, it has a significant impact on the vehicle longitudinal speed estimation, essential for the development of wheel slip-based traction control systems, so widely spreading. In this paper the problem of detecting a wheelie occurrence is discussed. Two algorithms, that employ only standard vehicle equipment sensors, are presented. Their parameter tuning procedure is described and experimental data are used to show their effectiveness, as well as for a performances comparison.
14:20-14:40	WeB14.2
Controller Design and Evaluation for Vehicle Run-Off-The-Road and Recovery, pp. 962-967
Jensen, Matthew	Florida Inst. of Tech.
Freeman, Paul (Tim)	Clemson Univ.
Wagner, John	Clemson Univ.
Alexander, Kim	Clemson Univ.
Run-off-the-road (ROR) and unsuccessful recovery events account for a significant percentage of fatal vehicle crashes. Changes to roadway infrastructure have reduced the severity of automotive crashes; however, it is driver performance which remains a primary contributing factor to these incidents. An onboard vehicle control system, using combined steering and differential braking, provides enhanced safety by removing driver error from the recovery process once the ROR event has been identified. In this study, two control algorithms, based on sliding mode and state flow theories, have been designed to automatically recover a vehicle from this dangerous scenario. A driver model was also developed based on experimental data gathered at a tire manufacturer's test track. The controllers were numerically evaluated for a common road departure and return situation with comparison to the empirical driver model. The results show that both controllers safely recovered the vehicle and outperformed the driver steering model. Specifically, the state flow controller reduced recovery time by 40% compared to the sliding mode controller, while the driver model resulted in vehicle spin out.
14:40-15:00	WeB14.3
Immersion and Invariance Control for Lateral Dynamics of Autonomous Vehicles, with Experimental Validation, pp. 968-973
Talj, Reine	Lab. HEUDIASYC, Univ. of Tech. of Compiègne
Tagne, Gilles	Heudiasyc, Univ. of Tech. of Compiègne
Charara, Ali	UMR CNRS 7253
Autonomous intelligent vehicles are under intensive development, especially this last decade. The autonomous navigation consists on developing three main key steps: 1. environment perception, 2. path planning and decision, and 3. vehicle control. This paper focus on the lateral control of intelligent vehicles ; it presents design and experimental validation of a vehicle lateral controller based on Immersion and Invariance (I&I) principle, to minimize the lateral displacement of the autonomous vehicle with respect to a given reference trajectory. The control input is the steering angle and the output is the lateral error displacement. The closed-loop system simulated on Matlab-Simulink has been compared to the experimental data acquired on our vehicle DYNA, a Peugeot 308, according to several driving scenarios. The experimental validation shows robustness and good performance of the proposed control approach.
15:00-15:20	WeB14.4
Optimal Vehicle Dynamics Control for Combined Longitudinal and Lateral Autonomous Vehicle Guidance, pp. 974-979
Katriniok, Alexander	RWTH Aachen Univ. Inst. of Automatic Control
Maschuw, Jan P.	RWTH Aachen Univ.
Christen, Frederic	Forschungsgesellschaft Kraftfahrwesen mbH Aachen
Eckstein, Lutz	RWTH Aachen Univ. Inst. of Automotive Engineering
Abel, Dirk	RWTH Aachen Univ.
This contribution proposes a model-based predictive control approach for combined longitudinal and lateral vehicle guidance. The controller, which has been designed for an automotive collision avoidance system, aims at following a desired evasion trajectory at the handling limits. Thereby, the trajectory following problem is decomposed in a path following and a velocity trajectory tracking problem using the wheel steering angle and the longitudinal acceleration as control inputs. There are two major advantages of this approach. First, the a priori knowledge of the evasion trajectory is explicitly incorporated into the computation of control inputs. Second, the combined transmission of longitudinal and lateral tire forces is considered in the sense of an integrated vehicle dynamics control approach. Experimental results show the potential of the introduced control scheme.
15:20-15:40	WeB14.5
On a Reduced-Order Two-Track Car Model Including Longitudinal and Lateral Load Transfer, pp. 980-985
Rucco, Alessandro	Univ. of Salento
Notarstefano, Giuseppe	Univ. of Lecce
Hauser, John	Univ. of Colorado at Boulder
In this paper we propose a novel reduced-order car model that captures some key features for aggressive maneuvering of complex car vehicles. The proposed model, called RigidCar, consists of a rigid body interacting with the ground at four contact points (two-track car) and including tire models and load transfer. The model does not include suspension models, thus keeping a reasonable level of complexity. Load transfer (both longitudinal and lateral) is taken into account by explicitly imposing the holonomic constraints and computing the reaction forces of the ground at the four contact points. Since the vehicle interacts with the ground at four contact points, it results to be a hyper-static structure so that reaction forces are not uniquely determined. We use the Principle of Least Work to get a compatibility equation and thus resolving the indeterminateness. We provide numerical computations validating the proposed model with respect to a multi-body virtual prototype on an aggressive ISO lane-change maneuver.
15:40-16:00	WeB14.6
Continuous-Time and Discrete-Time Switched H$_{infty}$ State Feedback Controllers: Application for a Robust Steering Vehicle Control, pp. 986-991
Lghani, Menhour	Mines ParisTech, Centre de Robotique, 60 boulevardSaint-Michel,
Koenig, Damien	Grenoble - Inp
D'Andrea-Novel, Brigitte	Ec. des Mines de Paris
Two control laws for switched uncertain linear systems are proposed: firstly, a continuous-time switched H$_{infty}$ state feedback control is given, and secondly, a discrete-time H$_infty$ state feedback control. These controllers are designed using a common Lyapunov function approach and switched Lyapunov function approach used for switched continuous-time and switched discrete-time linear systems respectively. All sufficient conditions of the existence of these controllers are proved and given in term of LMIs. These conditions are provided for an arbitrary switching rule. The lateral control via steering vehicle control is implemented to control simultaneously lateral deviation and yaw motion of the vehicle. In fact, a validation of these two controllers is performed using experimental data acquired by laboratory vehicle under high lateral accelerations.

WeB15	HG D7.2
Control of Airborne Wind Energy Systems (Invited Session)
Chair: Fagiano, Lorenzo	ETH Zurich
Co-Chair: Erhard, Michael	SkySails GmbH
Organizer: Fagiano, Lorenzo	ETH Zurich
Organizer: Gambier, Adrian	Fraunhofer Inst. for Wind Energy and Energy System Tech.
14:00-14:20	WeB15.1
Automatic Control of Tethered Wings for Airborne Wind Energy: Design and Experimental Results (I), pp. 992-997
Fagiano, Lorenzo	ETH Zurich
Zgraggen, Aldo	ETH Zurich
Morari, Manfred	ETH Zurich
Khammash, Mustafa H.	Univ. of California at Sta. Barbara
A new approach to control tethered wings for airborne wind energy is described. A fixed length of the lines is considered, and the aim of the control system is to obtain figure-eight crosswind trajectories. The proposed technique is based on the notion of the wing's "velocity angle" and, in contrast with most existing approaches, it does not require a measurement of the wind speed or of the effective wind at the wing's location. Moreover, the proposed approach features few parameters, whose effects on the system's behavior are very intuitive, hence simplifying tuning procedures. Experimental results are presented, obtained by extensively testing the approach with a small-scale prototype.
14:20-14:40	WeB15.2
Sensors and Navigation Algorithms for Flight Control of Tethered Kites (I), pp. 998-1003
Erhard, Michael	SkySails GmbH
Strauch, Hans	SkySails GmbH
We present the sensor setup and the basic navigation algorithm used for the flight control of the SkySails towing kite system. Starting with brief summaries on system setup and equations of motion of the tethered kite system, we subsequently give an overview of the sensor setup, present the navigation task and discuss challenges which have to be mastered. In the second part we introduce in detail the inertial navigation algorithm which has been used for operational flights for years. The functional capability of this algorithm is illustrated by experimental flight data. Finally we suggest a modification of the algorithms as further development step in order to overcome certain limitations.
14:40-15:00	WeB15.3
Reference Governor Design for Computationally Efficient Attitude and Tether Tension Constraint Enforcement on a Lighter-Than-Air Wind Energy System (I), pp. 1004-1010
Kalabic, Uros V.	Univ. of Michigan
Vermillion, Christopher	Altaeros Energies
Kolmanovsky, Ilya V.	Univ. of Michigan
In this paper, we propose a reference governor-based approach to guarantee enforcement of critical flight constraints on the Altaeros tethered, lighter-than-air wind energy system. While the high-altitude flight made available by the tethered system leads to significant increases in power production over traditional, tower-mounted systems, the freedom of motion resulting from the tethers and aerodynamic shell introduces critical attitude and tether tension constraints. To date, methods considered for enforcing these constraints have relied upon heuristic static maps or model predictive control (MPC). The former cannot guarantee transient constraint satisfaction, whereas the latter is computationally burdensome given Altaeros's current microcontroller capabilities. The approach pursued in this paper uses a reference governor, which is a computationally simple add-on to the existing controller that enforces transient and steady-state constraints. The methodology proposed in this paper is demonstrated through simulations on linear and nonlinear models of the longitudinal dynamics of the Altaeros system with wind gust disturbances.
15:00-15:20	WeB15.4
A Relaxation Strategy for the Optimization of Airborne Wind Energy Systems (I), pp. 1011-1016
Gros, Sébastien	KU Leuven
Zanon, Mario	KU Leuven
Diehl, Moritz	KU Leuven
Optimal control is recognized by the Airborne Wind Energy (AWE) community as a crucial tool for the development of the AWE industry. More specifically, the optimization of AWE systems for power generation is required to achieve the performance needed for their industrial viability. Models for AWE systems are highly nonlinear coupled systems. As a result, the optimization of power generation based on Newton-type techniques requires a very good initial guess. Such initial guess, however, is generally not available. To tackle this issue, this paper proposes a homotopy strategy based on the relaxation of the dynamic constraints of the optimization problem. The relaxed problem differs from the original one only by a single parameter, which is gradually modified to obtain the solution to the original problem.
15:20-15:40	WeB15.5
Control of Airborne Wind Energy Systems Based on Nonlinear Model Predictive Control & Moving Horizon Estimation (I), pp. 1017-1022
Gros, Sébastien	KU Leuven
Zanon, Mario	KU Leuven
Diehl, Moritz	KU Leuven
Among the several problems arising in the Airborne Wind Energy paradigm, an essential one is the control of the tethered airfoil trajectory during power generation. Tethered flight is a fast, strongly nonlinear, unstable and constrained process, motivating control approaches based on fast Non-linear Model Predictive Control. In this paper, a computationally efficient model is proposed, based on Differential-Algebraic equations. A control scheme based on Nonlinear Model Predictive Control (NMPC) and an estimator based on Moving Horizon Estimation (MHE) is proposed to handle the wind turbulences. In order to make a real-time application of Non-linear Model Predictive Control possible, a Real-Time Iteration scheme is proposed.
15:40-16:00	WeB15.6
Rotational Start-Up of Tethered Airplanes Based on Nonlinear MPC and MHE (I), pp. 1023-1028
Zanon, Mario	KU Leuven
Gros, Sébastien	KU Leuven
Diehl, Moritz	KU Leuven
The idea of Airborne Wind Energy (AWE) is to generate power by flying a tethered airfoil across the windflow. Tethered flight is a fast, strongly nonlinear, unstable and constrained process, motivating control approaches based on fast Nonlinear Model Predictive Control (NMPC) and state estimation approaches based on Moving Horizon Estimation (MHE). In particular, the start-up phase of AWE systems is an involved procedure, and starting and landing using NMPC has not been investigated yet. In this paper, a control strategy for starting-up AWE systems is proposed, based on a rotating carousel that is currently built at the KU Leuven. A computationally efficient 6-DOF control model for a small-scale, rigid airfoil is presented. We present and investigate a control scheme based on receding-horizon Nonlinear Model Predictive Control to track reference trajectories and Moving Horizon Estimation to estimate the actual system state and parameters. The MHE shceme is able to estimate also the wind speed, given no direct wind measurement.

WeC1	HG F1
Distributionally Robust Optimization (Tutorial Session)
Chair: Kuhn, Daniel	Imperial Coll. London
Co-Chair: Natarajan, Karthik	Singapore Univ. of Tech. and Design
Organizer: Kuhn, Daniel	Imperial Coll. London
Organizer: Natarajan, Karthik	Singapore Univ. of Tech. and Design
16:20-17:20	WeC1.1
Distributionally Robust Convex Optimization (I)*
Kuhn, Daniel	Imperial Coll. London
We propose a unifying framework for modeling and solving distributionally robust convex optimization problems. We introduce standardized ambiguity sets that contain all distributions with prescribed conic representable confidence sets and with mean values residing on an affine manifold. These ambiguity sets are highly expressive and encompass many ambiguity sets from the recent literature as special cases. They also allow us to characterize distributional families in terms of several classical and/or robust statistical indicators that have not yet been studied in the context of robust optimization. We determine sharp conditions under which distributionally robust optimization problems based on our standardized ambiguity sets are computationally tractable. We also provide tractable conservative approximations for problems that violate these conditions. Moreover, we discuss applications of distributionally robust optimization in chance-constrained programming, portfolio optimization with Value-at-Risk and optimal control.
17:20-18:20	WeC1.2
Distributionally Robust Discrete Optimization (I)*
Natarajan, Karthik	Singapore Univ. of Tech. and Design
We review techniques of conic programming methods that are used to find distributionally robust bounds for the expected optimal value of mixed integer linear programs over the set of all distributions with the given moment information. These methods also provide additional information on the probability that a binary variable attains a value of 1 in the optimal solution for 0-1 integer linear programs. This probability is defined as the persistency of a binary variable. We will provide an overview of the complexity results for these models, conic programming formulations that are readily implementable with standard solvers and simple applications. The main message that we hope to convey through this part of the review is that tools of conic programming provide important insights in the probabilistic analysis of discrete optimization problems. These tools lead to distributionally robust bounds with applications in activity networks, discrete choice models, random walks, scheduling problems, and newsvendor problems.

WeC2	HG F3
Model Predictive Control III (Regular Session)
Chair: Pannocchia, Gabriele	Univ. of Pisa
Co-Chair: Van Impe, Jan F.M.	Katholieke Univ. Leuven
16:20-16:40	WeC2.1
Scenario-Based MPC for Energy-Efficient Building Climate Control under Weather and Occupancy Uncertainty, pp. 1029-1034
Zhang, Xiaojing	ETH Zurich
Schildbach, Georg	ETH Zurich
Sturzenegger, David	ETH Zurich
Morari, Manfred	ETH Zurich
Heating, ventilation and air conditioning (HVAC) systems regulate comfort levels in buildings, but also consume a large amount of energy, which makes them an attractive target for efficiency improvements. In this paper, a novel technique called Randomized Model Predictive Control (RMPC) is investigated to improve the control of existing HVAC systems. RMPC uses weather and occupancy predictions to minimize the building's energy consumption. It accounts for the prediction uncertainties by basing its control actions on a given number of sampled uncertainty scenarios. The main advantage of RMPC over existing methods is the absence of a probabilistic disturbance model. This makes the handling of uncertainties straightforward, even if they are non-Gaussian or non-additive. Moreover, the method of removing adverse samples after solving the initial control problem (RMPC-SR) can lead to a further improvement in the control performance, up to a saturation limit. Although theoretical bounds for choosing the sample sizes are available, our simulations show that only a fraction of these numbers is required for a good performance of RMPC and RMPC-SR. The performance of RMPC and RMPC-SR is investigated through extensive simulations on different models, based on empirically collected data. The results demonstrate that both techniques are attractive alternatives to other Model Predictive Control methods, because they show a higher energy saving potential, and are computationally tractable.
16:40-17:00	WeC2.2
Explicit MPC of LPV Systems in the Controllable Canonical Form, pp. 1035-1040
Kvasnica, Michal	Slovak Univ. of Tech. in Bratislava
Szucs, Alexander	Slovak Univ. of Tech. in Bratislava
Fikar, Miroslav	Slovak Univ. of Tech.
Drgona, Jan	Slovak Univ. of Tech. in Bratislava
We exploit the controllable canonical form of single-input linear parameter-varying (LPV) systems to synthesize explicit Model Predictive Control (MPC) feedback laws. The nonlinear state-parameter dependence is first moved into the feedback term, followed by devising a suitable input constraint set. This allows the MPC problem to be formulated with a quadratic performance index, avoiding costly dynamic programming iterations. The resulting MPC optimization is shown to be convex and can be solved parametrically. The explicit solution takes a form of a piecewise affine (PWA) function which allows to implement the feedback law quickly even with limited computational resources.
17:00-17:20	WeC2.3
Constrained Decomposition Based Control Approach for Linear Parameter Varying Systems, pp. 1041-1046
Nguyen, Hoai-Nam	Tech.
Gutman, Per-Olof	Tech.
Olaru, Sorin	Supelec
Considering a constrained discrete-time linear time-varying system, this paper proposes a novel approach which aims at achieving high performance and enlarging the domain of attraction with respect to any single controller. The main idea of the paper is to use a linear decomposition principle together with a parameter dependent Lyapunov function. At each time instant a quadratic programming problem is solved on-line. Proofs of recursive feasibility and asymptotic stability are given.
17:20-17:40	WeC2.4
Sparse Quadratic Regulator, pp. 1047-1052
Jovanovic, Mihailo	Univ. of Minnesota
Lin, Fu	Univ. of Minnesota
We consider a control design problem aimed at balancing quadratic performance of linear systems with additional requirements on the control signal. These are introduced in order to obtain controls that are either sparse or infrequently changing in time. To achieve this objective, we augment a standard quadratic performance index with an additional term that penalizes either the $ell_1$ norm or the total variation of the control signal. We show that the minimizer of this convex optimization problem can be found by solving a two point boundary value problem (TPBVP) with non-differentiable nonlinearities. Furthermore, we employ alternating direction method of multipliers to determine the optimal controller iteratively from a sequence of linear TPBVPs. Examples are provided to illustrate the developed method.
17:40-18:00	WeC2.5
Model Predictive Control of a Pilot-Scale Distillation Column Using a Programmable Automation Controller, pp. 1053-1058
Huyck, Bart	KU Leuven
De Brabanter, Jos	K.U.Leuven
De Moor, Bart L.R.	Katholieke Univ. Leuven
Van Impe, Jan F.M.	Katholieke Univ. Leuven
Logist, Filip	Katholieke Univ. Leuven
Following the evolution of growing computing power of industrial embedded devices, this paper investigates the use of an online Model Predictive Control (MPC) algorithm on an embedded Programmable Automation Controller (PAC). The controller is tested on a pilot-scale binary distillation column to track reference temperatures. A major task in each model predictive control algorithm is the solution of a quadratic problem. The Hildreth quadratic programming algorithm is used for this purpose. It turns out that this algorithm is able to control the set-up running on the PAC hardware.
18:00-18:20	WeC2.6
Achieving a Large Domain of Attraction with Short-Horizon Linear MPC Via Polyhedral Lyapunov Functions, pp. 1059-1064
Grammatico, Sergio	ETH Zurich
Pannocchia, Gabriele	Univ. of Pisa
Polyhedral control Lyapunov functions (PCLFs) are exploited in this paper to propose a linear model predictive control (MPC) formulation that guarantees a “large” domain of attraction (DoA) even for short horizon. In particular, the terminal region of the proposed finite-horizon MPC formulation is chosen as a level set of an appropriate PCLF. For small dimensional systems, this terminal region can be explicitly computed as an arbitrarily close approximation to the entire (infinite-horizon) stabilizable set. Global stability of the origin is guaranteed by using an “inflated” PCLF as terminal cost. The proposed MPC scheme can be formulated as a (small dimensional) quadratic programming problem by introducing one additional scalar variable. Numerical examples show the main benefits and achievements of the proposed formulation in terms of trade-off between volume of the DoA, computational time and closed-loop performance.

WeC3	HG F5
Fluid Systems (Regular Session)
Chair: Duncan, Stephen	Univ. of Oxford
Co-Chair: Aarsnes, Ulf Jakob F.	NTNU
16:20-16:40	WeC3.1
Application of an Infinite-Dimensional Observer for Drilling Systems Incorporating Kick and Loss Detection, pp. 1065-1070
Hauge, Espen	Norwegian Univ. of Science and Tech.
Aamo, Ole Morten	NTNU
Godhavn, John-Morten	Statoil
We apply a PDE observer to a hydraulic system of a managed pressure drilling rig. The observer can be used to detect and quantify an in- or outflux. We show that the PDE model of the hydraulic system can be expressed as a 2x2 linear hyperbolic system of PDEs with spatially varying coefficients coupled with an ODE at the inlet boundary that models the in- or outflux. Using the method of backstepping, we design an observer which is exponentially stable at the origin in the L2-norm while relying on measurements taken at the outlet boundary, only. Simulation results verify the validity of the observer.
16:40-17:00	WeC3.2
Limits of Controller Performance in the Heave Disturbance Attenuation Problem, pp. 1071-1076
Aarsnes, Ulf Jakob F.	NTNU
Aamo, Ole Morten	NTNU
Hauge, Espen	Norwegian Univ. of Science and Tech.
Pavlov, Alexey	Statoil Res. center
The periodic heave movement of a floating drilling rig is transferred to an elastic pipe suspended in a drilling well. The pipe oscillates up and down in the well with the period of the heave motion pushing the mud in the well and exciting pressure fluctuations in the mud in the pipe. These dynamics are described by six coupled first-order linear partial differential equations of hyperbolic type. By using the Laplace transform, the system is solved in the frequency domain and the resulting irrational transfer functions, from the periodic heave disturbance and the controlled choke flow to the distributed pressure in the annulus around the pipe, are found for different boundary conditions. We show that the response can be improved by choosing an impedance matched boundary condition which can be achieved without the use of active control. We also derive the response when a theoretical 'perfect feed-forward control law', which yields full attenuation of the effect of the disturbance at a single point in the well-annulus, is applied. The limitations of this 'perfect point control' when considering an extended section of the well is shown.
17:00-17:20	WeC3.3
Passivity of Plane Poiseuille Flow, pp. 1077-1082
Zhao, Shi	Univ. of Oxford
Duncan, Stephen	Univ. of Oxford
This paper considers the passivity of plane Poiseuille flow, which is the incompressible flow observed between two parallel plates that are assumed to be of infinite extent. A model in which the flow is considered as the feedback connection between a linear time-invariant system and a static, memoryless nonlinear system is used. It is well known that the nonlinearity of plane Poiseuille flow is not only passive but lossless, which means that it does not generate or consume energy and the only effect of the nonlinearity is to move energy from one flow mode to another. However, little has been addressed about the passivity of the entire flow system. The primary aim of this paper is to find a subcritical Reynolds number below which the linearized flow is always strictly passive, which means that the origin of the full nonlinear system is globally asymptotically stable when the Reynolds number does not exceed the subcritical number. Our results show that the subcritical Reynolds number obtained by the passivity approach is equal to the energy Reynolds number, which is derived by the classical energy approach. This result indicates that the passivity approach is closely related to the energy approach and is a valuable tool to study the stability of fluid flows.
17:20-17:40	WeC3.4
Optimal Placement of Actuators and Sensors for Control of Nonequilibrium Dynamics, pp. 1083-1088
Sinha, Subhrajit	Iowa State Univ.
Vaidya, Umesh	Iowa State Univ.
Rajaram, Rajeev	Iowa State Univ.
In this paper, we provide a systematic convex programming-based approach for the optimal locations of static actuators and sensors for the control of nonequilibrium dynamics. The problem is motivated with regard to its application for control of nonequilibrium dynamics in the form of temperature in building systems and control of oil spill in oceanographic flow. The controlled evolution of a passive scalar field, modeling the temperature distribution or the density of oil dispersant, is governed by the linear advection partial differential equation (PDE) with spatially located actuators and sensors. Spatial locations of actuators and sensors are optimized to maximize the controllability and observability of the linear advection PDE. Linear transfer Perron-Frobenius and Koopman operators, associated with the advective velocity field, are used to provide analytical characterization for the controllable and observable spaces of the advection PDE. Set-oriented numerical methods are used for the finite dimensional approximation of the transfer operators and in the formulation of the optimization problem. Application of the framework is demonstrated for the optimal placement of actuators for the release of dispersant for oil spill control.
17:40-18:00	WeC3.5
Non-Linear Position Control of a Pneumatic Actuator with Closed-Loop Stiffness and Damping Tuning, pp. 1089-1094
Abry, Frédéric	Univ. de Lyon
Brun, Xavier	Insa de Lyon
Sesmat, Sylvie	Lab. Ampère, Univ. de Lyon
Bideaux, Eric	INSA Lyon, Lab. Ampere
This article proposes a new strategy for pneumatic cylinder control using an innovative control synthesis model of the system. A backstepping based control law is synthesized to take advantage of the system two degrees of freedom and a thorough tuning method of the closed-loop stiffness and damping is provided.
18:00-18:20	WeC3.6
Optimal Wells Scheduling of a Petroleum Reservoir, pp. 1095-1100
Hasan, Agus	Norwegian Univ. of Science and Tech.
Foss, Bjarne A.	Norwegian Univ. of Science & Tech.
Production optimization involving injection has been successful in improving oil production by increasing recovery. This technique is now combined with information obtained from gradient-based methods to calculate the optimal flow rate or bottom hole pressure (BHP) in life cycle optimization. It has been shown that if the objective functional is linear in the control and the only constraints are upper and lower bounds on the control, due to its particular structure, the problem will sometimes have bang-bang (on-off) optimal solutions. This paper discusses an optimal control approach to the life cycle optimization by calculating the optimal switching times given the optimal wells settings to be bang-bang controls. The optimal switching times are obtained from the adjoint method combined with the halving interval method. A numerical example using realistic reservoir settings is included to illustrate the application of the proposed method.

WeC4	HG F7
Uncertain Systems (Regular Session)
Chair: Kowalewski, Stefan	RWTH Aachen Univ.
Co-Chair: Henrion, Didier	LAAS-CNRS, Univ. Toulouse
16:20-16:40	WeC4.1
Comparison of Reachability Methods for Uncertain Linear Time-Invariant Systems, pp. 1101-1106
Ben Makhlouf, Ibtissem	RWTH Aachen
Haensch, Paul	RWTH Aachen Univ.
Kowalewski, Stefan	RWTH Aachen Univ.
In this paper, we evaluate a number of methods for computing reachable sets using on one hand approximations and on the other hand invariants. We address systems of the form ˙ x(t) = Ax(t)+ Bu(t) with uncertain but bounded input function u(t). We introduce our implementation based on zonotopes and use SpaceEx for support functions. Then, we go through our LMI-based implementation of ellipsoidal invariants and the canonical decomposition for large scale systems. Performance and accuracy of each method are evaluated using academic and practical examples.
16:40-17:00	WeC4.2
Finding an LFT Uncertainty Model with Minimal Uncertainty, pp. 1107-1113
Haggblom, Kurt E.	Abo Akademi Univ.
In this paper, we present a procedure for finding the best LFT uncertainty model by minimizing the H-infinity norm of the uncertainty set with respect to a nominal model subject to known input-output data. The main problem is how to express the data-matching constraints for convenient use in the optimization problem. For some uncertainty structures, they can readily be formulated as a set of linear matrix inequalities (LMIs), for some other structures, LMIs are obtained after certain transformations. There are also cases, when the constraints result in bilinear matrix inequalities (BMIs), which can be linearized to enable an efficient iterative solution. Essentially all LFT uncertainty structures are considered. An application to distillation modeling is included.
17:00-17:20	WeC4.3
Set Approximation Via Minimum-Volume Polynomial Sublevel Sets, pp. 1114-1119
Dabbene, Fabrizio	Pol. di Torino
Henrion, Didier	Czech Tech. Univ. in Prague
Motivated by problems of uncertainty propagation and robust estimation , we are interested in computing a polynomial sublevel set of fixed degree and minimum volume that contains a given semialgebraic set K. At this level of generality this problem is not tractable, even though it becomes convex e.g. when restricted to nonnegative homogeneous polynomials. Our contribution is to describe and justify a tractable L1-norm or trace heuristic for this problem, relying upon hierarchies of linear matrix inequality (LMI) relaxations when K is semialgebraic, and simplifying to linear constraints when K is a collection of samples, a discrete union of points.
17:20-17:40	WeC4.4
A New Observer-Based Stabilization Method for Linear Systems with Uncertain Parameters, pp. 1120-1125
Kheloufi, Houria	Mouloud Mammeri Univ.
Zemouche, Ali	Nancy-Univ.
Bedouhene, Fazia	Lab. de Mathématiques Pures et Appliquées(LMPA),Mouloud M
Boutayeb, M.	Lorraine Univ.
This paper deals with the problem of observer-based stabilization for linear systems with parameter uncertainties. A new design methodology is proposed thanks to a judicious use of the famous Young relation. This latter leads to a less restrictive synthesis condition, expressed in term of Linear Matrix Inequality~(LMI), than those available in the literature. A numerical example is provided in order to show the validity and superiority of the proposed method.
17:40-18:00	WeC4.5
Design of Interval Observers for LPV Systems Subject to Exogenous Disturbances, pp. 1126-1131
Thabet, Rihab El Houda	Univ. OF BORDEAUX 1
Raïssi, Tarek	Conservatoire National des Arts et Métiers
Combastel, Christophe	Ec.
Zolghadri, Ali	Univ. Bordeaux I
In this paper an interval observer for Linear-Parameter-varying (LPV) Systems is proposed. Usually, the design of such observers is based on monotone systems theory and the observation error should have a cooperative dynamics. In many cases, such a property is hard to satisfy. In this paper, a time-varying change of coordinates is used to overcome this limitation. Two sufficient conditions, according to the center and the radius dynamics of the proposed interval observer and ensuring the stability, are given. The efficiency of the proposed observer is illustrated through computer simulations.
18:00-18:20	WeC4.6
Fixed-Order H_infinity and H_2 Controller Design for Continuous-Time Polytopic Systems: An LMI-Based Approach, pp. 1132-1137
Sadabadi, Mahdieh Sadat	Ec. Pol. Federale de Lausanne (EPFL)
Karimi, Alireza	Ec. Pol. Federale
In this paper, a new approach for fixed-order H_infinity and H_2 dynamic output-feedback controller design of continuous-time systems with polytopic uncertainty is proposed. This approach is based on the use of some instrumental matrices, which operate as a tool to decouple the controller parameters and the Lyapunov matrices. The stability of the closed-loop system as well as H_infinity and H_2 performance criteria are expressed by a set of linear matrix inequalities based upon linearly parameter dependent Lyapunov matrices. Iterative procedure for updating the instrumental matrices using the previous controller can increase the performance of the approach. Simulation results demonstrate the effectiveness of the proposed method.

WeC5	HG E1.1
Event-Triggered Control (Regular Session)
Chair: Johansson, Karl Henrik	Royal Inst. of Tech.
Co-Chair: Dimarogonas, Dimos V.	Royal Inst. of Tech. (KTH)
16:20-16:40	WeC5.1
Extension and Evaluation of Model-Based Periodic Event-Triggered Control, pp. 1138-1144
Verhaegh, Jan	Eindhoven Univ. of Tech.
Gommans, Tom	Eindhoven Univ. of Tech.
Heemels, Maurice	Eindhoven Univ. of Tech.
Periodic event-triggered control (PETC) is a control strategy that combines ideas from conventional periodic sampled-data control and event-triggered control. By communicating periodically sampled sensor and controller data only when needed to guarantee stability and performance properties, PETC is capable of reducing the number of transmissions significantly, while still retaining a satisfactory closed-loop behavior. In this paper, we provide an extension of an existing model-based PETC strategy for linear systems by including an (approximate) disturbance model. This extension can further enhance communication savings in the presence of disturbances. In addition, we evaluate the extended model-based PETC strategy by comparing this strategy to the standard model-based PETC and to a model-based periodic time-triggered control (PTTC) strategy. In this PTTC strategy, data is transmitted at fixed sampling times. For the evaluation, we present techniques for stability and l₂-gain performance analysis for both the PETC strategy and the PTTC strategy. Finally, the advantage of the (extended) PETC strategy over the PTTC strategy will be demonstrated by providing numerical examples.
16:40-17:00	WeC5.2
Input-To-State Stability of Event-Based State-Feedback Control, pp. 1145-1150
Stoecker, Christian	Ruhr-Univ. Bochum
Lunze, Jan	Ruhr-Univ. Bochum
This paper investigates the stability of event-based state-feedback loops. Two analysis methods are proposed. The first method proves the event-based state-feedback loop to be input-to-state practically stable which means that the state converges to a vicinity of the origin. The size of this region depends upon the magnitude of the disturbance and the event threshold, which is a design parameter. The second method is tailored for event-based state-feedback loops with stable plant dynamics and it shows that the investigated system is input-to-state stable, which implies that for small disturbance magnitudes the size of the region to which the state converges is independent of the event threshold. This new result shows that asymptotic stability can be achieved by means of an event-based controller with constant event threshold which has been proven in literature only for decreasing event thresholds. Both analysis methods are applied to a benchmark example and the results are compared with an analysis method known from literature which shows that the second proposed analysis method yields less conservative results with respect to the ultimate bound than existing methods in literature.
17:00-17:20	WeC5.3
Self-Triggered Stabilization of Continuous Stochastic State-Feedback Controlled Systems, pp. 1151-1155
Anderson, Ross	Univ. of California, Santa Cruz
Milutinovic, Dejan	Univ. of California at Santa Cruz
Dimarogonas, Dimos V.	Royal Inst. of Tech. (KTH)
Event-triggered and self-triggered control, in which the time of update to the controls is based on either current or outdated sampled data, have recently been employed to reduce the computational load or resource consumption for distributed real-time control systems. In this work, we propose a self-triggered scheme for nonlinear controlled stochastic differential equations with additive noise terms. A self-triggering update condition is derived that guarantees stability in the p-th moment of the state distribution. We show that the length of the times between controller updates as computed from the proposed scheme is strictly positive and provide examples.
17:20-17:40	WeC5.4
Event-Triggered Model Predictive Control of Discrete-Time Linear Systems Subject to Disturbances, pp. 1156-1161
Lehmann, Daniel	Ruhr-Univ. Bochum
Henriksson, Erik	KTH Royal Inst. of Tech.
Johansson, Karl Henrik	Royal Inst. of Tech.
This paper presents an approach to event-triggered model predictive control for discrete-time linear systems subject to input and state constraints as well as exogenous disturbances. Stability properties are derived by evaluating the difference between the event-triggered implementation and the conventional time-triggered scheme. It is shown that the event-triggered implementation, in stationarity, is able to keep the state in an explicitly computable set given by the disturbance bound and the event threshold. Simulation results underline the effectiveness of the proposed scheme in terms of reducing the communication and computational effort while guaranteeing a desired performance.
17:40-18:00	WeC5.5
Event-Triggered Consensus of Heterogeneous Multi-Agent Systems with Double-Integrator Dynamics, pp. 1162-1167
Xue, Dong	Tech. Univ. München
Hirche, Sandra	Inst. of Automatic Control Engineering
This paper studies the second-order consensus problem for multi-agent systems (MASs) associated with an event-triggered update policy. We first investigate the topological properties of the heterogeneous position and velocity information graph to guarantee consensus. Then, consensus protocols using an event-triggered update law are analyzed for the centralized and distributed cases. In particular, we establish an event-triggered scheduling policy ensuring that the MAS exponentially achieves consensus. Furthermore, a positive lower bound for inter-event times is derived to guarantee the absence of Zeno behavior. Finally, simulations are provided to demonstrate the effectiveness of proposed approach.
18:00-18:20	WeC5.6
On the Trade-Off between Control Performance and Communication Cost for Event-Triggered Control Over Lossy Networks, pp. 1168-1174
Demirel, Burak	KTH Royal Inst. of Tech.
Gupta, Vijay	Univ. of Notre Dame
Johansson, Mikael	Royal Inst. of Tech.
This paper develops a theoretical framework for quantifying the trade-off between communication cost and control performance in event-triggered control over lossy networks. We consider a system where the communication between controller and actuator is dictated by a threshold-based event-triggering algorithm, and develop a Markov-chain model that describes the attempted and successful transmissions of control messages over the lossy communication channel. A particular feature of our model is that it considers retransmissions of unsuccessful messages and that it accounts for the delay associated with such retransmissions. A systematic framework for analyzing the trade-off between the communication rate and control performance and for optimal tuning of the event threshold emanates by combining this model with an analytical model of the closed-loop performance. Numerical examples demonstrate the effectiveness of the proposed framework.

WeC6	HG E1.2
Iterative Control (Regular Session)
Chair: Camacho, Eduardo F.	Univ. of Sevilla
Co-Chair: Alleyne, Andrew G.	Univ. of Illinois at Urbana-Champaign
16:20-16:40	WeC6.1
Learning Control in Spatial Coordinates for the Path-Following of Autonomous Vehicles, pp. 1175-1180
Consolini, Luca	Univ. of Parma
Verrelli, Cristiano Maria	Univ. Di Roma
On the basis of a kinematic third order nonlinear model of an autonomous vehicle, a space-learning control is designed for the tracking of planar curves whose uncertain curvature is L-periodic in the curvilinear abscissa. The behaviour of a human driver, who repetitively learns the correct action from the past experience, is mathematically reproduced. Simulation results demonstrate the effectiveness of the presented approach.
16:40-17:00	WeC6.2
Optimal Iterative Learning Control Design with Trial-Varying Initial Conditions, pp. 1181-1186
Son, Tong Duy	KU Leuven
Pipeleers, Goele	KU Leuven
Swevers, Jan	KU Leuven
In this paper we present an approach to deal with trial-varying initial conditions in norm-optimal iterative learning control (ILC). Varying initial conditions generally degrade the performance of conventional learning algorithms. We therefore introduce a worst-case optimization problem that accounts for trial-varying of initial conditions. The optimization is then reformulated as a convex minimization problem, which can be solved efficiently to generate the control signal. We investigate the relationship between the proposed approach and classical norm-optimal ILC; where we find that our methodology is equivalent to classical norm-optimal ILC with trial-varying parameters. Finally, simulation results of the presented technique are given.
17:00-17:20	WeC6.3
Repetitive Control to Counteract the Effect of People on Thermal Comfort Control, pp. 1187-1191
Álvarez Hervás, José Domingo	Univ. of Seville
Costa-Castello, Ramon	Univ. Pol. de Catalunya (UPC)
Castilla, Maria Del Mar	Univ. OF ALMERÍA
Camacho, Eduardo F.	Univ. of Sevilla
People usually spend most of the time inside buildings. Therefore, it is necessary to reach an optimal thermal comfort situation since thermal comfort has a direct effect on people's productivity. The use of appropriate control strategies can highly contribute to this purpose. Usually the own people's influence is not taken into account at the time to maintain an optimal thermal comfort. People enter and leave the building following certain pattern which is repeated from one day to another. This paper presents a Repetitive Control (RC) approach which can counteract this periodic behaviour. This controller anticipates the effects produced by the entries and outputs of the people. Moreover, this controller is complemented by a feedback controller in order to cancel the non periodic disturbances. Simulation results obtained from the application of this control strategy to a characteristic room of a building are included and commented.
17:20-17:40	WeC6.4
Iterative Learning Control of the Electrostatic Microbridge Actuator, pp. 1192-1197
Cichy, Blazej	Univ. of Zielona Gora
Galkowski, Krzysztof	Univ. of Zielona Gora
Rauh, Andreas	Univ. of Rostock
Aschemann, Harald	Univ. of Rostock
In this paper, we consider the control of an electrostatic microbridge actuator with a distributed electrostatic force input and distributed displacement sensing. A bounded desired trajectory is achieved by using an Iterative Learning Control (ILC) method based on discrete linear repetitive processes with the use of Linear Matrix Inequalities (LMI). Space and time discretization is accomplished by using a finite difference approach based on the so-called Crank-Nicolson method.
17:40-18:00	WeC6.5
Learning/Repetitive Control for Building Systems with Nearly Periodic Disturbances, pp. 1198-1203
Vinther, Kasper	Aalborg Univ.
Chandan, Vikas	Univ. of Illinois at Urbana Champaign
Alleyne, Andrew G.	Univ. of Illinois at Urbana-Champaign
In this paper, learning/repetitive control is proposed for improvement of existing feedback control loops for temperature regulation in buildings. A single zone office building is used as an example, with real weather data for Phoenix Arizona and realistic occupancy load schedules. Simulations have shown a decrease in the average set point tracking error of more than 50%, even without additional energy consumption. This can be achieved in situations where the load disturbances have enough repeatability and a repeatable-to-nonrepeatable ratio can be computed to determine if learning should be used and at which frequencies. Furthermore, the increased tightness in reference tracking could be used to lower energy consumption by moving the reference set point closer to the boundaries of the allowable temperature range.
18:00-18:20	WeC6.6
Transformation-Based Iterative Learning Control for Non-Collocated Sensing of a Galvanometer Scanner, pp. 1204-1209
Yoo, Han Woong	Delft Univ. of Tech.
Ito, Shingo	Vienna Univ. of Tech.
Verhaegen, Michel	Delft Univ. of Tech.
Schitter, Georg	Vienna Univ. of Tech.
A transformation-based iterative learning control (ILC) approach is proposed to achieve accurate image scanning for the non-collocated dynamics of a galvanometer scanner. The non-collocation between the encoder and beam scanning mirror results in a tracking error of the actual beam position although the encoder measurement matches the reference signal. The proposed ILC is extended from the conventional ILC design by adding a reference transformation filter, which is based on the transfer functions between the measured and the controlled output. An error analysis shows that the proposed method can reduce the error of the actual controlled output, especially for applications of a large tracking reference such as image scanning. Experimental results with the proposed ILC show a better tracking accuracy as compared to conventional ILC design with non-collocated sensing.

WeC7	HG E3
Applications of Optimization and Optimal Control (Regular Session)
Chair: Giglio, Davide	Univ. of Genova
Co-Chair: Summers, Sean	ETH Zurich
16:20-16:40	WeC7.1
Optimal Control of Inventory-Production Systems in Presence of Perturbations on the Inventory Level, pp. 1210-1217
Giglio, Davide	Univ. of Genova
Siri, Silvia	Univ. of Genova
A set of optimal control strategies is determined in this paper for a class of inventory-production systems, in order to optimize the production process in presence of perturbations on the inventory level. The inventory-production system has to satisfy a time-varying external demand, which is assumed to be a piecewise constant function of time; the optimization involves both production speed and the time instants at which the speed is changed. The inventory level, which is the state of the system, is supposed to be subject to some perturbations such that, at some time instants, it can be lower or higher than expected. Starting from the optimal (open-loop) solution found by the authors in a previous work, in this paper the optimal production speed is determined as an optimal (closed-loop) control strategy, which is a function of the system state. Such strategy minimizes the extra-cost that is paid by the system in consequence of a perturbation.
16:40-17:00	WeC7.2
Optimizing the Anaerobic Digestion of Microalgae in a Coupled Process, pp. 1218-1223
Bayen, Térence	Univ. Montpellier 2
Mairet, Francis	Univ. Técnica Federico Santa María
Martinon, Pierre	INRIA
Sebbah, Matthieu	INRIA Sophia-Antipolis Mediterranee, UMR INRA-SupAgro
This work is devoted to maximizing the production of methane in a bioreactor coupling an anaerobic digester and a culture of micro-algae limited by light. The decision parameter is the dilution rate which is chosen as a control, and we enforce periodic constraints in order to repeat the same operation every day. The system is gathered into a three-dimensional system taking into account a day-night model of the light in the culture of micro-algae. Applying Pontryagin maximum principle, the necessary conditions on optimal trajectories indicate that the control consists of bang and/or singular arcs. We provide numerical simulations by both direct and indirect methods, which show the link between the light model and the structure of optimal solutions.
17:00-17:20	WeC7.3
Optimal Pedestrian Evacuation Using Model Predictive Control, pp. 1224-1229
Blom Västberg, Oskar	Royal Inst. of Tech. KTH
Hu, Xiaoming	Royal Inst. of Tech.
Dong, Hairong	Beijing Jiaotong Univ.
During an emergency in a building complex, an effective evacuation is essential to avoid crowd disasters. This article presents a route guiding that minimize the evacuation time during the evacuating of pedestrians from a building. A dynamic network model, namely the point queue model, is used to form a linear programming problem whose solution is used as an evacuation plan. By continuously solving the problem using the new state as initial data a feedback control law is derived using Model Predictive Control. The control law is tested on a microscopic pedestrian simulation for a building with five rooms and one respectively two exits. The simulation shows that the control law provides efficient and seemingly optimal routing, taking the predicted future distribution of pedestrians into account. The control law further manages to handle minor errors in the provided layout information.
17:20-17:40	WeC7.4
Collocation Methods for Multi-Vehicle Trajectory Optimization, pp. 1230-1235
Turnbull, Oliver David Naismith	Univ. of Bristol
Richards, Arthur	Univ. of Bristol
Direct collocation offers an efficient way of transcribing optimal control problems to form nonlinear optimizations. Collocation is particularly attractive for variable time problems as the finishing time can be made a decision variable. However, this causes problems in coupled multi-vehicle problems, for example, where different vehicles may have different finishing times. This paper proposes a way of capturing coupling constraints - in particular, collision avoidance - between vehicles without requiring a common time of arrival. The approach exploits a recently-developed dualization approach for avoidance constraints, extended to act time as well as spatial dimensions.
17:40-18:00	WeC7.5
Extension of the Global Optimization Using Multi-Unit Extremum Seeking Control for Noisy Scalar Systems, pp. 1236-1241
Esmaeilzadeh Azar, Farhad	Ec. Pol. de Montreal
Perrier, Michel	Ec. Pol.
An efficient global optimization method based on multi-unit extremum seeking has been proposed recently for noise-free nonlinear continuous static, scalar systems. Herein, a monotonically decreasing offset is introduced between the inputs of two identical units and the estimated gradient by finite difference is controlled to zero by an integrator. This paper is the extension of the mentioned algorithm to globally optimize scalar objective functions with the presence of noise. The algorithm is applied on several noisy global optimization test problems to show the capability of this methodology.
18:00-18:20	WeC7.6
Optimization-Based Autonomous Remote Sensing of Surface Objects Using an Unmanned Aerial Vehicle, pp. 1242-1249
Haugen, Joakim	Norwegian Univ. of Science and Tech.
Imsland, Lars	Norwegian Univ. of Science and Tech.
This manuscript presents an optimization-based approach for path planning of an aerial mobile sensor that monitors a set of moving surface objects. The purpose of the optimization problem is to obtain feasible mobile sensor trajectories with an objective to minimize the uncertainty of the objects, represented as the trace of the state estimation error covariance. The dynamic optimization problem is discretized into a large-scale nonlinear programming (NLP) problem using the direct transcription method known as simultaneous collocation. The numerical simulation periodically provides desired sensor trajectories and thus illustrates the approach.

WeC8	HG E5
Cooperative Control (Regular Session)
Chair: Zhang, Fumin	Georgia Inst. of Tech.
Co-Chair: Georges, Didier	Grenoble Inst. of Tech.
16:20-16:40	WeC8.1
Energy-Information Tradeoffs in Motion and Sensing for Target Localization, pp. 1250-1255
Wu, Wencen	Georgia Inst. of Tech.
Zhang, Fumin	Georgia Inst. of Tech.
Wardi, Yorai	Georgia Inst. of Tech.
Recent study in behavioral ecology reveal that certain species trade-off energy and information during various routine activities. This paper explores a similar tradeoff between sensing and motion of mobile agents tasked with target localization. Specifically, the goal is to balance the quality of information collected by sensing agents with the kinetic energy expended by their motion relative to the target. Following common practice, the determinant of the Fisher Information Matrix is chosen to represent the information-quality, while the kinetic energy is quantified by the usual Newton's laws. We argue that the quality of information has a form reminiscent of potential energy, and with such interpretation, we mimic the basic laws of mechanics by assuming the conservation of the total energy, which is the sum of the potential energy and kinetic energy. In this framework, we formulate an optimal path problem that aims at increasing the information quality while reducing the kinetic energy along the path of the agents. To solve this problem, we apply the calculus of variations, only to discover that the optimal solutions (paths) satisfy the Euler-Lagrange equations that are well known in classical mechanics. This realization suggests that the resulting trajectories may exhibit some well-known phenomena in classical mechanics. Simulation results support this hypothesis.
16:40-17:00	WeC8.2
Synthesis of an Asynchronous Communication Protocol for Search and Rescue Robots, pp. 1256-1261
Wiltsche, Clemens	Univ. of Oxford
Ramponi, Federico Alessandro	Univ. di Brescia
Lygeros, John	ETH Zurich
We present a protocol for reliable communication between search and rescue robots, which is synthesized from a high-level temporal logic specification. The protocol ensures provably correct data transmission on an asynchronous point-to-point link in the presence of an adverse environment. We synthesize the protocol into correct-by-construction transceiver controllers that can be included as building blocks in a larger design. The viability of our approach of synthesizing controllers with clearly defined interfaces and the validity of our protocol is demonstrated by implementing controllers for robots searching cooperatively for a moving target.
17:00-17:20	WeC8.3
A Bondgraph Approach to Formation Control Using Relative State Measurements, pp. 1262-1267
Stacey, Geoff	Australian National Univ.
Mahony, Robert	Australian National Univ.
Corke, Peter	CSIRO
In this paper we apply port-Hamiltonian theory with the bondgraph modelling approach to the problem of formation control using partial measurements of relative positions. We present a control design that drives a group of vehicles to a desired formation without requiring inter-vehicle communications or global position and velocity measurements to be available. Our generic approach is applicable to any form of relative measurement between vehicles, but we specifically consider the important cases of relative bearings and relative distances. In the case of bearings, our theory closely relates to the field of image-based visual servo (IBVS) control. We present simulation results to support the developed theory.
17:20-17:40	WeC8.4
Spatio-Temporal Symmetries in Linear Systems with an Application to Formation Control, pp. 1268-1273
Consolini, Luca	Univ. of Parma
Tosques, Mario	Univ. of Parma
We study a class of controlled linear systems characterized by a property of spatio-temporal symmetry. We study its stabilizability and detectability properties. We present a design technique to locally stabilize spatio-temporally symmetric orbits of equivariant systems, that has applications in the control of identical interconnected systems. As an example, we present the stabilization of a formation of unicycle robots in cyclic pursuit.
17:40-18:00	WeC8.5
Formation Control Design for Car--Like Nonholonomic Robots Using the Backstepping Approach, pp. 1274-1279
Sadowska, Anna	TU Delft
Huijberts, Henri	Queen Mary, Univ. of London
In this paper we study the formation control problem for car-like robots which may be viewed as a more general type of robots than the usually considered unicycle type robots. We develop a controller using the backstepping approach and give conditions solving the formation control problem as well as the coordination control problem. We also present simulation results to demonstrate the applicability of the proposed controller.
18:00-18:20	WeC8.6
Optimal Location of Mobile Sensors for Environmental Monitoring, pp. 1280-1285
Georges, Didier	Grenoble Inst. of Tech.
This paper is devoted to the optimal location of a mobile sensor network for environmental monitoring applications. Effective pollution monitoring in the atmosphere or in hydraulic systems (rivers, lakes or seas) with a limited number of sensors is a very challenging issue. In this paper, we will only consider physical phenomena governed by the so-called advection-diffusion partial-differential equation (ADPDE). The paper goal is to derive a simple potential-based navigation strategy ensuring the optimal location of a mobile sensor network. The optimal location of the sensors is determined through the maximization of an observability index based on the observability gramian of the ADPDE derived from a spectral method. The application to the control of a 20-sensor network applied to a 2D air pollution model demonstrates the effectiveness of the proposed approach. The here-proposed methodology can be easily applied to the optimal location of static sensors.

WeC9	HG E7
Mechatronic Systems III (Regular Session)
Chair: Ozbay, Hitay	Bilkent Univ.
Co-Chair: Dasdemir, Janset	Yildiz Tech. Univ.
16:20-16:40	WeC9.1
On the Robust Controller Design for Hard Disk Drive Servo Systems with Time Delays, pp. 1286-1291
Yan, Peng	Shandong Univ.
Ozbay, Hitay	Bilkent Univ.
Due to the existence of various sources of delays, the dynamical model of HDD (Hard Disk Drive) servo systems is actually infinite dimensional, although most of the control algorithms simplified the model with Pade expansions or other finite dimensional approximations. In this paper, a robust loop shaping algorithm is developed for the HDD model with delays by using an H-inf synthesis approach for infinite dimensional systems. The H-inf controller is derived with a structure of an internal feedback loop including an FIR (Finite Impulse Response) filter and an IIR (Infinite Impulse Response) filter, which facilitates non-fragile implementations. Comparisons to other robust control methods are given and the advantages of this approach are demonstrated in terms of improvement of TMR (track misregistration)and tracking TPI (Track-per-Inch) capability.
16:40-17:00	WeC9.2
Smith Predictor Based Generalized PI Control for a Class of Input Delayed Nonlinear Mechanical Systems, pp. 1292-1297
Ramirez-Neria, Mario	CINVESTAV-IPN
Sira-Ramirez, Hebertt J.	CINVESTAV-IPN
Luviano-Juárez, Alberto	UPIITA - IPN Mexico
Rodriguez-Angeles, Alejandro	CINVESTAV
This article addresses the problem of Linear Active Disturbance Rejection Control for a class of nonlinear mechanical input time delayed systems. To solve the problem, it is necessary to obtain a set of predictive lumped disturbance estimations, which is carried out by a purely linear high gain observer of Extended Luenberger type, denoted as a Generalized Proportional Integral (GPI) Observer. This task is performed using an approximation of the predictive disturbance using a truncated Taylor series expansion, where the set of necessary values for the estimation are provided by the observer. Once the (approximate) disturbance canceling is made, the control task is reduced to a simple linear Smith Predictor PD control, using the simplified model of the mechanical system. Some experimental results show the effectiveness of the strategy in a trajectory tracking task for a perturbed flywheel.
17:00-17:20	WeC9.3
Formation Control of Force-Controlled Mobile Robots in a Spanning-Tree Topology, pp. 1298-1303
Dasdemir, Janset	Yildiz Tech. Univ.
Loria, Antonio	CNRS
We solve the formation-tracking control problem for mobile robots via linear control, under the assumption that each agent communicates only with one “leader” robot and with one follower. We assume that the system is force- controlled (hence we use the dynamic model) as opposed to velocity-controlled (in the kinematic-model case). As in the classical tracking control problem for nonholonomic systems, the swarm is driven by a fictitious robot which moves about freely and which is leader to one robot only. For the case of a fixed spanning-tree topology we show that persistency of excitation on the velocity of the virtual leader is sufficient and necessary to achieve consensus tracking.
17:20-17:40	WeC9.4
Smooth Frequency Domain Parametric Optimization in Loop-Shaping Control, pp. 1304-1309
Lantto, Erkki	Sulzer Pump Solutions Finland
Hölttä, Vesa	Sulzer Pump Solutions Finland
Zenger, Kai	Aalto Univ. School of Electrical Engineering
Tommila, Ville	Sulzer Pump Solutions Finland
This paper presents two tools for frequency domain parametric optimization. The first tool is an approximation of the square of the largest singular value. This approximation works for complex matrices and enables using smooth constrained optimization methods, like sequential quadratic programming, in the multi-input-multi-output control system design. The second tool is a cost effective controller parametrization, which is based on the Hilbert transform. We also give a simple example to help the reader to get started with this synthesis framework.
17:40-18:00	WeC9.5
Metric Visual-Inertial Navigation System Using Single Optical Flow Feature, pp. 1310-1316
Omari, Sammy	ETH Zurich
Ducard, Guillaume	I3S, UMR7271, CNRS, Univ. de Nice Sophia Antipolis
This paper introduces a state estimation framework that allows estimating the attitude, full metric speed and the orthogonal metric distance of an IMU-camera system with respect to a plane. The filter relies only on a single optical flow feature as well as gyroscope and accelerometer measurements. The underlying assumption is that the observed visual feature lies on a static plane. The orientation of the observed plane is not required to be known a priori and is also estimated at runtime. The estimation framework fuses visual and inertial measurements in an Unscented Kalman Filter (UKF). The theoretical limitations of the UKF are investigated using a nonlinear observability analysis based on Lie-derivatives. Simulations using realistic sensor noise values successfully demonstrate the performance of the filter as well as validate the findings of the observability analysis. It is shown that the state estimate is converging correctly, even in presence of substantial initial state errors. To the authors' knowledge, this paper documents for the first time the estimation of the heading and metric distance to a wall with no range sensor, relying solely on optical flow as the only exteroceptive sensing modality.
18:00-18:20	WeC9.6
Resonance-Shifting Integral Resonant Control Scheme for Increasing the Positioning Bandwidth of Nanopositioners, pp. 1317-1322
Namavar, Mohammad	Univ. of Aberdeen
Fleming, Andrew J.	Univ. of Newcastle
Aphale, Sumeet	Univ. of Aberdeen
The performance of precision mechatronic systems is restricted by their first dominant resonant mode. Damping techniques have be employed to suppress this resonance peak and improve the performance. To increase the bandwidth of the system feed-forward techniques have been used but they can be very sensitive to modeling errors as well as loading effects. In this paper a simple frequency shifting controller is introduced and is combined with Integral Resonant Control (IRC). Using these controllers can increase the bandwidth of the system to a desired amount. Systems with colocated sensor-actuator pairs exhibit the interesting property of pole-zero interlacing. IRC exploits this property by changing the pole-zero interlacing to zero-pole interlacing. The unique phase response of this class of systems enables a simple integral feedback controller to add substantial damping.

WeC10	HG D1.1
Linear Time-Varying Systems (Regular Session)
Chair: Mourllion, Benjamin	Univ. de Haute-Alsace
Co-Chair: Svaricek, Ferdinand	Univ. of the German Armed Forces, Munich
16:20-16:40	WeC10.1
Young Duality and Schmidt-Pair for Linear Systems, pp. 1323-1328
Mourllion, Benjamin	Univ. de Haute-Alsace
Birouche, Abderazik	Univ. de Haute Alsace
This paper studies the concepts of reachability, observability, controllability and constructibility focusing on the principle of duality with time reversal. This work takes into account linear time-varying systems in continuous time. A discussion is proposed about the Legendre transforms of energy functions and the significance of Schmidt-pairs.
16:40-17:00	WeC10.2
Solving Discrete-Time Game Theoretic Periodic Riccati Equations: An Iterative Procedure, pp. 1329-1334
Dragan, Vasile	Romanian Acad.
Aberkane, Samir	UHP, NANCY 1
Ivanov, Ivan	Sofia Univ. "St.Kl.Ohridski"
This paper addresses the problem of solving a class of periodic discrete-time Riccati equation with an indefinite sign of it's quadratic term. More precisely, we focus on the computation of the stabilizing solutions of discrete-time game theoretic periodic Riccati equations. A convergent iterative algorithm is proposed for this purpose.
17:00-17:20	WeC10.3
Necessary Conditions for Structural and Strong Structural Controllability of Linear Time-Varying Systems, pp. 1335-1340
Hartung, Christoph	Univ. of the German Armed Forces, Munich
Reissig, Gunther	Univ. of the Federal Armed Forces Munich
Svaricek, Ferdinand	Univ. of the German Armed Forces, Munich
In this paper, the notion of structural controllability of linear time-invariant systems is extended to the time-varying case dx/dt = A(t) x(t) + B(t) u(t). We provide two Examples which show that neither the conditions for structural controllability of time-invariant systems are necessary, nor the conditions for strong structural controllability of time-invariant systems are sufficient for the controllability of time-varying systems. We present a necessary condition for structural controllability of linear time-varying systems and in our main result a necessary condition for strong structural controllability of linear time-varying systems is given. In a previous work, this necessary condition for strong structural controllability of linear time-invariant systems was shown to be sufficient, so that the strong structural controllability of linear time-invariant systems is now characterized. We want to emphasize, that our results cover the single and the multi-input case.
17:20-17:40	WeC10.4
Derivative Based Control for LTV System with Unknown Parameters, pp. 1341-1346
Maalej, Sonia	LAGIS, CNRS, Ec. LILLE
Kruszewski, Alexandre	Ec. Centrale de LILLE
Belkoura, Lotfi	Univ. des Sciences et Tech. de Lille (Lille 1)
This paper deals with the robust stabilization of a class of linear time varying systems. Instead of using a state observer or having dynamic structure, the controller is based on output derivative estimation. This allows the stabilization of linear time varying systems with very large parameter variation and a small number of controller parameters. The proof of stability is based on the polytopic representation of the closed loop and Lyapunov conditions. The result is proposed in a LinearMatrix Inequality (LMI) form. The validity of this approach is illustrated by a second order system case of study.
17:40-18:00	WeC10.5
Model Invalidation for Repeated $ell_1$-Bounded Linear Time-Varying Uncertainty Models, pp. 1347-1352
Rödönyi, Gábor	Comp. and Automation Res. Inst.
Gaspar, Peter	Computer & Automation Inst. of HAS
Conditions are derived for the existence of a repeated linear time-varying (LTV) $ell_1$-bounded perturbation and some $ell_infty$-bounded disturbance signals that are able to describe experimental data. The uncertainty structure is motivated by formation control problems where complex dynamic systems with similar dynamics are moving together and safety can be measured in terms of the relative distances. It is shown that by solving linear programming problems a disturbance-perturbation trade-off curve useful for unfalsified robust performance computation can be determined. The invalidation results are demonstrated on the simulation example of a homogeneous vehicle platoon.
18:00-18:20	WeC10.6
A Discrete Time-Varying Internal Model Based Approach for High Precision Tracking, pp. 1353-1358
Zhang, Zhen	Tsinghua Univ.
Lu, Chao	Univ. of South Carolina
Leng, Tongtong	Shandong Univ.
Yan, Peng	Beihang Univ.
Ye, Peiqing	Tsinghua Univ.
In this paper, we consider the discrete time case of a time-varying internal model-based control design for high precision tracking of frequency-varying reference trajectories. Thanks to a recently proposed parallel time-varying internal model structure, the asymptotic tracking conditions for the design of internal model units are developed, and a low order robust stabilizer is synthesized. In a discrete time setting, the high precision tracking control architecture is deployed on a Voice Coil Motor (VCM) actuated servo gantry system, where various numerical results are given to validate the proposed method.

WeC11	HG D1.2
Predictive Control of UAVs (Regular Session)
Chair: Kerrigan, Eric C.	Imperial Coll. London
Co-Chair: Summers, Tyler H.	ETH Zurich
16:20-16:40	WeC11.1
Predictive Control for Spacecraft Rendezvous in an Elliptical Orbit Using an FPGA, pp. 1359-1364
Hartley, Edward Nicholas	Univ. of Cambridge
Maciejowski, Jan M.	Univ. of Cambridge
A field programmable gate array (FPGA)-based predictive controller for a spacecraft rendezvous manoeuvre is presented. A linear time varying prediction model is used to accommodate elliptical orbits, and a variable prediction horizon is used to facilitate finite time completion of manoeuvres. The resulting constrained optimisation problems are solved using a primal dual interior point algorithm. The majority of the computational demand is in solving a set of linear equations at each iteration of this algorithm. To accelerate this operation, a custom circuit is implemented, using a combination of Mathworks HDL Coder and Xilinx System Generator for DSP, and used as a peripheral to a MicroBlaze soft core processor. The system is demonstrated in closed loop by linking the FPGA with a simulation of the plant dynamics running in Simulink on a PC, using Ethernet.
16:40-17:00	WeC11.2
Nonlinear Predictive Control of Autonomous Soaring UAVs Using 3DOF Models, pp. 1365-1370
Liu, Yuyi	Imperial Coll. London
Longo, Stefano	Cranfield Univ.
Kerrigan, Eric C.	Imperial Coll. London
We design a nonlinear model predictive control (NMPC) system for a soaring UAV in order to harvest the energy from the atmospheric updrafts. Our control framework combines an online estimation with a heuristic search method to obtain the UAV optimal trajectory. To allow for real-time computation of the control commands we solve the optimal control problem using a 3 degrees-of-freedom (DOF) model but apply the inputs to a more realistic 6DOF model. Hence, we design a 3DOF-6DOF model interaction strategy. Simulations show how the control system succeeds in energy extraction in a challenging dynamic atmospheric environment while satisfying its real-time contraints.
17:00-17:20	WeC11.3
Trajectory-Tracking and Path-Following Controllers for Constrained Underactuated Vehicles Using Model Predictive Control, pp. 1371-1376
Alessandretti, Andrea	Inst. Superior Técnico - École Pol. Fédérale de Lau
Aguiar, A. Pedro	Faculty of Engineering, Univ. of Porto (FEUP)
Jones, Colin N	EPFL, Lausanne
This paper addresses the design of Model Predictive Control (MPC) laws to solve the trajectory-tracking problem and the path-following problem for constrained underactuated vehicles. By allowing an arbitrarily small asymptotic tracking error, we derive MPC laws where the size of the terminal set is only limited by the size of the system constraints. In fact, for the case of unconstrained inputs, the terminal set can be neglected and the resulting MPC controllers provide a global solution to the addressed constrained motion control problems. Simulation results are presented where the proposed MPC controllers are applied to 2-D and to 3-D moving vehicles.
17:20-17:40	WeC11.4
Fast Model Predictive Control of Miniature Helicopters, pp. 1377-1382
Kunz, Konstantin	ETH Zurich
Huck, Stephan Marc	ETH Zurich
Summers, Tyler H.	ETH Zurich
Model Predictive Control (MPC) is a well-developed and widely-used control design method, in which the control input is computed by solving an optimization problem at every sampling period. Traditionally, MPC has been associated with control of slow processes, with sampling times in the seconds/minutes/hours range, because an optimization problem must be solved online. However, dramatic increases in computing power and recent developments in code generation for convex optimization, which tailor to specific optimization problem structure, are allowing the use of MPC in fast processes, with sampling times in the millisecond range. In this paper, a MPC control design for a miniature remote-controlled coaxial helicopter is developed and experimentally validated. The nonlinear dynamic behavior of the helicopter was identified, simplified and approximated by a Linear Time Varying (LTV) model. A custom convex optimization solver was generated for the specific MPC problem structure and integrated into a controller, which was tested in simulation and implemented on a hardware testbed. A performance analysis shows that the MPC approach performs better than a tuned Proportional Integral Differential (PID) controller.
17:40-18:00	WeC11.5
A Model Predictive Controller for Quadrocopter State Interception, pp. 1383-1389
Mueller, Mark Wilfried	ETH Zurich
D'Andrea, Raffaello	ETH Zurich
This paper presents a method for generating quadrocopter trajectories in real time, from some initial state to a final state defined by position, velocity and acceleration in a specified amount of time. The end state captures the attitude to within a rotation about the thrust axis. Trajectory generation is done by formulating the trajectory of the quadrocopter in its jerk, in discrete time, and then solving a convex optimisation problem on each decoupled axis. Convex bounds are derived to include feasibility constraints with respect to the quadrocopter's total allowable thrust and angular rates.
18:00-18:20	WeC11.6
Waypoint Trajectory Planning in the Presence of Obstacles with a Tunnel-MILP Approach, pp. 1390-1397
Afonso, Rubens Junqueira Magalhães	ITA - Inst. Tecnológico de Aeronáutica
Galvão, Roberto Kawakami Harrop	Inst. Tecnológico de Aeronáutica
Kienitz, Karl Heinz	Inst. Tecnologico de Aeronautica
This work presents a waypoint trajectory planning technique for an autonomous vehicle in the presence of obstacles using a tunnel-MILP formulation for the avoidance constraints. Predictive Control is used to address the issues of dynamic constraint satisfaction and obstacle avoidance. However, the complexity of the optimization problem to be solved may escalate as the number of obstacles increases. To circumvent this issue, a tunnel-MILP approach is employed. Even so, the optimization problem may still be too complex, i. e., involve a large number of decision variables, to be computationally tractable within the relatively small sample time required by vehicle guidance applications. The number of decision variables is reduced via the pre-computation of waypoints during an off-line trajectory planning phase. In this manner, during the on-line control phase, the optimization problem to be solved needs only to compute a control solution to reach the next waypoint in the sequence, instead of the whole control solution to reach the target set from the current position. Simulation results are presented to show that the employment of the waypoint trajectory planning technique brings about benefits regarding the computational burden associated to the solution of the on-line optimal control problem.

WeC12	HG D3.2
Identification for Control (Regular Session)
Chair: Wahlberg, Bo	KTH
Co-Chair: Van den Hof, Paul M.J.	Eindhoven Univ. of Tech.
16:20-16:40	WeC12.1
Dual Control Approach for Zone Model Predictive Control, pp. 1398-1403
Zacekova, Eva	Department of Control Engineering, FacultyofElectricalEngineerin
Privara, Samuel	CzechTechnicalUniversityinPrague,FacultyofElectricalEngineering
Vana, Zdenek	Department of Control Engineering, FacultyofElectricalEngineerin
Cigler, Jiri	Czech Tech. Univ. in Prague, Faculty ofElectricalEngine
Over the last few years, model based predictive controller (MPC) has gained popularity in many industrial fields one of which is the building climate control. In order to work properly, the MPC needs a model of a controlled system which describes the reality as accurate as possible. In practice, the model used by the MPC often becomes inapplicable due to the change of either the operating point or other conditions which leads to control performance degradation. In such a situation, it is inevitable to re-identify the model. However, in a majority of cases, the data available for the re-identification are from the closed-loop and they do not contain enough information for the successful re-identification of the model. In this paper, a dual control algorithm based on the maximization of the smallest eigenvalue of the information matrix increase ensuring both the appropriately informative data and satisfaction of the control performance is presented. As the area of the interest is the building climate control, we offer the dual control algorithm for a specific class of the zone MPC which is widely used in this field.
16:40-17:00	WeC12.2
Physics-Based Modeling and Identification for HVAC Systems, pp. 1404-1409
Scotton, Francesco	KTH Royal Inst. of Tech.
Huang, Lirong	Swiss Federal Inst. of Tech.
Ahmadi, Seyed A.	KTH Royal Inst. of Tech.
Wahlberg, Bo	KTH Royal Inst. of Tech.
Heating, ventilation and air conditioning (HVAC) systems are among the largest energy consumers in many buildings. As is known, modeling and identification play important roles in the study of HVAC systems. A good model is very helpful for improving efficiency of the HVAC system. Very recently, a physics-based model of room temperature was proposed. Motivated by this inspiring work, this paper, based on the physical dynamical systems, proposes, identifies and validates three models for CO$_2$ concentration, temperature and humidity of a test-bed room, respectively. Particularly, our models take into account the effect of occupiers, since the indoor air quality (IAQ) is evidently affected by the number of occupiers. A test-bed has been set up for experiments in a laboratory room on KTH campus. Experimental results verify that our proposed method improves the performance of the physics-based linear parametric models.
17:00-17:20	WeC12.3
Model-Based and Data-Driven Model-Reference Control: A Comparative Analysis, pp. 1410-1415
Formentin, Simone	Pol. di Milano
van Heusden, Klaske	Univ. of British Columbia
Karimi, Alireza	Ec. Pol. Federale
In the scientific literature, two main approaches have been proposed for control system design from data. In the “model-based” approach, a model of the system is first derived from data and then a controller is computed based on the model. In the “data-driven” approach, the controller is directly computed from data. In this work, the above approaches are compared from a novel perspective. The main finding of the paper is that, although from the standard perspective of parameter variance analysis the model-based approach is always statistically more efficient, the data-driven controller might outperform the model-based solution for what concerns the final control cost.
17:20-17:40	WeC12.4
Robust and Adaptive Excitation Signal Generation for Input and Output Constrained Systems, pp. 1416-1421
Hägg, Per	KTH Royal Inst. of Tech.
Larsson, Christian A.	KTH Royal Inst. of Tech.
Hjalmarsson, Håkan	Royal Inst. of Tech.
Generating signals with a prespecified autocorrelation, available from some input design, while satisfying constraints on input and output signals of the systems which the signal is to be applied on is an important problem in system identification. This paper extends a recently proposed method for such signal generation. The method is modified to be robust to uncertainties in the system model. An adaptive formulation is also given which allows for improving the input design online. The adaptive method is compared to the true optimal input design on a simulation example. The robustness properties of the method are illustrated on an experimental setup.
17:40-18:00	WeC12.5
Predictor Input Selection for Two Stage Identification in Dynamic Networks, pp. 1422-1427
Dankers, Arne Geert	Delft Univ. of Tech.
Van den Hof, Paul M.J.	Eindhoven Univ. of Tech.
Bombois, Xavier	Delft Univ. of Tech.
Heuberger, Peter	Eindhoven Univ. of Tech.
Recently, the Two-Stage method has been proposed as a tool to obtain consistent estimates of modules embedded in dynamic networks. However, for this method the variables that are included in the predictor model are currently not considered as a user choice. In this paper it is shown that there is considerable freedom as to which variables can be included in the predictor model as inputs, and still obtain consistent estimates of the module of interest. Conditions that the choice of predictor inputs must satisfy are presented. The conditions could be used to find the smallest number of predictor inputs for instance. Algorithms are presented for checking the conditions and obtaining the estimates.
18:00-18:20	WeC12.6
Consistency Analysis of Orthogonal Projection Based Closed-Loop Subspace Identification Methods, pp. 1428-1432
Liu, Tao	Dalian Univ. of Tech.
Shao, Cheng	Dalian Univ. of Tech.
Based on the orthogonal projection approach, a few closed-loop subspace identification methods have been developed recently in the literature. To guarantee consistent estimation of these methods against measurement noise, this paper analyzes the rank condition for performing the corresponding singular value decomposition (SVD) involved in deriving the extended observability matrix and the lower triangular block-Toeliptz matrix of the plant state-space model, together with the requirement on the input excitation signal, such that no mistake will be made in picking up the corresponding eigenvalues for numerical computation. Also, no knowledge of the closed-loop controller is required for the computation of the model state matrices. An example from the literature is used to demonstrate the proposed conditions for consistent estimation.

WeC13	HG D5.2
Fault Detection and Identification III (Regular Session)
Chair: Merzouki, Rochdi	Ec. Pol. de Lille
Co-Chair: Sen, Siddhartha	Indian Inst. of Tech. Kharagpur
16:20-16:40	WeC13.1
A H-Infinity Approach to Robust Fault Estimation of Non-Linear Discrete-Time Systems, pp. 1433-1438
Witczak, Marcin	Univ. of Zielona Gora
Korbicz, Jozef	Univ. of Zielona Gora
Jozefowicz, Rafal	Univ. of Zielona Gora
The paper deals with the problem of robust fault estimation of non-linear discrete-time systems. In particular, it is shown how to employ the unknown input observer approach and the H-infinity strategy to design a robust fault estimation filter. The proposed approach is designed in such a way that a prescribed disturbance attenuation level is achieved with respect to the fault estimation error while guaranteeing the convergence of the observer. The resulting design procedure is relatively simple and boils down to solving a set of linear matrix inequalities, which can be efficiently achieved with modern computational packages. The final part of the paper presents an illustrative example which exhibits the performance of the proposed approach.
16:40-17:00	WeC13.2
Robust Fault Detection for Systems with Electronic Induced Delays: Application to the Rendezvous Phase of the MSR Mission (I), pp. 1439-1444
Fonod, Robert	Univ. of Bordeaux 1
Henry, David	Univ. Bordeaux I
Bornschlegl, Eric	European Space Res. and Tech. Centre
Charbonnel, Catherine	Thales Alenia Space
Two robust fault detection schemes are presented to detect faults affecting the thrust system of the chaser spacecraft involved in the rendezvous phase of the Mars Sample Return (MSR) mission. The idea of both proposed methods is to transform the unstructured uncertainty caused by the electronic induced delays into unknown inputs and decouple them by means of an eigenstructure assignment (EA) technique. The first method utilizes a Cayley-Hamilton theorem based transformation whereas the second relies on a first-order Pade approximation of the time delay. The performances of the proposed schemes are compared by a sensitivity/robustness analysis campaign of 4240 runs within the "high-fidelity" industrial simulator provided by Thales Alenia Space.
17:00-17:20	WeC13.3
Application of Principal Components Analysis to Improve Fault Detection and Diagnosis on Semiconductor Manufacturing Equipment (I), pp. 1445-1450
Thieullen, Alexis	LSIS - STMicroelectronics
Ouladsine, Mustapha	Univ. d'aix marseille III
Pinaton, Jacques	STMicroelectronics
With the evolutions in sensing technologies and the increasing use of advanced process control techniques, terabytes of data are recorded today during the manufacturing process of semiconductor devices. These large amount of data are then operated by Fault Detection and Classification (FDC) systems to assess the overall condition of production equipment. However, specific characteristics of semiconductor manufacturing suc as highly correlated parameters, time-varying behaviors, or the large number of operating conditions tend to limit the efficiency of current indicators to detect and diagnose a failure occurence. There is therefore a significant requirement for the development and application of new methodologies to improve detection efficiency while reducing the complexity of condition monitoring, without losing detailed insight for efficient failure analysis. In this paper, we use data pretreatment algorithms from signal processing and time series analysis, and Multiway Principal Components Analysis (MPCA) methods to accurately represent equipment behavior and process dynamics and thus overcome issues inherent to semiconductor manufacturing context. A realcase application on a plasma etcher from STMicroelectronics Rousset 8' fab is proposed to highlight benefits of these methods.
17:20-17:40	WeC13.4
Data-Driven Prognostics Based on Health Indicator Construction: Application to PRONOSTIA's Data (I), pp. 1451-1456
Medjaher, Kamal	FEMTO-ST / ENSMM
Zerhouni, Noureddine	FEMTO-ST Inst. UMR CNRS 6174 - UFC / ENSMM / UTBM
Baklouti, Jihène	FEMTO-ST Inst.
Failure prognostics can help improving the availability and reliability of industrial systems while reducing their maintenance cost. The main purpose of failure prognostics is the anticipation of the time of a failure by estimating the Remaining Useful Life (RUL). In this case, the fault is not undergone and the estimated RUL can be used to take appropriate decisions depending on the future exploitation of the industrial system. This paper presents a data-driven prognostic method based on the utilization of signal processing techniques and regression models. The method is applied on accelerated degradations of bearings performed under the experimental platform called PRONOSTIA. The purpose of the proposed method is to generate a health indicator, which will be used to calculate the RUL. Two acceleration sensors are used on PRONOSTIA platform to monitor the degradation evolution of the tested bearings. The vibration signals related to the degraded bearings are then compared to a nominal vibration signal of a nondegraded bearing (nominal bearing). The comparison between the signals is done by calculating a correlation coefficient (which is considered as the health indicator). The values of the calculated correlation coefficient are then fitted to a regression model which is used to estimate the RUL.
17:40-18:00	WeC13.5
Merging Bond Graph and Signed Directed Graph to Improve FDI Procedure (I), pp. 1457-1462
Chatti, Nizar	Pol. Lille
Ould Bouamama, Belkacem	Pol. Lille
Gehin, Anne-Lise	LAGIS
Merzouki, Rochdi	Ec. Pol. de Lille
The Fuel Cell (FC) is an ideal electrical power source. However, FC stacks and even more FC systems are vulnerable to faults (such as water flooding and membrane drying) that can cause the disruption or the permanent damage. To guarantee the safe operation of the FC systems, it is necessary to use systematic techniques to detect and isolate faults for the purpose of diagnosis. The problematic for the model-based Fault Detection and Isolation (FDI) of fuel cell is that the model is complex because of coupling multiple physical domains (electrochemical, electrical, thermofluidic...). This is why, we propose in this paper, the exploitation of the behavioral and structural properties of the Bond Graph (BG) as a multi domain power exchange and unified graphical modeling language for qualitative analysis of monitoring ability (using Signed Directed Graph properties). This is obtained after generation of the fault indicators from one part, and by dealing with an automatically built Signed Directed Graph (SDG) of the system, from another part. By combining qualitative method (based on Signed Graph) and quantitative method (fault indicator generation) using only one representation, an innovative approach to perform (single and multiple faults) diagnosis is proposed. The proposed contribution is illustrated by an application to a Proton Exchange Membrane Fuel Cell (PEMFC).
18:00-18:20	WeC13.6
A Reconfigurable Direct Control Allocation Method, pp. 1463-1468
Naskar, Asim Kr	Indian Inst. of Tech. Kharagpur
Patra, Sourav	Indian Inst. of Tech. Kharagpur
Sen, Siddhartha	Indian Inst. of Tech. Kharagpur
In the event of failure of actuators in an overactuated system a reconfigurable law is required to redistribute and coordinate the control effort among the remaining healthy actuators. In this paper, a modified lookup table based Direct Control Allocation method is proposed for reconfiguration of actuator command, for a single actuator fault, without updating the lookup tables. The proposed method saves both memory as well as time to prepare the tables. The method of reconfigurable Direct Control Allocation is then applied to a satellite launch vehicle model. The simulation results show satisfactory performance of the proposed method in the presence of a single Stuck-at-One type actuator fault.

WeC14	HG D7.1
Automotive Suspension and Traction Control (Regular Session)
Chair: Aschemann, Harald	Univ. of Rostock
Co-Chair: Göhrle, Christoph	Univ. of Stuttgart
16:20-16:40	WeC14.1
Suspension Control Strategy for a Fully Electrified Vehicle, pp. 1469-1474
Tudon-Martinez, Juan Carlos	Tecnológico de Monterrey
Varrier, Sébastien	Gipsa-Lab.
Morales-Menendez, Ruben	Tecnológico de Monterrey, Campus Monterrey
Ramírez-Mendoza, Ricardo A.	TEC de Monterrey, Campus Ciudad de Mexico
The design of chassis for a truck considers a wide set of conditions in which load and unload are the most typical variety. For years, flat ride rules have been applied to achieve comfort, by providing tuning rules in the suspension as today. For an electric vehicle however, the weight distribution is different as a result of power train and batteries. A suspension control strategy is proposed in a full size pick-up truck using magneto-rheological dampers. As a first principle, the estimation of the frequency of the road profile allows the control of comfort and road holding in normal driving situations (straight roads). This strategy includes the coupling joints among the four-wheel stations of the truck and it allows the control of vertical forces in order to minimize the chassis vertical acceleration. A second principle, it uses the continuous measurement of the steering wheel to configure the four semiactive damping forces in order to reduce the wheel tramp, wheel hop, front-end dive, and other causes of vehicle instability under risk driving conditions. Simulation results in CarSimTM of a full prototype show that the suspension performance of the new fully electrified truck is better in comfort (reduction up to 60% in the chassis vertical acceleration) and handling (reduction around 5% in the vehicle slip angle) than the original truck (before its modification).
16:40-17:00	WeC14.2
Hysteresis Compensation and Adaptive LQR Design for an Electro-Pneumatic Clutch for Heavy Trucks, pp. 1475-1480
Aschemann, Harald	Univ. of Rostock
Prabel, Robert	Univ. of Rostock
Schindele, Dominik	Univ. of Rostock
In this paper, a nonlinear model-based control design for an electro-pneumatic clutch for heavy trucks is presented. A clutch is required at start-up or during gear shifts to disconnect the combustion engine from the gear box. This automated actuator disburdens the driver and provides the necessary actuation force. The proposed feedback control represents an adaptive LQR design using extended linearisation techniques, where the controller gains are adapted by solving numerically the corresponding Riccati equation in real-time. In addition, an adaptive feedforward control is designed with the clutch position as controlled variable. The control structure is extended by a reduced-order observer that estimates an effective pressure, which accounts for model uncertainty as well as nonlinear friction. Moreover, an experimentally identified Bouc-Wen hysteresis model is used to improve the dynamic behaviour further. Thereby, high tracking accuracy is achievable for the piston position. The performance of the proposed control structure is pointed out by experimental results from a dedicated test rig.
17:00-17:20	WeC14.3
Variable-Geometry Suspension Design in Driver Assistance Systems, pp. 1481-1486
Nemeth, Balazs	MTA SZTAKI
Gaspar, Peter	Computer & Automation Inst. of HAS
The paper proposes variable-geometry suspension design to enhance road stability during vehicle maneuvers. The orientation of wheels is modified by a suspension actuator, which results in an additional steering angle and a camber angle. A detailed analysis shows that the variable-geometry suspension system affects both the steering and the camber angle. Consequently, the integration of steering and wheel tilting can be handled by the variable-geometry suspension system. It is also shown that the suspension construction affects the control design. The control system must guarantee various vehicle performances such as trajectory tracking, roll stability and geometry limits. The operation of the designed control system is illustrated through simulation examples.
17:20-17:40	WeC14.4
A LPV Suspension Control with Performance Adaptation to Roll Behavior, Embedded in a Global Vehicle Dynamic Control Strategy, pp. 1487-1492
Fergani, Soheib	GIPSA-Lab. Control Systems Dept, Grenoble Univ.
Sename, Olivier	INPG
Dugard, Luc	CNRS-INPG
This paper is concerned with a new LPV/Hinf coordination strategy that improves the vehicle stability using suspension, active steering and electro-mechanical braking actuators. The vehicle stability is evaluated through a stability index and a stability region, based on slip dynamics, that prove the efficiency of the proposed strategy. The new coordination technic aims, by monitoring the load transfer distribution of the vehicle while facing road irregularities, to tune the suspensions in the four corners of the vehicle and improve vertical performances. At the same time, lateral and longitudinal car dynamics are improved using braking and steering actuators scheduled through the stability index, based on slip dynamics. The proposed GCC (Global Chassis Control) provides a load allocation on the four corners of the vehicle, ensures a good coordination between different actuators, and improves the vehicle stability and car dynamics. Simulations are performed on a complex nonlinear full vehicle model, subject to critical driving situations, and show the reliability and the robustness of the proposed solution.
17:40-18:00	WeC14.5
Improving High Speed Road-Holding Using Actively Controlled Aerodynamic Surfaces, pp. 1493-1498
Corno, Matteo	Pol. di Milano
Bottelli, Stefano	Pol. di Milano
Panzani, Giulio	Univ. degli Studi di Trento
Tanelli, Mara	Pol. di Milano
Spelta, Cristiano	Univ. degli Studi di Bergamo
Savaresi, Sergio M.	Pol. Di Milano
Semi-active and active suspension technologies have shown that vehicles road holding can be enhanced, but any such improvement implies decreasing the overall comfort. In this paper, we introduce the use of active aerodynamic surfaces to improve vehicle road-holding without affecting comfort. An H infinity closed-loop controller is developed to minimize the tire deflection variations, which are responsible for the vehicle bad road holding. The controller is designed on a quarter-car model. Extensive simulations on a multi-body vehicle simulation suit assess the performance of the proposed control scheme, and an exhaustive sensitivity analysis shows the role of road roughness, velocity and airfoil design.
18:00-18:20	WeC14.6
Model Predictive Control of Semi-Active and Active Suspension Systems with Available Road Preview, pp. 1499-1504
Göhrle, Christoph	Univ. of Stuttgart
Schindler, Andreas	Audi AG
Wagner, Andreas	Audi AG
Sawodny, Oliver	Univ. of Stuttgart
Semi-active and active suspensions influencing the vertical dynamics of vehicles enable improvement of ride comfort and handling characteristics compared to vehicles with passive suspensions. If the road height profile in front of the car is measured using vehicle sensors, more efficient control strategies can be applied. First, regarding vehicles with continuously variable dampers: A Model Predictive Controller incorporating the nonlinear constraints of the damper characteristic is set up and compared to a controller using linear constraints. The approximate linear constraints are obtained by a prediction of passive vehicle behavior over the preview horizon using a linearized model. Additionally, a controller without input constraints and clipping of the damper force is applied. This results in a quadratic program without constraints, which can be solved efficiently. The result of applying the optimal force without constraints is also evaluated, which corresponds to an ideal high bandwidth actuator. Next, two controllers for vehicles with a low bandwidth active suspension and variable dampers are proposed. While the first approach optimizes only the actuator displacement combined with the damper's soft characteristic, the second approach optimizes both the damper force and the actuator displacement. Simulation results of the controllers and the active and semi-active suspensions over real road height profiles are presented.

WeC15	HG D7.2
Emerging Control Problems in Power Electronics Systems (Invited Session)
Chair: Mariethoz, Sebastien	ETH Zurich
Co-Chair: Warrington, Joseph	ETH Zurich
Organizer: Mariethoz, Sebastien	ETH Zurich
16:20-16:40	WeC15.1
Stability Analysis of Droop Controlled Inverters Via Dynamic Phasors and Contraction Theory (I), pp. 1505-1510
Mariani, Valerio	Univ. degli Studi del Sannio
Vasca, Francesco	Univ. of Sannio
Droop techniques are widely used in distributed generation systems for the control of parallel inverters operating in grid-connected and islanded modes. Droop controllers without communication among the units are based on local measurements of active and reactive powers and usually allow the synchronization of the inverters to the common bus. On the other hand, instabilities can also occur. This paper investigates the stability of grid-connected droop controlled inverters by tackling the problem with a large signal perspective. A dynamic phasor model represented in a reference frame synchronous with the inverter voltage is proposed. The contraction theory applied to the model allows to determine an estimate of the domain of attraction of the stable equilibrium point. Simulation results demonstrate the effectiveness of the proposed analysis.
16:40-17:00	WeC15.2
Multi-Core Parallelisation of Integer Optimisation Model Predictive Control for Power Electronic Applications (I), pp. 1511-1516
Peyrl, Helfried	ABB Corp. Res.
Richter, Stefan	ABB
Zanarini, Alessandro	ABB Schweiz AG - Corp. Res. Center, Segelhofstrasse 1K,
The deployment of integer optimisation based model predictive control (MPC) for the control of power electronic applications has been limited by the computational burden of the scenario tree exploration and the required fast cycle times. The recent technology trend towards multi-core control platforms offers new possibilities for enabling MPC for power converters. We propose a static scheduling method to efficiently parallelise the tree exploration for converter control using the example of Generalised Model Predictive Direct Torque Control (GMPDTC). We evaluate several scheduling strategies (both suboptimal and optimal) for distributing the work packages over the individual cores with respect to their effectiveness using an eight core platform from Freescale. Moreover, the proposed approach has the advantage that the code has a small memory footprint and every improvement of the sequential code will directly result in an improvement of the parallel version.
17:00-17:20	WeC15.3
Constrained Real-Time Optimal Control for Cyclic AC PWM Systems (I), pp. 1517-1522
Fischer, Claudia	ETH Zurich
Mariethoz, Sebastien	ETH Zurich
Morari, Manfred	ETH Zurich
The paper introduces a constrained optimal control approach for cyclic AC PWM systems for which computation of the optimal modulation with respect to some cost criterion is too complex for real-time execution. For such systems it is possible to precompute optimal cyclic steady state solutions for different operating points and create a look-up table for real-time implementation. As the operating point changes, direct application of the corresponding optimal cyclic steady state solution causes adverse effects, such as increased system losses. The proposed optimal control approach allows to reduce these effects by doing fast transitions between precomputed optimal solutions, bringing the system to the new optimal steady state operation within half a cycle. The approach effectiveness is demonstrated on a multisource DCDC converter benchmark example.
17:20-17:40	WeC15.4
Extending Affine Control Policies to Hybrid Systems: Robust Control of a DC-DC Buck Converter (I), pp. 1523-1528
Vujanic, Robin	ETH Zurich
Schmitt, Marius	ETHZ
Warrington, Joseph	ETH Zurich
Morari, Manfred	ETH Zurich
Robust model predictive control has been used to determine sequences of inputs to dynamical systems that maximize a certain objective under uncertain dynamics or constraints. These sequences also ensure that hard state and input constraints are satisfied for any realization of the uncertainty. Performing an optimization over open loop control inputs in such circumstances is well known to lead to conservative input choices, leading to poor performance or even a failure to find a feasible input sequence. However, this conservativeness can be reduced by including recourse in the problem formulation. For linear systems with bounded disturbances, optimization over control laws that are affine in the disturbance measurements has emerged as a valid trade-off between performance and computational expense. In this paper we extend the idea of affine recourse to hybrid systems, and propose a new control approach that chooses integer decision variables over the control horizon in conjunction with continuous decision variables that are subject to recourse. Using the example of a buck power converter, we illustrate the efficacy of our approach in comparison to an implementation of affine control policies based on a linearized model.
17:40-18:00	WeC15.5
Actuator Performance Evaluation Using LMIs for Optimal HVDC Placement (I), pp. 1529-1534
Fuchs, Alexander	ETH Zurich
Morari, Manfred	ETH Zurich
This work concerns the closed loop performance evaluation of actuators for constrained linear systems. The main motivation is the optimal placement of High Voltage Direct Current links (HVDC) in a meshed AC power system, but the method is also applicable to other actuator selection problems. The goal is the computation of a performance measure, that can be used to rank different candidate actuators according to their behavior after a disturbance. A measure is proposed, that is not based on simulations, but is computed for general disturbances, using Linear Matrix Inequalities (LMIs). It accounts for global limitations of the system and the actuators. As required for the decision process, it can be chosen to have a physical interpretation. The approach is illustrated by placing HVDC links in a power system with two areas oscillating against each other.
18:00-18:20	WeC15.6
Dynamic Analysis and Control of Dc/dc Boost Converters Used in Stand-Alone PV Systems, pp. 1535-1540
Krommydas, Konstantinos	Univ. of Patras
Alexandridis, Antonios	Univ. of Patras
A stand-alone photovoltaic (PV) system connected to a resistance load through a dc/dc boost converter is considered. Taking into account the PV source voltage-current nonlinear dependence as well as the dc/dc boost nonlinear dynamic model, an extensive system analysis is conducted. It is proven that as the load power becomes smaller than the maximum power input provided by the PV source, the system acquires the minimum phase property with respect to the output voltage, while the domain of attraction of the stable zero dynamics of the system becomes larger. Exploiting this property, a simple adaptive-based direct output voltage controller is implemented, which acts independently from the system parameters and the equilibrium guaranteeing stability and convergence to the desired operating point. Extensive simulations results verify the theoretical analysis and the system good performance.

Technical Program for Thursday July 18, 2013

ThPT1	HG F30, F1, F7
Plenary 2: Control for Smart Grids: Applications and Opportunities in the Customer Domain (Plenary Session)
Chair: Canudas de Wit, Carlos	CNRS-LAG-Grenoble
08:00-09:00	ThPT1.1
Control for Smart Grids: Applications and Opportunities in the Customer Domain*
Samad, Tariq	Honeywell Lab.

ThA1	HG F1
Control of Electrical Energy Systems (Tutorial Session)
Chair: Hiskens, Ian A.	Univ. of Michigan
Organizer: Hiskens, Ian A.	Univ. of Michigan
09:20-10:00	ThA1.1
Overview of Power System Controls (I)*
Hiskens, Ian A.	Univ. of Michigan
This lecture will provide an overview of current power system operational practices. The various levels of control associated with frequency regulation will be discussed, and concepts of reserve will be explained. Voltage regulation will be considered, and the dynamic mechanisms that drive voltage collapse will be presented. The talk will briefly discuss the changing nature of power system control, and in particular will consider issues and opportunities arising from large-scale integration of renewable sources, non-disruptive load control, and wide-spread storage. An overview of preventative versus corrective control will be provided, and MPC-based corrective control strategies will be discussed.
10:00-10:40	ThA1.2
Energy Hub Concepts and Controls (I)*
Andersson, Goran	Swiss Federal Inst. of Tech.
Energy hub concepts will be introduced in this lecture. It will be shown that energy hubs provide a framework for coupling multiple energy carriers, such as electricity, natural gas and heat. They are typically constructed from interconnections of five fundamental building blocks: input sources and storage, converters, and output sources and storage. The lecture will describe modeling formalisms that enable flexible construction of large, interconnected hub networks. These energy hub network models underpin the development of analysis tools for optimally scheduling energy resources over a finite horizon, and for assessing cascading failure mechanisms that involve multiple infrastructures.
10:40-11:20	ThA1.3
Demand Response Strategies for Enhanced Control of Power Systems (I)*
Callaway, Duncan	UC Berkeley
Responsive load control offers a particularly effective approach to compensating for the variability inherent in large-scale renewable generation. Expansive communications networks and advances in distributed control algorithms facilitate precise, non-disruptive forms of load control. The presentation will discuss various approaches to coordinated control of large numbers of highly distributed and diverse loads. The technology underpinning ubiquitous load control will be examined, and technical challenges will be considered. A range of applications will be presented.

ThA2	HG F3
Model Predictive Control IV (Regular Session)
Chair: Pannocchia, Gabriele	Univ. of Pisa
Co-Chair: Domahidi, Alexander	Automatic Control Lab. ETH Zurich
09:20-09:40	ThA2.1
Multi-Objective Predictive Control for Non Steady-State Operation, pp. 1541-1546
Maree, Johannes Philippus	Norwegian Univ. of Science and Tech.
Imsland, Lars	Norwegian Univ. of Science and Tech.
The concept of dynamic-mean Pareto optimality is introduced for multi-objective Model Predictive Control. Dynamic-mean Pareto optimal solutions are obtained by solving a free initial state and final time optimal control problem. Subsequently, we propose a receding horizon tracking formulation with dynamic-mean Utopia set-points. A Dynamic-mean Utopia point is defined as the intersection of average minima, of underlying performance indices, over a dynamic horizon. The latter is compared with recently proposed steady-state Utopia tracking and Pareto optimally weighted Economic MPC. Incorporating dynamic-mean Utopia set-points in a tracking formulation, one attains economic performance at least equal to that of steady-state Utopia tracking, and, performance close to that of Pareto optimal, weighted Economic MPC. The latter is illustrated for a CSTR numerical case example.
09:40-10:00	ThA2.2
Code Generation for Embedded Second-Order Cone Programming, pp. 1547-1552
Chu, Eric	Stanford Univ.
Parikh, Neal	Stanford Univ.
Domahidi, Alexander	Automatic Control Lab. ETH Zurich
Boyd, Stephen P.	Stanford Univ.
This paper describes a framework for generating easily verifiable code to solve convex optimization problems in embedded applications by transforming them into equivalent second-order cone programs. In embedded applications, it is critical to be able to verify code correctness, but it is also desirable to be able to rapidly prototype and deploy high-performance solvers for different problems. To balance these two requirements, we propose a code generation system that takes high-level descriptions of convex optimization problems and generates code that maps the parameters in the original problem to data in an equivalent second-order cone program, which is then solved by a single, external solver that can be verified once and for all. A novel aspect is that we restrict the parameters in the original problem to only appear in affine functions, which lets us map the parameters to problem data without performing any floating point operations. As a result, the generated code is lightweight, fast, and trivial to verify. The approach thus marries the benefits of high-level parser/solvers with custom, high-performance, high-reliability solvers for embedded applications.
10:00-10:20	ThA2.3
A Parsimonious Algorithm for the Solution of Continuous-Time Constrained LQR Problems with Guaranteed Convergence, pp. 1553-1558
Pannocchia, Gabriele	Univ. of Pisa
Mayne, David Q.	Imperial Coll. London
Rawlings, James B.	Univ. of Wisconsin at Madison
Mancuso, Giulio Mosè	Scuola Superiore Sant'Anna
This paper presents an efficient computational method for solving the input-constrained, continuous time, infinite horizon, linear quadratic regulator problem to within a user specified tolerance. The infinite dimensional input trajectory is approximated with a piecewise linear function on a finite time discretization to ensure input constraint satisfaction. This approximate problem is then a standard finite dimensional quadratic program and is solved by conventional methods, generating an upper bound for the optimal value function. The finite time discretization is then refined by subdividing the intervals estimated to cause the largest decrease in the cost function. Convergence of the solution of this discretized problem towards the optimal continuous-time solution, as the discretization is refined, is proved. Exploiting the strict convex- ity of the original infinite dimensional problem, the gradient of the cost function with respect to the continuous-time input can be computed to generate a lower bound for the optimal cost. For computational efficiency, a lower bound for the solution of the discretized control at a very fine discretization can be used instead. The algorithm terminates when the difference between the upper and lower bounds meets a user supplied tolerance.
10:20-10:40	ThA2.4
Model Predictive Control for a Thermostatic Controlled System, pp. 1559-1564
Shafiei, Seyed Ehsan	Aalborg Univ.
Rasmussen, Henrik	Aalborg Univ.
Stoustrup, Jakob	Aalborg Univ.
This paper proposes a model predictive control scheme to provide temperature set-points to thermostatic controlled cooling units in refrigeration systems. The control problem is formulated as a convex programming problem to minimize the overall operating cost of the system. The foodstuff temperatures are estimated by reduced order observers and evaporation temperature is regulated by an algorithmic suction pressure control scheme. The method is applied to a validated simulation benchmark. The results show that even with the thermostatic control valves, there exists significant potential to reduce the operating cost.
10:40-11:00	ThA2.5
A Supervisory Control Approach in Economic MPC Design for Refrigeration Systems, pp. 1565-1570
Shafiei, Seyed Ehsan	Aalborg Univ.
Stoustrup, Jakob	Aalborg Univ.
Rasmussen, Henrik	Aalborg Univ.
A model predictive control at supervisory level is proposed for refrigeration systems including distributed local controllers. Prediction of the electricity price and outdoor temperature are assumed available. The control objective is to minimize the overall energy cost within the prediction horizon. The method is mainly developed for demand-side management in the future smart grid, but a simpler version can be applied in the current electricity market. Due to the system nonlinearity, the minimization is in general a complicated nonconvex optimization problem. A new supervisory control structure as well as an algorithmic pressure control scheme is presented to rearrange the problem to facilitate convex programming. A nonlinear continuous time model validated by real data is employed to simulate the system operation. The results show a considerable economic saving as well as a trade-off between the saving level and the design complexity.
11:00-11:20	ThA2.6
A Flexible Low Cost Embedded System for Model Predictive Control of Industrial Processes, pp. 1571-1576
Lima, Daniel Martins	Univ. Federal de Santa Catarina
Americano da Costa, Marcus V.	Federal Univ. of Santa Catarina
Normey-Rico, Julio Elias	Federal Univ. of Santa Catarina
Model Predictive Control is an algorithm commonly used in the petrochemical and chemical sectors to control large processes. Its success can be traced to the fact that it is MIMO control algorithm with intrinsic dead-time compensation, capacity to handle the process' constraints and with tuning parameters that are easily understandable in the time domain. Nevertheless, its usage is limited to large processes where the investment in the currently expensive advanced control systems are economically viable. Thus, this work proposes a low cost embedded MPC controller targeted at companies with medium and small processes that could have production improvements with the use of MPC. This paper details the current embedded system prototype and the results obtained with a hardware-in-the-loop experiment of an ethanol distillation process simulator.

ThA3	HG F5
Estimation and Control of DPS I (Invited Session)
Chair: Demetriou, Michael A.	Worcester Pol. Inst.
Co-Chair: Meurer, Thomas	Christian-Albrechts-Univ. Kiel
Organizer: Demetriou, Michael A.	Worcester Pol. Inst.
Organizer: Meurer, Thomas	Christian-Albrechts-Univ. Kiel
09:20-09:40	ThA3.1
Spatial Proportional-Integral-Derivative Penalization of Distributed Consensus Filters for Spatially Distributed Processes (I), pp. 1577-1582
Demetriou, Michael A.	Worcester Pol. Inst.
The main thrust of this work is on the penalization of the pairwise state estimates used to enforce consensus in spatially distributed filters. It is assumed that a spatially distributed process has a network of in-domain sensors with spatially distributed filters corresponding to each sensor in the network. To better improve the agreement of the distributed filters, the spatial gradient of the pairwise difference of state estimates is used as a means to penalize their disagreement. Additionally, a proportional penalization and an integral penalization for the pairwise differences are also examined in order to lay down the foundation for a spatial proportional- integral-derivative penalization of the spatially distributed filters. Addressing the partial connectivity issue, a condition that resembles the Lagrangian potential for infinite dimensional systems is given in terms of the inner product of the state errors and their pairwise differences. In a forward looking approach, the extension to a more general class of partial differential equations, written as evolution equations in an appropriate Hilbert space, are examined and the conditions regarding the network connectivity are expressed as conditions on the inner product of the consensus operator and the pairwise difference of the state estimation errors.
09:40-10:00	ThA3.2
Estimation of the Distributed Temperature of a SI Engine Catalyst for Light-Off Strategy (I), pp. 1583-1590
Bresch-Pietri, Delphine	MIT
Leroy, Thomas	IFP
Petit, Nicolas	MINES ParisTech
This paper proposes a model for the internal temperature of a SI engine catalyst, aiming at designing a prediction-based light-off strategy. Due to its elongated geometry where a gas stream is in contact with a spatially distributed monolith, the system under consideration is inherently a distributed parameter system. This paper advocates an approach which is based on a one-dimensional distributed parameter model, coupled with an advection-diffusion equation accounting for the distributed heat generation resulting from pollutant conversion. Following recent works, this heat supply is shown here to be equivalent to an inlet temperature entering the system at a virtual entry point inside the catalyst. This new input has a static gain depending on the state of the system, which introduced a coupling. Taking advantage of the low-pass filter characteristic of the system, an estimate of this model is designed and results into a time-varying input-delay system whose dynamics parameters (time constant, delay, gains) are obtained through a simple analytic reduction procedure. A corresponding prediction-based light-off strategy is proposed and illustrated in simulations exploiting experimental data.
10:00-10:20	ThA3.3
A Contribution to Parameter Identification in Infinite-Dimensional Systems (I), pp. 1591-1596
Knüppel, Torsten	Tech. Univ. Dresden
Woittennek, Frank	Tech. Univ. Dresden
Parameter identification in infinite-dimensional systems with lumped measurements is considered. The proposed method is based on a convolutional input-output relation of the underlying system of partial differential equations. The derivation of this input-output relation is presented in a framework involving (ultra-)distributions. The unknown parameters are determined by minimizing an associated error functional.
10:20-10:40	ThA3.4
A Backstepping Boundary Observer for a Class of Linear First-Order Hyperbolic Systems (I), pp. 1597-1602
Di Meglio, Florent	MINES ParisTech
Vazquez, Rafael	Escuela Superior de Ingenieros, Univ. Sevilla
Krstic, Miroslav	Univ. of California at San Diego
We investigate the boundary observer design problem for a class of linear first-order hyperbolic systems on a finite space domain with spatially varying parameters. The system features one negative transport speed and an arbitrary fixed number n of positive transport speeds. Using a backstepping approach, the distributed states are estimated from a single boundary measurement, as illustrated in presented numerical simulations.
10:40-11:00	ThA3.5
State Estimation for Parabolic PDEs with Reactive-Convective Non-Linearities (I), pp. 1603-1608
Jadachowski, Lukas	Vienna Univ. of Tech.
Meurer, Thomas	Christian-Albrechts-Univ. Kiel
Kugi, Andreas	Vienna Univ. of Tech. (VUT)
An extended Luenberger observer is proposed for the solution of the state estimation problem for semi-linear parabolic PDEs with reactive-convective non-linearities. Here, the backstepping method is applied to the linearised observer error dynamics to determine the observer gains. This, however, requires a successive evaluation of the so-called Hopf-Cole transformation allowing to transform the PDE of the linearised observer error into a normal form, for which backstepping can be directly used. Moreover, the computational efficiency of determination of the gains is improved by combining the direct numerical solution approach with the sample-and-hold implementation. Finally, the observer error convergence is analysed both theoretically and by means of numerical simulations.
11:00-11:20	ThA3.6
A Parallel Algorithm for Optimum Monitoring Network Design in Parameter Estimation of Distributed Systems (I), pp. 1609-1614
Ucinski, Dariusz	Univ. of Zielona Gora
Baranowski, Przemyslaw	Univ. of Zielona Gora
The design of a network of observation nodes in a spatial domain is addressed. The observations are to be used to estimate unknown parameters of a distributed parameter system. Given a finite number of possible sites at which to locate a sensor, the problem is formulated as the selection of the gauged sites so as to minimize a convex criterion defined on the Fisher information matrix associated with the estimated parameters. The search for an optimal solution to this binary optimization problem is performed through solving a relaxed problem in which a constrained discrete probability distribution on the set of all allowable sites is sought. The main contribution here consists in properly parallelizing this solution using the parallel variable distribution approach. As a result, each processor minimizes a convex function subject to linear constraints through the use of a simplicial decomposition algorithm. The resulting individual solutions are then synchronized by finding their optimal convex combination.

ThA4	HG F7
Stochastic Filtering (Regular Session)
Chair: Hanebeck, Uwe	Karlsruhe Inst. of Tech. (KIT)
Co-Chair: Mahdianfar, Hessam	Norwegian Univ. of Science and Tech. (NTNU)
09:20-09:40	ThA4.1
An Empirical Method to Fuse Partially Overlapping State Vectors for Distributed State Estimation, pp. 1615-1620
Sijs, Joris	TNO
Hanebeck, Uwe	Karlsruhe Inst. of Tech. (KIT)
Noack, Benjamin	Karlsruhe Inst. of Tech. (KIT)
State fusion is a method for merging multiple estimates of the same state into a single fused estimate. Dealing with multiple estimates is one of the main concerns in distributed state estimation, where an estimated value of the desired state vector is computed in each node of a networked system. Most solutions for distributed state estimation currently available assume that every node computes an estimate of the (same) global state vector. This assumption is impractical for systems observing large-area processes, due to the sheer size of the process state. A feasible solution is one where each node estimates a part of the global state vector, allowing different nodes in the network to have overlapping state elements. Although such an approach should be accompanied by a corresponding state fusion method, existing solutions cannot be employed as they merely consider fusion of two different estimates with emph{equal} state representations. Therefore, an empirical solution is presented for fusing two state estimates that have partially overlapping state elements. A justification of the proposed fusion method is presented, along with an illustrative case study for observing the temperature profile of a large rod, though a formal derivation is future research.
09:40-10:00	ThA4.2
Multiple Model Adaptive Estimation for Open Loop Unstable Plants, pp. 1621-1626
Hassani, Vahid	Inst. Superior Técnico
Pascoal, Antonio Manuel	Inst. Superior Tecnico
Aguiar, A. Pedro	Faculty of Engineering, Univ. of Porto (FEUP)
Athans, Michael	Inst. Superior Tecnico
This paper addresses the problem of adaptive state and parameter estimation of open loop unstable plants using a multiple model structure. A state estimate is obtained as a probabilistically weighted sum of the estimates produced by a bank of individual observers. Model identification and convergence of the dynamic weights in the Multiple Model Adaptive Estimation (MMAE) for open-loop unstable plants are analyzed and the effect of the control action (by a controller in the loop) is studied. In the present paper we show that the techniques introduced in MMAE for open-loop stable plants and in the absence of control action are applicable to open-loop unstable plants with a stabilizing controller in the loop. A distance-like pseudo norm between the true plant and the identified model is developed and furthermore we show that the model identified is the one that is the closest to the true plant model in the defined norm among all models in the bank. The performance and convergence of the MMAE procedure are illustrated with Monte-Carlo simulation runs using the model of an inverted pendulum installed on a system of masses, springs, and dampers.
10:00-10:20	ThA4.3
On the Relation of Nonanticipative Rate Distortion Function and Filtering Theory, pp. 1627-1632
Charalambous, Charalambos D.	Univ. of Cyprus
Stavrou, Photios A.	Univ. of Cyprus
In this paper the relation between nonanticipative rate distortion function (RDF) and Bayesian filtering theory is investigated using the topology of weak convergence of probability measures on Polish spaces. The relation is established via an optimization on the space of conditional distributions of the so-called directed information subject to fidelity constraints. Existence of the optimal reproduction distribution of the nonanticipative RDF is shown, while the optimal nonanticipative reproduction conditional distribution for stationary processes is derived in closed form. The realization procedure of nonanticipative RDF which is equivalent to joint-source channel matching for symbol-by-symbol transmission is described, while an example is introduced to illustrate the concepts.
10:20-10:40	ThA4.4
Equivalent Linearization Kalman Filter with Application to Cubic Sensor Problems, pp. 1633-1638
Katayama, Tohru	Ritsumeikan Univ.
We revisit the equivalent linearization technique to clarify a relationship between the extended Kalman filter (EKF) and equivalent linearization Kalman filter (EqKF). By deriving the equivalent gain for a static nonlinearity, we show that the equivalent linearization for the EqKF is a global method, though the first-order linearization for the EKF is a local one. Then, we consider discrete-time cubic sensor problems and analyze the Kalman gains and filtered covariances of respective filters, showing that the EqKF is quite close to the Gaussian filter (GF). Moreover, numerical results are included to compare the performances of the EqKF, EKF and GF.
10:40-11:00	ThA4.5
Particle Filters for Max Plus Systems, pp. 1639-1644
Ferreira Cândido, Renato Markele	Univ. Estadual de Campinas
Santos-Mendes, Rafael	State Univ. of Campinas
Hardouin, Laurent	Univ. of Angers
Maia, Carlos Andrey	Univ. Federal de Minas Gerais
The main objective of this article is to synthesize a particle filter algorithm for max-plus systems. It is presented a brief introduction to the max-plus approach for Discrete Event Systems. Next, the fundamentals of the particle filter and the way in which they can be applied to max-plus systems are presented. It leads to the algorithm for particles filtering. Lastly, some examples are given. The results shows the accuracy of the method and the improvements in comparison with the deterministic observer.
11:00-11:20	ThA4.6
Joint Unscented Kalman Filter for State and Parameter Estimation in Managed Pressure Drilling, pp. 1645-1650
Mahdianfar, Hessam	Norwegian Univ. of Science and Tech. (NTNU)
Pavlov, Alexey	Statoil Res. center
Aamo, Ole Morten	NTNU
Drilling into offshore deep water, high-pressure high-temperature reservoirs is a very challenging process. The most important task in these drilling operations is to control bottomhole pressure. Many automatic control systems for drilling operations are based on models calculating wellbore pressure, flow and downhole hydraulics. Closed loop control systems, for example Managed Pressure Drilling, are examples of systems that may involve such real-time calculations. Therefore a high degree of accuracy in pressure and flow predictions is crucial to the performance of automatic drilling applications. In this paper the key uncertain model parameters and the bottom-hole pressure are estimated using joint unscented Kalman filter based on only available top-side measurements. The results of simulations show accurate estimation of the bottom-hole pressure and uncertain parameters, even in transient periods for example the scenario of pipe connection operations, where there is no available bottom-hole pressure measurement, and flow through the bit.

ThA5	HG E1.1
Analysis of Hybrid Systems (Regular Session)
Chair: Trenn, Stephan	Univ. of Kaiserslautern
Co-Chair: Mattsson, Per	Uppsala Univ.
09:20-09:40	ThA5.1
What to Do When Hybrid Systems “freeze” Due to an Interconnection?, pp. 1651-1656
Dashkovskiy, Sergey	Univ. of Applied Sciences Erfurt
Kosmykov, Michael	Univ. of Bremen
Promkam, Ratthaprom	Univ. of Bremen
The paper addresses the question of a composition and decomposition of hybrid systems. Motivated by a simple example, we propose an extended definition of a hybrid system that allows for a natural and simple way to consider an interconnection of several hybrid systems as one hybrid system and to decompose one large hybrid system in a composition of several ones. A stability result and existence of solutions for the new framework are shown.
09:40-10:00	ThA5.2
A Sum-Of-Squares Approach to the Analysis of Zeno Stability in Polynomial Hybrid Systems, pp. 1657-1662
Murti, Chaitanya	Illinois Inst. of Tech.
Peet, Matthew Monnig	Arizona State Univ.
Hybrid dynamical systems can exhibit many unique phenomena, such as Zeno behavior. Zeno behavior is the occurrence of infinite discrete transitions in finite time. Zeno behavior has been likened to a form of finite-time asymptotic stability, and corresponding Lyapunov theorems have been developed. In this paper, we propose a method to construct Lyapunov functions to prove Zeno stability of compact sets in cyclic hybrid systems with parametric uncertainties in the vector fields, domains and guard sets, and reset maps utilizing sum-of-squares programming. This technique can easily be applied to cyclic hybrid systems without parametric uncertainties as well. Examples illustrating the use of the proposed technique are also provided.
10:00-10:20	ThA5.3
Controllability of a Class of Bimodal Discrete-Time Piecewise Linear Systems, pp. 1663-1668
Yurtseven, Evren	SKF
Camlibel, Kanat	Univ. of Groningen
Heemels, Maurice	Eindhoven Univ. of Tech.
In this paper we will provide algebraic necessary and sufficient conditions for the controllability/reachability/null controllability of a class of bimodal discrete-time piecewise linear systems including several instances of interest that are not covered by existing works which focus primarily on the planar case. In particular, the class is characterized by a continuous right-hand side, a scalar input and a transfer function from the control input to the switching variable with at most two zeroes whereas the state can be of any dimension. To arrive at the main result, we will make use of geometric control theory for linear systems and a novel result on controllability for input-constrained linear systems with non-convex constraint sets.
10:20-10:40	ThA5.4
The Bang-Bang Funnel Controller: Time Delays and Case Study, pp. 1669-1674
Liberzon, Daniel	Univ. of Illinois at Urbana-Champaign
Trenn, Stephan	Univ. of Kaiserslautern
We investigate the recently introduced bang-bang funnel controller with respect to its robustness to time delays. We present slightly modified feasibility conditions and prove that the bang-bang funnel controller applied to a relative-degree-two nonlinear system can tolerate sufficiently small time delays. A second contribution of this paper is an extensive case study, based on a model of a real experimental setup, where implementation issues such as the necessary sampling time and the conservativeness of the feasibility assumptions are explicitly considered.
10:40-11:00	ThA5.5
State Estimation in Linear Time-Invariant Systems with Unknown Impulsive Inputs, pp. 1675-1680
Mattsson, Per	Uppsala Univ.
Medvedev, Alexander V.	Uppsala Univ.
This paper deals with state estimation in linear time-invariant systems subject to unknown impulsive input signals. A solution based on a linear impulsive observer and a finite-memory convolution operator is suggested. The problem arises e.g. in the context of systems with intrinsic pulse-modulated feedback that have recently been applied to mathematical modeling of endocrine systems with pulsatile hormone secretion. Simulation results illustrating the performance of the proposed method are provided.
11:00-11:20	ThA5.6
Static Anti-Windup Scheme for a Class of Homogeneous Dwell-Time Hybrid Controllers, pp. 1681-1686
Fichera, Francesco	CNRS
Prieur, Christophe	CNRS
Tarbouriech, Sophie	LAAS-CNRS
Zaccarian, Luca	CNRS
In this paper, some tools for the analysis of a class of hybrid control systems for a continuous-time plant with input magnitude saturation are presented. We show that certain continuous-time LMI-based techniques can be extended to this setting when a dwell-time property is satisﬁed by the hybrid loop. Moreover, for the case in which the hybrid controller has the same order as the plant, the synthesis of a static direct linear anti-windup (DLAW) compensator is proposed. The need of including the ﬂow and jump sets in the analysis and synthesis conditions leads to LMIs for global results, that is, when the plant is exponentially stable and BMIs otherwise.

ThA6	HG E1.2
Stability of Nonlinear Systems I (Regular Session)
Chair: Heath, William Paul	Univ. of Manchester
Co-Chair: Perruquetti, Wilfrid	Ec. Centrale de Lille
09:20-09:40	ThA6.1
Basic Properties and Stability of Fractional Order Reset Control Systems, pp. 1687-1692
Hosseinnia, S. Hassan	Univ. de Extremadura
Tejado, Ines	Univ. Técnica de Lisboa, Inst. Superior Técnico
Vinagre, B. M.	Univ. de Extremadura
Reset control is introduced to overcome limitations of linear control. A reset controller includes a linear controller which resets some of states to zero when their input is zero or certain non-zero values. This paper studies the application of the fractional-order Clegg integrator (FCI) and compares its performance with both the commonly used first order reset element (FORE) and traditional Clegg integrator (CI). Moreover, stability of reset control systems is generalized for the fractional-order case. Two examples are given to illustrate the application of the stability theorem.
09:40-10:00	ThA6.2
Extension of the Krylov-Bogoliubov Method and Its Application to the Decay Rate Analysis of Nonlinear Control Algorithms, pp. 1693-1698
Röthig, Andreea	Darmstadt Univ. of Tech. Inst. of Automatic Contr
An extension of the Krylov-Bogoliubov method and of the harmonic linearization method for approximating second order nonlinear systems with fast variation of damping factor and frequency of oscillation is revisited and a solution method involving a power series approach is proposed, which makes it easy for the extended asymptotic method to be applied to other damped nonlinear systems. Furthermore, the application of the method to the analysis of the decay rate of the nonlinear solution is presented. A performance measure for the analysis of the decay rate for nonlinear systems is proposed and exemplified on two nonlinear control algorithms, linear constrained control and constrained soft variable structure control with implicit Lyapunov functions.
10:00-10:20	ThA6.3
A Generalisation of the Nonlinear Small-Gain Theorem for Systems with Abstract Initial Conditions, pp. 1699-1704
Liu, Jing	Univ. of Southampton
French, Mark	Univ. of Southampton
We consider the development of a general nonlinear small-gain theorem for systems with abstract initial conditions. Systems are defined in a set theoretic manner from input-output pairs on a doubly infinite time axis, and a general construction of the initial conditions (i.e. a state at time zero) is given in terms of an equivalence class of trajectories on the negative time axis. By using this formulation, an ISS-type nonlinear small-gain theorem is established with complete disconnection between the stability property and the existence, uniqueness properties. We provide an illustrative example.
10:20-10:40	ThA6.4
On ISS and Iiss Properties of Homogeneous Systems, pp. 1705-1710
Bernuau, Emmanuel	Ec. Lille
Polyakov, Andrey	INRIA Lille Nord-Europe
Efimov, Denis	INRIA - LNE
Perruquetti, Wilfrid	Ec. Centrale de Lille
Several conditions are proposed to check input-to-state stability (ISS) and integral input-to-state stability (iISS) properties for generic nonlinear systems applying the weighted homogeneity concept (global or local). The advantages of this result is that, under some mild conditions, the system robustness can be established as a function of the degree of homogeneity.
10:40-11:00	ThA6.5
On Multipliers for Bounded and Monotone Nonlinearities, pp. 1711-1716
Carrasco, Joaquin	Univ. of Manchester
Heath, William Paul	Univ. of Manchester
Lanzon, Alexander	Univ. of Manchester
Recent results in equivalence between classes of multipliers for slope-restricted nonlinearities are extended to multipliers for bounded and monotone nonlinearities. This extension requires a slightly modified version of the Zames--Falb theorem and a more general definition of phase--substitution. The results in this paper resolve apparent contradictions in the literature on classes of multipliers for bounded and monotone nonlinearities.

ThA7	HG E3
Algebraic and Geometric Methods (Regular Session)
Chair: Aguiar, A. Pedro	Faculty of Engineering, Univ. of Porto (FEUP)
Co-Chair: Biggs, James Douglas	Univ. of Strathclyde
09:20-09:40	ThA7.1
Quadratic Hamiltonians on Non-Euclidean Spaces of Arbitrary Constant Curvature, pp. 1717-1721
Biggs, James Douglas	Univ. of Strathclyde
This paper derives explicit solutions for Riemannian and sub-Riemannian curves on non-Euclidean spaces of arbitrary cross-sectional curvature. The problem is formulated in the context of an optimal control problem on a 3-D Lie group and an application of Pontryagin's maximum principle of optimal control leads to the appropriate quadratic Hamiltonian. It is shown that the regular extremals defining the necessary conditions for Riemannian and sub-Riemannian curves can each be expressed as the classical simple pendulum. The regular extremal curves are solved analytically in terms of Jacobi elliptic functions and their projection onto the underlying base space of arbitrary curvature are explicitly derived in terms of Jacobi elliptic functions and an elliptic integral.
09:40-10:00	ThA7.2
Sufficiency of a Necessary Condition for Local Observability of Discrete-Time Polynomial Systems, pp. 1722-1727
Kawano, Yu	Osaka Univ.
Ohtsuka, Toshiyuki	Osaka Univ.
We consider local observability of discrete-time polynomial systems. When testing local observability of nonlinear systems, the observability rank condition is commonly used. However, the rank condition is only a sufficient condition in general. For this problem, recently, in terms of commutative algebra and algebraic geometry, a necessary condition for local observability of the polynomial systems has been derived. In this paper, we show the necessary condition is a sufficient condition for local observability for almost all initial states. Based on this result, we derive a necessary and sufficient condition for local observability.
10:00-10:20	ThA7.3
Input-Output Linearization by Dynamic Output Feedback, pp. 1728-1733
Kaldmäe, Arvo	Inst. of Cybernetics at TUT
Kotta, Ülle	Inst. of Cybernetics at TUT
The problem of input-output linearization by dynamic output feedback of discrete-time multi-input multi-output nonlinear control systems is considered in this paper. The system is described by the set of higher order input-output difference equations. First, an algorithm is given to find a set of functions, which are then used to provide sufficient solvability conditions and the equations of the dynamic output feedback. The results are specified for the special subclass of systems, called ANARX systems and comparisons with the single-input single-output case is made.
10:20-10:40	ThA7.4
A New Algorithm for Linearization up to Multi-Output Injection for a Class of Systems with Implicitly Defined Outputs, pp. 1734-1739
Rodrigues, Sergio	RICAM-OEAW, Linz
Aguiar, A. Pedro	Faculty of Engineering, Univ. of Porto (FEUP)
Given a class of nonlinear systems with implicitly defined outputs, we provide a new algorithm to find appropriate local coordinates, in which the resulting system takes a desired target form that is state-affine, up to output and input injection. Once in the target form, it is possible to construct a state-space observer with linear, possibly time-varying, error dynamics.
10:40-11:00	ThA7.5
Stationary Trajectories, Singular Hamiltonian Systems and Ill-Posed Interconnection, pp. 1740-1745
Jugade, Shriram	Indian Inst. of Tech. Bombay
Pal, Debasattam	Indian Inst. of Tech. Guwahati
Kalaimani, Rachel Kalpana	Indian Inst. of Tech. Bombay
Belur, Madhu N.	Indian Inst. of Tech. Bombay
We study Hamiltonian systems, namely, systems comprising of trajectories which are `stationary' with respect to a quadratic performance index: they play a central role in many optimal control problems. A typical assumption in the literature is that of `regularity': the resulting first-order dynamical system is a regular state space system and not a singular descriptor system. In this paper we show that the first order representation of a Hamiltonian is a singular descriptor system if and only if the interconnection of a related MIMO system G(s) with its dual (i.e. its adjoint) is ill-posed. We address the possibility of existence of inadmissible initial conditions, i.e. initial conditions that give rise to impulsive solutions. We characterize conditions on G(s) under which the corresponding singular Hamiltonian system has inadmissible initial conditions. Under suitable simplifying assumptions, which amount to studying an extreme case of ill-posedness, our main result states that there exist no inadmissible initial conditions if and only if the skew-symmetric part of the first moment about s equal to infinity of the transfer matrix G(s) is nonsingular; a condition we show that is opposite to that for G(s) to be an all-pass filter. As a corollary, ill-posed interconnection of a square MIMO system with odd number of inputs (in particular, SISO systems) with its adjoint always contains in inadmissible initial conditions.
11:00-11:20	ThA7.6
Almost Global Attitude Stabilization of a Rigid Body for Both Internal and External Actuation Schemes, pp. 1746-1751
Bayadi, Ramaprakash	Indian Inst. of Tech. Bombay
Banavar, Ravi N.	Indian Inst. of Tech.
Recent developments in the attitude control of a rigid body include the development of almost globally stabilizing coordinate-free control laws. However, all of these results focus on external actuation. In this paper, we first observe that the frequently used way of stabilization by error functions can also be given a Hamiltonian interpretation. We then show that there exists a class of control laws which are applicable for both external and internal actuation.

ThA8	HG E5
Networks and Consensus (Regular Session)
Chair: Hadjicostis, Christoforos	Univ. of Cyprus
Co-Chair: Hendrickx, Julien M.	UCL
09:20-09:40	ThA8.1
Distributed Formation of Balanced and Bistochastic Weighted Digraphs in Multi-Agent Systems, pp. 1752-1757
Charalambous, Themistoklis	Royal Inst. of Tech. (KTH)
Hadjicostis, Christoforos	Univ. of Cyprus
We propose two distributed algorithms, one for solving the weight-balance problem and another for solving the bistochastic matrix formation problem, in a distributed system whose components (nodes) can exchange information via interconnection links (edges) that form an arbitrary, possibly directed, strongly connected communication topology (digraph). Both distributed algorithms achieve their goal asymptotically and operate iteratively by having each node adapt the (nonnegative) weights on its outgoing edges based on the weights of its incoming links. The weight-balancing algorithm is shown to admit geometric convergence rate, whereas the second algorithm, which is a modification of the weight-balancing algorithm, leads asymptotically to a bistochastic digraph with geometric convergence rate for a certain set of initial values. The two algorithms perform better than existing approaches, as illustrated by the examples we provide.
09:40-10:00	ThA8.2
Singularly Perturbed Algorithms for Dynamic Average Consensus, pp. 1758-1763
Kia, Solmaz	Univ. of California San Diego
Cortes, Jorge	Univ. of California, San Diego
Martinez, Sonia	Univ. of California at San Diego
This paper proposes two continuous-time dynamic average consensus algorithms for networks with strongly connected and weight-balanced interaction topologies. The proposed algorithms, termed 1st-Order-Input Dynamic Consensus (FOI-DC) and 2nd-Order-Input Dynamic Consensus (SOI-DC), respectively, allow agents to track the average of their dynamic inputs within an O(epsilon)-neighborhood with a pre-specified rate. The only requirement on the set of reference inputs is having continuous bounded derivatives, up to second order for FOI-DC and up to third order for SOI-DC. The correctness analysis of the algorithms relies on singular perturbation theory for non-autonomous dynamical systems. When dynamic inputs are offset from one another by static values, we show that SOI-DC converges to the exact dynamic average with no steady-state error. Simulations illustrate our results.
10:00-10:20	ThA8.3
A GTS Scheduling for Consensus Problems Over IEEE 802.15.4 Wireless Networks, pp. 1764-1769
Hayashi, Naoki	Osaka Univ.
Takai, Shigemasa	Osaka Univ.
This paper presents consensus problems over IEEE 802.15.4 wireless networks. The IEEE 802.15.4 protocol defines the Medium Access Control (MAC) and the physical layer (PHY) for Low-Rate Wireless Personal Area Networks (LR-WPANs). The IEEE 802.15.4 MAC protocol supports the beacon-enabled mode that enables real-time communication by allocating Guaranteed Time Slots (GTSs) to a designated node in a network. Despite a number of work on cooperative control, consensus problems over wireless networks remains less well understood. In this paper, we propose a modified version of the standard IEEE 802.15.4 protocol where each agent communicates through a PAN coordinator with a superframe structure. We consider a branch-and-bound GTS scheduling algorithm for non-preemptive communication tasks under the modified IEEE 802.15.4 protocol. The simulation result shows that each agent can successfully communicate with each other and achieve a consensus by the proposed GTS scheduling algorithm.
10:20-10:40	ThA8.4
On Distributed Cluster Consensus for Multiple Double-Integrator Agents, pp. 1770-1775
Qin, Jiahu	The Australian National Univ.
Yu, Changbin	Australian National Univ.
Anderson, Brian D.O.	Australian National Univ.
This paper investigates the consensus for multiple interacting clusters of double-integrator agents under two different frameworks, viz, the framework that all agents share the same position and velocity interaction topology and the framework that the position and velocity topologies are modeled by totally independent graphs. Different cluster consensus algorithms are designed and analyzed accordingly. A consistent structural result is shown for both frameworks that cluster consensus can be reached if the interaction topologies satisfy some connectivity assumptions and further, compared to the interactions among different clusters, the interactions within each cluster are sufficiently strong. Some lower bounds for such strengths are specified as well.
10:40-11:00	ThA8.5
A Distributed Randomized Algorithm for Relative Localization in Sensor Networks, pp. 1776-1781
Ravazzi, Chiara	Pol. di Torino
Frasca, Paolo	Pol. di Torino
Ishii, Hideaki	Tokyo Inst. of Tech.
Tempo, Roberto	Pol. di Torino
This paper regards the relative localization problem in sensor networks. We propose for its solution a distributed randomized algorithm, which is based on input-driven consensus dynamics and features pairwise “gossip” communications and updates. Due to the randomness of the updates, the state of this algorithm oscillates in time around a certain limit value. We show that the time-average of the state asymptotically converges, in the mean-square sense, to the least-squares solution of the localization problem. Furthermore, we describe an update scheme ensuring that the time-averaging process is accomplished in a fully distributed way.
11:00-11:20	ThA8.6
Large Network Consensus Is Robust to Packet Losses and Interferences, pp. 1782-1787
Frasca, Paolo	Pol. di Torino
Hendrickx, Julien M.	UCL
In many randomized consensus algorithms, the constraint of average preservation may not be enforced at every time step, resulting in an error between the average of the initial conditions and the current average. We have recently shown that under mild conditions on the distribution of the update matrices, the mean square error has an upper bound inversely proportional to the size of the network. In this work, we consider the case of consensus with packet losses and interferences. Using an extension of our results taking correlations into account, we show that the MSE induced by losses and interferences can be estimated by such a bound: hence we argue that larger networks are naturally more robust, in terms of accuracy, to packet losses and interferences. Our results hold for general networks, without restrictive assumptions on its topology.

ThA9	HG E7
Quadrotor Control (Regular Session)
Chair: Silvestre, Carlos	Inst. Superior Tecnico
Co-Chair: Tzes, Anthony	Univ. of Patras
09:20-09:40	ThA9.1
Time-Optimal Quadrotor Flight, pp. 1788-1792
Van Loock, Wannes	KU Leuven
Pipeleers, Goele	KU Leuven
Swevers, Jan	KU Leuven
Solving the time-optimal path planning problem, where a system is brought from an initial to terminal state in minimal time while obeying geometric and dynamic constraints, has been an active area of research for many years. Very often the problem is divided into a high-level path planning stage where a feasible geometric path is determined and a low-level path following stage where system dynamics are taken into account. This paper combines both approaches for differentially flat systems into a single optimization problem. The geometric path is represented as a convex combination of two or more feasible paths and the dynamics of the system can subsequently be projected onto the path which leads to a single input system. The resulting optimization problem is transformed into a fixed end-time optimal control problem that can be initialized easily. Throughout the paper, the quadrotor, a challenging non-linear system, is used to illustrate the proposed approach.
09:40-10:00	ThA9.2
Hybrid Model Predictive Flight Mode Conversion Control of Unmanned Quad-TiltRotors, pp. 1793-1798
Papachristos, Christos	Univ. of Patras
Alexis, Kostas	Eidgenössische Tech. Hochschule Zürich
Tzes, Anthony	Univ. of Patras
In this paper the autonomous flight mode conversion control scheme for a Quad-TiltRotor Unmanned Aerial Vehicle is presented. This convertible UAV type has the capability for flying both as a helicopter as well as a fixed-wing aircraft type, by adjusting the orientation of its tilt-enabled rotors. Thus, a platform combining the operational advantages of two commonly distinct aircraft types is formed. However, its autonomous mid-flight conversion is an issue of increased complexity. The approach presented is based on an innovative control scheme, developed based on hybrid systems theory. Particularly, a piecewise affine modeling approximation of the complete nonlinear dynamics is derived and serves as the model for control over which a hybrid predictive controller that provides global stabilization, optimality and constraints satisfaction is computed. The effectiveness of the proposed control scheme in handling the mode conversion from helicopter to fixed-wing (and conversely) is demonstrated via a series of simulation studies. The proposed control scheme exceeds the functionality of the aforementioned flight-mode conversion and is also able to handle the transition to intermediate flight-modes with rotors slightly tilted forward in order to provide a forward force component while flying in close to helicopter-mode.
10:00-10:20	ThA9.3
An Iterative Learning Scheme for High Performance, Periodic Quadrocopter Trajectories, pp. 1799-1804
Hehn, Markus	ETH Zurich
D'Andrea, Raffaello	ETH Zurich
Quadrocopters allow the execution of high-performance maneuvers under feedback control. However, repeated execution typically leads to a large part of the tracking errors being repeated. This paper evaluates an iterative learning scheme for an experiment where a quadrocopter flies in a circle while balancing an inverted pendulum. The scheme permits the non-causal compensation of periodic errors when executing the circular motion repeatedly, and is based on a Fourier series decomposition of the repeated tracking error and compensation input. The convergence of the learning scheme is shown for the linearized system dynamics. Experiments validate the approach and demonstrate its ability to significantly improve tracking performance.
10:20-10:40	ThA9.4
Admittance Control for Physical Human-Quadrocopter Interaction, pp. 1805-1810
Augugliaro, Federico	ETH Zurich
D'Andrea, Raffaello	ETH Zurich
This paper analyzes the application of admittance control to quadrocopters, focusing on physical human-vehicle interaction. Admittance control allows users to define the apparent inertia, damping, and stiffness of a robot, providing an intuitive way to physically interact with it. In this work, external forces acting on the quadrocopter are estimated from position and attitude information and then input to the admittance controller, which modifies the vehicle reference trajectory accordingly. The reference trajectory is appropriately tracked by an underlying position and attitude controller. The characteristics of the overall control scheme are investigated for the near-hover case. Experimental results complement the paper, demonstrating the suitability of the method for physical human-quadrocopter interaction.
10:40-11:00	ThA9.5
Experimental Validation of a Nonlinear Quadrotor Controller with Wind Disturbance Rejection, pp. 1811-1816
Cabecinhas, David	Inst. Superior Tecnico
Cunha, Rita	Inst. Superior Técnico
Silvestre, Carlos	Inst. Superior Tecnico
This paper addresses the problem of designing and experimentally validating a controller for steering a quadrotor vehicle along a trajectory, while rejecting wind disturbances. The proposed solution consists of a nonlinear adaptive state feedback controller for thrust and torque actuation that asymptotically stabilizes the closed-loop system in the presence of constant force disturbances, used to model the wind action, and ensures that the actuation does not grow unbounded as a function of the position errors. A prototyping and testing architecture, developed to streamline the implementation and the tuning of the controller, is also described. Experimental results are presented to demonstrate the performance and robustness of the proposed controller.
11:00-11:20	ThA9.6
A Novel Leader-Following Strategy Applied to Formations of Quadrotors, pp. 1817-1822
Roldão, Valter	Inst. Superior Técnico
Cunha, Rita	Inst. Superior Técnico
Cabecinhas, David	Inst. Superior Tecnico
Silvestre, Carlos	Inst. Superior Tecnico
Oliveira, Paulo Jorge	Inst. Superior Técnico
This paper presents a strategy for formation control of autonomous vehicles using a leader-following approach. A trajectory planner prescribes the motion of a group of virtual vehicles, using a Lyapunov-based nonlinear controller that stabilizes the position of the leader in the reference frame of the virtual vehicles, at a predefined distance vector. This strategy differs from the standard approach of defining the desired distance vector in an inertial frame and can be used to obtain rich formation trajectories with varying curvatures between vehicles. By imposing adequate constraints on the motion of the virtual vehicles, the planner naturally guarantees the generation of valid formation trajectories, without requiring the parametrization of the space curve described by the leader. The trajectories are generated online and provided to a trajectory tracking controller specifically designed for quadrotor vehicles. Results of experimental tests are presented demonstrating the performance of the proposed solution for formation control of autonomous vehicles.

ThA10	HG D1.1
Linear Parameter-Varying Systems (Regular Session)
Chair: Puig, Vicenc	Univ. Pol. de Catalunya
Co-Chair: Bokor, Jozsef	Hungarian Acad. of Sciences
09:20-09:40	ThA10.1
H-Infinity LPV Filtering for Discrete-Time Linear Systems Subject to Additive and Multiplicative Uncertainties in the Measurement, pp. 1823-1828
Lacerda, Márcio Júnior	Univ. of Campinas
Tognetti, Eduardo S.	Univ. of Brasilia
Oliveira, Ricardo C. L. F.	Univ. of Campinas
Peres, Pedro L. D.	Univ. of Campinas
This paper is concerned with the problem of H-infinity linear parameter-varying (LPV) filter design for discrete-time linear systems where the measurement of the scheduling parameters may be affected by additive and multiplicative uncertainties. By conveniently modeling the uncertainties and the time-varying parameters, new robust linear matrix inequality (LMI) conditions for the existence of a full order LPV filter assuring a prescribed H-infinity performance, irrespective of the uncertainties affecting the measures, are given. The design procedure can simultaneously handle time-invariant uncertainties and arbitrary time-varying parameters as well. The problem is solved through LMI relaxations based on homogeneous polynomial matrices of arbitrary degree. A numerical experiment illustrates the performance of the proposed LPV filter when compared to other filters obtained with methods from the literature.
09:40-10:00	ThA10.2
A Shifting Pole Placement Approach for the Design of Parameter-Scheduled State-Feedback Controllers, pp. 1829-1834
Rotondo, Damiano	UPC
Nejjari, Fatiha	Univ. Pol. de Catalunya
Puig, Vicenc	Univ. Pol. de Catalunya
In this paper, the problem of designing a parameter-scheduled state-feedback controller is investigated. The main novelty and contribution of this paper is the extension of the classical regional pole placement problem, that will be referred to as shifting pole placement, to the design of parameter scheduled controller taking advantage of polytopes and LMIs properties. By introducing some parameters, or using existing ones, the controller can be designed in such a way that different values of these parameters imply different regions where the closed-loop poles are situated. The problem is analyzed in both linear time-invariant (LTI) and parameter-varying (LPV) cases, and some results obtained in simulation are shown so as to demonstrate the effectiveness of the proposed approach.
10:00-10:20	ThA10.3
Reduced-Order Two-Parameter Plpv Controller for the Rejection of Nonstationary Harmonically Related Multisine Disturbances, pp. 1835-1842
Ballesteros, Pablo	Clausthal Univ. of Tech.
Shu, Xinyu	Clausthal Univ. of Tech.
Heins, Wiebke	Clausthal Univ. of Tech.
Bohn, Christian	Clausthal Univ. of Tech.
A reduced-order gain-scheduling controller based on two time-varying parameters for the rejection of harmonic disturbances with time-varying frequencies is presented. The frequencies are harmonically related and assumed to be known. The control design results in a discrete-time controller with an order which is two times the number of frequency components of the multisine disturbance. Two gain-scheduling parameters are used independently of the number of harmonic components and a triangle is considered as the polytope, therefore the controller is obtained by interpolation between three controllers calculated for the vertices of the polytope. The resulting controller structure is therefore very simple. Experimental results on an active vibration control test bed are used to validate the controller.
10:20-10:40	ThA10.4
Continuity of the Maximal Negative Graph Spaces, pp. 1843-1847
Szabo, Zoltan	MTA SZTAKI
Biro, Zsolt	MTA SZTAKI
Bokor, Jozsef	Hungarian Acad. of Sciences
In the framework of LPV and qLPV controller design one often encounter the problem of finding a maximal negative graph subspace of a parameter dependent non-singular indefinite matrix. For practical reason it is necessary that these subspaces depend continuously on the parameters. The paper provides an exhaustive condition for the existence of such a continuous solution.
10:40-11:00	ThA10.5
LPV Observer Design and Damping Control of Container Crane Load Swing, pp. 1848-1853
Mendez Gonzalez, Antonio	Hamburg Univ. of Tech.
Hoffmann, Christian	Hamburg Univ. of Tech.
Radisch, Christian	Hamburg Univ. of Tech.
Werner, Herbert	Hamburg Univ. of Tech.
This paper presents advances in the damping of container crane load swing via hoisting modulation based on linear parameter-varying (LPV) control techniques. We propose controllers based on constant, as well as parameter-dependent Lyapunov functions and formulate our problem in the linear fractional transformation (LFT) framework. The dynamics of a nonlinear observer are included into the generalized plant. Simulation and experimental results are compared to previous work using a polytopic LPV approach, as well as to earlier work based on the concept of resonant coupling control realized by a reduced normal form approach. The comparison indicates, that including the observer dynamics in the synthesis comes at the price of reduced performance, which is alleviated by the use of parameter-dependent Lyapunov functions. Furthermore, an a posteriori analysis verifies that the controller is robust against erroneously estimated scheduling signals and thus provides closed-loop guarantees for stability and performance for the new controller.
11:00-11:20	ThA10.6
Finite Time Robust Filtering for Time-Varying Uncertain Polytopic Linear Systems, pp. 1854-1859
Borges, Renato Alves	Univ. of Brasilia
Ishihara, João Yoshiyuki	Univ. of Brasília
Rocha, Iara Guimarães	Univ. of Brasilia
Risso, Lara de Oliveira	Univ. of Brasilia
Kussaba, Hugo Tadashi	Univ. of Brasília
In this paper, the problem of robust filter design for uncertain continuous-time systems is investigated in the context of finite time stability. The filter is obtained in order to guarantee that the augmented system is finite time stable. The system is considered time varying with the parameters modeled by a polytope. The design conditions obtained by means of Lyapunov functions are expressed as linear matrix inequalities. A complete order filter is obtained by the solution of a factibility problem. A numerical example is provided.

ThA11	HG D1.2
Multi-Agent Systems (Regular Session)
Chair: Allgower, Frank	Univ. of Stuttgart
Co-Chair: Werner, Herbert	Hamburg Univ. of Tech.
09:20-09:40	ThA11.1
Robust Design of Sparse Relative Sensing Networks, pp. 1860-1865
Schuler, Simone	Univ. of Stuttgart
Zelazo, Daniel	Tech. - Israel Inst. of Tech.
Allgower, Frank	Univ. of Stuttgart
This paper considers the robust design of sparse relative sensing networks subject to a given Hinf-performance constraint. The topology design considers heterogenous agents over weighted graphs. We develop a robust counterpart to the uncertain optimization problem and formulate the sparsity constraint via a convex l1-relaxation. We also demonstrate how this relaxation can be used to embed additional performance criteria, such as the maximization of the algebraic connectivity of the relative sensing network.
09:40-10:00	ThA11.2
Convergence Bounds for Discrete-Time Second-Order Multi-Agent-Systems, pp. 1866-1871
Eichler, Annika	Hamburg Univ. of Tech. (TUHH)
Werner, Herbert	Hamburg Univ. of Tech.
This paper presents convergence bounds for discrete-time second-order multi-agent systems with undirected or directed communication graphs. As has been shown before, the convergence depends on the eigenvalues of the Laplace matrix of the communication graph. For each eigenvalue (or eigenvalue pair) analytic bounds for the parameter set are given to render the protocol for that eigenvalue pair stable. In addition it is shown examplarily, that for the case of normalized Laplacian, the stabilizing solution set for the whole topology is non-empty.
10:00-10:20	ThA11.3
Stability of Continuous-Time Consensus Algorithms for Switching Networks with Bidirectional Interaction, pp. 1872-1877
Matveev, Alexey S.	St.Petersburg Univ.
Novinitsyn, Igor	St.Petersburg State Univ.
Proskurnikov, Anton	St.-Petersburg State Univ.
We consider the problem of reaching consensus in a network of the first-order continuous-time agents with switching topology. The coupling gains are neither piecewise-ontinuous, nor separated from zero, nor symmetric, though the network topology is assumed to be undirected. The main result of the paper provides conditions on the network topology that are necessary and sufficient for reaching consensus.
10:20-10:40	ThA11.4
Distributed Stabilization of 3D Circular Formations, pp. 1878-1883
El-Hawwary, Mohamed I.	Cairo Univ.
The paper presents distributed control design to stabilize circular formations of steered particles in three-dimensional space. In formation, the particles are required to follow equal radius circular paths with common orientation, but not necessarily common center. The formation is given by specifying desired separations of the centers of the circular paths and desired relative headings. The information exchange between the particles is modeled by a directed graph which is assumed to have a spanning tree. Control design is based on a hierarchical approach utilizing a reduction principle for asymptotic stability of closed sets.
10:40-11:00	ThA11.5
Consensus of Multi-Agent Systems with Nonuniform Non-Differentiable Time-Varying Delays, pp. 1884-1889
Savino, Heitor J.	Univ. Federal de Minas Gerais
Cota, Anna Paula	Univ. Federal de Minas Gerais
Souza, Fernando de Oliveira	Federal Univ. of Minas Gerais
Pimenta, Luciano Cunha de Araújo	Federal Univ. of Minas Gerais
Mendes, Eduardo Mazoni Andrade Marçal	Federal Univ. of Minas Gerais
Mozelli, Leonardo	Federal Univ. of São João Del-Rei
In this paper the consensus problem for continuous time multi-agent systems in the presence of time-delay is addressed. A novel sufficient condition for the case of nonuniform non-differentiable time-varying delays with minimum value greater than zero and a method to compute an estimate of the convergence rate are given. Simulation examples are given to show the performance of the proposed method.
11:00-11:20	ThA11.6
H_infinity Almost Regulated Synchronization and H_infinity Almost Formation for Heterogeneous Networks under External Disturbances, pp. 1890-1895
Peymani Foroushani, Ehsan	Norwegian Univ. of Sci. and Tech.
Grip, Håvard Fjær	NTNU
Saberi, Ali	Washington State Univ.
Wang, Xu	New York Univ.
Fossen, Thor I.	Norwegain Univ. of Sci and Tech.
We introduce the notions of H_infinity almost regulated synchronization and H_infinity almost formation for multi-agent systems subject to external disturbances and under directed interconnection structures. We assume that agents are linear, right-invertible and introspective with non-identical dynamics. The objective is to suppress the impact of disturbances on the synchronization error dynamics in terms of the H_infinity norm of the corresponding closed-loop transfer function. Inspired by the time-scale structure assignment techniques based on the singular perturbation theory, a family of observer-based protocols is introduced to achieve synchronization with any desired accuracy.

ThA12	HG D3.2
System Identification I (Regular Session)
Chair: Diversi, Roberto	Univ. of Bologna
Co-Chair: Ramos, Jose A.	Nova Southeastern Univ.
09:20-09:40	ThA12.1
Identification of Errors-In-Variables Models As a Quadratic Eigenvalue Problem, pp. 1896-1901
Diversi, Roberto	Univ. of Bologna
Soverini, Umberto	Univ. of Bologna
The paper proposes a new approach for identifying linear dynamic errors-in-variables (EIV) models, whose input and output are affected by additive white noise. The method is based on a nonlinear system of equations consisting of part of the compensated normal equations and of a set of high order Yule-Walker equations. This system allows mapping the EIV identification problem into a quadratic eigenvalue problem that, in turn, can be mapped into a linear generalized eigenvalue problem. The system parameters are thus estimated without requiring the use of iterative identification algorithms. The effectiveness of the method has been tested by means of Monte Carlo simulations and compared with those of other EIV identification methods.
09:40-10:00	ThA12.2
Algebraic Parameter Estimation of a Multi-Sinusoidal Waveform Signal from Noisy Data, pp. 1902-1907
Ushirobira, Rosane	Inria Lille - Nord Europe & Univ. de Bourgogne
Perruquetti, Wilfrid	Ec. Centrale de Lille
Mboup, Mamadou	Univ. de Reims Champagne Ardenne
Fliess, Michel	Ec. Pol.
In this paper, we apply an algebraic method to estimate the amplitudes, phases and frequencies of a biased and noisy sum of complex exponential sinusoidal signals. Let us stress that the obtained estimates are integrals of the noisy measured signal: these integrals act as time varying filters. Compared to usual approaches, our algebraic method provides a more robust estimation of these parameters within a fraction of the signal's period. We provide some computer simulations to demonstrate the efficiency of our method.
10:00-10:20	ThA12.3
Errors-In-Variables Identification of Thermal Models for Many-Core Computing Systems, pp. 1908-1915
Diversi, Roberto	Univ. of Bologna
Tilli, Andrea	Univ. of Bologna
Bartolini, Andrea	Univ. of Bologna - DEIS
Benini, Luca	Univ. of Bologna - DEIS
A crucial issue in next generation many-core computing systems is the dynamic thermal control for run-time performance optimization. To this end, scalable, compact and effective thermal models are essential. In this paper, we present an algorithm based on the Frisch scheme to obtain identification of distributed core-centric interacting models, dealing with very noisy temperature measurements and high process noises. We apply the proposed solution to an Intel's Single-chip-Cloud-Computer (SCC), a many-core prototype with 48 cores.
10:20-10:40	ThA12.4
A Local Approach Framework for Black-Box and Gray-Box LPV System Identification, pp. 1916-1921
Vizer, Daniel	Univ. of Tech. and Ec. of Budapest, Univ. d
Mercère, Guillaume	Poitiers Univ.
Prot, Olivier	Univ. de Limoges
Ramos, Jose A.	Nova Southeastern Univ.
This paper presents a comparison of two techniques dedicated to the identification of LPV systems by using local experiments only. Such an approach can be justified by the fact that, in many practical cases, exciting the scheduling variables persistently is not conceivable for safety/economic reasons. According to the prior information available on the system, a black-box and a gray-box model-based technique are described and compared through a simulation example. More precisely, a new version of the algorithm suggested in [1] is compared with a gray-box model-based technique consisting in interpolating local re-structured LTI state-space models, whose basis coherence is ensured thanks to prior knowledge about the system to identify. This contribution shows that prior information can be really helpful when the problem of coherent basis selection arises.
10:40-11:00	ThA12.5
A Generalized Vector Observation for Discrete-Time Varying Delay Systems Identification, pp. 1922-1927
Bedoui, Saïda	Gabes Univ. National Engineering school of Gabes
Ltaief, Majda	National School of Engineering of Gabes Tunisia
Kamel, Abderrahim	National School of Engineers of Gabes
This paper suggests an extended recursive least squares algorithm for simultaneous identification of an unknown time delay and dynamic parameters of discrete-time varying delay systems. The basic idea is to obtain a new formulations allowing to admit the unknown time varying delay in the parameter vector. The recursive least-squares method is then used to deal with the identification problem. A simulation study is included to illustrate the merit of our algorithm.
11:00-11:20	ThA12.6
Mixed Frequency Structured AR Model Identification, pp. 1928-1933
Zamani, Mohsen	Australian National Univ.
Felsenstein, Elisabeth	Vienna Univ. of Tech.
Anderson, Brian D.O.	Australian National Univ.
Deistler, Manfred	Tech. Univ. of Vienna
This paper is concerned with identifiability of an underlying high frequency multivariate stable singular AR system from mixed frequency observations. Such problems arise for instance in economics when some variables are observed monthly whereas others are observed quarterly. In particular, this paper studies stable singular AR systems where the covariance matrix associated with the vector obtained by stacking observation vector, $y_t$, and its lags from the first lag to the $p$-th one ($p$ is the order of the AR system), is also singular. To deal with this, it is assumed that the column degrees of the associated polynomial matrix are known. We consider first that there are given nonzero unequal column degrees and we show generic identifiability of the system and noise parameters. Then we extend the results to allow zero column degrees corresponding to fast components. In this case, we first show generic identifiability of the subsystem of the components with nonzero column degree. Then we show how to obtain those components of the parameter matrices of the components corresponding to zero column degree by regression.

ThA13	HG D5.2
Fault Diagnosis I (Regular Session)
Chair: Raimondo, Davide Martino	Univ. of Pavia
Co-Chair: Reppa, Vasso	Univ. of Cyprus
09:20-09:40	ThA13.1
Semidefinite Programming Relaxation of Optimum Active Input Design for Fault Detection and Diagnosis: Model-Based Finite Horizon Prediction, pp. 1934-1939
Kim, Kwang-Ki	Univ. of Illinois
Braatz, Richard D.	Massachusetts Inst. of Tech.
This paper establishes optimal/suboptimal active fault detection and diagnosis (FDD) methods in which semidefinite programming relaxation is used and the optimality criteria are information theoretic measures of the statistical distance between probability distributions. The design problems are formulated as optimizations in which an optimal sequence of inputs within a prediction horizon is computed for maximizing the statistical discrimination of different models of fault scenarios. Three different measures for the degree of statistical distinguishability between two hypothesized stochastic dynamical system models are considered and their mathematical properties that are related to Bayesian hypothesis tests are studied. The resulting input design problems are non-convex and we propose associated convex relaxation methods that can be solved in polynomial time using interior point methods. In addition, an upper bound on the sub-optimality of the proposed convex relaxation is presented for the case when there is only the input amplitude constraint, and randomized algorithms are presented to compute a suboptimal solution from an optimal solution of the convex relaxation problem. Numerical simulations with an aircraft model are provided to illustrate and demonstrate the presented methods of optimal input design for FDD.
09:40-10:00	ThA13.2
Optimum Input Design for Fault Detection and Diagnosis: Model-Based Prediction and Statistical Distance Measures, pp. 1940-1945
Kim, Kwang-Ki	Univ. of Illinois
Raimondo, Davide Martino	Univ. of Pavia
Braatz, Richard D.	Massachusetts Inst. of Tech.
This paper proposes optimization-based active fault detection and diagnosis (FDD) methods. An optimal input sequence is computed for maximizing discrimination between system models of fault scenarios in a statistical sense. Two different measures quantifying the degree of distinguishability between two stochastic LTI system models are considered, and their geometric properties are investigated. Their connection to the generalized likelihood ratio tests are also presented. Constrained open- and closed-loop feedback input design methods using model-based prediction are presented. Constraints on the predicted controlled output trajectory are imposed for ensuring operational safety as well as the input constraints that correspond to hardware limitations. Receding horizon method is used to implement the computed inputs.
10:00-10:20	ThA13.3
Fault Diagnosis in Managed Pressure Drilling Using Nonlinear Adaptive Observers, pp. 1946-1951
Willersrud, Anders	Univ. of Science and Tech. (NTNU)
Imsland, Lars	Norwegian Univ. of Science and Tech.
A bank of nonlinear adaptive observers is used for fault diagnosis in oil and gas drilling where managed pressure drilling (MPD) is applied. The particular fault considered is formation of a pack-off, causing increased friction in one part of the annulus. The process model is a simplified hydraulics model with a Newtonian fluid. All states in the model are assumed measurable, an assumption based on planned implementation of the wired drill pipe measurement technology. A fault detection observer is used to detect that a pack-off is being formed somewhere in the annulus. Then a set of fault isolation and approximation observers, one for each possible fault, is used to isolate the location of the pack-off and estimating its magnitude. Isolation is done by using residuals of annular friction estimation. The method for fault diagnosis is illustrated in a simulation study.
10:20-10:40	ThA13.4
Multiple Sensor Fault Detection and Isolation for Large-Scale Interconnected Nonlinear Systems, pp. 1952-1957
Reppa, Vasso	Univ. of Cyprus
Polycarpou, Marios M.	Univ. of Cyprus
Panayiotou, Christos	Univ. of Cyprus
This paper presents the design of a methodology for detecting and isolating multiple sensor faults in large-scale interconnected nonlinear systems. For each of the interconnected subsystems, we design a local sensor fault diagnosis (LSFD) agent responsible for multiple sensor fault detection and isolation in the local sensor set. The multiple sensor fault detection is realized through a bank of modules, monitoring smaller groups of sensors that belong to the local sensor set. The detection of faults in sensor groups is conducted using robust analytical redundancy relations, formulated by structured residuals and adaptive thresholds. The isolation of multiple faulty sensors in the local sensor set is realized by integrating the decisions of the LSFD agent's modules and applying a reasoning-based combinatorial decision logic. The simulation example of an automated highway system is used to illustrate the application of the multiple SFDI methodology.
10:40-11:00	ThA13.5
A Generative Approach to Qualitative Trend Analysis for Batch Process Fault Diagnosis, pp. 1958-1963
Villez, Kris	Eawag
Rengaswamy, Raghunathan	Clarkson Univ. Texas Tech. Univ.
Most of the existing methods for qualitative trend analysis are based on discriminative models. A disadvantage of such models is that many heuristic rules or local search methods are needed. Recently, an effort has been made to develop a globally optimal method for qualitative trend analysis. This method is based on a generative (rather than discriminative) model and has shown to lead to excellent performance. However, this method comes at an extreme computational demand which renders the methods unlikely for on-line application. In this work, an alternative method, while still generative in nature, is proposed which is shown to deliver the same performance while reducing the computational demand considerably.
11:00-11:20	ThA13.6
Isolating Faulty Variables for Fault Propagation Using Bayesian Decision Theory, pp. 1964-1969
Liu, Jialin	National Tsing Hua Univ.
Wong, David Shan Hill	National Tsing Hua Univ.
Chen, Ding-Sou	China Steel Corp.
Isolating fault variables is a crucial step to provide the information that which variables are responsible for the fault for diagnosing the root causes of a process fault. In chemical processes, process faults rarely show a random behavior; on the contrary, they will be propagated to varying variables due to the actions of the process controllers. During the evolution of a fault, the task of isolating faulty variables needs to be concerned with the faulty variables decided in the previous data; in addition, the current decisions should influence the isolation results for the next sample when the fault is constantly occurring. In the presented work, an unsupervised data-driven fault isolation method was developed based on Bayesian decision theory. The proposed approach successfully located the faulty variables that were individually responsible for the simultaneous occurrence of multiple sensor faults and a process fault.

ThA14	HG D7.1
Control of Internal Combustion Engines (Regular Session)
Chair: Aschemann, Harald	Univ. of Rostock
Co-Chair: Tayamon, Soma	Uppsala Univ.
09:20-09:40	ThA14.1
A Multiple Model-Based Controller for NOx Reduction in a Selective Catalytic Reduction System, pp. 1970-1975
Tayamon, Soma	Uppsala Univ.
Zambrano, Darine	Univ. De Los Andes
This paper deals with the design of a multiple model based controller for the nitrogen oxide emissions (NOx), from vehicles using the selective catalyst as an aftertreatment system. The selective catalyst reduction (SCR) system is nonlinear, since the chemical reactions involved are highly depending on the operating points. Local linear models were used for identification. Local internal controllers are integrated into a global controller. A Bayesian approach is used to mix the local output of the controllers. A detailed simulator is used for the multiple model identification and testing the controller. For validation, experimental data based on a standard transient test developed for Euro VI testing are used in the simulator. Results obtained for this control approach are compared to one model controller.
09:40-10:00	ThA14.2
Influence of Transitions between SI and HCCI Combustion on Driving Cycle Fuel Consumption, pp. 1976-1981
Nuesch, Sandro	Univ. of Michigan
Hellstrom, Erik	Univ. of Michigan
Jiang, Li	Robert Bosch LLC
Stefanopoulou, Anna G.	Univ. of Michigan
A model for a vehicle equipped with a multimode combustion engine is formulated and calibrated to experiments. A study is performed for the FTP-75 driving cycle about the influence of the mode switch strategy on the engine operation and fuel consumption, using simulation results and driving cycle measurements. It is found that the fuel penalty for mode transitions cancels a significant part of the theoretical benefits of advanced combustion modes. A smoothing strategy, where a mode switch is delayed, is introduced and it is shown that, by computing the optimal wait time, the negative influence of mode transitions on the overall efficiency can be reduced.
10:00-10:20	ThA14.3
Easily Adaptable Model of Test Benches for Internal Combustion Engines, pp. 1982-1987
Blumenschein, Josef	JKU Linz
Schrangl, Patrick	JKU Linz
Passenbrunner, Thomas Ernst	Johannes Kepler Univ. Linz
Trogmann, Hannes	Johannes Kepler Univ. Linz
Del Re, Luigi	Johannes Kepler Univ. Linz
Internal combustion engine test benches allow to simulate the operation of an engine under various conditions and are essential in engine development in the automotive industry. Models of such test benches are often used in the design and improvement of test bench controls and to simulate their behavior. Therefore, the paper summarizes a modeling procedure based on a simple model structure. In particular, the individual components of a test bench are described and their models are combined. A torque observer is used to compensate the lack of torque measurements. Finally, a model for a test bench equipped with a heavy duty engine is determined and validated with measurements.
10:20-10:40	ThA14.4
Model-Predictive On-Off Control of a Combustion-Heating-System for Vehicles, pp. 1988-1993
Butt, Saif Siddique	Univ. of Rostock
Prabel, Robert	Univ. of Rostock
Aschemann, Harald	Univ. of Rostock
In this paper, a control-oriented model of a combustion-heating system for vehicles such as buses, coaches and trucks is developed. Based on this model, a novel model predictive on-off strategy is proposed for the control of the combustion-heating system. Experiments are carried out to obtain a fair comparison of the predictive on-off controller with a classical on-off controller. The classical on-off controller and the predictive on-off controller are evaluated with respect to fuel economy, reduction in the emission of pollutants and the number of on-off switching cycles of the combustion valve connected to the heater. A discrete-time Extended Kalman Filter is employed for the estimation of state variables, heat losses and the volumetric flow. Experimental results show that the predictive on-off control strategy leads to a superior performance in terms of fuel economy and switching action of the valve as compared to the classical on-off controller.
10:40-11:00	ThA14.5
Common Rail Injection System Controller Design Using Input-To-State Linearization and Optimal Control Strategy with Integral Action, pp. 1994-1999
Rivas, Maria	Gipsa Lab.
Sename, Olivier	INPG
Witrant, Emmanuel	Univ. Joseph Fourrier
Caillol, Christian	Univ. of Orleans
Higelin, Pascal	Univ. of Orleans
The goal is to develop a controller for a common rail injection system where the calibration process for the implementation is systematic and simpler. A 0D nonlinear model of the common rail system is designed. The common rail model contains strong nonlinearities which might be hard to handle from the control point of view. In order to overcome this constraint, an input-to-state linearization is applied to the common rail nonlinear model. This procedure yields with a virtual linear system, where two optimal Linear Quadratic Regulator LQR strategies are applied. Very good results are obtained with the LQR, the calibration process is dramatically simplified and the controller stability is ensured.
11:00-11:20	ThA14.6
Piecewise Affine Modeling of NOx Emission Produced by a Diesel Engine, pp. 2000-2005
Vereshchaga, Yana	Johannes Kepler Univ. Linz
Stadlbauer, Stephan	Johannes Kepler Univ. Linz
Bako, Laurent	Ec. des Mines de Douai
Del Re, Luigi	Johannes Kepler Univ. Linz
Nitrogen oxides (NOx) emission control is a critical task of engine control as far as raw emissions and aftertreat- ment is concerned. Hardware NOx sensor are now available, but they still have several limitations, starting from their costs. Against this background, a lot of work both inside companies and in the academy has been done to develop virtual sensors for these emissions. In this paper, we propose a minimum-complexity model in the form of a piecewise affine model, and show that its performance is comparable to more complex models and better than conventional sensors if dynamic response is concerned.

ThA15	HG D7.2
Control of Smart Grids and Micro Grids (Regular Session)
Chair: Piroddi, Luigi	Pol. di Milano
Co-Chair: Glielmo, Luigi	Univ. of Sannio
09:20-09:40	ThA15.1
Stability Concerns for Indirect Consumer Control in Smart Grids, pp. 2006-2013
Juelsgaard, Morten	Aalborg Univ.
Andersen, Palle	Aalborg Univ.
Wisniewski, Rafael	Section for Automation and Control, Aalborg Univ.
Demand side management will be an important tool for maintaining a balanced electrical grid in the future, when the penetration of volatile resources, such as wind and solar energy increases. Recent research focuses on two different management approaches, namely direct consumer control by an external third party, and indirect consumer control through incentives and price signals. In this work we present a simple formulation of indirect control, where the behavior of each consumer, is governed by local optimization of energy consumption. The local optimization accounts for both cost of energy and distribution losses, as well as any discomfort incurred by consumers from any shift in energy consumption. Our work will illustrate that in the simplest formulation of indirect control, the stability is greatly affected of both the behavior of consumers, and the number of consumers to include. We will show how instability is related to the local optimization problem of the consumer, and the information made available to him.
09:40-10:00	ThA15.2
Stochastic Model Predictive Control for Economic/Environmental Operation Management of Microgrids (I), pp. 2014-2019
Parisio, Alessandra	Royal Inst. of Tech. Stockholm, Sweden
Glielmo, Luigi	Univ. of Sannio
Microgrids are subsystems of the distribution grid which comprises generation capacities, storage devices and controllable loads, operating as a single controllable system that can operate either connected or isolated from the utility grid. In this work, microgrid management system is developed in a stochastic framework. It is seen as a constraint-based system that employs forecasts and stochastic techniques to manage microgrid operations. Uncertainties due to fluctuating demand and generation from renewable energy sources are taken into account and a two-stage stochastic programming approach is applied to efficiently optimize microgrid operations while satisfying a time-varying request and operation constraints. Mathematically, the stochastic optimization problem is stated as a mixed-integer linear programming problem, which is solved in an efficient way by using commercial solvers. The stochastic problem is incorporated in a Model Predictive Control (MPC) scheme to further compensate the uncertainty though the feedback mechanism. Simulations show the effective performance of the proposed approach.
10:00-10:20	ThA15.3
Fuzzy Demand Forecasting in a Predictive Control Strategy for a Renewable-Energy Based Microgrid, pp. 2020-2025
Ávila, Fernanda	Univ. de Chile
Saez, Doris	Univ. de Chile
Jiménez-Estévez, Guillermo	Univ. de Chile
Reyes, Lorenzo	Univ. de Chile
Nuñez, Alfredo	Delft Univ. of Tech.
In model based control approaches for the dynamic operation of renewable-energy based microgrid, an accurate demand forecast is crucial. However, the high level of uncertainties in the system and non-linearities make the task of prediction not easy. In this context, we propose the use of a stable Takagi & Sugeno (T&S) fuzzy model to perform the demand forecasting in a real-life microgrid located in Huatacondo, Chile. Based on real-data from the microgrid, located in northern Chile, the T&S fuzzy model was identified and compared with an adaptive neural network, showing the T&S fuzzy model better open-loop prediction capabilities. To increase the prediction capability, an analysis of the amount of historical data needed, and the frequency required for training purposes was also done. For the case study, it is suggested to use a large amount of data rather than increasing the training frequency.
10:20-10:40	ThA15.4
An Approximate Dynamic Programming Approach to the Energy Management of a Building Cooling System (I), pp. 2026-2031
Ceriani, Nicola Maria	Pol. di Milano
Vignali, Riccardo	Pol. DI MILANO
Piroddi, Luigi	Pol. di Milano
Prandini, Maria	Pol. di Milano
This paper is concerned with optimal energy management of micro-grids. The goal is to show that the problem of minimizing the operating costs of a micro-grid by coordinating and scheduling its components can be formulated as a constrained optimal control problem for a stochastic hybrid system. This, in turn can be addressed through the Dynamic Programming (DP) approach, and the resulting DP equations solved through approximate DP techniques. A simple case study of a building cooling system with two chillers serving a cooling load is presented to this purpose.
10:40-11:00	ThA15.5
A Game Theory Framework for Active Power Injection Management with Voltage Boundary in Smart Grids, pp. 2032-2037
Cavraro, Guido	Univ. of Padova
Badia, Leonardo	Univ. of Padova
Smart grids are a novel paradigm for energy distribution, where instead of the traditional directed flow from a producer to the consumers, several micro-generators are spread throughout the network. We focus on the problem of coordinating the active power injected into the grid among the micro-generators. Each of them aims at injecting the maximum amount of power, satisfying some operative constraint such as voltage boundaries, but a tradeoff must be found among these conflicting objectives. First, we characterize the active power increment region, i.e., the set of all the increments of injected power that, depending on the grid state, satisfy the voltage boundary. Based on this finding, we frame the problem within game theory and propose a distributed approach that achieves a fair share of the active power injection, while at the same time satisfying the voltage boundary.
11:00-11:20	ThA15.6
Voltage Stability and Robustness for Microgrid Systems (I), pp. 2038-2043
Wang, Le Yi	Wayne State Univ.
Polis, Michael	Oakland Univ.
Wang, Caisheng	Wayne State Univ.
Lin, Feng	Wayne State Univ.
Voltage stability is of essential importance for power grids. The emergence of distributed energy generators, controllable loads, and local-area energy storage capabilities have introduced new scenarios for distribution networks inwhich classical frameworks for voltage stability may be inadequate. This paper introduces a control-theoretic framework for studying voltage stability and its robustness, as well as optimal power management in distribution systems composed of networked microgrids. The framework involves descriptions of the loads and generators by nonlinear state space models and the network connections by a set of topology-based algebraic equations. The integration of the combined system leads to a general nonlinear state space model for the microgrid systems. Simplified microgrids are used to illustrate the concepts.

ThST1	HG F1, F3
Semi-Plenary 3: The Design and Control of Airborne Wind Turbines (Semi-Plenary Session)
Chair: Jones, Colin N	EPFL, Lausanne
11:30-12:15	ThST1.1
The Design and Control of Airborne Wind Turbines*
Vander Lind, Damon	Macani Power

ThST2	HG F7, F5
Semi-Plenary 4: Demand-Side Modeling, Estimation and Control in Electric Power Systems (Semi-Plenary Session)
Chair: Johansson, Karl Henrik	Royal Inst. of Tech.
11:30-12:15	ThST2.1
Demand-Side Modeling, Estimation and Control in Electric Power Systems*
Callaway, Duncan	UC Berkeley

ThB1	HG F1
Partial Control of Agents on Networks and Applications to Rerouting a Subset of Drivers on Freeways (Tutorial Session)
Chair: Bayen, Alexandre M.	Univ. of California at Berkeley
Co-Chair: Reilly, Jack	U.C. Berkeley
Organizer: Bayen, Alexandre M.	Univ. of California at Berkeley
Organizer: Reilly, Jack	U.C. Berkeley
13:40-14:40	ThB1.1
Nash Equilibria and Stackelberg Games on Horizontal Queueing Networks (I)*
Reilly, Jack	U.C. Berkeley
We study inefficiencies in horizontal queueing networks due to the selfish behavior of agents, by comparing social optima to Nash equilibria. We expand on the studies on routing games which traditionally model congestion with latency functions that increase with the flow on a particular link. This type of latency function cannot capture congestion effects on horizontal queueing networks. Latencies on horizontal queueing networks increase as a function of density, and flow can decrease with increasing latencies. We present a novel algorithm, quadratic in the number of links, for computing the Nash equilibrium that minimizes total cost (best Nash equilibrium). Price of stability is shown to be sensitive to changes in demand when links are near capacity. We also investigate possible strategies to reduce the inefficiency of Nash equilibria by studying the Stackelberg routing game: assuming a coordinator has control over a fraction of the flow on the network, what is a good way of routing the compliant flow so that the induced Nash equilibrium is closer to the social optimum than the initial Nash equilibrium? We show that the optimal Stackelberg strategy can be computed in polynomial time for this class of latency functions on parallel networks. This contrasts with results for monotonically increasing latency functions, where optimal Stackelberg strategies are NP-hard to compute. Increasing the compliance rate α improves the cost of the equilibrium, and we are interested in particular in the Stackelberg threshold, i.e. the minimal compliance rate that achieves a strict improvement. We derive the optimal Stackelberg cost as a function of the compliance rate α, and obtain, in particular, the expression of the Stackelberg threshold. Numerical results illustrate the results and demonstrate the effectiveness of partial rerouting strategies on parellel routes.
14:40-15:40	ThB1.2
Adjoint-Based Optimization of PDE Systems on Networks with Applications to Ramp Metering and Partial Rerouting (I)*
Bayen, Alexandre M.	Univ. of California at Berkeley
A common cause of congestion on freeways is an excess of flow from onramps, which leads to reduced capacity over a large stretch of the freeway. By intelligently limiting the flow onto the freeway from onramps, congestion can be mitigated. This problem can be posed as a finite-horizon, optimal control problem, where the ramp-metering rates across different ramps at different times constitutes the control parameters. Unfortunately, the combination of nonlinear flow dynamics based on a system of first-order PDE's and the dimensionality of the problem make this problem difficult for standard optimization algorithms. We present a gradient-based algorithm based on the adjoint method which leads to efficient solutions of this problem, while also accurately modeling the underlying dynamics. With the growing popularity of personal navigation systems in vehicles, routing advice based on centrally-aggregated demand information is becoming increasingly feasible. One issue that arises when considering this problem is that not all vehicles on the road will be equipped with a navigation device, or be willing to follow centrally-given advice. Furthermore, accurate demand information is unavailable for non-participating vehicles, making it difficult for a central control to account for their effect on the system. The same networked PDE's adjoint method can be applied to solve for optimal rerouting strategies with limited participation. We propose a model for studying optimal rerouting for a subset of users on a multiple source, multiple sink model. A multi-class traffic model is used to divide participating users by origin-destination pairs, while non-participating users are modeled as a single class with behavior learned from pervasive static sensor data. The finite-horizon optimal control problem is efficiently solved using the adjoint method.

ThB2	HG F3
Nonlinear Model Predictive Control (Regular Session)
Chair: Glad, S. Torkel	Linkoping Univ.
Co-Chair: Baldea, Michael	Univ. of Texas at Austin
13:40-14:00	ThB2.1
Economic Model Predictive Control with Self-Tuning Terminal Weight, pp. 2044-2049
Muller, Matthias A.	Univ. of Stuttgart
Angeli, David	Univ. of Firenze
Allgower, Frank	Univ. of Stuttgart
In this paper, we propose an economic model predictive control (MPC) framework with a self-tuning terminal weight, which builds on a recently proposed MPC algorithm with a generalized terminal state constraint. First, given a general time-varying terminal weight, we derive an upper bound on the closed-loop average performance which depends on the limit value of the predicted terminal state. After that, we derive conditions for a self-tuning terminal weight such that bounds for this limit value can be obtained. Finally, we propose several update rules for the self-tuning terminal weight and analyze their respective properties. We illustrate our findings with several examples.
14:00-14:20	ThB2.2
Robust Stabilizing Output Feedback Nonlinear Model Predictive Control by Using Passivity and Dissipativity, pp. 2050-2055
Yu, Han	Univ. of Notre Dame
Zhu, Feng	Univ. of Notre Dame
Xia, Meng	Univ. of Notre Dame
Antsaklis, Panos J.	Univ. of Notre Dame
Motivated by the passivity-based nonlinear model predictive control (NMPC) scheme reported in [1], in this paper, we propose a robust stabilizing output feedback NMPC scheme by using passivity and disspativity. Model discrepancy between the nominal model and the real system is characterized by comparing the outputs for the same excitation function, and with this kind of characterization, we are able to compare the supply rate between the nominal model and the real system based on their passivity indices. Then, by introducing specific stabilizing constraint based on the passivity indices of the nominal model into the MPC, we show that our proposed NMPC scheme can stabilize the real system to be controlled.
14:20-14:40	ThB2.3
Nonlinear Model Predictive Control Using Feedback Linearization and Local Inner Convex Constraint Approximations, pp. 2056-2061
Simon, Daniel	Linköping Univ.
Löfberg, Johan	Linköping Univ.
Glad, S. Torkel	Linkoping Univ.
Model predictive control (MPC) is one of the most popular advanced control techniques and is used widely in industry. The main drawback with MPC is that it is fairly computationally expensive and this has so far limited its practical use for nonlinear systems. To reduce the computational burden of nonlinear MPC, Feedback Linearization together with linear MPC has been used successfully to control nonlinear systems. The main drawback is that this results in an optimization problem with nonlinear constraints on the control signal. In this paper we propose a method to handle the nonlinear constraints that arises using a set of dynamically generated local inner polytopic approximations. The main benefits of the proposed method is guaranteed recursive feasibility and convergence.
14:40-15:00	ThB2.4
Concurrent Nonlinear Predictive Control and Economic Management of Energy-Integrated Systems, pp. 2062-2067
Baldea, Michael	Univ. of Texas at Austin
Touretzky, Cara R.	Univ. of Texas at Austin
We consider a class of process systems with significant energy recovery. Extending our previous results concerning the two time scale dynamics of such systems, we demonstrate that the fast component of their dynamics is asymptotically stable in most physically-relevant cases. Using this result, we develop a composite controller design, consisting of a linear control system for the fast dynamics and a MISO nonlinear model predictive controller for the slow dynamics, and demonstrate that it guarantees exponential stability for the overall system. Subsequently, we establish a parallel between this approach and economic model predictive control, and show its implications in optimal energy management. We illustrate our theoretical developments with a simulation case study.
15:00-15:20	ThB2.5
Discrete-Time Incremental ISS: A Framework for Robust NMPC, pp. 2068-2073
Bayer, Florian	Univ. of Stuttgart
Bürger, Mathias	Univ. of Stuttgart
Allgower, Frank	Univ. of Stuttgart
In this paper, Incremental Input-to-State Stability is studied as a system theoretic framework to address the challenges of robust nonlinear model predictive control. In the first part of the paper, a Lyapunov framework for Incremental Input-to-State Stability of nonlinear discrete-time dynamical systems is established. In the second part, Incremental Input-to-State Stability is shown to lead to an efficient MPC method for disturbed nonlinear systems. Based on the Incremental Input-to-State Stability Lyapunov function, a tightening of the constraints is proposed. Satisfaction of the tightened constraints can be guaranteed under the disturbances. By this concept, a robust nonlinear model predictive control problem is handled and the effectiveness is shown through an example from the literature.
15:20-15:40	ThB2.6
Simulation Based Evaluation of a Nonlinear Model Predictive Controller for Friction Stir Welding of Nuclear Waste Canisters, pp. 2074-2079
Nielsen, Isak	Linköping Univ.
Garpinger, Olof	Lund Univ.
Cederqvist, Lars	Swedish Nuclear Fuel and Waste Management Company
The Swedish nuclear waste will be stored in copper canisters and kept isolated deep under ground for more than 100,000 years. To ensure reliable sealing of the canisters, friction stir welding is used. To repetitively produce high quality welds, it is vital to use automatic control of the process. This paper introduces a nonlinear model predictive controller for regulating both plunge depth and stir zone temperature, which has not been presented in literature before. Further, a nonlinear process model has been developed and used to evaluate the controller in simulations of the closed loop system. The controller is compared to a decentralized solution, and simulation results indicate that it is possible to achieve higher control performance using the nonlinear model predictive controller.

ThB3	HG F5
Estimation and Control of DPS II (Invited Session)
Chair: Demetriou, Michael A.	Worcester Pol. Inst.
Co-Chair: Meurer, Thomas	Christian-Albrechts-Univ. Kiel
Organizer: Demetriou, Michael A.	Worcester Pol. Inst.
Organizer: Meurer, Thomas	Christian-Albrechts-Univ. Kiel
13:40-14:00	ThB3.1
On Some Open Questions in Bilinear Quantum Control (I), pp. 2080-2085
Boscain, Ugo V.	SISSA-ISAS
Chambrion, Thomas	Univ. de Lorraine
Sigalotti, Mario	INRIA Saclay
The aim of this paper is to provide a short introduction to modern issues in the control of infinite dimensional closed quantum systems, driven by the bilinear Schrodinger equation. The first part is a quick presentation of some of the numerous recent developments in the fields. This short summary is intended to demonstrate the variety of tools and approaches used by various teams in the last decade. In a second part, we present four examples of bilinear closed quantum systems. These examples were extensively studied and may be used as a convenient and efficient test bench for new conjectures. Finally, we list some open questions, both of theoretical and practical interest.
14:00-14:20	ThB3.2
Parametric State Feedback Design for Linear Infinite-Dimensional Systems (I), pp. 2086-2091
Mohr, Andreas	Univ. Erlangen-Nürnberg
Deutscher, Joachim	Univ. Erlangen-Nürnberg
In this paper, a parametric approach for state feedback controllers for a class of linear infinite-dimensional systems is presented. Therein, besides the closed-loop eigenvalues, the corresponding parameter vectors are introduced as design parameters. This allows not only to assign finitely many eigenvalues, but also the corresponding eigenvectors. The presented approach is applied to solve a partial eigenstructure assignment problem for linear time-delay systems with point-delays. An example for the usefulness of the proposed design method is then given by a partial input-output decoupling of a simple time-delay system.
14:20-14:40	ThB3.3
$H_{infty}$-Control with State Feedback of an Inclined Cable (I), pp. 2092-2097
Baudouin, Lucie	LAAS-CNRS ; Univ. de Toulouse
Neild, Simon	Univ. of Bristol, Department of Mechanical Engineering
Wagg, David	Univ. of Bristol, Department of Mechanical Engineering
In this study we consider the robust control of an inclined cable modelled using partial differential equation and subjected to external disturbances. This paper focuses on the construction of a standard linear infinite dimensional state space system and an H-infinity feedback control with full observation of the state.
14:40-15:00	ThB3.4
Passivity-Based Control of Implicit Port-Hamiltonian Systems (I), pp. 2098-2103
Macchelli, Alessandro	Univ. of Bologna - Italy
The main contribution of this paper is the generalisation of well-known energy-based control techniques (i.e., energy-balancing passivity-based control and passivity-based control with state modulated source), to the case in which the plant is a port-Hamiltonian system in implicit form. A typical situation is when (part of) the system is obtained from the spatial discretization of an infinite dimensional port-Hamiltonian system: in this case, the dynamics is not given in standard input-state-output form, but as a set of DAEs. Consequently, the control by energy-shaping has to be extended to deal with dynamical systems with constraints. The general methodology is discussed with the help of a simple but illustrative example, i.e. a transmission line interconnected with an RLC circuit.
15:00-15:20	ThB3.5
On an Approximation with Prescribed Zeros of SISO Abstract Boundary Control Systems (I), pp. 2104-2109
Iftime, Orest V.	Univ. of Groningen
Ionescu, Tudor Corneliu	Imperial Coll. London
Finite-dimensional approximations of partial differential equations are used not only for simulation, but also for controller design. Modal truncation and numerical approximation are common practical methods for approximating distributed parameter systems. The modal approximation preserves the exact, low-order poles of the original system. However, the zeros of modal approximations may differ significantly from those of the original distributed parameter system. In particular, right half-plane zeros, which are not present in the original infinite-dimensional model, may appear in modal truncations. In this paper we consider a boundary control system and propose a moment matching based approximation which preserves a prescribed set of zeros. To illustrate the advantages of the method, we consider its application to the heat equation with Neumann boundary control at the right end (HENBCR). Although the modal approximation provides good error bounds for the HENBCR, it contains non-minimum phase zeros which lead to erroneous predictions. The moment matching approach sketched in this paper yields an approximation of the HENBCR with minimum phase zeros only. We consider that the numerical example is very interesting and convincing for the reader. Further theoretical analysis will be addressed in the full paper.
15:20-15:40	ThB3.6
On the Control of Spin-Boson Systems (I), pp. 2110-2115
Boscain, Ugo V.	SISSA-ISAS
Mason, Paolo	CNRS, Lab. des Signaux et Systèmes, Supélec
Panati, Gianluca	Univ. degli Studi di Roma "La Sapienza"
Sigalotti, Mario	INRIA Saclay
In this paper we study the so-called spin-boson system, namely a spin-1/2 particle in interaction with a distinguished mode of a quantized bosonic field. We control the system via an external field acting on the bosonic part. Applying geometric control techniques to the Galerkin approximation and using perturbation theory to guarantee non-resonance of the spectrum of the drift operator, we prove approximate controllability of the system, for almost every value of the interaction parameter.

ThB4	HG F7
Observers for Nonlinear Systems (Regular Session)
Chair: Richard, Jean-Pierre	Ec. Centrale de Lille
Co-Chair: Maree, Johannes Philippus	Norwegian Univ. of Science and Tech.
13:40-14:00	ThB4.1
On Reduced-Order Interval Observers for Time-Delay Systems, pp. 2116-2121
Efimov, Denis	INRIA - LNE
Perruquetti, Wilfrid	Ec. Centrale de Lille
Richard, Jean-Pierre	Ec. Centrale de Lille
The estimation problem for uncertain time-delay systems is addressed. A design method of reduced-order interval observers is proposed. The observer estimates the set of admissible values (the interval) for the state at each instant of time. The cases of known fixed delays and uncertain time-varying delays are analyzed. The proposed approach can be applied to linear time-delay systems and nonlinear time-delay systems in the output canonical form. The framework efficiency is demonstrated on examples of nonlinear systems.
14:00-14:20	ThB4.2
State Observer for a First-Order Plant under Intrinsic Pulse-Modulated Feedback: A Case Study, pp. 2122-2127
Churilov, Alexander	St.Petersburg State Marine Tech. Univ.
Medvedev, Alexander V.	Uppsala Univ.
Zhusubaliev, Zhanybai	South-West State Univ.
Dynamical behaviors in a recently proposed observer for linear time-invariant systems under intrinsic pulse-modulated feedback are studied. A special case of scalar continuous dynamics is considered as it is not covered by the previously presented mathematical analysis. Notably, the lowest non-trivial differential order of the continuous part of the plant results in a more complex dynamics of the observer. In fact, the convergence of the observer becomes dependent on the observer initialization, which phenomenon does not exist in the case of the second and higher order continuous dynamics. As an alternative to the static gain observer, an integral feedback observer is suggested that exhibits global convergence for certain values of the observer gain and types of the periodic solution in the observed plant. Extensive bifurcation analysis is though necessary to select a proper observer gain.
14:20-14:40	ThB4.3
A Novel Adaptive Unscented Kalman Filter Attitude Estimation and Control System for a 3U Nanosatellite, pp. 2128-2133
Li, Junquan	York Univ.
Post, Mark	York Univ.
Lee, Regina	York Univ.
A novel adaptive unscented Kalman filter (AUKF) based estimation algorithm is proposed for a 3U Cubsat. This small satellite employs a three axis magnetometer and three MEMS gyroscopes as well as three magnetic torque rods and one reaction wheel on the pitch axis. Unlike the existing UKF, in this paper, an n+1 sigma set is used to estimate the nanosatellite attitude instead of 2n+1 sigma points as in a conventional UKF. Numerical Simulation results validate the performance of the proposed adaptive Kalman filter. There is no need for linearization of the nonlinear dynamics of the system. The estimated result tracks satellite attitude during the damping and stable control stages. Euler angles, gyro bias, and angular velocity of the satellite are estimated using this proposed AUKF with good convergence time and estimation accuracy.
14:40-15:00	ThB4.4
A Convergence Result for the Unscented Kalman-Bucy Filter Using Contraction Theory, pp. 2134-2139
Maree, Johannes Philippus	Norwegian Univ. of Science and Tech.
Imsland, Lars	Norwegian Univ. of Science and Tech.
Jouffroy, Jerome	Univ. of Southern Denmark
This paper applies contraction theory to establish sufficient conditions for contraction, hence, exponential convergence of the unscented Kalman-Bucy Filter. It follows that regions of contraction can subsequently be defined, given such sufficient conditions. Both state, and measurement models are It^o type stochastic differential equations. By employing a virtual/actual system framework, a special relation is established between sigma-point dynamics, and observed process states, with respect to contraction and convergence. The proposed theory is illustrated on an isothermal, non-linear CSTR process.
15:00-15:20	ThB4.5
Cascade Observer for a Class of Nonlinear Systems with Output Delays, pp. 2140-2145
Fall, Mamadou Lamine	ENSICAEN, Univ. de Caen, CNRS
Farza, Mondher	Univ. DE CAEN, ENSICAEN
M'Saad, Mohammed	ENSICAEN
Pigeon, Eric	GREYC
Gehan, Olivier	ENSICAEN
This paper proposes a state observer with a cascade structure for a class of nonlinear systems in the presence of delayed output measurements. The first system in the cascade allows to estimate the delayed state while each of the remaining ones is a predictor. Each predictor estimates the state of the preceding one with a prediction horizon equal to a fraction of the time delay in such a way that the state of the last predictor is an estimate of the system actual state. The design of the observer is achieved by assuming a set of conditions under which the exponential convergence of the estimation error to zero is established, namely the system is uniformly observable for any any input and its nonlinearities are globally Lipschitz. Of particular interest, it is shown that the number of the systems in the cascade depends on the magnitudes of the considered delay and the Lipschitz constant. The performance of the proposed observer and its main properties are compared with those of two existing observers through a typical bioreactor model.
15:20-15:40	ThB4.6
State Observers for a Class of Continuous Nonlinear Systems with Discrete Measurements, pp. 2146-2151
Fall, Mamadou Lamine	ENSICAEN, Univ. de Caen, CNRS
Farza, Mondher	Univ. DE CAEN, ENSICAEN
M'Saad, Mohammed	ENSICAEN
Pigeon, Eric	GREYC
Gehan, Olivier	ENSICAEN
Mosrati, Ridha	Univ. de Caen, ERPCB
In this paper, we investigate the possibility of designing an observer for a class of continuous-time dynamical systems with non-uniformly sampled measurements. More specifically, we propose an observer with a time varying gain witch converges exponentially under some conditions on the sampling partition diameter. The proposed observer is an impulsive system since it is described by a set of differential equations with instantaneous state impulses corresponding to the measured samples and their estimates. As it is customarily done in the literature, we show that such an impulsive system can be split into two subsystems and be put under the form of a hybrid system which is designed using a continuous-time observer together with an inter-sample output predictor. Simulations results involving a typical bioreactor are given to show the effectiveness of the proposed observer

ThB5	HG E1.1
Switched Linear Systems (Regular Session)
Chair: Valcher, Maria Elena	Univ. di Padova
Co-Chair: Raisch, Joerg	Tech. Univ. Berlin
13:40-14:00	ThB5.1
A New Stability Result for Switched Linear Systems, pp. 2152-2156
Kouhi, Yashar	Max Planck Insitute
Bajcinca, Naim	Max Planck Insitute
Raisch, Joerg	Tech. Univ. Berlin
Shorten, Robert	Nat. Univ. of Ireland
In this paper we extend the result by Shorten and Narendra on common quadratic Lyapunov functions for pairs of matrices in companion form. Specifically, we show that their result extends to more general matrix pairs provided that an associated transfer function matrix is symmetric. Examples are given to illustrate the usefulness of this result.
14:00-14:20	ThB5.2
Measurable Disturbance Rejection with Quadratic Stability in Continuous-Time Linear Switching Systems, pp. 2157-2162
Zattoni, Elena	Alma Mater Studiorum - Univ. of Bologna
Marro, Giovanni	Univ. of Bologna
It is well-known that disturbances with different features (i.e., inaccessible, measurable, or previewed disturbances) must be handled with the appropriate compensation schemes: namely, those which better exploit the available information. In particular, this work is focused on rejection of disturbances accessible for measurement in continuous-time linear switching systems, with the requirement that the compensated system be quadratically stable under arbitrary switching. A dynamic feedforward switching compensator is designed on the assumption that the plant be quadratically stable under arbitrary switching. This assumption can be relaxed to quadratic stabilizability by linear state feedback and by linear output injection, provided that a measurement dynamic feedback stabilizer is also devised. The proposed techniques apply to linear switching systems whose modes may be either left-invertible or not. The methodology adopted is based on the use of the geometric approach enhanced with stability notions which are typically considered in linear switching systems.
14:20-14:40	ThB5.3
On Averaging for Switched Linear Differential Algebraic Equations, pp. 2163-2168
Iannelli, Luigi	Univ. of Sannio in Benevento
Pedicini, Carmen	Univ. of Sannio, Benevento
Trenn, Stephan	Univ. of Kaiserslautern
Vasca, Francesco	Univ. of Sannio
Averaging is an effective technique which allows the analysis and control design of nonsmooth switched systems through the use of corresponding simpler smooth averaged systems. Approximation results and stability analysis have been presented in the literature for dynamic systems described by switched ordinary differential equations. In this paper the averaging technique is shown to be useful also for the analysis of switched systems whose modes are represented by means of differential algebraic equations (DAEs). An approximation result is derived for a simple but representative homogenous switched DAE with periodic switching signals and two modes. Simulations based on a simple electrical circuit model illustrate the theoretical result.
14:40-15:00	ThB5.4
On the Stability and Stabilizability of a Class of Continuous-Time Positive Switched Systems with Rank One Difference, pp. 2169-2174
Fornasini, Ettore	Univ. of Padova
Valcher, Maria Elena	Univ. di Padova
Given a single-input continuous-time positive system, described by a pair (A,b), with A a diagonal matrix, we investigate under what conditions there exist state-feedback laws u(t) = c'x(t) that make the resulting controlled system positive and asymptotically stable, namely A+b c' Metzler and Hurwitz. In the second part of the paper we assume that the state-space model switches among different state-feedback laws c_i, i=1,2,...,p, each of them ensuring the positivity, and show that the asymptotic stability of the switched system is equivalent to the asymptotic stability of all the subsystems, while its stabilizability is equivalent to the existence of an asymptotically stable subsystem.
15:00-15:20	ThB5.5
On the Numerical Solution of the Control Problem of Switched Linear Systems, pp. 2175-2179
Vargas, Alessandro N	Univ. Tec Federal do Parana, UTFPR
Ishihara, João Yoshiyuki	Univ. of Brasília
Do Val, Joao B.R.	UNICAMP - FEEC
This paper presents a method to compute an epsilon-optimal solution of the control problem of switched linear systems. A difficulty that emerges in the evalution of the optimal solution is that the cardinality of the solution set increases exponentially as long as the time-horizon increases linearly, which turns the problem intractable when the horizon is sufficiently large. We propose a numerical method to overcome such difficulty, in the sense that our approach allows the evalution of epsilon-optimal solutions with corresponding sets that do not increase exponentially.
15:20-15:40	ThB5.6
On-Line Switching Signal Estimation of Switched Linear Systems with Measurement Noise, pp. 2180-2185
Lee, Chanhwa	Seoul National Univ.
Ping, Zhaowu	Seoul National Univ.
Shim, Hyungbo	Seoul National Univ.
In this paper, we study a switching signal estimation problem for continuous time switched linear control systems with measurement noise. Inspired by the work of (Battistelli,2011), we first propose a generalized minimum distance criterion to estimate the active mode of the plant with inputs. Then we propose an implementable on-line robust switching signal estimation algorithm to detect the switching time with guaranteed precision, where some update condition and threshold condition are checked all the time. Under the update and threshold conditions, we detect the switching time within a predetermined time interval after the switching occurred.

ThB6	HG E1.2
Stability of Nonlinear Systems II (Regular Session)
Chair: Normand-Cyrot, Marie-Dorothée	CNRS-Supélec
Co-Chair: Turner, Matthew C.	Univ. of Leicester
13:40-14:00	ThB6.1
Stabilization of Nonlinear Discrete-Time Dynamics in Strict Feedforward Form, pp. 2186-2191
Monaco, Salvatore	Univ. di Roma La Sapienza
Normand-Cyrot, Marie-Dorothée	CNRS-Supélec
The paper discusses stabilization of nonlinear discrete-time dynamics exhibiting strict feedforward structures. An iterative design procedure is proposed which makes use of average passivity based controllers. This concept has been recently introduced by the authors because the usual definition in unsuitable in the absence of direct input-output link. Due to the triangular structure of strict feedforward dynamics, it is possible, at each step of the procedure and through coordinates change, to reset the problem as the one of average passivity based controller design. The complete controller is derived at the last step and a Lyapunov function is constructed. An example concludes the paper.
14:00-14:20	ThB6.2
High-Order Zames-Falb Multiplier Analysis Using Linear Matrix Inequalities, pp. 2192-2197
Turner, Matthew C.	Univ. of Leicester
Sofrony, Jorge	Univ. nacional de colombia
This paper proposes an algorithm for stability analysis of systems containing slope-restricted nonlinearities using high-order Zames-Falb multipliers. The main innovation in this paper is the use of a new congruence transformation which enables multipliers of twice the order of the linear part of the system to be used in a linear-matrix-inequality (LMI) framework for stability analysis. Although the use of such high-order multipliers increases the algorithms computational requirements, various numerical examples show that the resulting stability bounds are sometimes less conservative than using other similar approaches.
14:20-14:40	ThB6.3
Controller Synthesis for Incremental Stability: Application to Symbolic Controller Synthesis, pp. 2198-2203
Zamani, Majid	Delft Univ. of Tech.
Van De Wouw, Nathan	Eindhoven Univ. of Tech.
Incremental stability is a property of dynamical and control systems, requiring the stability and convergence of trajectories with respect to each other, rather than with respect to an equilibrium point or a particular trajectory. Most design techniques providing controllers rendering control systems incrementally stable have two main drawbacks: they can only be applied to control systems in parametric-strict-feedback or strict-feedback form, and they require the control systems to be smooth. In this paper, we propose a controller design technique that is applicable to larger classes of control systems, including a class of non-smooth control systems. Moreover, we propose a recursive way of constructing incremental Lyapunov functions which have been identified as a key tool enabling the construction of finite abstractions of nonlinear control systems. The effectiveness of the proposed results in this paper is illustrated by synthesizing a controller rendering a non-smooth control system incrementally stable as well as constructing its finite abstraction, using the computed incremental Lyapunov function. Finally, using the constructed finite abstraction, we synthesize another controller for the incrementally stable closed-loop system enforcing the satisfaction of logic specifications, difficult (or even impossible) to enforce using conventional techniques.
14:40-15:00	ThB6.4
On an Extension of Homogeneity Notion for Differential Inclusions, pp. 2204-2209
Bernuau, Emmanuel	Ec. Lille
Efimov, Denis	INRIA - LNE
Perruquetti, Wilfrid	Ec. Centrale de Lille
Polyakov, Andrey	INRIA Lille Nord-Europe
The notion of geometric homogeneity is extended for differential inclusions. This kind of homogeneity provides the most advanced coordinate-free framework for analysis and synthesis of nonlinear discontinuous systems. Theorem of L. Rosier on a homogeneous Lyapunov function existence for homogeneous differential inclusions is presented. An extension of the result of Bhat and Bernstein about the global asymptotic stability of a system admitting a strictly positively invariant compact set is also proved.
15:00-15:20	ThB6.5
On the Stability of Lyapunov Exponents of Discrete Linear Systems, pp. 2210-2213
Czornik, Adam	Silesian Tech. Univ.
Nawrat, Aleksander	Silesian Tech. Univ.
Niezabitowski, Michal	Silesian Univ. of Tech. of AutomaticControl
In this paper we present a sufficient condition for continuity of Lyapunov exponents of discrete time-varying linear system. Basing on the result we show that Lyapunov exponents of time-invariant systems depend continuously on the time-varying perturbations.
15:20-15:40	ThB6.6
Synchronization Control of Fractional-Order Discrete-Time Chaotic Systems, pp. 2214-2219
Liao, Xiaozhong	Beijing Inst. of Tech.
Gao, Zhe	Beijing Inst. of Tech.
Huang, Hong	Beijing Inst. of Tech.
This paper investigates the bifurcation phenomenon of fractional-order discrete-time systems and proposes the nonlinear control to synchronize two fractional-order discrete-time systems. By taking a finite truncation, fractional-order discrete-time chaotic systems are constructed. Chaotic phenomenon is also determined by the fractional order. By adjusting the fractional-order appropriately, the chaotic areas can be controlled. The synchronization control is designed for the same and different fractional-orders in the drive and response systems. Finally, the bifurcation phenomenon and synchronization of fractional-order Logistic system are studied. The proposed method can be extended to synchronize other fractional-order discrete-time systems.

ThB7	HG E3
Idempotent Methods in Control and Estimation (Invited Session)
Chair: McEneaney, William	Univ. of California, San Diego
Co-Chair: Gaubert, Stephane	INRIA and Ec. Pol.
Organizer: Sridharan, Srinivas	Univ. of California San Diego
Organizer: McEneaney, William	Univ. of California, San Diego
Organizer: Gaubert, Stephane	INRIA and Ec. Pol.
13:40-14:00	ThB7.1
Deterministic Filtering for Optimal Attitude Estimation on $SO(3)$ Using Max-Plus Methods (I), pp. 2220-2225
Sridharan, Srinivas	Univ. of California San Diego
McEneaney, William	Univ. of California, San Diego
In this article we introduce the use of recently developed min/max-plus techniques in order to solve the optimal attitude estimation problem in filtering for nonlinear systems on the special orthogonal ($SO(3)$) group. This work helps synthesize deterministic filters for nonlinear systems -- i.e. optimal filters which estimate the system state using a related optimal control problem. The technique indicated herein is validated using a set of optimal attitude estimation example problems on $SO(3)$.
14:00-14:20	ThB7.2
Contraction of Riccati Flows Applied to the Convergence Analysis of the Max-Plus Curse of Dimensionality Free Method (I), pp. 2226-2231
Qu, Zheng	CMAP, Pol.
Max-plus based methods have been recently explored for solution of first-order Hamilton-Jacobi-Bellman equations by several authors. In particular, McEneaney's curse-of-dimensionality free method applies to the equations where the Hamiltonian takes the form of a (pointwise) maximum of linear/quadratic forms. In previous works of McEneaney and Kluberg, the approximation error of the method was shown to be $O(1/(Ntau))$+$O(sqrt{tau})$ where $tau$ is the time discretization step and $N$ is the number of iterations. Here we use a recently established contraction result of the indefinite Riccati flow in Thompson's metric to show that under different technical assumptions, still covering an important class of problems, the total error incorporating a pruning procedure of error order $tau^2$ is $O(e^{-alpha Ntau})+O(tau)$ for some $alpha>0$ related to the contraction rate of the indefinite Riccati flow.
14:20-14:40	ThB7.3
Max-Plus Based Computation of Nonlinear L2-Gain Performance Bounds Using a Piecewise Affine-Quadratic Basis (I), pp. 2232-2237
Zhang, Huan	The Univ. of Melbourne
Dower, Peter M.	The Univ. of Melbourne
Nonlinear L2-gain is a generalization of the well-known finite L2-gain robust stability property for nonlinear systems. Computation of tight performance bounds associated with this nonlinear L2-gain property is key to avoiding conservatism in its application, for example in small-gain based design. In previous work, a number of max-plus eigenvector methods have been proposed to facilitate this computation. Those methods have each employed quadratic basis functions, which have been shown to lead to a specific computational issue concerning continuity of the associated Hamiltonian. In this paper, an alternative piecewise affine-quadratic basis is proposed that allows the development of a refined max-plus eigenvector method that avoids this computational issue.
14:40-15:00	ThB7.4
A Max-Plus Based Fundamental Solution to a Class of Linear Regulator Problems with Non-Quadratic Terminal Payoff (I), pp. 2238-2243
Zhang, Huan	The Univ. of Melbourne
Dower, Peter M.	The Univ. of Melbourne
This paper studies a class of linear regulator problem where the terminal payoff function is not necessarily quadratic. The value function for this problem is generally not quadratic and thus it can not be reduced to solving the corresponding matrix Riccati equation as for the standard linear quadratic regulator (LQR) problem. The computational method of direct iteration using the dynamic programming equations is computationally expensive. In this paper, a new computational method based on max-plus techniques is developed for this problem which is demonstrated to be more efficient and more accurate. In particular, three max-plus fundamental solutions are obtained which can be used as the kernel of max-plus integration with respect to the max-plus dual of the terminal payoff to generate the value function of the linear regulator problem.
15:00-15:20	ThB7.5
Certification of Inequalities Involving Transcendental Functions: Combining SDP and Max-Plus Approximation (I), pp. 2244-2250
Allamigeon, Xavier	INRIA Saclay -- Ile-de-France and CMAP, Ec. Pol.
Gaubert, Stephane	INRIA and Ec. Pol.
Magron, Victor, Liev	INRIA/Pol.
Benjamin, Werner	Ec. Pol.
We consider the problem of certifying an inequality of the form $f(x)geq 0$, $forall xin K$, where $f$ is a multivariate transcendental function, and $K$ is a compact semi-algebraic set. We introduce a certification method, combining semi-algebraic optimization and max-plus approximation. We assume that $f$ is given by a syntaxic tree, the constituents of which involve semi-algebraic operations as well as some transcendental functions like $cos$, $sin$, $exp$, etc. We bound some of these constituents by suprema or infima of quadratic forms (max-plus approximation method, initially introduced in optimal control), leading to semi-algebraic optimization problems which we solve by semi-definite relaxations. The max-plus approximation is iteratively refined and combined with branch and bound techniques to reduce the relaxation gap. Illustrative examples of application of this algorithm are provided, explaining how we solved tight inequalities issued from the Flyspeck project (one of the main purposes of which is to certify numerical inequalities used in the proof of the Kepler conjecture by Thomas Hales).
15:20-15:40	ThB7.6
On Fourier Transform, Parseval Equality, and the Inversion Formula in Idempotent Analysis (I), pp. 2251-2256
Avantaggiati, Antonio	Sapienza, Univ. di Roma
Loreti, Paola	Sapienza, Univ. di Roma
In the paper we revisit some known remarkable formulas and their idempotent versions highlighting the role of functional spaces where the analogues take place and we discuss the different roles of R_min and R_max. More precisely, we focus on the space where the analogue of the Fourier transform takes place, the Legendre transform in R_min and the concave conjugate transform in R_max, then we give the analogues of some classical theorems.

ThB8	HG E5
Networked Systems (Regular Session)
Chair: Bitmead, Robert	Univ. of California San Diego
Co-Chair: Sarlette, Alain	Ghent Univ.
13:40-14:00	ThB8.1
Performance Cost of Adaptation in Mobile Wireless Power Control, pp. 2257-2262
Ha, Minh Hong	Univ. of California, San Diego
Bitmead, Robert	Univ. of California San Diego
Power control in mobile communications is an adaptive control system which functions routinely without direct human intervention. In this paper we examine the cost in terms of power usage of the one-parameter system identification problem of estimating the signal-to-noise ratio and thereby setting subsequent transmissions at the appropriate power level to accommodate the error detection and recovery. This is performed with an explicit formulation of the density function of the signal-to-noise ratio estimate and its dependence on the number of training samples used in this phase - it is a non-central $chi^2$ density if the noise power is known, else it is $F$-distributed. The objective is to develop a total cost function based on transmission power under which the price paid for adaptation becomes apparent. Our example is based on GSM mobile telephony systems and is part of a broader study into the costs associated with adaptation.
14:00-14:20	ThB8.2
Network Localization by Shadow Edges, pp. 2263-2268
Oliva, Gabriele	Univ. degli Studi Roma Tre
Panzieri, Stefano	Univ. degli Studi Roma Tre
Pascucci, Federica	Univ. degli Studi Roma Tre
Setola, Roberto	Univ. Campus BioMedico of Rome
Localization is a fundamental task for sensor networks. Traditional network localization approaches allow to obtain localized networks requiring the nodes to be at least tri-connected (in 2D), i.e., the communication graph needs to be globally rigid. In this paper we exploit, besides the information on the neighbors sensed by each robot/sensor, also the information about the lack of communication among nodes. The result is a framework where the nodes need to be at least bi-connected and the communication graph has to be rigid. This is possible considering a novel typology of link, namely Shadow Edge, that accounts for the lack of communication among nodes and allows to reduce the uncertainty associated to the position of the nodes.
14:20-14:40	ThB8.3
A Fixed-Neighbor, Distributed Algorithm for Solving a Linear Algebraic Equation, pp. 2269-2273
Mou, Shaoshuai	Yale Univ.
Morse, A. Stephen	Yale Univ.
This paper presents a distributed algorithm for solving a linear algebraic equation of the form $Ax=b$ where $A$ is an $ntimes n$ nonsingular matrix and $b$ is an $n-$vector. The equation is solved by a network of $n$ agents assuming that each agent knows exactly one distinct row of the partitioned matrix $matt{A&b}$, the current estimates of the equation's solution generated by its neighbors, and nothing more. Each agent recursively updates its estimate of $A^{-1}b$ by utilizing the current estimates generated by each of its neighbors. Neighbor relations are characterized by a simple, undirected graph $mathbb{G}$ whose vertices correspond to agents and whose edges depict neighbor relations. It is shown that for any nonsingular matrix $A$ and any connected graph $mathbb{G}$, the proposed algorithm causes all agents' estimates to converge exponentially fast to the desired solution $A^{-1}b$.
14:40-15:00	ThB8.4
Performance Limitations for Distributed Systems Via Spatial-Frequency Bode Integrals, pp. 2274-2279
Sarlette, Alain	Ghent Univ.
The Bode integral expresses a standard performance limitation for (almost) any controller that asymptotically stabilizes a linear time-invariant system. For the control of distributed systems, spatial invariance allows to write one such "Bode time-integral" per spatial frequency. The present paper inverses the roles of spatial and temporal independent variables in this latter viewpoint. By transposing the notions of controller, causality, and asymptotic stability to the spatial variable, we obtain and interpret "Bode space-integrals", one per temporal frequency. The result directly connects to the notion of string stability.
15:00-15:20	ThB8.5
Node Knock-Out Based Structure Identification in Networks of Identical Multi-Dimensional Subsystems, pp. 2280-2285
Suzuki, Masayasu	Japan Science and Tech. Agency
Takatsuki, Nobuki	Tokyo Inst. of Tech.
Imura, Jun-ichi	Tokyo Inst. of Tech.
Aihara, Kazuyuki	The Univ. of Tokyo
In this paper, based on the node knock-out procedure, for a networked system consisting of identical multi-dimensional subsystems in which the network structure is unknown and the number of input/output nodes is less than that of subsystems, we propose a novel method to identify the strength of the interaction between nodes even if we do not know any information on the other nodes.
15:20-15:40	ThB8.6
A Distributed Control Law with Guaranteed Convergence Rate for Identically Coupled Linear Systems, pp. 2286-2291
Wang, Qingling	Harbin Inst. of Tech.
Yu, Changbin	Australian National Univ.
Gao, Huijun	Harbin Inst. of Tech.
Liu, Fangzhou	Res. Inst. of Intelligent Control and Systems, Harbin In
This paper investigates the stabilization and optimization problems for a group of identically linear agents with undirected interaction topology. It is shown that a distributed control law based on local measurements and relative information exchanged from neighboring agents can be designed for each agent to enable the agent states to be stabilized. Furthermore, due to the use of a parametric Lyapunov approach, the designed distributed control law guarantees not only optimization performance at a network level but also a convergence rate for the group of agents. Finally, a simulation example is provided to demonstrate the advantage as well as the effectiveness of the proposed method.

ThB9	HG E7
Robotics (Regular Session)
Chair: Cadenat, Viviane	LAAS-CNRS
Co-Chair: Moustris, George	National Tech. Univ. of Athens
13:40-14:00	ThB9.1
Generalization of Bernstein's Problem Toward Autonomous Action Development of Artificial Muscle Based Robots, pp. 2292-2298
Suzuki, Masakazu	Tokai Univ.
The Bernstein's problem, asking how to cope with the redundant degrees-of-freedom of musculoskeletal systems in motion control, is straightforwardly extended to more general redundancy resolution problems. What element actions should be coordinated to achieve a complex task, or to realize multi-robot cooperative work? In this article they are unified and formulated as Generalized Bernstein Problem for developmental realization of complex actions, and then the ease criterion toward a smart solution for the problem is proposed.
14:00-14:20	ThB9.2
Internal and External Force-Based Impedance Control for Cooperative Manipulation, pp. 2299-2304
Heck, Dennis	Eindhoven Univ. of Tech.
Kostic, Dragan	SEGULA Tech. Nederland BV
Denasi, Alper	Tech. Univ. Eindhoven
Nijmeijer, Hendrik	Eindhoven Univ. of Tech.
An asymptotically stable cascaded control algorithm is proposed for cooperative manipulation of a common object. This algorithm controls motion and internal forces of the object, as well as the contact forces between the object and environment. The motion of each manipulator is controlled using an inverse dynamics type of controller. Only knowledge of the kinematics of the manipulated object is required, since the interaction forces and moments between the object and manipulators are measured. The internal stresses in the object are controlled based on enforced impedance relationships between the object and each manipulator. The internal forces and moments are computed using the object kinematics. Contact with the environment is controlled with an enforced impedance relationship between the object and the environment. For both internal and external forces, reference trajectories can be specified. Asymptotic stability of each controller is proven using Lyapunov stability theory and LaSalle's invariance principle. Guidelines are suggested to compute control parameters of the internal impedance parameters. Merits of the control algorithm are demonstrated in simulations.
14:20-14:40	ThB9.3
Management of Visual Signal Loss During Image Based Visual Servoing, pp. 2305-2310
Durand Petiteville, Adrien	LAAS/CNRS; Univ. de Toulouse UPS
Durola, Sylvain	LAAS/CNRS; Univ. de Toulouse UPS
Cadenat, Viviane	Laas/cnrs
Courdesses, Michel	CNRS/LAAS and Univ. de Toulouse , UPS, INSA, INP, ISAE
In this paper, we address the problem of the total visual features loss during visual servoing. We present a new method allowing to reconstruct these features even if the image is completely unavailable. The proposed method has been developed for a 6 degree-of-freedom (DOF) calibrated camera and a static landmark of interest which can be characterized by point features. It relies on a predictor/corrector pair coupled with a depth estimation algorithm. Numerous simulation results are provided and show the relevance of the proposed approach.
14:40-15:00	ThB9.4
Active Motion Compensation in Robotic Cardiac Surgery, pp. 2311-2316
Moustris, George	National Tech. Univ. of Athens
Mantelos, Andreas Ioannis	National Tech. Univ. of Athens
Tzafestas, Costas	National Tech. Univ. of Athens
Motion compensation is a prominent application in robotic beating heart surgery, with significant potential benefits for both surgeons and patients. In this paper we investigate an activate assistance control scheme on a simple tracking task, which helps the surgeon guide the robot on a predefined reference. The control is implemented on top of a shared control system, which serves as a basis for implementing higher level controllers. Experiments with a trained surgeon are also presented, which show the positive effect of the approach.
15:00-15:20	ThB9.5
Control of a Swinging Juggling Robot, pp. 2317-2322
Fontana, Flavio	ETH Zurich
Reist, Philipp	ETH Zurich
D'Andrea, Raffaello	ETH Zurich
We present a control strategy for a robot that juggles a ball with a single actuated paddle that is attached to the tip of a pendulum-like mechanism. The robot juggles the ball from side-to-side by striking the ball with the paddle when the pendulum reaches its peak angles. Sustained juggling is only possible if the pendulum motion is synchronized to the ball motion. We propose adapting the paddle motion to achieve synchronization. Specifically, we exploit the dynamic coupling between the pendulum and the paddle, which is essentially a moving mass at the tip of the pendulum. Optimal control is used to compute paddle motions that synchronize the pendulum to the ball. Feedback is introduced with a lookup table that maps a measured state to an appropriate paddle motion. In experiments, the proposed feedback strategy enables the robot to juggle at various amplitudes.

ThB10	HG D1.1
Robust Control of Linear Systems I (Regular Session)
Chair: Bokor, Jozsef	Hungarian Acad. of Sciences
Co-Chair: Grammatico, Sergio	ETH Zurich
13:40-14:00	ThB10.1
Eigenvalue Assignment for Componentwise Ultimate Bound Minimisation in LTI Discrete-Time Systems, pp. 2323-2330
Heidarihaei, Rahmatollah	Univ. of Newcastle
Seron, Maria M.	The Univ. of Newcastle
Braslavsky, Julio H.	Commonwealth Scientific and Industrial Res.
Haimovich, Hernan	Univ. Nacional de Rosario
We address optimal eigenvalue assignment in order to obtain minimum ultimate bounds on every component of the state of a linear time-invariant (LTI) discrete-time system in the presence of non-vanishing disturbances with known constant bounds. As opposed to some continuous-time cases where ultimate bounds can be made arbitrarily small by applying feedback with sufficiently high gain so that the closed-loop eigenvalues are sufficiently fast, the ultimate bound of a discrete-time system with an additive bounded disturbance can never be made smaller than some set that depends on the disturbance bound, even if all closed-loop eigenvalues are set at zero (the fastest possible in discrete-time). In this context, our contribution is twofold: (a) we single out cases where feedback that may not assign all closed-loop eigenvalues at zero achieves the minimum possible ultimate bound for some component of the system state, and (b) by employing an existing componentwise ultimate bound computation formula, we find a class of systems for which assigning all closed-loop eigenvalues at zero indeed yields minimum ultimate bounds. An intermediate result---and our third contribution---in the derivation of (b) is the obtention of the Jordan decomposition that minimises the componentwise ultimate bound formula employed.
14:00-14:20	ThB10.2
A Universal Class of Non-Homogeneous Control Lyapunov Functions for Linear Differential Inclusions, pp. 2331-2336
Grammatico, Sergio	ETH Zurich
Blanchini, Franco	Univ. degli Studi di Udine
Caiti, Andrea	Univ. of Pisa
The constrained stabilization of Linear Differential Inclusions (LDIs) via non-homogeneous control Lyapunov functions (CLFs) is addressed in this paper. We consider the class of ``merging'' CLFs, which are composite functions whose gradient is a positive combination of the gradients of two given parents CLFs. In particular, we consider the constructive merging procedure based on recently-introduced composition via R-functions, which represents a parametrized trade-off between the two given CLFs. We show that this novel class of non-homogeneous Lyapunov functions is ``universal'' for the stabilization of LDIs, besides some equivalence results between the control-sharing property under constraints, i.e. the existence of a single control law which makes simultaneously negative the Lyapunov derivatives of the two given CLFs, and the existence of merging CLFs. We also provide an explicit stabilizing control law based on the proposed merging CLF. The theoretical results are finally applied to a perturbed constrained double integrator system.
14:20-14:40	ThB10.3
Time-Varying Generalisations of the Gap and Nu-Gap Metrics Induce the Same Topology in Continuous Time, pp. 2337-2342
Khong, Sei Zhen	Univ. of Melbourne
Cantoni, Michael	Univ. of Melbourne
Recent robust stability analysis results for linear time-varying feedback interconnections are based on a time-varying generalisation of the nu-gap metric. The causality of closed-loop mappings is dealt with explicitly, rather than via well-posedness assumptions as is common in the literature. Here, an alternative time-varying gap metric is defined. It is shown that this gives rise to correspondong robust stability results. It is also established that the time-varying gap metric induces the same topology as the generalised nu-gap metric, this being the coarsest under which closed-loop stability and performance are both robust properties.
14:40-15:00	ThB10.4
Robust Control of Flexible High-Speed Rotors Via Mixed Uncertainties, pp. 2343-2350
Riemann, Bernd	TU Darmstadt
Sehr, Martin	TU Darmstadt, Inst. for Mechatronic Systems in Mechanical En
Schittenhelm, Rudolph Sebastian	TU Darmstadt, Inst. for Mechatronic Systems in Mechanical En
Rinderknecht, Stephan	Tech. Univ. Darmstadt
Active vibration control for flexible high-speed rotors tends to be a particularly challenging problem due to the influence of gyroscopic terms, resulting in the need for speeddependent system models. This paper addresses robust control of such systems, using Linear Fractional Transformations (LFTs) for decoupling the system model into speed-dependent and - independent components in LFT feedback. Based on the resulting LFT decomposition, the speed-dependent terms are efficiently reduced in order and considered uncertain with respect to the rotational speed of the shaft. The resulting perturbations are augmented by complex, additive uncertainties and explicitly used for control synthesis. Defining semi-modal performance measures, the perturbed openloop systems are well-suited for mixed µ synthesis techniques. In particular, (D,G)-K and µ-K algorithm, both enabling explicit treatment of mixed perturbations, are investigated in approaching the robust vibration attenuation problem across the range of operating speeds.
15:00-15:20	ThB10.5
Sufficiency of Vertex Matrix Check for Robust Stability of Interval Matrices Via the Concept of Qualitative Robustness, pp. 2351-2356
Yedavalli, Rama K.	Ohio State Univ.
Devarakonda, Nagini	The Ohio State Univ.
This paper revisits the issue of robust stability analysis of linear interval parameter matrices, which used to be a highly active research topic in the eighties and nineties.The reason for this revived interest in this topic is that the recent research by the authors on Qualitative Stability, a topic of interest in the field of population/community dynamics in ecology is shown to shed considerable insight with possible new results in the robust stability of matrix families. Thus in this paper, we expand on the two notions of robustness introduced recently by the authors, namely `Qualitative Robustness' and `Quantitative Robustness' and investigate their interdependence. Specifically, it is shown that for a class of matrix families with specified `Qualitative Robustness' indices, it is sufficient to check the stability of only `vertex' matrices (i.e. an extreme point solution) to guarantee the robust stability of the entire interval matrix family. This is indeed deemed important and significant because with this result, we can easily identify for which `interval matrix families' we need to resort to more sophisticated stability check algorithms, and for which families we can get away with a ` vertex matrix check' (i.e. an `extreme point solution'). It turns out that this class of `qualitative stable' matrices that admit `vertex solution' for its `quantitative robustness' is quite large. Thus the results of this paper offer new insight into the nature of interactions and interconnections in a matrix family on its robust stability. Encouraged by the results of this paper, continued research is underway in using this interdependence of `qualitative robstness' and `quantitative robustness' in the design of robust controllers for engineering systems.
15:20-15:40	ThB10.6
Probabilistic $mu$ for Rank-One and Perturbed Rank-One Matrices, pp. 2357-2361
Mánfay, Máté	MTA SZTAKI, Central European Univ.
Balas, Gary J.	Univ. of Minnesota
Bokor, Jozsef	Hungarian Acad. of Sciences
Gerencsér, László	MTA SZTAKI
The structured singular value $mu$ has been widely studied for uncertain dynamical systems. Recently a great attention is paid to the probabilistic $mu$ problem. Instead of computing the conservative worst-case $mu$ we are interested in the probabilistic distribution of $mu,$ given a probability distribution on the set of uncertainties. Traditionally this problem is solved by Monte Carlo algorithms. In this paper we propose analytic methods to compute the probabilistic $mu$ for rank-one and perturbed rank-one matrices. We expect that these results will provide an algorithm that is not as computationally expensive as the linear cut algorithm in cite{prob_lincut}.

ThB11	HG D1.2
Control Over Communication Channels (Regular Session)
Chair: Ishii, Hideaki	Tokyo Inst. of Tech.
Co-Chair: Sinopoli, Bruno	Carnegie Mellon Univ.
13:40-14:00	ThB11.1
Transient Response of Minimum Variance Control Over a Gaussian Communication Channel, pp. 2362-2367
Freudenberg, James S.	Univ. of Michigan
Middleton, Richard H.	The Univ. of Newcastle
Recently, a finite horizon minimum variance control problem was proposed using feedback over a Gaussian communication channel. Because only the terminal state is penalized, it was shown that linear communication and control strategies are optimal and achieve the information theoretic minimum cost. However, because the transient state is not penalized, the transient behavior can be poor. In the present paper, we show that if there is at most one open loop unstable plant pole, then the transient response will remain bounded as the control horizon tends to infinity, and will approach a value determined by the solution to a certain algebraic Riccati equation.
14:00-14:20	ThB11.2
Stabilization of Uncertain Systems with Finite Data Rates and Markovian Packet Losses, pp. 2368-2373
Okano, Kunihisa	Tokyo Inst. of Tech.
Ishii, Hideaki	Tokyo Inst. of Tech.
This research addresses stabilization of uncertain systems over data rate constrained and lossy channels. While many of the existing works assume that the packet loss process is independent and identically distributed, we model it as a two-state Markov chain, which can deal with more practical situations including bursty dropouts. For parametrically uncertain plants, a necessary condition and a sufficient condition for mean square stability are derived. These conditions are represented by the product of the eigenvalues of the nominal plants, the data rate, the transition probabilities of the channel states, and the upper bounds of uncertainties. In particular, for scalar plants, the conditions coincide with each other.
14:20-14:40	ThB11.3
LQG Cheap Control Subject to Packet Loss and SNR Limitations, pp. 2374-2379
Chiuso, Alessandro	Univ. di Padova
Schenato, Luca	Univ. of Padova
Laurenti, Nicola	Univ. of Padova
Zanella, Andrea	Univ. of Padova
In this paper we consider the problem of controlling unstable stochastic linear systems in the presence of a communication channel between the sensors and the actuators. We propose an LQG architecture that separates the problem of designing suitable regulators for controlling the plant, referred to as Plant encoder/decoders, from the problem of designing encoder/decoder for the communication channel. We provide a mathematical model that takes into account the most important features of today's wireless communication protocols such as quantization errors, limited channel capacity, decoding delay and packet loss, while still being amenable to analytic treatment. We then restrict our discussion to a special class of linear plant encoder/decoders and to a channel with signal-to-noise (SNR) limitations and packet loss only, and we derive stability conditions and optimal parameters for the controller design in the cheap-control setting. Through this analysis we are able to recover several results available in the literature that treated packet loss and quantization error separately.
14:40-15:00	ThB11.4
LQG Control with Markovian Packet Loss, pp. 2380-2385
Mo, Yilin	Carnegie Mellon Univ.
Garone, Emanuele	Univ. degli Studi della Calabria
Sinopoli, Bruno	Carnegie Mellon Univ.
This paper is concerned with the optimal LQG control of a system through lossy data networks. In particular we will focus on the case where control commands are issued to the system over a communication network where packets may be randomly dropped according to a two-state Markov chain. Under these assumptions, the optimal finite-horizon LQG problem is solved by means of dynamic programming arguments. The infinite horizon LQG control problem is explored and conditions to ensure its convergence are investigated. Finally it is shown how the results presented in this paper can be employed in the case that also the observation packet may be dropped. A numerical simulation shows the relationship between the convergence of the LQG cost and the value of the parameters of the Markov chain.
15:00-15:20	ThB11.5
On Stabilization Over Gaussian Interference Channel, pp. 2386-2391
Zaidi, Ali	Royal Inst. of Tech. (KTH)
Oechtering, Tobias J.	Royal Inst. of Tech. (KTH)
Skoglund, Mikael	Royal Inst. of Tech.
The problem of feedback stabilization of LTI plants over a Gaussian interference channel is considered. Two plants with arbitrary distributed initial states are monitored by two separate sensors which communicate their measurements to two separate controllers over a Gaussian interference channel under average transmit power constraints. The necessary conditions for the mean square-stabilization over a memoryless symmetric Gaussian interference channel are derived. These conditions are shown to be tight for some system parameters. Further it is shown that linear and memory-less sensing and control schemes are optimal for stabilization in some special cases.
15:20-15:40	ThB11.6
Continuous-Time Indefinite Linear Quadratic Optimal Control with Random Input Gains, pp. 2392-2397
Chen, Wei	The Hong Kong Uni. of Sci. and Tech.
Qiu, Li	Hong Kong Univ. of Sci. & Tech.
In this paper, the indefinite linear quadratic (LQ) optimal control of continuous-time linear time-invariant systems with random input gains is studied. One main novelty of this work is the use of channel/controller co-design framework which bridges and integrates the design of the channels and controller. The co-design is carried out by the twist of channel resource allocation, i.e., the channel capacities can be allocated among the input channels by the control designer subject to an overall capacity constraint. With the channel/controller co-design, necessary and sufficient conditions for the well-posedness as well as the attainability of the indefinite LQ problem concerned are obtained. The optimal control law is given by a linear state feedback associated with the mean-square stabilizing solution of a modified algebraic Riccati equation.

ThB12	HG D3.2
System Identification II (Regular Session)
Chair: Picci, Giorgio	Univ. di Padova
Co-Chair: Rojas, Cristian R.	KTH Royal Inst. of Tech.
13:40-14:00	ThB12.1
An EM-Based Estimation Algorithm for a Class of Systems Promoting Sparsity, pp. 2398-2403
Godoy, Boris I.	The Univ. of Newcastle
Carvajal, Rodrigo	Univ. Tecnica Federico Santa Maria
Aguero, Juan C.	The Univ. of Newcastle
In this paper we propose a Maximum a Posteriori (MAP) approach for estimating a random sparse parameter vector in the presence of nonlinearities of unknown parameters. In this Bayesian approach, the a priori probability distribution for the parameter vector is utilised as a mechanism to promote sparsity. We solve this identification problem by using a generalized Expectation Maximization algorithm in a MAP framework.
14:00-14:20	ThB12.2
A Spectral Estimation Case Study in Frequency-Domain by Subspace Methods, pp. 2404-2409
Akcay, Huseyin	Anadolu Univ.
Turkay, Semiha	Anadolu Univ.
In this paper, the properties of two recently proposed frequency-domain subspace-based algorithms to estimate discrete-time cross-power spectral density (cross-PSD) and auto-power spectral density (auto-PSD) matrices of vector auto-regressive moving-average and moving-average (ARMAMA) models from sampled values of the Welch cross-PSD and auto-PSD estimators on uniform grids of frequencies, are illustrated by numerical and real-life application examples. The latter is concerned with the modeling of acoustic spectra for detecting faults in induction motors.
14:20-14:40	ThB12.3
A Sparse Estimation Technique for General Model Structures, pp. 2410-2414
Rojas, Cristian R.	KTH Royal Inst. of Tech.
Wahlberg, Bo	KTH Royal Inst. of Tech.
Hjalmarsson, Håkan	Royal Inst. of Tech.
In this paper, a general sparse estimator is proposed, based on the maximum likelihood / prediction error method (or any root-N-consistent estimator). This procedure does not rely on the convexity of the cost function of the underlying estimator (in case such estimator is an M-estimator), and it provides an automatic tuning of the (implicit) regularization parameter. The idea behind the proposed method is a three step procedure, where the first step consists in a standard root-N-consistent estimation, the second step seeks for the sparsest estimate in a neighborhood of the initial estimate, and the last step is a refinement based on the sparseness pattern estimated in the second step. A rigorous statistical analysis is provided, which establishes conditions for consistency, asymptotic variable selection and the so-called Oracle property. A simulation example is given to demonstrate the performance of the method.
14:40-15:00	ThB12.4
Identification of Finite Dimensional Linear Stochastic Systems Driven by Levy Processes, pp. 2415-2420
Manfay, Mate	MTA SZTAKI
Gerencsér, László	MTA SZTAKI
We study the problem of identifying a finite dimensional linear stochastic SISO system driven by a Levy process. The latter are widely used in modelling financial time series. In a number of important examples the density function of the innovation term is unknown, but its characteristic function is explicitly known, possibly up to a few unknown parameters. In this paper we present and analyze a novel identification method that exploits the information on the characteristic function of the noise. It is obtained by adapting adapting the empirical characteristic function method (ECF for short) developed for i.i.d. samples. We will show that the new method may be more efficient in estimating the system parameters than a plain prediction error method.
15:00-15:20	ThB12.5
Modeling Random Flocks through Generalized Factor Analysis, pp. 2421-2426
Bottegal, Giulio	Univ. of Padova
Picci, Giorgio	Univ. di Padova
In this paper, we study modeling and identification of stochastic systems by Generalized Factor Analysis models. Although this class of models was originally introduced for econometric purposes, we present some possible applications of engineering interest. In particular, we show that there is a natural connection between Generalized Factor Analysis models and multi-agents systems. The common factor component of the model has an interpretation as a flocking component of the system behavior.
15:20-15:40	ThB12.6
Recursive Estimation Algorithm for <i>l</i><sub>1</sub>-Norm Approximation in Dynamic Systems with Nonoptimality Levels, pp. 2427-2432
Matasov, Alexander	Lomonosov Moscow State Univ.
Akimov, Pavel A.	Lomonosov Moscow State Univ.
State estimation problem for a linear discrete dynamic system is considered. Some components of the state vector can abruptly change under the influence of rare uncontrolled input pulses in the right-hand side of equations. In this case, l₁-norm approximation (least absolute deviations method) gives better results than the standard l₂-norm approximation (least squares method). A recursive estimation algorithm for finding l₁-norm approximation in case of large amount of measurements is presented. To make numerical procedure more reliable, a nonoptimality level for current iteration is constructed. An example from inertial navigation verifies the effectiveness of proposed approach.

ThB13	HG D5.2
Fault Diagnosis II (Regular Session)
Chair: Mahulea, Cristian	Univ. of Zaragoza
Co-Chair: Ferrari, Riccardo	Danieli Automation S.p.A.
13:40-14:00	ThB13.1
Fault Detection and Isolation on a Three Tank System Using Differential Flatness, pp. 2433-2438
Martinez Torres, Cesar	Bordeaux 1 IMS
Lavigne, Loic	Univ. Bordeaux1
Cazaurang, Franck	Univ. Bordeaux I
Alcorta-García, Efraín	Univ. Autonoma De Nuevo Leon
Diaz-Romero, David Alejandro	Univ. Autonoma de Nuevo Leon
This paper addresses a fault detection and isolation technique for differential flat systems. For such nonlinear systems, it is possible to find a set of variables, named flat outputs such that states and control inputs can be expressed as functions of flat outputs and their derivatives. Flat systems properties are used to detect and isolate faults and the nonuniqueness property of the set of flat outputs is used for increase the number of residues and improve the fault isolation, the proposed approach will be applied on a classical three tank system.
14:00-14:20	ThB13.2
Online Diagnosis of PEMFC by Analyzing Individual Cell Voltages, pp. 2439-2444
Li, Zhongliang	Univ. of Aix-Marseille
Outbib, Rachid	lsis
Hissel, Daniel	Univ. of Franche-Comte
Giurgea, Stefan	Univ. of Tech. Belfort-Montbeliard (UTBM)
Polymer Electrolyte Membrane Fuel Cell (PEMFC) is a promising power source for a wide range of applications. Fault diagnosis, especially online fault diagnosis, is an essential issue to promote the development and widespread use of PEMFC technology. This paper proposes a diagnosis approach for large PEMFC stack. In this approach, flooding fault is concerned; individual cell voltages are chosen as original variables for diagnosis. A dimension reduction method Fisher linear discrimination (FDA) is adopted to extract the features from the cell voltage composed vectors. After that, a classification methodology, Gaussian mixture model (GMM) is applied for fault detection. Flooding experiments were conducted on a 20-cell stack to test the approach. The obtained results showed that data points can be classified to different states of health with a high accuracy. It is also verified that the real-time implementation of the algorithm is feasible.
14:20-14:40	ThB13.3
An Algebraic Approach for Robust Fault Detection of Input-Output Elastodynamic Distributed Parameter Systems, pp. 2445-2452
Ferrari, Riccardo	Danieli Automation S.p.A.
Parisini, Thomas	Imperial Coll. & Univ. of Trieste
Polycarpou, Marios M.	Univ. of Cyprus
This paper deals with the problem of designing a robust fault detection methodology for a class of input-output, uncertain dynamical distributed parameter systems, namely mechanical elastodynamic systems, which are representative of a whole class of problems related to on-line health monitoring of mechanical and civil engineering structures. The proposed approach does not require full state measurements and is robust to measuring, modeling and numerical errors, thanks to a time varying detection threshold. In order to avoid the problems associated with classical discretization techniques for distributed parameter systems, which can lead to numerical errors difficult to bound a priori, and thus higher thresholds, a suitable structure-preserving algebraic approach, called Cell Method, will be employed. This method consists in writing the equations of a distributed parameter system directly in discrete form, avoiding the usual discretization process and leading to a symplectic, that is energy preserving, numerical scheme.
14:40-15:00	ThB13.4
Data-Based Causality Detection from a System Identification Perspective, pp. 2453-2458
Marques, Vinicius Moura	Federal Univ. of the Espírito Santo
Munaro, Celso Jose	Federal Univ. of Espirito Santo
Shah, Sirish L	Univ. of Alberta
The problem of detecting causality, from routine operating data, is reviewed from a system identification perspective. It is shown that even simple examples from the literature under Granger causality analysis do not have adequate model fit. As an alternative, this study uses the system identification platform to capture causality from process data. For example, the model inadequacy test is considered an important reason to reject a given causal relationship. The rich framework of system identification techniques and the choice of models to deal with exogenous variables and nonlinearities are shown to be an extremely suitable foundation to detect casual relationships. The utility of the proposed approach is illustrated by several benchmark examples including the analysis of routine operating data in an industrial case study.
15:00-15:20	ThB13.5
Fault Diagnosis Graph of Time Petri Nets, pp. 2459-2464
Wang, Xu	Univ. de Zaragoza
Mahulea, Cristian	Univ. of Zaragoza
Silva, Manuel	Univ. De Zaragoza
This paper proposes an online approach for the fault diagnosis for time discrete event systems, which are modeled by time Petri net. The observation is given by a subset of transitions whose occurrence is always observable. Faults correspond to a subset of the transitions whose firing are not observable. According to the most of the literature on discrete event systems, we define three fault states, namely N, F and U, corresponding to normal, fault and uncertain states, respectively. The proposed approach use a Fault Diagnosis Graph (FDG). It is adapted from state class graph by keeping only the necessary information for computation of the fault states and removing the unnecessary states. Some algorithms to compute after each observation only part of the FDG required to update the diagnosis states are given.
15:20-15:40	ThB13.6
State Estimation of Unknown Input Fuzzy Bilinear Systems: Application to Fault Diagnosis, pp. 2465-2470
Saoudi, Dhikra	ESSTT
Chadli, Mohammed	Univ. de Picardie-Jules Verne
Benhadj Braiek, Naceur	Ec. Pol. de Tunisie
This article aims the observer synthesis for a class of nonlinear systems and affected by unknown inputs, represented under the multimodel bilinear formulation. Sufficient conditions to design an unknown input fuzzy bilinear observer are given in Linear Matrix Inequalities (LMIs) terms. The paper studies also the problem of fault detection and isolation. An unknown input fuzzy bilinear fault diagnosis observer design is proposed. Numerical example is given to illustrate the effectiveness of the given results.

ThB14	HG D7.1
Modelling and Control of Fuel Cells (Invited Session)
Chair: Kunusch, Cristian	IRI (UPC-CSIC)
Co-Chair: Usai, Elio	Univ. degli Studi di Cagliari
Organizer: Kunusch, Cristian	IRI (UPC-CSIC)
Organizer: Ocampo-Martinez, Carlos	Tech. Univ. of Catalonia (UPC)
13:40-14:00	ThB14.1
A Multiagent Model for PEM Fuel Cell Microscopic Simulation (I), pp. 2471-2476
Gechter, Franck	Univ. of Tech. of Belfort-Montbeliard
Bouquain, David	Univ. of Tech. of Belfort-Montbeliard
Gao, Fei	Univ. of Tech. of Belfort-Montbeliard
Koukam, Abderafiaa	Univ. of Tech. of Belfort-Montbeliard
Miraoui, Abdellatif	Univ. de Tech. de Belfort-Montbéliard
Since several years, multi-agent systems have been used for wide range of application areas and particularly in systems modelling and simulation. Those systems are based on the interaction of several individual components, called agents, which are proactive considering interaction with neighbours and with the surrounding environment. The IRTES-SeT laboratory develops a PEM-Fuel Cell model for detailed layer level simulation based on multi-agent paradigm. Whatever the level of abstraction considered, modelling the electrochemical process requires at least: an electrochemical model, a fluidic model, and a thermal model. These elements are integrated into our simulation model at the microscopic level. The goal of this paper is to present a multi-physical micro-level PEM-Fuel Cell model by using a multi-agent approach.
14:00-14:20	ThB14.2
Differential Flatness-Based Observer Design for a PEM Fuel Cell Using Adaptive-Gain Sliding Mode Differentiators (I), pp. 2477-2482
Liu, Jianxing	Univ. of Tech. of Belfort-Montbeliard (UTBM)
Laghrouche, Salah	Supélec / CNRS
Wack, Maxime	Lab. systems and Transport, Univ. of Tech. of Belfort-
In this paper, we propose a differential flatness-based observer for a PEM fuel cell air-feed system. The proposed observer uses the measurements of supply manifold air pressure and compressor mass flow rate in order to estimate the oxygen partial pressure, nitrogen partial pressure and the compressor speed. An adaptive-gain Second Order Sliding Mode differentiator based on super-twisting algorithm is designed to estimate the time derivatives of the system outputs when the boundary of higher time derivative is unknown. Then, the states are estimated based on these values. The objective is to minimize the number of embedded sensors in order to get a precise and economic solution. The feasibility and effectiveness of the proposed approach is demonstrated through simulation results obtained from a nonlinear fuel cell system.
14:20-14:40	ThB14.3
Experimental Validation of Equilibria in Fuel Cells with Dead-Ended Anodes (I), pp. 2483-2488
Chen, Jixin	Univ. of Michigan
Siegel, Jason	Univ. of Michigan
Matsuura, Toyoaki	Univ. of Michigan
Stefanopoulou, Anna G.	Univ. of Michigan
Yesilyurt, Serhat	Sabanci Univ.
This paper investigates the nitrogen blanketing front during the dead-ended anode (DEA) operation of a PEM fuel cell. It is found that the dynamic evolution of nitrogen accumulation in the DEA of a PEM fuel cell eventually arrives to a steady-state, which suggests the existence of equilibrium. We use a multi-component model of the two-phase, one-dimensional (along-channel) system to analyze this phenomenon. Specifically, the model is first verified with experimental observations, and then utilized to show the evolution toward equilibrium. The full order model is reduced to a second-order partial differential equation (PDE) with one state, which can be used to predict and analyze the observed steady state DEA behavior. The parametric study is performed focusing on the influence of the cathode pressure on the existence of equilibrium in the DEA condition.
14:40-15:00	ThB14.4
Experimental Validation of Interval-Based Sliding Mode Control for Solid Oxide Fuel Cell Systems (I), pp. 2489-2494
Doetschel, Thomas	Univ. of Rostock
Rauh, Andreas	Univ. of Rostock
Aschemann, Harald	Univ. of Rostock
Senkel, Luise	Univ. of Rostock, Chair of Mechatronics
The utilization of solid oxide fuel cells (SOFCs) as well as other high-temperature fuel cells for a decentralized power supply demands for reliable and guaranteed stabilizing control strategies, that are capable of providing electrical and thermal energy. Moreover, it is essential to increase the number of possible thermal cycles of SOFCs by means of control laws which reduce temperature gradients in the space coordinates of a stack module during transient operating conditions. For this purpose, suitable control-oriented system models have been identified in previous work and parameterized reliably by means of both local and global optimization procedures. By exploiting these models, which can be extended to account for parameters that are subject to bounded uncertainty, interval-based sliding mode controllers can be derived. These controllers simultaneously adjust the mass flow and temperature of air supplied to the cathode of the SOFC. In this paper, a real-time capable implementation of the corresponding interval-based sliding mode controller and selected experimental results are presented for an SOFC test rig available at the Chair of Mechatronics at the University of Rostock.
15:00-15:20	ThB14.5
Observer-Based Output Feedback Control of a PEM Fuel Cell System by High-Order Sliding Mode Technique (I), pp. 2495-2500
Pisano, Alessandro	Univ. di Cagliari
Salimbeni, Domenico	Univ. of Cagliari, Dept of Electrical and Electronic Engine
Usai, Elio	Univ. degli Studi di Cagliari
Rakhtala Rostami, Seyed Mehdi	Faculty of electrical and computer engineering, Babol industrial
Ranjbar, Noei	Faculty of electrical and computer engineering, Babol industrial
This paper deals with an high-order sliding-mode approach to the observer-based output feedback control of a PEM fuel cell system comprising a compressor, a supply manifold, the fuel-cell stack and the return manifold. The suggested scheme assumes the availability for measurements of readily accessible quantities such as the compressor angular velocity, the load current, and the supply and return manifold pressures. The control task is formulated in term of regulating the oxygen excess ratio (which is estimated by the observer) to a suitable set-point value by using, as adjustable input variable, the compressor supply voltage. The treatment is based on a nonlinear modeling of the PEM fuel cell system under study. Simulations results showing the feasibility and satisfactory performance of the proposed approach are provided
15:20-15:40	ThB14.6
On the Implementation of an Adaptive Extremum Seeking Algorithm for Hydrogen Minimization in PEM Fuel Cell Based Systems (I), pp. 2501-2506
Kunusch, Cristian	IRI (CSIC)
Castanos, Fernando	Cinvestav del IPN
This work presents initial experimental results of an adaptive sliding-mode extremum seeker that minimizes the hydrogen consumption in a fuel cell based system. The extremum seeker is based on the classical steepest-descent method, the main challenge being the fact that the gradient of the objective function is unknown. The gradient is estimated by means of a sliding-mode adaptive estimator. The strategy is applied in experimental practical situations in a fuel cell test bench, this allows to asses the performance of the scheme as well as the difficulties that arise in real applications.

ThB15	HG D7.2
Generation and Load Side Control for Power Systems with Stochastic Uncertainty (Invited Session)
Chair: Callaway, Duncan	UC Berkeley
Co-Chair: Kamgarpour, Maryam	ETH Zurich
Organizer: Margellos, Konstantinos	ETH Zurich
Organizer: Kamgarpour, Maryam	ETH Zurich
Organizer: Callaway, Duncan	UC Berkeley
13:40-14:00	ThB15.1
A Comparative Study of Stochastic Unit Commitment and Security-Constrained Unit Commitment Using High Performance Computing (I), pp. 2507-2512
Papavasiliou, Anthony	Catholic Univ. of Louvain
Oren, Shmuel	UC Berkeley
The large-scale integration of renewable resources has recently raised interest in systematic methods for committing locational reserves in order to secure the system against contingencies and the unpredictable and highly variable fluctuation of renewable energy supply, while accounting for power flow constraints imposed by the transmission network. In this paper we compare two approaches for committing locational reserves: stochastic unit commitment and a hybrid approach of scenario-based security-constrained commitment. Parallel algorithms are developed for solving the resulting models, based on Lagrangian relaxation and Benders decomposition. The proposed algorithms are implemented in a high performance computing environment and the performance of the resulting policies is tested against a reduced model of the California ISO interconnected with the Western Electricity Coordinating Council.
14:00-14:20	ThB15.2
Stochastic Unit Commitment and Reserve Scheduling: A Tractable Formulation with Probabilistic Certificates (I), pp. 2513-2518
Margellos, Konstantinos	ETH Zurich
Rostampour, Vahab	ETH Zurich
Vrakopoulou, Maria	ETH Zurich
Prandini, Maria	Pol. di Milano
Andersson, Goran	Swiss Federal Inst. of Tech.
Lygeros, John	ETH Zurich
The increased penetration of renewable energy sources to the network highlights the necessity of constructing stochastic variants of the standard unit commitment and reserve scheduling problems. Earlier approaches to such problems are either restricted to ad-hoc methodologies (at the expense of a suboptimal solution), or lead to computationally intractable formulations. In this paper we provide a unified framework to deal with such planning problems for systems with uncertain generation, while providing a-priori probabilistic certificates for the robustness properties of the resulting solution. Our methodology is based on a mixture of randomized and robust optimization and leads to a tractable problem formulation. To illustrate the performance of the proposed methodology we apply it to the IEEE 30-bus network, and compare it by means of Monte Carlo simulations against an algorithm based on a deterministic variant of the unit commitment problem.
14:20-14:40	ThB15.3
Energy Arbitrage with Thermostatically Controlled Loads (I), pp. 2519-2526
Mathieu, Johanna	ETH Zurich
Kamgarpour, Maryam	ETH Zurich
Lygeros, John	ETH Zurich
Callaway, Duncan	UC Berkeley
We investigate the potential for aggregations of residential thermostatically controlled loads (TCLs), such as air conditioners, to arbitrage intraday wholesale electricity market prices via non-disruptive direct load control. Since wholesale electricity prices reflect power system conditions, arbitrage provides a service to the grid, helping to balance real-time supply and demand. While previous work on the energy arbitrage problem has used simple energy storage models, we use high fidelity TCL-specific models which allow us to understand and quantify the full capabilities and constraints of these time-varying systems. We explore two optimization/control frameworks for solving the arbitrage problem, both based on receding horizon linear programming. Since we find that the first approach requires significant computation, we develop a second approach involving decomposition of the optimal control problem into separate optimization and control problems. Simulation results show that TCLs could save on the order of 10% of wholesale energy costs via arbitrage, with savings decreasing with price forecast error.
14:40-15:00	ThB15.4
Predictive Control of Buildings for Demand Response with Dynamic Day-Ahead and Real-Time Prices, pp. 2527-2534
Vrettos, Evangelos	Power Systems Lab. Swiss Federal Inst. of Tech.
Lai, KuanLin	Department of Electrical Engineering and Information Tech.
Oldewurtel, Frauke	ETH Zurich
Andersson, Goran	Swiss Federal Inst. of Tech.
This paper proposes a Model Predictive Control (MPC) scheme to control residential buildings with space heating/cooling loads, an Electric Water Heater (EWH), photovoltaics (PV) and battery storage in a time-varying electricity price environment. The controller uses models for the system components as well as predictions for future disturbances such as weather conditions, occupancy and electricity prices to find the building operation that minimizes electricity costs over the prediction horizon while respecting user comfort constraints. The building operation is investigated under three different price scenarios: (1) a simple day-night tariff for end-customers, (2) a day-ahead dynamic tariff reflecting the wholesale market marginal costs and (3) a real-time dynamic tariff. Residential buildings response to such price signals is investigated in a case study and their potential for Demand Response (DR) programs is evaluated.
15:00-15:20	ThB15.5
State-Space Modelling of Hysteresis-Based Control Schemes (I), pp. 2535-2540
Kundu, Soumya	Univ. of Michigan
Hiskens, Ian A.	Univ. of Michigan
The paper develops a state-space model for the aggregate power drawn by a group of plug-in electric vehicle (PEV) chargers under hysteresis-based charging. The aggregate response of the PEV charging loads to changes in the hysteresis deadband is nonlinear, requiring detailed analysis to accurately capture the dynamics. Related work, which focused on thermostatically controlled loads (TCLs), addressed state-space modelling by dividing the hysteresis deadband into equally sized bins and keeping track of the inter-bin migration of loads. A system of PEV chargers can be treated similarly. A new modelling paradigm has been developed to allow for fast variation of the hysteresis deadband. This model tracks the distribution of PEV chargers within the hysteresis deadband, with the inter-bin migration of chargers used to capture the aggregate dynamics.
15:20-15:40	ThB15.6
Smart Grid Dispatch Strategy for ON/OFF Demand-Side Devices, pp. 2541-2548
Biegel, Benjamin	Aalborg Univ.
Andersen, Palle	Aalborg Univ.
Pedersen, Tom S.	Aalborg Univ.
Nielsen, Kirsten Mølgaard	Aalborg Univ.
Stoustrup, Jakob	Aalborg Univ.
Hansen, Lars Henrik	Dong Energy A/S
We consider an aggregator managing a portfolio of runtime and downtime constrained ON/OFF demand-side devices. The devices are able to shift consumption in time within certain energy limitations. We show how the aggregator can manage the portfolio of devices to collectively provide upward and downward regulation. Two control strategies are presented enabling the portfolio to provide regulating power while respecting the runtime, downtime, and energy constraints of the devices. The first strategy is a predictive controller requiring complete device information; this controller is able to utilize the full flexibility of the portfolio but can only handle a small number of devices. The second strategy is an agile controller requiring less device information; this controller is able to handle a large number of devices but not able to utilize the full flexibility of the portfolio.

ThC1	HG F1
Cyber-Secure Control (Tutorial Session)
Chair: Sandberg, Henrik	Royal Inst. of Tech. (KTH)
Co-Chair: Sinopoli, Bruno	Carnegie Mellon Univ.
Organizer: Sandberg, Henrik	KTH Royal Inst. of Tech.
Organizer: Sinopoli, Bruno	Carnegie Mellon Univ.
16:00-16:40	ThC1.1
Theory for Secure Control Systems (I)*
Sinopoli, Bruno	Carnegie Mellon Univ.
The area of control systems design has been typically concerned with performance criteria satisfying basic stability requirements. Within this framework, several control techniques have been developed, such as robust control to deal with model uncertainties and disturbance rejection, and fault tolerant control to ensure graceful performance degradation against sensors and actuators failures. Concerns about security for traditional control systems have been non-existent as they have traditionally been closed systems where all the components, namely sensors, actuators, computers and communication devices were used exclusively for the control problem at hand, without any connection to the outside world. This provided an air gap that guaranteed security by physical isolation. Modern control systems rely upon computation and communication networks, and they are connected to the outside world. The integration of control systems and networks may increase system performance, but, at the same time, it introduces novel vulnerabilities that undermine their reliability. In contrast with traditional control system, modern control systems are prone to failures and coordinated attacks against their physical infrastructure, as well as cyber attacks against their data management and communication layer. In this talk I will provide an overview of recent advances the control community has made toward addressing the issue of security in control systems as well as new interesting directions. I will concentrate on robustness results. In particular I will review novel analytical techniques to study the effect of adversarial attacks on control systems as well as detection techniques and resilient control.
16:40-17:00	ThC1.2
Game Theory for Secure Control (I)*
Langbort, Cedric	UIUC
We summarize various applications of game theory to problems motivated by cyber-security of (networked) controlled systems. We cover several scenarios, occurring at various levels of the control loop (from the coding level of the physical network to the strategic interaction between stakeholders), which all correspond to different types of games (simple matrix games, stochastic, and differential games) and point to some future directions. Throughout, we emphasize the subtleties resulting from the possible presence of partial and/or asymmetric information between the players.
17:00-17:40	ThC1.3
Secure Control and Applications in Power Systems (I)*
Sandberg, Henrik	KTH Royal Inst. of Tech.
Power networks are examples of large-scale critical infrastructures where secure operation is an absolute necessity. Power networks are routinely managed over supervisory control and data acquisition (SCADA) systems. Today, very few of the control loops that are closed over the SCADA system are automatic. Instead most loops are managed by a human operator, who sends out set-points to lower level controllers, for example. For the operator to obtain a state of awareness and to provide decision support, the SCADA systems often include energy management systems (EMS), which feature applications such as state estimators (SE), bad data detection (BDD), and optimal power flow (OPF). In this tutorial lecture, we will give an introduction to these applications from a secure control point of view. Towards the end, we will also be forward-looking and discuss secure control in the perspective of the Smart Grid. The Smart Grid will feature many automatic distributed control systems, which will put very high demands on security.
17:40-18:00	ThC1.4
Air-Gap to Always-Connected: An Industry Perspective on Security for Control Systems (I)*
Griffin, Robert	RSA Security
The transformations in information technologies, threat landscape and user devices have profound implications for control systems. Fundamental architectural approaches such as air gaps are now ineffective because of requirements for integration with data analytics environments. Aggressive attacks use social engineering tactics to make human agents the bridge across system boundaries in order to introduce stealthy and persistent malware. Meanwhile, opportunities for shared threat intelligence and analytics on massive volumes of security-related information offer new models for control system security. This presentation discusses the issues and opportunities to which security for control systems must respond, as well as the architectural models that are emerging in areas such as Smart Grid in order to address these issues and opportunities.

ThC2	HG F3
Applications of Model Predictive Control (Regular Session)
Chair: De Keyser, Robin M.C.	Univ. of Gent
Co-Chair: Casale Brunet, Simone	EPFL
16:00-16:20	ThC2.1
A Lego Mindstorms NXT Experiment for Model Predictive Control Education, pp. 2549-2554
Canale, Massimo	Pol. di Torino
Casale Brunet, Simone	EPFL
This paper presents an educational framework based on the Lego Mindstorms NXT robotic platform used to outline both the theoretical and practical aspects of the Model Predictive Control theory. The case of a two-wheeled inverted pendulum is considered as at-size scenario. For such a system, starting from its mathematical modeling, an established design methodology is presented aiming to outline step-by-step the predictive controller implementation on a low-power architecture. The effectiveness of this multidisciplinary approach is illustrated along this presentation and demonstrated with experimental results.
16:20-16:40	ThC2.2
Performance Improvement of an NMPC Problem by Search Space Reduction and Experimental Validation to a PEM Fuel Cell System, pp. 2555-2560
Ziogou, Chrysovalantou	Centre for Res. and Tech. Hellas (CERTH)
Georgiadis, Michael C.	Univ. of Western Macedonia, DepartmentofEngineeringInformat
Pistikopoulos, Efstratios N.	Imperial Coll.
Voutetakis, Spyridon	Centre for Res. and Tech. Hellas (CERTH)
Papadopoulou, Simira	Alexander Tech. Educational Inst.
The current work addresses the control issues that arise during the operation of a fuel cell system based on a novel combination of two Model Predictive Control (MPC) strategies, explicit and Nonlinear MPC (NMPC). The proposed framework relies on an NMPC formulation that uses a simultaneous direct transcription dynamic optimization method that recasts the multivariable control problem into a nonlinear programming problem (NLP) using a warm-start initialization method. The performance of the optimizer is improved by a search space reduction (SSR) technique which is based on a piecewise affine approximation (PWA) of the variable's feasible space, derived offline by a multi-parametric Quadratic Programming (mpQP) method. The behavior of the explicit NMPC (exNMPC) framework is initially explored by a simulation study and subsequently it is experimentally verified through the online deployment to the fuel cell unit, demonstrating excellent response in terms of computational effort and accuracy with respect to the control objectives.
16:40-17:00	ThC2.3
Centralised and Decentralised Control of the Broken River, pp. 2561-2566
Foo, Mathias	Asia Pacific Center for Theoretical Physics (APCTP)
Ooi, Su Ki	NICTA
Weyer, Erik	Univ. of Melbourne
In this paper centralised Model Predictive Control (MPC), tuned in two different ways, and a decentralised control scheme are proposed for the control of the Broken River in Victoria, Australia. The control objective is to improve water resource management for the benefit of irrigators and the environment. The controllers are designed based on simple time delay and integrator delay models. The controllers are evaluated in realistic simulation scenarios and compared to manual operation. The use of control offers increased operational flexibility with a significant potential for substantial water savings, improved level of service to irrigators and improved environmental benefits.
17:00-17:20	ThC2.4
Thermal Storage Power Balancing with Model Predictive Control, pp. 2567-2572
Halvgaard, Rasmus	Tech. Univ. of Denmark
Poulsen, Niels Kjølstad	Tech. Univ. of Denmark
Madsen, Henrik	Tech. Univ. of Denmark
Jørgensen, John Bagterp	Tech. Univ. of Denmark
The method described in this paper balances power production and consumption with a large number of thermal loads. Linear controllers are used for the loads to track a temperature set point, while Model Predictive Control (MPC) and model estimation of the load behavior are used for coordination. The total power consumption of all loads is controlled indirectly through a real-time price. The MPC incorporates forecasts of the power production and disturbances that influence the loads, e.g. time-varying weather forecasts, in order to react ahead of time. A simulation scenario demonstrates that the method allows for the integration of flexible thermal loads in a smart energy system in which consumption follows the changing production.
17:20-17:40	ThC2.5
Nonlinear Predictive Control of an Evaporator for Bioethanol Production, pp. 2573-2578
Ipanaque Alama, William	Univ. de Piura
Oliden, José Carlos	Univ. de Piura
Manrique, José	Univ. de Piura
Hernandez, Andres	Ghent Univ.
Dutta, Abhishek	Ghent Univ.
De Keyser, Robin M.C.	Univ. of Gent
Production of biofuel has a positive environmental and economic impact; therefore, the interest for accurate modeling and more advanced control techniques has grown considerably over the last years. The reason is that it allows optimizing the productivity while the energy consumption is minimized. This paper presents the modeling, simulation and nonlinear control of a double-effect evaporation process to obtain bioethanol from sugarcane juice. This is achieved by controlling the juice concentration at the output of the last evaporator stage. Nonlinear Model Predictive Control (NMPC) has been successfully implemented, following the Extended Prediction Self-Adaptive Control (EPSAC) approach. This algorithm has been chosen among the different methods as it requires less computational load. The effectiveness of the proposed EPSAC controller is presented and compared against PI and Generalized Predictive Control (GPC), through simulation.
17:40-18:00	ThC2.6
Governor Principles for Increased Safety on Vessels with Diesel-Electric Propulsion, pp. 2579-2584
Veksler, Aleksander	Norweigian Univ. of Sci. & Tech.
Johansen, Tor Arne	Norweigian Univ. of Sci. & Tech.
Skjetne, Roger	Norwegian Univ. of Science And Tech.
Mathiesen, Eirik	Kongsberg Maritime AS
In this paper, a governor for marine diesel-electric power plants with a normal and an emergency mode is proposed. In normal mode, variations in the electric frequency are weighed against wear-and-tear of the generator due to thermic variations in the engine. In emergency mode, the governor briefly disregards the wear-and-tear concerns, and attempts to maintain the frequency as steady as possible. This leads to reduced risk of blackout due to larger margins to the underfrequency condition and more reliable connection of additional generating sets to the electric grid. The emergency mode is entered under abnormal conditions only, the overall wear-and-tear resulting from this addition is negligible. The governor is implemented as a moving horizon controller.

ThC3	HG F5
Sampled Data Control and Delay Systems (Regular Session)
Chair: Griggs, Wynita M.	NUIM
Co-Chair: Hetel, Laurentiu	LAGIS UMR CNRS 8219
16:00-16:20	ThC3.1
On the Stability of Input-Affine Nonlinear Systems with Sampled-Data Control, pp. 2585-2590
Omran, Hassan	Ec. Centrale de Lille
Hetel, Laurentiu	LAGIS UMR CNRS 8219
Richard, Jean-Pierre	Ec. Centrale de Lille
Lamnabhi-Lagarrigue, Francoise	CNRS-EECI
This paper is dedicated to the stability analysis of nonlinear sampled-data systems, which are affine in the input. Assuming that a stabilizing continuous-time controller exists and it is implemented digitally, we intend to provide sufficient asymptotic/exponential stability conditions for the sampled-data system. This allows to find an estimate of the upper bound on the asynchronous sampling periods. The stability analysis problem is formulated both globally and locally. The main idea of the paper is to address the stability problem in the framework of dissipativity theory. Furthermore, the result is particularized for the class of polynomial input-affine sampled-data systems, where stability may be tested numerically using sum of squares decomposition and semidefinite programming.
16:20-16:40	ThC3.2
Digital Stabilization of Finite Sampled Nonlinear Dynamics with Delays: The Unicycle Example, pp. 2591-2596
Tanasa, Valentin	Univ. Paris 11
Monaco, Salvatore	Univ. di Roma La Sapienza
Normand-Cyrot, Marie-Dorothée	CNRS-Supélec
The paper illustrates through the example of a mobile robot, how discretization makes easier the design of a predictor based stabilizer for nonlinear dynamics with delayed input, admitting finite sampled equivalent models.
16:40-17:00	ThC3.3
Characterisations of the "Mixed" Small Gain and Passivity Property for Linear Systems in Discrete Time, pp. 2597-2602
Griggs, Wynita M.	NUIM
Ordóñez-Hurtado, Rodrigo H.	Univ. of Chile
Sajja, Shravan	NUI Maynooth
Lanzon, Alexander	Univ. of Manchester
Shorten, Robert	Nat. Univ. of Ireland
Characterisations of "mixed" systems are presented in a discrete-time setting. First, a feedback stability result based on the Nyquist stability theorem is presented. Second, an eigenvalue-based characterisation of "mixed" systems based on their state-space data is derived. The results are analogous to previous results presented for the continuous-time case and provide a foundation for further study concerning the discretisation of "mixed" systems.
17:00-17:20	ThC3.4
A Robust Polytopic Approach for State-Dependent Sampling, pp. 2603-2608
Fiter, Christophe	CNRS - Ec. Centrale de Lille
Hetel, Laurentiu	LAGIS UMR CNRS 8219
Perruquetti, Wilfrid	Ec. Centrale de Lille
Richard, Jean-Pierre	Ec. Centrale de Lille
This work aims at decreasing the number of sampling instants in state feedback control for perturbed linear time invariant systems. The approach is based on linear matrix inequalities obtained thanks to Lyapunov-Razumikhin stability conditions and convexification arguments that guarantee the exponential stability for a chosen decay-rate. First, the method enables to perform a robust stability analysis regarding time-varying sampling and to maximize a lower-bound estimate of the maximal allowable sampling interval, by computing the adequate Lyapunov-Razumikhin function. Then, it makes it possible to design a state-dependent sampling control scheme that enlarges even further the maximal allowable sampling intervals.
17:20-17:40	ThC3.5
Stability of Discrete-Time Systems with Stochastically Delayed Feedback, pp. 2609-2614
Gomez, Marcella Mary	California Inst. of Tech.
Orosz, Gabor	Univ. of Michigan, Ann Arbor
Murray, Richard	Caltech
This paper investigates the stability of linear systems with stochastic delay in discrete time. Stability of the mean and second moment of the non-deterministic system is determined by a set of deterministic discrete-time equations with distributed delay. A theorem is provided that guarantees convergence of the state with convergence of the second moment, assuming that delays are identically independently distributed. The theorem is applied to a scalar equation where the stability of the equilibrium is determined.
17:40-18:00	ThC3.6
Numerical Simulation and Identification of Fractional Systems Using Digital Adjustable Fractional Order Integrator, pp. 2615-2620
Djouambi, Abdelbaki	Univ. Oum El Bouaghi
Charef, Abdelfatah	Univ. Constantine
Voda, Alina	UJF
A new method for numerical simulation and parameter identification of fractional order models is presented in this paper. A Digital Adjustable Fractional Integrator is proposed to improve the numerical simulation of fractional order systems. The main feature of this tool consists in: the important reduction of simulation run time, operate as a one-step forward predictor and, consequently, it can be used for real-time numerical simulation and identification of different fractional order systems using the same structure. The consistence, convergence and stability of the proposed method are proved. Finally, some numerical results are presented to demonstrate that the proposed approximation is a computationally efficient method.

ThC4	HG F7
State Estimation (Regular Session)
Chair: Zolghadri, Ali	Univ. Bordeaux I
Co-Chair: Bharani Chandra, Kumar Pakki	Univ. of Leicester
16:00-16:20	ThC4.1
An Iterative Partition-Based Moving Horizon Estimator for Large-Scale Linear Systems, pp. 2621-2626
Schneider, René	RWTH Aachen Univ.
Scheu, Holger	RWTH Aachen Univ.
Marquardt, Wolfgang	Aachener Verfahrenstechnik - Prozesstechnik
We transfer the ideas behind sensitivity-driven distributed model predictive control (c.f. Scheu and Marquardt, 2011) to the moving horizon state estimation problem and present a novel decentralized state estimation algorithm, namely, sensitivity-driven partition-based moving horizon estimation (S-PMHE). We discuss convergence and optimality of S-PMHE for the case of given positive-definite arrival cost weights. Finally, we demonstrate the method on a numerical example.
16:20-16:40	ThC4.2
Observer-Based Stabilization of Uncertain Linear Systems with Recycle: An LMI Approach, pp. 2627-2632
Parada, Miguel	Tech. Univ. Berlin
Borges, Renato Alves	Univ. of Brasilia
Sbarbaro, Daniel G.	Univ. de Concepcion
Peres, Pedro L. D.	Univ. of Campinas
This paper deals with the control of processes having recycle streams. These processes can be modeled as state delayed systems. In this work the problem of robust design of an observed-based controller for these systems is presented. The system is assumed to have norm-bounded uncertainties which are independent in every matrix involved in its state space realization. To ensure asymptotic stability of the closed loop system, a Lyapunov functional is used to obtain delay-independent design conditions. The controller design is accomplished by means of a convex optimization procedure formulated using linear matrix inequalities (LMIs). Numerical examples are provided to illustrate the main characteristics of the proposed design method.
16:40-17:00	ThC4.3
Distributed Bounded-Error State Estimation for Partitioned Systems Based on Practical Robust Positive Invariance, pp. 2633-2638
Riverso, Stefano	Univ. degli Studi di Pavia
Rubini, Daria	Univ. degli Studi di Pavia
Ferrari Trecate, Giancarlo	Univ. degli Studi di Pavia
In this paper we propose a novel partition-based state estimator for linear discrete-time systems composed of physically coupled subsystems affected by bounded disturbances. The proposed scheme is distributed in the sense that each local state estimator exploits suitable pieces of information from parent subsystems. Moreover, differently from schemes based on moving horizon estimation, it does not require the on-line solution to optimization problems. Our method also guarantees the satisfaction of constraints on local estimation errors. We achieve our aims exploiting the notion of practical robust positive invariance developed in Rakovic et al., 2011. As an example, we illustrate the use of the distributed state estimator for reconstructing the states of a power network system.
17:00-17:20	ThC4.4
Cubature $H_infty$ Information Filter, pp. 2639-2644
Bharani Chandra, Kumar Pakki	Control Res. Group, Dept. of Engineering, Univ. of Leic
Gu, Dawei	Univ. of Leicester
Postlethwaite, Ian	Northumbria Univ.
This paper presents a state estimation algorithm referred to as a cubature $H_infty$ information filter (C$H_infty$IF) for nonlinear systems. The proposed algorithm is developed from a cubature Kalman filter, an $H_infty$ filter and an extended information filter. The C$H_infty$IF is a derivative free filter, where the information state vector and information matrix are propagated rather than the state vector and error covariance matrix. Furthermore, the C$H_infty$IF is extended for multi-sensor state estimation. The efficacy of the C$H_infty$IF is demonstrated by a simulation example of a permanent magnet synchronous motor in the presence of Gaussian and non-Gaussian noises.
17:20-17:40	ThC4.5
Adaptive Observer-Based Sinusoid Identification: Structured and Bounded Unstructured Measurement Disturbances, pp. 2645-2650
Chen, Boli	Imperial Coll. London
Pin, Gilberto	Univ. of Trieste (Italy)
Parisini, Thomas	Imperial Coll. & Univ. of Trieste
The paper deals with an adaptive observer methodology for estimating the parameters of an unknown sinusoidal signal from a measurement perturbed by structured and unstructured uncertainties. The proposed technique makes it possible to handle measurement signals affected by structured uncertainties like, for example, bias and drifts which are typically present in applications. The stability of the estimator with respect to bounded additive disturbances is addressed by Input-to-State Stability arguments. The effectiveness of the proposed technique is shown through numerical simulations where comparisons with some recently proposed algorithms are also provided.
17:40-18:00	ThC4.6
On Interval Observer Design for Time-Invariant Discrete-Time Systems, pp. 2651-2656
Efimov, Denis	INRIA - LNE
Perruquetti, Wilfrid	Ec. Centrale de Lille
Raïssi, Tarek	Conservatoire National des Arts et Métiers
Zolghadri, Ali	Univ. Bordeaux I
The problem of interval state observer design is addressed for time-invariant discrete-time systems. Two solutions are proposed: the first one is based on a similarity transformation synthesis, which connects a constant matrix with its nonnegative representation ensuring the observation error positivity. The second contribution shows that in the discrete-time case the estimation error dynamics always can be represented in a cooperative form without a transformation of coordinates. The corresponding observer gain can be found as a solution of the formulated LMIs. The performances of the proposed observers are demonstrated through computer simulations.

ThC5	HG E1.1
Optimization for Hybrid Systems (Regular Session)
Chair: Daoutidis, Prodromos	Univ. of Minnesota
Co-Chair: Hempel, Andreas Berndt	ETH Zurich
16:00-16:20	ThC5.1
Every Continuous Piecewise Affine Function Can Be Obtained by Solving a Parametric Linear Program, pp. 2657-2662
Hempel, Andreas Berndt	ETH Zurich
Goulart, Paul J.	ETH Zurich
Lygeros, John	ETH Zurich
It is well-known that solutions to parametric linear or quadratic programs are continuous piecewise affine functions of the parameter. In this paper we prove the converse, i.e. that every continuous piecewise affine function can be identified with the solution to a parametric linear program. In particular, we provide a constructive proof that every piecewise affine function can be expressed as the linear mapping of the solution to a parametric linear program with at most twice as many variables as the dimension of the image of the piecewise affine function. Our method is illustrated via two small numerical examples.
16:20-16:40	ThC5.2
Suboptimality Bounds for Linear Quadratic Problems in Hybrid Linear Systems, pp. 2663-2668
Kouhi, Yashar	Max Planck Insitute
Bajcinca, Naim	Max Planck Insitute
Sanfelice, Ricardo G.	Univ. of Arizona
A method for computation of lower and upper bounds for the linear quadratic cost function associated to a class of hybrid linear systems is proposed. The optimization problem involves state space constraints and switches between the continuous and discrete dynamics at fixed time instances on the boundaries of the flow and jump sets. Our approach computes a quadratic suboptimal cost parameterized by initial and end state variables of all time intervals. Then, the unknown parameters are determined via solving constrained quadratic programming problems.
16:40-17:00	ThC5.3
Dynamic Real-Time Optimization and Control of a Hybrid Energy System, pp. 2669-2674
Trifkovic, Milana	Univ. of Minnesota
Marvin, W. Alex	Univ. of Minnesota
Sheikhzadeh, Mehdi	Lambton Coll.
Daoutidis, Prodromos	Univ. of Minnesota
A proactive energy management strategy for a stand-alone hybrid renewable energy system is presented. The study was motivated by the system built in Lambton College (Sarnia, Ontario, Canada) which includes photovoltaic arrays, wind turbine, battery, electrolyzers, hydrogen storage tanks, and fuel cells. The control architecture consists of two levels of hierarchy: (1) an optimal predictive scheduling at the supervisory level; (2) system unit control at the low level. A "day-ahead" approach is followed at the supervisory level and a bidirectional communication between the supervisory, proactive control, and the low level control layer. The proposed energy management strategy accounts for external (i.e. weather and demand) and internal disturbances.
17:00-17:20	ThC5.4
Hybrid MPC Approach to Reconfiguration of Building Heating System, pp. 2675-2680
Siroky, Jan	Univ. of West Bohemia in Pilsen, Faculty of AppliedSciences
Cigler, Jiri	Czech Tech. Univ. in Prague, Faculty ofElectricalEngine
Ferkl, Lukas	Czech Tech. Univ. in Prague
This paper describes the application of Hybrid Model Predictive Control (HMPC) to a building heating system. The hybrid model contains continuous variables corresponding to physical quantities as well as discrete variables serving as indices of a Linear Time Invariant (LTI) model in action. Two LTI models are considered, each describing different configurations of the building heating system. The application of HMPC allows efficient handling of disturbances by reconfiguration of the heating system. The proposed HMPC strategy improves comfort and reduces energy demands. An exact solution of HMPC is computationally demanding; therefore, three suboptimal solutions are suggested. These take into account specifics of the building heating system and significantly reduce the computational complexity. All presented strategies are compared by means of a numerical simulations using real weather data and a model of a real building.
17:20-17:40	ThC5.5
Modeling and Predictive Control of Nonlinear Hybrid Systems Using Disaggregation of Variables – a Convex Formulation, pp. 2681-2686
Nandola, Nareshkumar Naranbhai	ABB Corp. Res. Center
Puttannaiah, Karan	Arizona State Univ.
The current work is motivated by the need of achieving global solution and better computational efficiency for control of any arbitrary nonlinear hybrid dynamical systems (NHDS). In this work, we present a novel modeling and corresponding model predictive control (MPC) formulation for NHDS. The proposed modeling approach relies on disaggregation of polynomials of binary variables that appear in the multiple partially linearized (MPL) model. In particular, we use auxiliary continuous variables and linear constraints to model these polynomials and represent the MPL model in a linear fashion. Subsequently, disaggregation of the variables based multiple models are used to formulate the MPC law for NHDS. The MPC formulation takes similar form as multiple mixed logical dynamical (MMLD) model based MPC and yields a convex MIQP optimization problem. Moreover, the proposed modeling approach results in a compact model than the corresponding MMLD model as it refrains from adding any extra binary variables. Therefore, offers certain computational advantage when used for the predictive control of NHDS. The efficacy of the proposed solution is demonstrated on a three-tank benchmark hybrid system
17:40-18:00	ThC5.6
Low-Rank and Sparse Optimization for GPCA with Applications to SARX System Identification, pp. 2687-2692
Konishi, Katsumi	Kogakuin Univ.
This paper proposes a low-rank and sparse optimization approach to generalized principal component analysis (GPCA) problems. The GPCA problem has a lot of applications in control, system identification, signal processing, and machine learning, however, is a kind of combinatorial problems and NP hard in general. This paper formulates the GPCA problem as a low-rank and sparse optimization problem, that is, matrix rank and $l_0$ norm minimization problem, and proposes a new algorithm based on the iterative reweighed least squares (IRLS) algorithm. This paper applies this algorithm to the system identification problem of switched autoregressive exogenous (SARX) systems, where the model order of each submodel is unknown. Numerical examples show that the proposed algorithm can identify the switching sequence, system order and parameters of submodels simultaneously.

ThC6	HG E1.2
Quantum Control (Regular Session)
Chair: Petersen, Ian R.	Australian Defence Force Acad.
Co-Chair: Albertini, Francesca	Univ. Di Padova
16:00-16:20	ThC6.1
Markov Operators on Cones and Non-Commutative Consensus, pp. 2693-2700
Gaubert, Stephane	INRIA and Ec. Pol.
Qu, Zheng	CMAP, Pol.
The analysis of classical consensus algorithms relies on contraction properties of Markov matrices with respect to the Hilbert semi-norm (infinitesimal version of Hilbert's projective metric) and to the total variation norm. We generalize these properties to the case to operators on cones. This is motivated by the study of ``non-commutative consensus'', i.e., of the dynamics of linear maps leaving invariant cones of positive semi-definite matrices. Such maps appear in quantum information (Kraus maps), and in the study of matrix means. We give a characterization of the contraction rate of an abstract Markov operator on a cone, which extends classical formulae obtained by Doe blin and Dobrushin in the case of Markov matrices. In the special case of Kraus maps, we relate the absence of contraction to the positivity of the ``zero-error capacity'' of a quantum channel. We finally show that a number of decision problems concerning the contraction rate of Kraus maps reduce to finding a rank one matrix in linear spaces satisfying certain conditions and discuss complexity issues.
16:20-16:40	ThC6.2
Algebraic Conditions for Indirect Controllability in Quantum Coherent Feedback Schemes, pp. 2701-2706
Albertini, Francesca	Univ. Di Padova
D'Alessandro, Domenico	Iowa State Univ.
In coherent feedback control schemes a target quantum system S is put in contact with an auxiliary system A and the coherent control can directly affect only A. The system S is controlled indirectly through the interaction with A. The system S is said to be indirectly controllable if every unitary transformation can be performed on the state of S with this scheme. In this paper we show how indirect controllability of S is equivalent to complete controllability of the combined system S+A, if the dimension of A is greater than 2. In the case where the dimension of A is equal to 2, it is possible to have indirect controllability without having complete controllability of S+A and we give sufficient conditions for this to happen. We conjecture that these conditions are also necessary. The results of the paper extend previous results known in the literature relaxing assumptions on the dimensions and on the initial conditions of the systems involved.
16:40-17:00	ThC6.3
Quantum Popov Robust Stability Analysis of an Optical Cavity Containing a Saturated Kerr Medium, pp. 2707-2711
Petersen, Ian R.	Australian Defence Force Acad.
This paper applies results on the robust stability of nonlinear quantum systems to a system consisting an optical cavity containing a saturated Kerr medium. The system is characterized by a Hamiltonian operator which contains a non-quadratic term involving a quartic function of the annihilation and creation operators. A saturated version of the Kerr nonlinearity leads to a sector bounded nonlinearity which enables a quantum small gain theorem to be applied to this system in order to analyze its stability. Also, a non-quadratic version of a quantum Popov stability criterion is presented and applied to analyze the stability of this system.
17:00-17:20	ThC6.4
A Stochastic Lyapunov Feedback Technique for Propagator Generation of Quantum Systems on U(n), pp. 2712-2716
Silveira, Hector Bessa	Federal Univ. of Santa Catarina
Pereira da Silva, Paulo Sergio	Univ. of Sao Paulo
Rouchon, Pierre	ENSMP
This work treats the problem of generating any desired goal propagator for a driftless quantum system that evolves on the unitary group U(n). The physical relevance of such control problem is the realization of arbitrary quantum gates in quantum computers. Assuming only the controllability of the system, the paper constructs explicit stochastic control laws that assure global asymptotic convergence of the propagator of the system towards the goal propagator. The purpose of introducing a stochastic behaviour in the controls is to speed up convergence. The control strategy can be rigorously proved based on Lyapunov feedback and stochastic techniques. The controls laws rely on a reference trajectory that crosses the desired goal propagator in a time-periodic fashion and such that its corresponding linearised system generates the Lie algebra u(n). Their existence is ensured by the Return Method of Coron, and standard Fourier series results allows them to be explicitly constructed.
17:20-17:40	ThC6.5
Coherent Quantum Filtering for Physically Realizable Linear Quantum Plants, pp. 2717-2723
Vladimirov, Igor	Univ. of New South Wales at the AustralianDefenceForceAcade
Petersen, Ian R.	Australian Defence Force Acad.
The paper is concerned with a problem of coherent (measurement-free) filtering for physically realizable (PR) linear quantum plants. The state variables of such systems satisfy canonical commutation relations and are governed by linear quantum stochastic differential equations, dynamically equivalent to those of an open quantum harmonic oscillator. The problem is to design another PR quantum system, connected unilaterally to the output of the plant and playing the role of a quantum filter, so as to minimize a mean square discrepancy between the dynamic variables of the plant and the output of the filter. This coherent quantum filtering (CQF) formulation is a simplified feedback-free version of the coherent quantum LQG control problem which remains open despite recent studies. The CQF problem is transformed into a constrained covariance control problem which is treated by using the Frechet differentiation of an appropriate Lagrange function with respect to the matrices of the filter.
17:40-18:00	ThC6.6
Quantum Implemention of an LTI System with the Minimal Number of Additional Quantum Noise Inputs, pp. 2724-2727
Vuglar, Shanon Leigh	Univ. of New South Wales (ADFA)
Petersen, Ian R.	Australian Defence Force Acad.
Physical Realizability addresses the question of whether it is possible to implement a given LTI system as a quantum system. It is in general not true that a given synthesized quantum controller described by a set of stochastic differential equations is equivalent to some physically meaningful quantum system. However, if additional quantum noises are permitted in the implementation it is always possible to implement an arbitrary LTI system as a quantum system. In this paper we give an expression for the exact number of noises required to implement a given LTI system as a quantum system. Furthermore, we focus our attention on proving that this is a minimum, that is, it is not possible to implement the system as a quantum system with a smaller number of additional quantum noises.

ThC7	HG E3
Constrained Control (Regular Session)
Chair: Prandini, Maria	Pol. di Milano
Co-Chair: Duncan, Stephen	Univ. of Oxford
16:00-16:20	ThC7.1
Further Results on Saturated Globally Stabilizing Linear State Feedback Control Laws for Single-Input Neutrally Stable Planar Systems, pp. 2728-2733
Yang, Tao	Royal Inst. of Tech.
Stoorvogel, Anton A.	Univ. of Twente
Saberi, Ali	Washington State Univ.
Johansson, Karl Henrik	Royal Inst. of Tech.
It is known that for single-input neutrally stable planar systems, there exists a class of saturated globally stabilizing linear state feedback control laws. The goal of this paper is to characterize the dynamic behavior for such a system under arbitrary locally stabilizing linear state feedback control laws. On the one hand, for the continuous-time case, we show that all locally stabilizing linear state feedback control laws are also globally stabilizing control laws. On the other hand, for the discrete-time case, we first show that this property does not hold by explicitly constructing nontrivial periodic solution for a particular system. We then show for an example that there exists more globally stabilizing linear state feedback control laws than well known ones in the literature.
16:20-16:40	ThC7.2
Prioritization Schemes for Reference and Command Governors, pp. 2734-2739
Kalabic, Uros V.	Univ. of Michigan
Chitalia, Yash	Univ. of Michigan
Julia, Buckland	Ford Motor Company
Kolmanovsky, Ilya V.	Univ. of Michigan
Reference and command governors are predictive add-on schemes that are applied to closed-loop systems and guarantee constraint enforcement while tracking desired reference inputs. This paper introduces two methods for using these predictive methods in the presence of prioritized constraints. The ﬁrst method handles the case of “soft” constraints by using a slack variable to relax the constraints. The second method prioritizes the reference inputs, enforcing the constraints by modifying the lower priority reference inputs ﬁrst. Two examples are reported consisting of a constrained spring-mass-damper system and an F-16 aircraft with actuator constraints.
16:40-17:00	ThC7.3
Stochastic Constrained Control: Trading Performance for State Constraint Feasibility, pp. 2740-2745
Deori, Luca	Dipartimento di Elettronica e Informazione, Pol. di Milan
Garatti, Simone	Pol. di Milano
Prandini, Maria	Pol. di Milano
In this paper, we address finite-horizon control for a stochastic linear system subject to constraints on the control and state variables. A control design methodology is proposed where the appropriate trade-off between the minimization of the control cost (performance) and the satisfaction of the state constraints (safety) can be decided by introducing appropriate chance-constrained problems depending on some parameter to be tuned. From an algorithmic viewpoint, a computationally tractable randomized approach to find approximate solutions which are guaranteed to be feasible for the original chance-constrained problem is proposed. A numerical example concludes the paper.
17:00-17:20	ThC7.4
The Linear Quadratic Regulator with Chance Constraints, pp. 2746-2751
Schildbach, Georg	ETH Zurich
Goulart, Paul J.	ETH Zurich
Morari, Manfred	ETH Zurich
This paper is concerned with the design of linear state feedback control laws for linear systems with additive Gaussian disturbances. The objective is to find the feedback gain that minimizes a quadratic cost function in closed-loop operation, while observing chance constraints on the input and/or the state. It is shown that this problem can be cast as a semi-definite program (SDP), in which the chance constraints appear as linear or bilinear matrix inequalities. Both individual chance constraints (ICCs) and joint chance constraints (JCCs) can be considered. In the case of ICCs only, the resulting SDP is linear and can be solved efficiently as a convex optimization program. In the presence of JCCs the SDP becomes bilinear, however it can still be solved efficiently by an iterative algorithm, at least to a local optimum. The application of the method is demonstrated for several numerical examples, underscoring its flexibility and ease of implementation.
17:20-17:40	ThC7.5
Anti-Windup Compensation for Electron Beam Stabilisation Control Systems on Synchrotrons with Rate Constrained Actuators, pp. 2752-2757
Gayadeen, Sandira	Univ. of Oxford
Duncan, Stephen	Univ. of Oxford
In modern synchrotron machines, electrons travelling at relativistic speeds in a closed circular path are bent by strong electromagnetic fields, which cause the electrons to lose energy in the form of synchrotron radiation. In order to achieve optimum performance, electron beam stability is a crucial parameter for modern synchrotrons. In particular, sub-micron stability is now a common requirement for the vertical position of the beam and to achieve the required performance, beam stabilisation feedback systems are used. A common nonlinearity encountered with the actuators in synchrotron feedback systems are the slew rate limits that are included in the circuits that apply power to the magnets in order to limit voltage changes. The large dimensions of synchrotron feedback systems and fast sample rates mean that robust Model Predictive Control (MPC) is not feasible. Therefore, for this application, anti-windup techniques for rate constrained nonlinearities are appropriate. The approach in this paper is anti-windup synthesis based on Internal Model Control (IMC) where it is demonstrated how IMC anti-windup synthesis for static constraints can be extended to rate constraints to improve constrained performance and guarantee stability. An Integral Quadratic Constraint (IQC) framework is used to analyse the robust stability of the system in the presence of both rate constraints and an infinity norm bounded uncertainty. Robust stability tests results and simulation of the anti-windup performance using machine data of the implementation of the control design on the Storage Ring of the UK's national synchrotron facility, Diamond Light Source are presented.
17:40-18:00	ThC7.6
Null-Controllable Set Computation for a Class of Constrained Bilinear Systems, pp. 2758-2763
Schulze Darup, Moritz	Ruhr-Univ. Bochum
Mönnigmann, Martin	Ruhr-Univ. Bochum
We present a method for the exact computation of null-controllable sets for single-input bilinear systems with input and state constraints. The proposed approach is based on first linearizing the bilinear system exactly, and applying known methods for the recursive calculation of null-controllable sets subsequently. While these steps are obvious from a conceptual point of view, it has to be taken into account that the constraints are transformed in the exact linearization step. Unfortunately, the transformed constraints are in general non-convex, even if the original constraints are convex. We show how to represent the transformed constraints in terms of a finite number of convex sets, which is instrumental for the computer-based evaluation of the null-controllable sets.

ThC8	HG E5
Distributed Control and Optimization (Invited Session)
Chair: Necoara, Ion	Univ. Pol. Bucharest
Co-Chair: Zeilinger, Melanie N.	Ec. Pol. Fédéral de Lausanne (EPFL)
Organizer: Zeilinger, Melanie N.	UC Berkeley
Organizer: Necoara, Ion	Univ. Pol. Bucharest
16:00-16:20	ThC8.1
Robust Distributed Model Predictive Control of Linear Systems (I), pp. 2764-2769
Conte, Christian	ETH Zurich
Zeilinger, Melanie N.	UC Berkeley
Morari, Manfred	ETH Zurich
Jones, Colin N	EPFL, Lausanne
In this work, synthesis and closed-loop operation of robust distributed model predictive control (MPC) for linear systems using distributed optimization is discussed. Previous work has shown that a nominal MPC controller for this setup can be synthesized and operated in a purely distributed manner. This paper extends this concept to linear systems subject to additive bounded disturbance. It is shown how well-established robust MPC approaches can be applied to distributed systems. The main focus of the paper is on a thorough discussion of computational issues arising from distributed synthesis and closed-loop operation of existing robust MPC controllers. In particular, techniques for distributed synthesis of structured robust positive invariant sets and distributed constraint tightening are proposed. The paper is concluded by a numerical example which illustrates the functionality and performance of the proposed techniques.
16:20-16:40	ThC8.2
Distributed Model Predictive Control for Energy Distribution, pp. 2770-2776
Scherer, Helton Fernando	Univ. Federal de Santa Catarina
Pasamontes Romera, Manuel	Univ. de Almeria
Álvarez Hervás, José Domingo	Univ. of Seville
Guzman, Jose Luis	Univ. of Almeria
Camponogara, Eduardo	Univ. Federal de Santa Catarina
Normey-Rico, Julio Elias	Federal Univ. of Santa Catarina
This work presents a distributed model predictive control scheme for energy distribution. The energy is supposed to be supplied by a renewable power system whose energy production is limited and demanded by several consumers. Therefore, in some cases the produced energy cannot fulfill the energy requirements of the consumers. The proposed controller allows to distribute equitably the produced energy to these consumers without harming any of them. Moreover, simulated results based on a real case are presented in order to assess the proposed control strategy.
16:40-17:00	ThC8.3
Hierarchical Task Allocation for Multi-Agent Systems Encoded by Stochastic Reachability Specifications (I), pp. 2777-2782
Kariotoglou, Nikolaos	ETH Zurich
Summers, Sean	ETH Zurich
Raimondo, Davide Martino	Univ. of Pavia
We consider the problem of satisfying a set of objectives over a collection of agents. For a single agent, the optimal solution can be obtained via a stochastic reachability framework where optimal control policies come along with a performance metric, defined as the probability of successfully achieving a specified objective. As the number of agents increases, the approach quickly becomes computationally expensive and often intractable. We propose a method which includes an advisory controller that allocates tasks among agents based on their ability of handling individual objectives. This ability is encoded by the stochastic reachability performance metrics. The proposed method is tailored to an autonomous surveillance system composed of pan-tilt-zoom (PTZ) cameras and verified experimentally.
17:00-17:20	ThC8.4
Fast Coordinated Model Predictive Control of Large-Scale Distributed Systems with Single Coupling Constraint, pp. 2783-2788
Spudic, Vedrana	Univ. of Zagreb
Baotic, Mato	Univ. of Zagreb
In this paper we describe an efficient implementation of a discrete-time model predictive controller for coordinated control of large-scale distributed systems. The approach is based on parametrization and splitting of the underlying optimization problem into local optimization problems coupled via one coordination problem. The local problems are solved off-line using the multi-parametric optimization approach. The coordination problem is solved on-line at every time step, on a centralized control hardware. Due to the properties of local parametric solutions the coordination problem has a specific structure that can be utilized to solve the on-line problem very efficiently. In case when there is only one linear coupling constraint between subsystems, the on-line computation time can be drastically decreased. The efficiency of proposed controller implementation is depicted on a design of the optimal wind farm controller applicable for implementation on very large wind farms.
17:20-17:40	ThC8.5
A Random Coordinate Descent Algorithm for Large-Scale Sparse Nonconvex Optimization (I), pp. 2789-2794
Patrascu, Andrei	Automatic Control and Systems Engineering Department, Univ.
Necoara, Ion	Univ. Pol. Bucharest
In this paper we develop a random coordinate descent method suitable for solving large-scale sparse nonconvex optimization problems with composite objective function. Under the typical assumptions of nonconvexity of the smooth part of the objective function and separability and convexity of the nonsmooth part (e.g. $l_1$ regularization, box indicator functions or others), we derive an algorithm with a very simple and cheap iteration. We prove sublinear convergence rate for our method to a stationary point. Numerical results show that our algorithm performs favourably in comparison to other algorithms on large-scale sparse nonconvex problems, e.g. the eigenvalue complementarity problem arising in different areas such as stability of dynamical systems, distributed control and resonance frequency of mechanical structures with friction.
17:40-18:00	ThC8.6
From Non-Cooperative to Cooperative Distributed MPC: A Simplicial Approximation Perspective, pp. 2795-2800
Bürger, Mathias	Univ. of Stuttgart
Notarstefano, Giuseppe	Univ. of Lecce
Allgower, Frank	Univ. of Stuttgart
The paper deals with the coordination of dynamical systems by distributed model predictive control. We consider a set-up in which the subsystems dynamics are decoupled, while the subsystems outputs are coupled by some constraint. Starting from a well established non-iterative and non-cooperative architecture, we provide a novel interpretation for this noncooperative scheme as a simplicial approximation of a convex program. Thanks to this novel interpretation, we are able to show why the existing algorithm, while guaranteeing feasibility, fails to compute an optimal solution to the centralized problem. Furthermore, by exploiting the simplicial approximation structure, we are able to propose a novel algorithm. The proposed algorithm inherits all the properties of the existing one, namely little communication, and feasibility. Furthermore, increasing the communication among the subsystems between two control updates improves the performance of the algorithm, regaining in the limit optimality in a cooperative sense.

ThC9	HG E7
Vehicle Formation Control (Regular Session)
Chair: Morbidi, Fabio	Inria, Grenoble - Rhone-Alpes
Co-Chair: Kim, Yoonsoo	Gyeongsang National Univ.
16:00-16:20	ThC9.1
On the Control of the Algebraic Connectivity and Clustering of a Mobile Robotic Network, pp. 2801-2806
Morbidi, Fabio	Inria, Grenoble - Rhone-Alpes
In this paper two related problems are studied: the control of the algebraic connectivity and clustering of a network of single-integrator agents. A steepest-descent algorithm is presented for the first problem, so that a smooth approximation of the algebraic connectivity of the underlying undirected communication graph converges to an assigned value. For the second problem, a new gradient-based control strategy is proposed to automatically partition the mobile robotic network into two predefined groups: our spectral-clustering method leverages a continuous-time power-iteration algorithm on the normalized Laplacian matrix which provides an estimate of its Fiedler vector at each time instant. The results of numerical simulations are provided to illustrate our theoretical findings.
16:20-16:40	ThC9.2
Consensus Control of Linear Multi-Agent Systems under Directed Dynamic Topology, pp. 2807-2812
Qin, Jiahu	The Australian National Univ.
Yu, Changbin	Australian National Univ.
This paper aims to extend the nonnegative matrix theory, which is widely employed for multiple integrator agents, to deal with the consensus control of generic linear multi-agent systems (MASs) under directed dynamic topology. It is finally shown that the exponential consensus can be reached under very relaxed conditions, i.e., the directed interaction topology is only required to be repeatedly jointly rooted and the exponentially unstable mode of each individual system is weak enough. Moreover, a least convergence rate and a bound for the unstable mode of the individual agent system, both of which are independent of the switching mode, can be explicitly specified.
16:40-17:00	ThC9.3
Cooperative Translation Control Based on Consensus with Reference Velocity: A Source-Seeking Application, pp. 2813-2818
Briñón Arranz, Lara	Gipsa-Lab.
Seuret, Alexandre	LAAS
This paper deals with the design of cooperative control laws for nonlinear multi-agent systems. On the first hand, the control objectives are to ensure that a group of agents reaches a time-varying circular formation characterized by some external references representing the position of the center of the formation and its derivatives. In order to reduce the amount of information to be shared, the second part of the paper considers the situation where only the velocity of the center of the formation is available to each agent. Then a distributed consensus algorithm is provided in order to satisfy the same control objectives. Finally, an application to the source-seeking problem is proposed to emphasize previous contribution. These results are supported through computer simulations.
17:00-17:20	ThC9.4
Collision-Free Vehicle Formation Control under Arbitrarily Switching Network Topology, pp. 2819-2823
Kim, Yoonsoo	Gyeongsang National Univ.
This paper is concerned with collision-free vehicle formation control when the inter-vehicle network topology is time-varying. Like typical collision-free formation control approaches, the proposed control approach involves two control phases: one for converging to a desired formation (no collision avoidance considered); the other for achieving collision avoidance. However, the proposed control approach has two distinctive features. First, the proposed control approach uses a fixed set of time-invariant control gains for achieving the desired formation but allows time-varying control gains only for avoiding collision. This feature lends itself well to the increase (and the decrease) of the controller's reliability (and complexity) during the most of manoeuvring periods. Second, the proposed control approach provably works for arbitrarily switching network topologies. Theoretical and numerical evidences are provided to justify these two features.
17:20-17:40	ThC9.5
Distributed Control of Swarm Motions As Continua Using Homogeneous Maps and Agent Triangulation, pp. 2824-2830
Rastgoftar, Hossein	Univ. of Central Florida
Jayasuriya, Suhada	Univ. of Central Florida
In this paper, presented is an approach for the control of swarm motions that treats a swarm as a deformable body or a continuum. By considering a special class of motion maps between a current configuration prescribed at a given time t and a desired configuration prescribed at a subsequent time t+∆t called a homogeneous map two strategies are proposed for the control of a swarm or a multi-agent system (MAS). In the first strategy, MAS motion control is achieved with no communication among agents in the idealized case where the required map from any t to t+∆t may be pre-determined. It is the case when a desired swarm task can be scripted ahead of time with some certainty. The second control strategy is based on three leaders prescribing the motion map for a desired swarm task which is then propagated to the followers via a local inter-agent communication protocol. The proposed communication protocol exploits some special features of homogeneous maps. It achieves the desired MAS motion control with three leaders and minimum inter-agent communications. Simulation results validate the effectiveness of the proposed strategies.
17:40-18:00	ThC9.6
Consensus-Based Source-Seeking with a Circular Formation of Agents, pp. 2831-2836
Briñón Arranz, Lara	Gipsa-Lab.
Schenato, Luca	Univ. of Padova
This paper deals with the source-seeking problem in which a group of autonomous vehicles must locate and follow the source of some signal based on measurements of the signal strength at different positions. As recently suggested, the gradient of the signal strength can be approximated by a circular formation of agents via a simple weighted average of the signal measured by the agents. Using this result, we propose a distributed source-seeking algorithm based on a consensus method which is guaranteed to steer the circular formation towards the source location using the estimated gradient direction. The proposed algorithm is provided with two tunable parameters that allow for a tradeoff between speed of convergence, noise filtering and formation stability. The benefit of using consensus-based algorithms resides in a more realist discrete time control of the agents and in asynchronous communication resilient to delays which is particularly relevant for underwater applications. The analytic results are finally complemented with numerical simulations.

ThC10	HG D1.1
Robust Control of Linear Systems II (Regular Session)
Chair: Glizer, Valery Y.	ORT Braude Coll.
Co-Chair: Vujanic, Robin	ETH Zurich
16:00-16:20	ThC10.1
On Practical Fixed Order H-Infinity Loop-Shaping, pp. 2837-2842
Serge, Hirwa	SUPELEC
Feyel, Philippe	Sagem Défense Sécurité
Duc, Gilles	Ec. Supérieure D' Electricité
Sandou, Guillaume	Ec. Superieure d Electricite
In this paper, a fixed order H-infinity synthesis is used to directly shape the open loop transfer function so that it matches as closely as possible, in the singular values sense, a desired frequency gain which captures both performance and robust stability objectives. First, a loop shaping weight is automatically computed, and then it is used in the controller synthesis by solving a well-suited four-block H-infinity control problem with a recently developed nonsmooth H-infinity synthesis routine. Two industrial applications are provided to illustrate the potential of the approach for low order robust controller design: a first one on the control of a piezo-actuated optronic mechanism, a second one dealing with inertial Line Of Sight stabilization loops.
16:20-16:40	ThC10.2
Solution of a Singular H-Infinity Control Problem for Linear Systems with State Delays, pp. 2843-2848
Glizer, Valery Y.	ORT Braude Coll.
An H-infinity control problem for linear systems with point-wise and distributed state delays is considered. The case, where the output equation does not contain the control variable, is treated. In this case, the problem is singular, i.e., the game-theoretic Riccati equation approach is not applicable to its solution. A method of solution of this problem, based on its regularization and asymptotic solution of the regularized problem, is proposed.
16:40-17:00	ThC10.3
Active Disturbance Rejection Control Based on a Simultaneous Adaptive Observer and a Time Varying Parameter Identifier, pp. 2849-2854
Luviano-Juárez, Alberto	UPIITA - IPN Mexico
Chairez, Isaac	UPIBI - IPN
Design of adaptive observers has been an active field in the systems theory. Until know, most of the existing solutions use a class of regular form with a set of unknown parameters. A different scheme used an estimation of external perturbations that can be compensated by the adaptive gain associated to the observer. In this paper, the presence of external perturbations for a model defined by a chain of integrators is considered. The adaptive observer used an identifier to obtain the time varying parameters used by the observer. Simultaneously, an adaptive gain associated to the observer adjusts the observer trajectories to provide the convergence between the states of the uncertain nonlinear system and the ones associated to the estimator suggested in this paper. Once the states of the uncertain system were obtained, a simple feedback controller was able to reject actively the perturbations that affect the nonlinear system. A simple third order uncertain system was evaluated in a numerical simulation for proving the performance of this observer/identifier. The same system was controlled using the estimated trajectories provided by the observer.
17:00-17:20	ThC10.4
I-PD Controller Design Based on Generalized KYP Lemma for Ball and Plate System, pp. 2855-2860
Mochizuki, Shuichi	Meiji Univ.
Ichihara, Hiroyuki	Meiji Univ.
A ball-on-plate balancing system, the ball and plate, has a camera to catch a ball position and a plate whose angle of inclination is limited. This paper proposes a design method of PID control based on the generalized Kalman-Yakubovich-Popov lemma for the ball and plate. The design method has two features: the first one is that a structure of the controller is called I-PD to suppress a large input signal against a major change of the reference signal; the second one is that a filter is introduced into the feedback loop to reduce an influence of the noise measurement from the camera. Both simulation and experiment evaluate the effectiveness of the design method.
17:20-17:40	ThC10.5
Robust Stabilization Via Disturbance Observer with Noise Reduction, pp. 2861-2866
Jo, Nam H.	Soongsil Univ.
Shim, Hyungbo	Seoul National Univ.
In this paper, we propose a robust controller using a modified disturbance observer (DOB). This modification is asked because the conventional DOB structure does not provide with any means to tune the high-frequency response. Since the measurement noise is significant in the high-frequency range, the performance against the noise in the conventional DOB control system cannot be better than that of the closed-loop system without DOB. Inspired by the new structure given in (Xie, 2010), we propose a modified disturbance observer structure, and present a necessary and sufficient condition for robust stability of the actual closed-loop system. Illustrative examples are included to show the effectiveness of the proposed design.
17:40-18:00	ThC10.6
On Quantifying Tolerable Closed-Loop Uncertainty in Frequency Domain, pp. 2867-2872
Li, Yuping	The Univ. of Melbourne
Given a linear plant and a feedback controller it is natural to ask: How much uncertainty can be tolerated by the closed-loop while achieving a specified level of performance? In this paper, a characterization of this question is formulated in terms of a constrained optimization problem; the cost reflects the size of non-constant weights used to quantify system uncertainty across frequency and the constraint is a structured singular value characterization of the required level of robust performance. In the case of unstructured uncertainty the problem can be solved via a family of problems that are convex pointwise in frequency. An iterative algorithm is developed for the case of structured uncertainty.

ThC11	HG D1.2
Benchmark on Adaptive Regulation: Rejection of Unknown/time-Varying Multiple Narrow Band Disturbances (Invited Session)
Chair: Landau, Ioan Dore	CNRS
Co-Chair: Castellanos Silva, Abraham	GIPSA-Lab.
Organizer: Landau, Ioan Dore	CNRS
Organizer: Airimitoaie, Tudor-Bogdan	Gipsa-Lab. Univ. de Grenoble
Organizer: Castellanos Silva, Abraham	GIPSA-Lab.
Organizer: Buche, Gabriel	GIPSA-Lab. CNRS
Organizer: Martinez, John-Jairo	Gipsa-Lab. INP-Grenoble
Organizer: Alma, Marouane	Univ. joseph fourier
Organizer: Karimi, Alireza	Ec. Pol. Federale
16:00-16:20	ThC11.1
An Active Vibration Control System As a Benchmark on Adaptive Regulation (I), pp. 2873-2878
Landau, Ioan Dore	CNRS
Castellanos Silva, Abraham	GIPSA-Lab.
Airimitoaie, Tudor-Bogdan	IRISA / Inria Rennes Bretagne Atlantique
Buche, Gabriel	GIPSA-Lab. CNRS
Noë, Mathieu	Paulstra S.A.; Vibrachoc Div.
The adaptive regulation is an important issue with a lot of potential for applications in active suspension, active vibration control, disc drives control and active noise control. One of the basic problem from the "control system" point of view is the rejection of multiple unknown and time varying narrow band disturbances without using an additional transducer for getting information upon the disturbances. An adaptive feedback approach has to be considered for this problem. Industry needs a "state of the art" in the field based on a solid experimental verification on a system using a current used technology. The paper present a benchmark problem for suppression of multiple unknown and/or time-varying vibrations and an associated active vibration control system using an inertial actuator on which the experimental verifications will be done. The objective is to minimize the residual force by applying an appropriate control effort through the inertial actuator. The system does not use any additional transducer for getting in real time information upon the disturbances. The benchmark has three levels of difficulties and the associated control performance specifications are presented. The results of the contibutors have been evaluated on a simulator and on the experimental facilities located at GIPSA-LAB, Grenoble, France. An extensive comparison of the results obtained by various approaches is presented.
16:20-16:40	ThC11.2
Adaptive Attenuation of Disturbance Formed As a Sum of Sinusoidal Signals Applied to a Benchmark Problem (I), pp. 2879-2884
Aranovskiy, Stanislav	Umea Univ.
The problem of adaptive attenuation of a disturbance formed as a finite sum of unknown sinusoidal signals is solved for a discrete-time plant with unstable zeros. It is assumed that a reliable model of the plant is known and the system is internally stable. We propose to construct a control signal as a linear combination of outputs of carefully chosen filters. The coefficients of the combination are tuned via an on-line identification based on the plant model. However, our approach avoids constructing an inverse of the plant model. The technique is applied to a case study on a challenging benchmark example in the field of active vibration control. Attenuation of a disturbance formed as a sum of up to three unknown/time-varying sinusoidal signals is demonstrated via simulation and experimental studies.
16:40-17:00	ThC11.3
Direct and Indirect Adaptive Regulation Strategies for Rejection of Time Varying Narrow Band Disturbances Applied to a Benchmark Problem (I), pp. 2885-2890
Airimitoaie, Tudor-Bogdan	IRISA / Inria Rennes Bretagne Atlantique
Castellanos Silva, Abraham	GIPSA-Lab.
Landau, Ioan Dore	CNRS
The paper will compare the performances which can be obtained using a direct adaptive regulation scheme based on the Youla-Kucera (YK) parametrization of the controller and using an adaptive finite impulse response (FIR) filter for implementing the internal model of the disturbance with a new indirect adaptive regulation scheme. The main features of this new scheme are: (i) the use of adaptive Band-stop Filters (BSFs) tuned at the frequencies of the disturbance instead of the Internal Model Principle (IMP) and (ii) a procedure for direct identification of frequencies contained in the disturbance. The use of adaptive BSFs allows to introduce the desired attenuation of the disturbance (instead of total rejection) and allows to get a much better shaping of the output sensitivity function (to meet the specification for the tolerated amplification outside the frequencies of the disturbance). The two approaches are comparatively evaluated on the benchmark simulator and on the benchmark active vibration control system.
17:00-17:20	ThC11.4
Adaptive Regulation of Time Varying Disturbances Via Weighted Robust Estimation and Automatic Controller Tuning (I), pp. 2891-2896
Fang, Huazhen	Dept. of MAE, UCSD
de Callafon, Raymond A.	Univ. of California, San Diego
This paper proposes a methodology to adaptively reduce time varying and narrow band harmonic disturbances via the estimation of a feedback controller transfer function parameterized in a Youla-Kucera parametrization. The proposed adaptive feedback regulation simultaneously minimizes the variance of a performance (output) signal and a control (input) signal in real-time. Uncertainty on the plant dynamics is taken into account by including a frequency weighting on the control signal and the solution is formulated as a Recursive Least Squares minimization. The methodology is applied to a mechanical vibration control benchmark that is part of a collaborative invited session at the conference to demonstrate how the proposed adaptive feedback regulation can effectively reduce unknown harmonic disturbances with time-varying frequency contents.
17:20-17:40	ThC11.5
Selective Model Inversion and Adaptive Disturbance Observer for Rejection of Time-Varying Vibrations on an Active Suspension (I), pp. 2897-2903
Chen, Xu	Univ. of California, Berkeley
Tomizuka, Masayoshi	UC Berkeley/NSF
This paper presents an adaptive control scheme for identifying and rejecting unknown and/or time-varying narrow-band vibrations. We discuss an idea of selective model inversion (SMI) for a (possibly non-minimum phase) plant dynamics at multiple narrow frequency regions, so that vibrations can be estimated and canceled by feedback. By taking advantage of the structure of the disturbance model, we can reduce the adaptation to identify the minimum amount of parameters, achieve accurate parameter estimation under noisy environments, and flexibly reject the narrow-band disturbances with clear tuning intuitions. Evaluation of the proposed algorithm is performed via simulation and experiments on a benchmark system for active vibration control.
17:40-18:00	ThC11.6
Nonlinear Learning-Based Adaptive Control for Electromagnetic Actuators, pp. 2904-2909
Benosman, Mouhacine	Mitsubishi Electric Res. Lab.
Atinc, Gokhan M.	Univ. of Illinois at Urbana Champaign
We present in this paper our preliminary results on the problem of learning-based adaptive trajectory tracking control for electromagnetic actuators. First, we develop a nominal nonlinear backstepping controller that stabilizes the tracking errors asymptotically and globally. Second, we robustify the nominal controller using a model-free learning technique, namely, multiparameter extremum seeking, to estimate the uncertain model parameters. In this sense we are proposing to solve an adaptive control problem with model-free learning-based algorithms. We show the performance of the proposed controller on a numerical example.

ThC12	HG D3.2
Nonlinear System Identification (Regular Session)
Chair: Findeisen, Rolf	Univ. of Magdeburg
Co-Chair: Nie, Xiaokai	The Univ. of Sheffield
16:00-16:20	ThC12.1
An Indirect Model Selection Algorithm for Nonlinear Active Noise Control, pp. 2910-2915
Morici, Simone	Dipartimento di Elettronica e Informazione, Pol. di Milan
Spiriti, Emanuele	Dipartimento di Elettronica e Informazione, Pol. di Milan
Piroddi, Luigi	Pol. di Milano
Model structure selection is a crucial task in applications where nonlinear black-box models are used, in order to reduce the model size and the associated computational effort. One such application is Active Noise Control (ANC), where nonlinear effects arise due e.g. to saturation and distortion of microphones and loudspeakers. Both parameter estimation and model selection are complex in the general nonlinear case if standard algorithms of the Least Mean Squares (LMS) type are used, due to the inherent difficulties in the gradient calculation when the secondary path is nonlinear. A model selection method is here proposed that employs a gradient-free parameter estimation algorithm to tackle the secondary path issue. A virtualization scheme is used to estimate the performance of the model subject to various different structural modifications, in order to select the most appropriate one to apply to the actual control filter. Some simulation examples are discussed to show the effectiveness of the algorithm.
16:20-16:40	ThC12.2
A New Approach to Solving the Inverse Frobenius-Perron Problem, pp. 2916-2920
Nie, Xiaokai	The Univ. of Sheffield
Coca, Daniel	The Univ. of Sheffield
This paper proposes a new matrix method to solve the inverse problem for the Frobenius-Perron equation. The method can be used to construct a piecewise linear Markov transformation, which approximates the evolution of an unknown dynamical system, based on a sequence of observed probability density functions generated by the system. This particular nonlinear system identification problem is solved using a three-step approach which involves determining the Markov partition, the matrix representation of the Frobenius-Perron operator and finally the corresponding point transformation. A numerical example is used to demonstrate the applicability of the approach.
16:40-17:00	ThC12.3
Outlier Analysis in Set-Based Estimation for Nonlinear Systems Using Convex Relaxations, pp. 2921-2926
Streif, Stefan	Otto-von-Guericke-Univ. Magdeburg
Karl, Matthias	Otto-von-Guericke-Univ. Magdeburg
Findeisen, Rolf	Univ. of Magdeburg
Set-based estimation for nonlinear systems is a useful tool to handle sparse and uncertain data. The tool provides outer bounds on feasible parameter sets and reachable states, as well as provable inconsistency certificates for entire parameter regions. In case of errors in the data such as outliers or incorrect a priori assumptions on variable uncertainties, set-based approaches can, however, lead to poor estimates or even rejection of a consistent model. We present a set-based approach to systematically identify outliers or incorrect variable uncertainty assumptions. The basic idea is to detect outliers by quantifying the influence they have on the inconsistency of an underlying feasibility problem. The results build on a set-based estimation framework that employs convex relaxations. Specifically we derive model consistency measures and sensitivity measures that combine the sensitivity information stored in the Lagrange dual variables. An algorithm is developed that iteratively detects outliers that contribute most to inconsistency. The algorithm terminates once the data and model are no longer proved inconsistent. The approach is illustrated with an example.
17:00-17:20	ThC12.4
Non Linear Electromechanical Cart Characterization Using Minimal Modeling Approach, pp. 2927-2932
Wongvanich, Napasool	Univ. of Canterbury
Hann, Christopher Eric	Univ. of Canterbury
Sirisena, Harsha	Univ. of Canterbury
This paper presents a minimal modelling methodology for capturing highly non-linear dynamics in an electromechanical cart system. The second order differential equation model describing the cart system is reformulated in terms of integrals to enable a fast method for identification of both constant and time varying parameters. The model is identified based on a single experimental proportional step response and is validated on a proportional-derivative (PD) controlled step input for a range of gains. Two models with constant damping and time varying non-linear damping were considered. The fitting accuracy for each model was tested on three separate data sets corresponding to three proportional gains. The three data sets gave similar non-linear damping models and in all cases the non-linear model gave smaller fitting errors than thelinear model. For the PD control responses, the constant damping model gave average percentage prediction errors of 9.3% and the non-linear model gave errors of 3.7%. The non-linear model also provided significantly better PD control design. These results demonstrate the ability of the proposed method to accurately capture significant non-linearities in the data. Computationally, the proposed algorithm is shown to be significantly faster than standard non-linear regression.
17:20-17:40	ThC12.5
Three Free Data Sets for Development and Benchmarking in Nonlinear System Identification, pp. 2933-2938
Wigren, Torbjorn	Uppsala Univ.
Schoukens, Johan	Vrije Univ. Brussels
System identification is a fundamentally experimental field of science in that it deals with modeling of system dynamics using measured data. Despite this fact many algorithms and theoretical results are only tested with simulations at the time of publication. One reason for this may be a lack of easily available live data. This paper therefore presents three sets of data, suitable for development, testing and benchmarking of system identification algorithms for nonlinear systems. The data sets are collected from laboratory processes that can be described by block – oriented dynamic models, and by more general nonlinear difference and differential equation models. All data sets are available for free download.
17:40-18:00	ThC12.6
Guaranteed Robust Optimal Experiment Design for Nonlinear Dynamic Systems, pp. 2939-2944
Telen, Dries	Katholieke Univ. Leuven
Houska, Boris	Univ. of Heidelberg
Logist, Filip	Katholieke Univ. Leuven
Diehl, Moritz	KU Leuven
Van Impe, Jan F.M.	Katholieke Univ. Leuven
This paper is about optimal experiment design for uncertain nonlinear dynamic processes. We are interested in designing experiments which allow to identify the unknown states and parameters of a differential equation from noisy measurements. Here, unpredictable process noise or structural model-plant mismatches can be an additional complication. In this case, robustness aspects have to be taken into account as the experiment has to be planned under incomplete information. The paper discusses problem formulations and numerical solution approaches for this type of robust optimal experiment design problems under various assumptions on the uncertainties. The corresponding techniques are illustrated in the design of experiments for a fedbatch bioreactor.

ThC13	HG D5.2
Fault Tolerant Systems (Regular Session)
Chair: Schiavoni, Nicola	Pol. di Milano
Co-Chair: Puig, Vicenc	Univ. Pol. de Catalunya
16:00-16:20	ThC13.1
Fault-Tolerant Pole-Placement in Single-Integrator Networks, pp. 2945-2950
Locatelli, Arturo	Pol. di Milano
Schiavoni, Nicola	Pol. di Milano
The paper faces a fault-tolerant control problem for a network of agents with single-integrator internal dynamics, which share information on their states according to an arbitrary topology. Specifically, we aim at designing a decentralized regulator able to place the closed-loop dominant poles near prespecified locations. Furthermore, this goal has to be achieved even in the presence of faults of the transmitting and receiving apparatuses of the single agents. We prove a necessary and sufficient solvability condition for our pole-placement problem, and show that it can be solved if and only if a simpler fault-tolerant stabilization problem admits a solution. Then, we give an explicit formula for a class of possible regulators.
16:20-16:40	ThC13.2
Passive and Active FTC Comparison for Polytopic LPV Systems, pp. 2951-2956
Rotondo, Damiano	UPC
Nejjari, Fatiha	Univ. Pol. de Catalunya
Puig, Vicenc	Univ. Pol. de Catalunya
Fault-tolerant control (FTC) allows to preserve performance and stability despite the presence of faults. The literature considers two main groups of techniques: the passive and the active FTC techniques. In case of the passive techniques, the fault is taken into account as a system perturbation, so that the control law has fault capabilities that allow the system to cope with the fault presence. On the other hand, in the case of the active FTC techniques, the control law uses some information given by a Fault Detection and Isolation (FDI) module, so that through some automatic adjustment in the control loop, the fault is tolerated with minimum performance degradation. In this paper, a linear parameter-varying (LPV)/linear matrix inequalities (LMIs)-based technique is used to achieve fault tolerance and to compare benefits and drawbacks of passive and active FTC. The proposed approach is applied to a two-wheel differential robot.
16:40-17:00	ThC13.3
Fault Tolerant Control for Polynomial Linear Parameter Varying (LPV) Systems Applied to the Stabilization of a Riderless Bicycle, pp. 2957-2962
Brizuela Mendoza, Jorge Aurelio	cenidet
Astorga-Zaragoza, Carlos	National Center for Res. and Tech. Development
Zavala-Rio, Arturo	Inst. Potosino de Investigacion Cientifica y Tecnologica
Canales Abarca, Francisco	ABB
Reyes-Reyes, Juan	Inst. Tecnologico de Zacatepec
This paper presents the results of a Fault Tolerant Control based on observers for polynomial LPV systems. The main contribution lies in observers, controller, Fault Diagnosis and Isolation unit and Fault Tolerant Control (FTC) design procedure, based on the solution of Parameterized Linear Matrix Inequalities (PLMI) with application to a riderless bicycle dynamics. Further, unlike previous works, this approach takes for its design only the measured outputs provided by the in-built bicycle prototype sensors, eliminating the necessity of additional computations for the control law. The previous fact is viewed as an additional contribution in this development.
17:00-17:20	ThC13.4
Design of Multi-Objective Control Systems with Optimal Failure Tolerance, pp. 2963-2968
Mesbah, Ali	Massachusetts Inst. of Tech.
Braatz, Richard D.	Massachusetts Inst. of Tech.
The selection of the control structure, which is a specification of the interconnection of measurements, exogenous inputs, and manipulated variables, is a critical step in the design of a control system. This paper presents a general internal model control structure with multiple degrees-of-freedom in which each controller is independently designed. The control system is shown to remain optimal when controllers are taken off-line due to component failures, without requiring re-design of any of the on-line controllers. The optimality of the proposed approach is demonstrated for the control of a simulated thin-film process for a variety of component failures.
17:20-17:40	ThC13.5
Dynamic Modeling and Control of Power-Split HEV with Multi-Phase Electric Machines under Fault Condition, pp. 2969-2975
Grossi, Federica	Univ. of Modena and Reggio Emilia
Fei, Marco	Univ. of Modena and Reggio Emilia
Zanasi, Roberto	Univ. of Modena and Reggio Emilia
The aim of this paper is to provide a dynamic model of the Toyota Hybrid System (THS) for simulation and control purposes in the case of fault of one of the electric machines. The two three-phase electric machines that are commonly used in THS are here replaced by two five-phase machines. The model of the whole system is realized by using the Power-Oriented Graphs modeling technique and includes the dynamics of the engine, electric machines, planetary gear, transmission and vehicle. A rule-based control strategy is used to operate the vehicle in different operation modes and a fault-tolerant control is applied in the case of electrical machine failure. Simulation results are given in both healthy and fault condition to show the effectiveness of the dynamic model and the robustness of the proposed control.
17:40-18:00	ThC13.6
Robust Fault Tolerant Application for HVAC System Based on Combination of Online SVM and ANN Black Box Model, pp. 2976-2981
Dehestani, Davood	Univ. of Tech. Sydney (UTS)
Su, Steven	Univ. of Tech. Sydney (UTS)
Ying, Guo	CSIRO
Nguyen, Hung	Univ. of Tech. Sydney (UTS)
Efficient heat, ventilating and air-conditioning (HVAC) systems is one of the big challenges today around the world. The fault detection and isolation (FDI) play a significant role in the monitoring, repairing and maintaining of technical systems for the final destination of cost reduction. FDI makes it possible to reduce total cost effective of maintenance and thus increase the capacity utilization rates of equipment. Reduction of energy wasting in the system by on time fault detection is another goal. Therefore, this work proposes a new fault detector based on a black box Neural Network (ANN) model and online support vector machines (SVM) classifier which integrates a dimension reduction scheme to analyze the failure of air fan supply and dampers fault. The key advantage of this algorithm is to make robustness for SVM to recognize a faulty condition with unexpected sensors values. The NN generate a high accurate model which is based reference for SVM classifier. Now by using this black box model we make possibility of robustness for SVM to increase detection probability. Finally, a series of faulty experimental data are applied to evaluate the effectiveness of the robust classifier. Final results show that online SVM can detect accurately the air supply fan fault and damper fault of a HVAC system with minimum usage data. It is also outperforms offline SVM on such energy systems for classification.

ThC14	HG D7.1
Maritime Systems (Regular Session)
Chair: Picci, Giorgio	Univ. di Padova
Co-Chair: Guerreiro, Bruno J. N.	Inst. Superior Tecnico
16:00-16:20	ThC14.1
GAS Tightly Coupled LBL/USBL Position and Velocity Filter for Underwater Vehicles, pp. 2982-2987
Batista, Pedro	Inst. Superior Técnico
Silvestre, Carlos	Inst. Superior Tecnico
Oliveira, Paulo Jorge	Inst. Superior Técnico
This paper presents a novel navigation filter for estimation of linear motion quantities based on a combined Long Baseline / Ultra Short Baseline (LBL/USBL) acoustic positioning system with application to underwater vehicles. The filtering algorithm does not resort to any algebraic inversion techniques and no linearizations are carried out whatsoever. In this way, the nonlinear sensor-based system dynamics are considered to their full extent and globally asymptotically stable (GAS) error dynamics are achieved. Finally, it is shown, under simulation environment, that the filter achieves very good performance in the presence of sensor noise.
16:20-16:40	ThC14.2
GES Tightly Coupled Attitude Estimation Based on a LBL/USBL Positioning System, pp. 2988-2993
Batista, Pedro	Inst. Superior Técnico
Silvestre, Carlos	Inst. Superior Tecnico
Oliveira, Paulo Jorge	Inst. Superior Técnico
Typical attitude estimation solutions for underwater vehicles rely on magnetometers, which are prone to magnetic field distortions. This can preclude its use in intervention scenarios, in the vicinity of objects with strong magnetic signatures, severely endangering not only the intervention mission but also the operation of the underwater vehicle. This paper presents a novel attitude estimation solution, based on a combined Long Baseline / Ultra Short Baseline (LBL/USBL) acoustic positioning system, with application to underwater vehicles. The range and range differences of arrival obtained with the LBL/USBL are directly embedded in the estimator dynamics, without any linearization whatsoever, and globally exponentially stable (GES) error dynamics are achieved. Simulation results evidence good performance of the proposed solution.
16:40-17:00	ThC14.3
System Identification for Tide Prediction in the Venetian Lagoon, pp. 2994-2999
Parise, Francesca	Automatic Control Lab. ETH Swiss Federal Inst. of Tech.
Picci, Giorgio	Univ. di Padova
An assessment of a statistical model currently used for the prediction of high tides in the Venetian lagoon is presented. The model is analyzed from several points of view and is compared with state-space descriptions of smaller order which seem to provide slightly better results. Moreover the relevance of additional external inputs to the model, that is additional meteorological agents, is discussed.
17:00-17:20	ThC14.4
A Study of Ship's Mooring Method with Controllable Pitch Propeller (CPP) by Applying Generalized Minimum Variance Control, pp. 3000-3005
Doi, Masayoshi	Hiroshima Inst. of Tech.
Nagamoto, Kazuhisa	Yuge National Coll. of Maritime Tech.
Takehira, Tetsuya	Japan Air Self-Defense Foorce
The purpose of this study is to improve a mooring system with Controllable Pitch Propeller (CPP). Because of the slip of propeller, velocity response of CPP has time delay. In this study, the ship's propelling force with the CPP's angle was identified on an actual ship and delay time was observed by analyzing the data of propelling force. Then the CPP angle is controlled by using Generalized Minimum Variance Control (GMVC) to obtain better performance for the mooring system.
17:20-17:40	ThC14.5
Trajectory Tracking Nonlinear Model Predictive Control for Autonomous Surface Craft, pp. 3006-3011
Guerreiro, Bruno J. N.	Inst. Superior Tecnico
Silvestre, Carlos	Inst. Superior Tecnico
Cunha, Rita	Inst. Superior Técnico
Pascoal, Antonio Manuel	Inst. Superior Tecnico
This paper presents a solution to the problem of trajectory tracking control for autonomous surface craft (ASC) in the presence of ocean currents. The proposed solution is rooted in nonlinear model predictive control (NMPC) techniques and addresses explicitly state and input constraints. Whereas state saturation constraints are added to the underlying optimization cost functional as penalties, input saturation constraints are made intrinsic to the nonlinear model used in the optimization problem, thus reducing the computational burden of the resulting NMPC algorithm. Simulation results, obtained with a nonlinear dynamic model of a prototype ASC, show that the NMPC strategy adopted yields good performance in the presence of constant currents. Experimental results are also provided to validate the real-time implementation of the proposed techniques.
17:40-18:00	ThC14.6
Modeling, Identification and Control of a Boat Parking Assistance System, pp. 3012-3017
Berretta, Daniele	Pol. di Milano
Urbano, Nazario	Pol. di Milano
Formentin, Simone	Pol. di Milano
Boniolo, Ivo	Pol. di Milano
De Filippi, Pierpaolo	Pol. di Milano
Savaresi, Sergio M.	Pol. Di Milano
In this paper, the problem of designing a Boat Parking Assistance (BPA) system for a small-scale vessel is addressed. A control-oriented model is derived from the physics underlying the system and gray-box identification is carried out by designing suitable experiments. A 3 DOF cascade control scheme is then implemented to achieve semi-automatic parking and station-keeping. The latter is also shown to be a solid groundwork for future research on fully automatic maneuver- ing. The proposed strategy is finally tested on stationkeeping and parking of a real vessel.

ThC15	HG D7.2
Control of Electrical Networks (Regular Session)
Chair: Rantzer, Anders	Lund Univ.
Co-Chair: Damm, Gilney	Lab. IBISC - CNRS/Evry Univ.
16:00-16:20	ThC15.1
A Decomposition Approach to Multi-Region Optimal Power Flow in Electricity Networks, pp. 3018-3024
Chasparis, Georgios	Lund Univ.
Rantzer, Anders	Lund Univ.
Jörnsten, Kurt	Norwegian School of Ec.
We present a decomposition approach to a class of social welfare optimization problems for optimal power flow in multi-region electricity networks. The electricity network is decomposed into multiple regions which decide independently over the amount of power produced within the region and exchanged with neighboring regions. We decompose the overall power flow (or social welfare) optimization into region-based optimization problems (namely, power flow game), which is based on the introduction of dual variables representing nodal and link prices. Due to the interdependencies between regions' utilities, the social welfare maximizer might not necessarily correspond to a Nash equilibrium of the power flow game. We derive conditions under which the social welfare maximizer is a Nash equilibrium of the game, and investigate uniqueness of Nash equilibria. Finally, we examine whether convergence to the social welfare maximizer may occur under natural best-response dynamics.
16:20-16:40	ThC15.2
A Statistically Robust Payment Sharing Mechanism for an Aggregate of Renewable Energy Producers, pp. 3025-3031
Nayyar, Ashutosh	Univ. of California, Berkeley
Poolla, Kameshwar	Univ. of California at Berkeley
Varaiya, Pravin P.	Univ. of California at Berkeley
Variability of supply is a fundamental difficulty associated with renewable resources in the electricity market. One way of mitigating this difficulty is to aggregate a diverse collection of resources in order to exploit the negative correlations that may exist among them. We consider a aggregation scheme where individual renewable energy producers offers day-ahead contracts to an aggregate manager which in turn participates in a two stage electricity market. The net payments received by the aggregate manager from the market have to be fairly distributed among the participants in the aggregate. Since the actual power supplied by the aggregate is random, its net payments are also random. The problem of sharing these random payments is complicated by the fact that different participants may have different statistical models for the payments because they have different statistical models for their (and other producers') net generation. We propose a simple payment sharing mechanism that is independent of the statistical models of the participants. We show that our payment sharing mechanism ensures that individual producers are better off in the aggregate than on their own. Further, under certain conditions, aggregation provides the social benefit of increasing the amount of renewable energy available in the day-ahead market.
16:40-17:00	ThC15.3
Dynamic Pricing in Consolidated Ancillary Service Markets, pp. 3032-3037
Taylor, Joshua	Univ. of Toronto
Nayyar, Ashutosh	Univ. of California, Berkeley
Callaway, Duncan	UC Berkeley
Poolla, Kameshwar	Univ. of California at Berkeley
Regulation is a costly aspect of power system operation, which has long been inadequately and inconsistently incentivized via ill-formulated pricing schemes, as recently acknowledged by regulators. We propose a pricing formulation in which regulation prices are the optimal dual multipliers or costate of an optimal control problem. More precisely, we use the linear quadratic regulator to formulate a regulation pricing policy. We then construct a Vickrey-Clarke-Groves mechanism to induce honest participation among selfish agents. We apply the formulation to a scenario combining traditional frequency regulation and the California Independent System Operator's Flexible Ramping Product.
17:00-17:20	ThC15.4
Risk-Limiting, Market-Based Power Dispatch and Pricing, pp. 3038-3045
Yo, Masaki	Keio Univ.
Ono, Masahiro	Keio Univ.
Williams, Brian	MIT
Adachi, Shuichi	Keio Univ.
In this paper, we propose a novel electricity dispatch and pricing mechanism that enables a grid with intermittent energy sources to dispatch energy within a user-specified risk bound in a decentralized manner. The key idea is to trade the standard deviation of future electricity supply and demand. A dispatchable power provider, such as a peaking power plant or a storage facility, adjusts its output in real time to reduce a certain amount of standard deviation of the future supply-demand balance. In the proposed market mechanism, such a dispatchable power provider can ``sell" the standard deviation at the price specified by a market. The market-clearing prices of the mean and the standard deviation of electricity are found through the Walrasian auction. This approach allows each power provider to specify the probability density function (pdf) of the amount of energy that it has to generate in the future. As a result, a power provider can quantitatively limit the risk of power shortage by imposing chance constraints in a decentralized manner. It can also evaluate the expected cost of future energy production as well as the cost of deviating its output by using the pdf. The decentralized risk-limiting dispatch and pricing problem are solved at each time step with a receding time horizon. We demonstrate the capabilities of the proposed approach by simulations using real data.
17:20-17:40	ThC15.5
Risk Limiting Dispatch with Optimal Curtailing in Active Distribution Networks, pp. 3046-3052
Georgiev, Daniel	Univ. of West Bohemia
Janecek, Eduard	Univ. of West Bohemia
Renewable energy sources often provide intermittent power at distributed locations in transmission and distribution networks. Efficient utilization of these sources must consider economics, computation, and reliability in managing network resources. A new computational tool for efficient dispatch of intermittent sources is developed using the risk limiting operational paradigm. Optimal dispatch of regulation and load following ancillary services is computed using current estimates of future random energy production. Substitution of intermittent power with firm power through a curtailment strategy is used to avoid regulation costs in excess of current firm power prices. The underlying mathematical framework is a non-convex optimal power flow problem, which is shown to have an exact convex relaxation under a set of realistic assumptions. The methodology is successfully tested on an IEEE 30-bus test system. Several operational effects of source uncertainty are captured. For instance, optimal solutions are shown to create power flows that mutually compensate intermittencies from different sources to minimize regulation requirements.
17:40-18:00	ThC15.6
Adaptive Control Scheme for Maximum Power Point Tracking of a Photovoltaic System Connected to the Grid, pp. 3053-3058
Jaramillo-Lopez, Fernando	Lab. des Signaux et Systemes, Supelec.
Damm, Gilney	Lab. IBISC - CNRS/Evry Univ.
Kenne, Godpromesse	Univ. de Dschang
Lamnabhi-Lagarrigue, Francoise	CNRS-EECI
An adaptive control scheme for maximum power point tracking of a single-phase single-stage photovoltaic system connected to the grid is presented. The maximum power point depends on temperature and solar irradiance, ambient conditions that are time-varying and difficult to measure. Two solutions are presented. Each solution derive an estimator that approximate three different parameters. These parameters are functions of solar irradiance and temperature. In this manner, we eliminate the necessity of climatic sensors. The first solution, uses an adaptive estimator that is able to estimate constant parameters, and the second one uses a sliding mode estimator that is capable of estimate time-varying parameters. A complete analysis was done taking into account the nonlinearities showed by the closed-loop system. The Lyapunov redesign technique was used to derive a controller that gives globally asymptotically stable trajectories of the closed-loop system. Computer simulations are presented to compare the performance of both estimators and also to show the good performance of the controller.

Technical Program for Friday July 19, 2013

FrPT1	HG F30, F1, F7
Plenary 3: Synthetic Biology: From Parts to Modules to Therapeutic Systems (Plenary Session)
Chair: Lygeros, John	ETH Zurich
08:00-09:00	FrPT1.1
Synthetic Biology: From Parts to Modules to Therapeutic Systems*
Weiss, Ron	M.I.T.

FrA1	HG F1
Systems and Synthetic Biology: Case Studies and Computational Challenges (Tutorial Session)
Chair: Koeppl, Heinz	ETH Zurich
Organizer: Koeppl, Heinz	ETH Zurich
09:20-10:00	FrA1.1
Tutorial on Synthetic Biology (I)*
Benenson, Yaakov	ETH Zurich
In this brief tutorial I will present an overview of the Synthetic Biology discipline. Research efforts in the field can be subdivided into three overlapping themes. First, there is an effort toward faster and cheaper synthesis of DNA sequences of ever increasing size. Related activities include new technologies for large-scale whole-genome mutagenesis in prokaryotes and efficient, rapid and high-throughput genome editing in human cells. These enabling technologies drive two additional thrusts: creating artificial genetic pathways for manufacturing of useful chemicals; and building functional pathways that augment, alter or "rewire" native processes in individual cells, organs, whole organisms and even ecosystems. Successful technology development does not imply automatic solution to the pathway engineering challenge. Indeed, the key question of synthetic biology is an inverse of an age-old problem in biology and it concerns designing the genotype for a pre-defined phenotype, such as a set of specifications for a desired metabolic or regulatory pathway. I will show a few examples where this challenge has been successfully addressed by researchers in the field.
10:00-10:40	FrA1.2
Applied Modeling and Theory for Stochastic Descriptions of Biochemical Networks (I)*
Khammash, Mustafa H.	Univ. of California at Sta. Barbara
Considerable experimental evidence has indicated that stochastic fluctuations are present in the cellular environment and has implicated them as a key source of cell-cell variability among genetically identical cells. Furthermore, studies of engineered genetic circuits such as toggle switches or oscillators, have revealed large stochastic effects. Stochasticity is therefore an inherent feature of biological dynamics, and as such, should be the subject of in-depth investigation and analysis. Such a study of stochastic properties in genetic systems involves the formulation of a correct representation of molecular noise, followed by the formulation of mathematically sound approximations for these representations. It also involves devising efficient computational algorithms capable of tackling the complexity of the dynamics involved. In this tutorial, we review a number of these techniques and provide compelling examples that illustrate the richness of phenomena that can result from the interaction of dynamics and noise in genetic networks. Specifically, we will explore the formulation of stochastic chemical kinetics leading to a general continuous time infinite discrete-state Markov Chain model of stochastic dynamics. For this model we introduce the Forward Kolmogorov Equation (Chemical Master Equation) and approaches for its solution. We also discuss sample path generation methods as well as approximate methods including SDE approximations and equations for describing statistical moments. Application of these methods to the parameter estimation and control of gene circuits will also be discussed.
10:40-11:20	FrA1.3
Differential Equation Models in Systems and Synthetic Biology (I)*
Craciun, Gheorghe	Univ. of Wisconsin-Madison
We discuss mathematical models of biochemical reaction networks based on systems of ordinary differential equations. In practice, these equations are almost always nonlinear and have many unknown parameters. We describe how we can relate functional properties of the biochemical network (like its capacity to function as a switch or an oscillator) with and properties of cycles in a graph associated to dynamical system, called the species-reactions graph. We also describe criteria for persistence and permanence of solutions (a positive solution is called persistent if no variable approaches zero) and for global convergence of solutions to an unique equilibrium, independent of initial conditions.

FrA2	HG F3
Embedded Optimization for Control and Estimation I – Numerical Method Development (Invited Session)
Chair: Zeilinger, Melanie N.	Ec. Pol. Fédéral de Lausanne (EPFL)
Co-Chair: Domahidi, Alexander	Automatic Control Lab. ETH Zurich
Organizer: Zeilinger, Melanie N.	UC Berkeley
09:20-09:40	FrA2.1
Stochastic Optimal Control in the Perspective of the Wiener Chaos (I), pp. 3059-3064
Huschto, Tony	Heidelberg Univ.
Sager, Sebastian	OVGU Magdeburg
We propose a novel and generic methodology for solving continuous finite-horizon stochastic optimal control problems. We develop innovative ideas for approximating controlled stochastic differential equations within the Wiener Chaos framework and expand them to reformulate stochastic optimal control problems directly into deterministic ones. Within our approach we present how to preserve the feedback character of the optimal Markov decision rules. This transformation allows the use of state-of-the-art methods of solving deterministic optimal control problems in the broad context of stochastic differential equations. We illustrate this new methodology by a numerical example and show the validity of the developed reformulations with huge computational savings compared to standard approaches to stochastic optimal control.
09:40-10:00	FrA2.2
Stabilizing Embedded MPC with Computational Complexity Guarantees (I), pp. 3065-3070
Rubagotti, Matteo	Nazarbayev Univ.
Patrinos, Panagiotis	IMT Inst. for Advanced Studies Lucca
Bemporad, Alberto	IMT Inst. for Advanced Studies Lucca
This paper describes a model predictive control (MPC) approach for discrete-time linear systems with hard constraints on control and state variables. The finite-horizon optimal control problem is formulated as a quadratic program (QP), and solved using a recently proposed dual fast gradient- projection method. More precisely, in a finite number of iterations of the mentioned optimization algorithm, a solution with bounded levels of infeasibility and suboptimality is determined for an alternative problem. This solution is shown to be a feasible suboptimal solution for the original problem, leading to exponential stability of the closed-loop system. The proposed strategy is particularly useful in embedded control applications, for which real-time constraints and limited computing resources can impose tight bounds on the possible number of iterations that can be performed within the scheduled sampling time.
10:00-10:20	FrA2.3
ECOS: An SOCP Solver for Embedded Systems (I), pp. 3071-3076
Domahidi, Alexander	Automatic Control Lab. ETH Zurich
Chu, Eric	Stanford Univ.
Boyd, Stephen P.	Stanford Univ.
In this paper, we describe the embedded conic solver (ECOS), an interior-point solver for second-order cone programming (SOCP) designed specifically for embedded applications. ECOS is written in low footprint, single-threaded, library-free ANSI-C and so runs on most embedded platforms. The main interior-point algorithm is a standard primal-dual Mehrotra predictor-corrector method with Nesterov-Todd scaling and self-dual embedding, with search directions found via a symmetric indefinite KKT system, chosen to allow stable factorization with a fixed pivoting order. The indefinite system is solved using Davis' SparseLDL package, which we modify by adding dynamic regularization and iterative refinement for stability and reliability, as is done in the CVXGEN code generation system, allowing us to avoid all numerical pivoting; the elimination ordering is found entirely symbolically. This keeps the solver simple, only 750 lines of code, with virtually no variation in run time. For small problems, ECOS is faster than most existing SOCP solvers; it is still competitive for medium-sized problems up to tens of thousands of variables.
10:20-10:40	FrA2.4
Fast Auto Generated ACADO Integrators and Application to MHE with Multi-Rate Measurements (I), pp. 3077-3082
Quirynen, Rien	KU Leuven
Gros, Sébastien	KU Leuven
Diehl, Moritz	KU Leuven
Algorithms for real-time, embedded optimization need to run within tight computational times, and preferably on embedded control hardware for which only limited computational power and memory is available. A computationally demanding step of these algorithms is the model simulation with sensitivity generation. This paper presents an implementation of code generation for Implicit Runge-Kutta (IRK) methods with efficient sensitivity generation, which outperforms other solvers for the targeted applications. The focus of this paper will be on the extension of the proposed tool to the integration of index-1 Differential Algebraic Equations (DAE), and continuous output functions, which are crucial for e.g. performing sensor fusion with measurements provided at very high sampling rates. The new tool is provided with a powerful MATLAB interface. It is illustrated in simulation for the trajectory estimation of a mechanical system modeled by complex Differential-Algebraic equations, using sensor information provided at fast, multi-rate sampling frequencies.
10:40-11:00	FrA2.5
Evaluation of Piecewise Affine Control Law Via Graph Traversal (I), pp. 3083-3088
Herceg, Martin	Swiss Federal Inst. of Tech. - ETH Zurich
Mariethoz, Sebastien	ETH Zurich
Morari, Manfred	ETH Zurich
This paper presents a constructive algorithm for the effective evaluation of the piecewise affine (PWA) control laws generated in explicit model predictive control (MPC). One of the main limitations of explicit MPC is the construction of the search trees that are used to efficiently compute the PWA control law in real-time. The key idea developed in the paper consists in replacing these search trees by graphs that are directly generated while solving the multiparametric problems. It is shown that if the parameter space is explored in the breadth-first search (BFS) fashion, the bounds for the worst case runtime implementation can be obtained after the construction of an explicit solution. The proposed method traverses the graph in linear time that corresponds to a breadth of the graph on the average. Effectiveness of the approach is demonstrated numerically on a power electronics benchmark.
11:00-11:20	FrA2.6
A Hierarchical Time-Splitting Approach for Solving Finite-Time Optimal Control Problems (I), pp. 3089-3094
Stathopoulos, Georgios	Ec. Pol. Federale de Lausanne (EPFL)
Keviczky, Tamas	Delft Univ. of Tech.
Wang, Yang	Stanford Univ.
We present a hierarchical computation approach for solving finite-time optimal control problems using operator splitting methods. The first split is performed over the time index and leads to as many subproblems as the length of the prediction horizon. Each subproblem is solved in parallel and further split into three by separating the objective from the equality and inequality constraints respectively, such that an analytic solution can be achieved for each subproblem. The proposed solution approach leads to a nested decomposition scheme, which is highly parallelizable. We present a numerical comparison with standard state-of-the-art solvers, and provide analytic solutions to several elements of the algorithm, which enhances its applicability in fast large-scale applications.

FrA3	HG F5
Biomedical Systems (Regular Session)
Chair: Raisch, Joerg	Tech. Univ. Berlin
Co-Chair: Banaei Khosroushahi, Reza	Univ. of Alberta
09:20-09:40	FrA3.1
Partial Stability of Controlled SEIR Epidemic Models, pp. 3095-3100
Ibeas, Asier	Univ. Autonoma de Barcelona
de la Sen, Manuel	Univ. del Pais Vasco
Alonso-Quesada, Santiago	UPV /EHU
Nistal Riobello, Raul	Ehu Univ. of Basque Country
This paper proposes to use the concept of partial stability instead of that of global stability to analyze the dynamic behavior of epidemic models. The rationale behind this is that it is only needed, from a practical point of view, to ensure the boundedness of the infectious and infected subpopulations in order to get the disease under control. Thus, the partial stability of a controlled SEIR epidemic model is studied in the vaccination-free case. It will be proved that the infectious and infected subpopulations may still be bounded when the total population increases and global stability is not achieved. In addition, a feedback-based vaccination policy is designed based on the concept of partial stability of the closed-loop system. Some numerical simulations illustrating the usefulness of the designed control law along with a comparison with previous vaccination laws complete the paper.
09:40-10:00	FrA3.2
Nonlinear Joint-Angle Feedback Control of Electrically Stimulated and λ -Controlled Antagonistic Muscle Pairs, pp. 3101-3107
Klauer, Christian	Tech. Univ. Berlin
Raisch, Joerg	Tech. Univ. Berlin
Schauer, Thomas	Tech. Univ. Berlin
In order to control human limb movements in neuroprosthetic systems, a nonlinear, model-based control strategy for torque generation by antagonistic muscle pairs is presented. The controller based on exact linearization methods enables the tracking of reference joint torque profiles and the generation of pre-defined muscular co-contractions. The controller adjusts the desired recruitment levels λ (lambda) of both muscles that are controlled by two underlying λ (lambda)-Controllers automatically compensating muscular fatigue. Estimates on λ are obtained from electrical stimulation evoked Electromyography (EMG)-measurements. Short-term increases of the muscular co-contractions are used to achieve an exact tracking of the accelerating joint torque also in the presence of actuations variable constraints which result from the fact that only positive muscular contractions can occur. The linearization-based controller can serve as underlying controller in cascaded control schemes for the joint-angle or velocity. The feasibility of the proposed approach was demonstrated in a simulation study as well as by a joint-angle control experiment for a healthy subject. Concluding, the authors expect that the proposed approach has a great potential for the future control of artificial limb movements also because of the possibility of modulating the system stiffness by controlled co-contractions of the muscles.
10:00-10:20	FrA3.3
Emulation of Ventricular Suction in a Hybrid Mock Circulation, pp. 3108-3112
Ochsner, Gregor	ETH Zurich
Amacher, Raffael	ETH Zurich
Schmid Daners, Marianne	ETH Zurich
Ventricular suction occurs when a turbodynamic ventricular assist device (VAD) tries to pump more blood from the ventricle than is available. Suction causes a stagnation of the blood flow and can damage the heart muscle and must therefore be avoided in clinical practice. A hybrid mock circulation is a test bench for VADs, which is based on a hardware-in-the-loop concept. It consists of a numerical model of the human blood circulation and a numerical-hydraulic interface, which allows interaction between the numerical model and a VAD to be tested. This paper shows how ventricular suction is implemented in the hybrid mock circulation to allow testing of suction detection algorithms and physiological VAD controllers. Experimental results show that suction can be emulated as desired.
10:20-10:40	FrA3.4
Reducing Domain Structural Complexity in PDE Backstepping Boundary Observer Design Using Conformal Mapping, pp. 3113-3118
Banaei Khosroushahi, Reza	Univ. of Alberta
Marquez, Horacio J.	Univ. of Alberta
In this paper we study backstepping boundary observer design for a class of distributed parameter systems defined in a cylindrical domain with prescribed boundary conditions. We show that applying standard PDE backstepping boundary observer design results in a hyperbolic PDE with four independent variables which is difficult to solve both numerically and analytically. By introducing a modified domain structure using conformal mapping, we obtain a simplified design process which is reduced to solving a two-dimensional PDE for the kernel function. Finally, we sketch the complete treatment to solve the partial differential equation describing the kernel equation and then perform a simulation study to verify the L² performance of the designed observer. This technique has a direct application in soft-sensor design for internal temperature measurement in systems subjected to dynamic heat distribution, particularly microfluidic Lab-on-a-Chip (LOC) devices.
10:40-11:00	FrA3.5
Close to Reality Evaluation of a PID Control Algorithm for Blood Glucose Regulation in Diabetic Goettingen Minipigs, pp. 3119-3124
Lunze, Katrin	RWTH Aachen Univ.
Zimmermann, Markus	RWTH Aachen Univ.
Walter, Marian	RWTH Aachen Univ. ,Helmholtz Inst.
Leonhardt, Steffen	RWTH Aachen Univ.
Diabetes mellitus is a widespread metabolic disease, which currently requires manual treatment. Until now, no commercial closed-loop insulin therapy system has been presented yet due to many control restrictions. To reduce safety risks for the patients, adequate control methods should be tested in animal trials, prior to clinical studies, which can be regarded as an intermediate step between in silico and in vivo control performance evaluation. In this paper, diabetic Goettingen minipigs serve as proxy for the human metabolism. After having developed a physiological and mathematical minipig model based on preceding animal trials, this paper describes a control algorithm which has been designed with the focus on real-life application. In particular, several control restrictions and requirements are taken into consideration such as delayed insulin effect on blood glucose concentration and fast response to blood glucose rise. The control robustness is evaluated with respect to three individual minipig models simulating three meals with different glucose content. It is shown that the controller responds to the disturbances similarly as a natural pancreas. Although blood glucose undershoots could not be completely avoided, no critical blood glucose drop appeared.
11:00-11:20	FrA3.6
Optimal Control of Influenza Epidemic Model with Virus Mutations, pp. 3125-3130
Gubar, Elena	St. Petersburg State Univ.
Zhu, Quanyan	Univ. of Illinois at Urbana Champaign
Strains of influenza viruses spread in human populations during every season of epidemics. As the infected population size increases, the virus can mutate itself and grow in its strength. The traditional epidemic SIR model does not capture the mutations of viruses, and hence the model is not sufficient to study epidemics where the virus mutate at the same timescale as the epidemic process. In this work, we establish a novel framework to study the epidemic process with mutations of influenza viruses, which couples the SIR model with replicator dynamics used to describe virus mutations. We formulate an optimal control problem to study the optimal strategies for medical treatment and quarantine. We obtain structural results for the optimal strategies and use numerical examples to illustrate our results.

FrA4	HG F7
Set Based Estimation and Control (Regular Session)
Chair: Olaru, Sorin	Supelec
Co-Chair: Raimondo, Davide Martino	Univ. of Pavia
09:20-09:40	FrA4.1
Active Fault Diagnosis Using Moving Horizon Input Design (I), pp. 3131-3136
Raimondo, Davide Martino	Univ. of Pavia
Braatz, Richard D.	Massachusetts Inst. of Tech.
Scott, Joseph K.	Massachusetts Inst. of Tech.
This article considers the design of an input signal for improving the diagnosability of faults from process measurements. Previous work has focused on open-loop input design. In particular, deterministic methods are available for computing an input that guarantees fault diagnosis within a specified time horizon, whenever such an input exists. Here, two closed-loop approaches are considered that use feedback in order to reduce the length and/or cost of the required input, while maintaining this guarantee. The first method uses an existing open-loop input design method within a receding horizon framework. The second method approximates the first by an explicit feedback law in order to reduce online computations.
09:40-10:00	FrA4.2
Analysis of Vehicle Actuators Based on Reachable Sets, pp. 3137-3142
Nemeth, Balazs	MTA SZTAKI
Gaspar, Peter	Computer & Automation Inst. of HAS
The paper examines the reachability characteristics of vehicle actuators in order to analyse their abilities for the entire vehicle system. A computation method for an outer approximation of a reachable set is proposed. Simulation experiments are used in order to formulate the shape of the reachable sets of actuators. The evaluation of the shape provides preliminary information in the approximation of the reachable set. The upper bounds of the reachable set are calculated based on the constrained LMI feasibility problem in the nonlinear optimization task. The aim of the analysis is to provide a theoretical basis for the coordination of the actuators. The method is illustrated through the influence of the steering and the brake control systems at various velocities and road conditions.
10:00-10:20	FrA4.3
Set-Membership State Estimation for Discrete Time Piecewise Affine Systems Using Zonotopes (I), pp. 3143-3148
Tabatabaeipour, Seyed Mojtaba	Tech. Univ. of Denmark
Stoustrup, Jakob	Aalborg Univ.
This paper presents a method for guaranteed state estimation of discrete time piecewise affine systems with unknown but bounded noise and disturbance. Using zonotopic set representations, the proposed method computes the set of states that are consistent with the model, observation, and bounds on the noise and disturbance such that the real state of the system is guaranteed to lie in this set. Because in piecewise affine systems, the state space is partitioned into a number of polyhedral sets, at each iteration the intersection of the zonotopes containing a set-valued estimation of the states with each of the polyhedral partitions must be computed. We use an analytic method to compute the intersection as a zonotope and minimize the size of the intersection. A numerical example is provided to illuminate the algorithm.
10:20-10:40	FrA4.4
Consensus Control Protocols for Nonlinear Dynamical Systems Via Hybrid Stabilization of Sets, pp. 3149-3154
Haddad, Wassim M.	Georgia Inst. of Tech.
Nersesov, Sergey G.	Villanova Univ.
Ghasemi, Masood	Villanova Univ.
In this paper, we develop a hybrid control framework for addressing multiagent consensus control protocols for general nonlinear dynamical systems using stabilization of sets. The proposed framework develops a novel class of fixed-order, energy-based hybrid controllers that combine a logical switching architecture with the continuous system dynamics to guarantee that a system generalized energy function whose zero level set characterizes a specified system formation is strictly decreasing across switchings. The proposed approach addresses general nonlinear dynamical systems and is not limited to systems involving single integrator dynamics for consensus control.
10:40-11:00	FrA4.5
Hyperplane Arrangements in Mixed-Integer Programming Techniques. Collision Avoidance Application with Zonotopic Sets (I), pp. 3155-3160
Stoican, Florin	Norwegian Univ. of Science and Tech. (NTNU)
Prodan, Ionela	SUPELEC
Olaru, Sorin	Supelec
The current paper addresses the problem of minimizing the computational complexity of optimization problems with non-convex and possibly non-connected feasible region of polyhedral type. Using hyperplane arrangements and Mixed-Integer Programming we provide an efficient description of the feasible region in the solution space. Moreover, we exploit the geometric properties of the hyperplane arrangements and adapt this description in order to provide an efficient solution of the mixed-integer optimization problem. Further, a zonotopic representation of the sets appearing in the problem is considered, The advantages of this representation are highlighted and exploited.
11:00-11:20	FrA4.6
An Interpolation-Based Robust MPC Algorithm Using Polyhedral Invariant Sets, pp. 3161-3166
Bumroongsri, Pornchai	Department of Chemical Engineering, Faculty of Engineering, Chul
Kheawhom, Soorathep	Chulalongkorn Univ.
This article presents an interpolation-based robust MPC algorithm for uncertain polytopic discrete-time systems using polyhedral invariant sets. Two nested polyhedral invariant sets are constructed off-line by solving robust constrained model predictive control optimization problems. The first one is a large set constructed to cover all of the desired operating spaces. The second one is a small target set constructed to drive the terminal state into. The real-time control law is calculated by linear interpolation between the two state feedback gains corresponding to these nested pre-computed polyhedral invariant sets. At each sampling instant, only a computationally low-demanding optimization problem is needed to be solved on-line. The controller design is illustrated with an example. The proposed algorithm can achieve good control performance while on-line computation is still tractable.

FrA5	HG E1.1
Switched Systems I (Regular Session)
Chair: Millerioux, Gilles	Henri Poincare Univ. of Nancy
Co-Chair: Rodrigues, Luis	Concordia Univ.
09:20-09:40	FrA5.1
Scalable Decay Factor and ISS Gain for Disturbed Linear Polytopic Discrete-Time Systems, pp. 3167-3172
Millerioux, Gilles	Univ. de Lorraine
Bloch, Gerard	CRAN
In this paper, Input-to-State Stability conditions are proposed for disturbed linear polytopic discrete-time systems. The conditions allow to optimize, independently in a certain extent, both the bounds on the decay rate and on the ISS gain, two central quantites. Indeed, the decay factor characterizes the transient behaviour of the state while the ISS gain characterizes the sensitivity with respect to the disturbances. The conditions are expressed in terms of tractable Matrix Inequalities. It is shown that the conditions are more general than existing ones proposed so far in the literature and thus less conservative. The conditions hold both for analysis or synthesis purposes. Two illustrative examples addressing the problem of polytopic observer design are given.
09:40-10:00	FrA5.2
Measures and LMIs for Optimal Control of Piecewise-Affine Systems, pp. 3173-3178
Abdalmoaty, Mohamed Rasheed-Hilmy	KTH
Henrion, Didier	Czech Tech. Univ. in Prague
Rodrigues, Luis	Concordia Univ.
This paper considers the class of deterministic continuous-time optimal control problems (OCPs) with piecewise-affine (PWA) vector field, polynomial Lagrangian and semialgebraic input and state constraints. The OCP is first relaxed as an infinite-dimensional linear program (LP) over a space of occupation measures. This LP, a particular instance of the generalized moment problem, is then approached by an asymptotically converging hierarchy of linear matrix inequality (LMI) relaxations. The relaxed dual of the original LP returns a polynomial approximation of the value function that solves the Hamilton-Jacobi-Bellman (HJB) equation of the OCP. Based on this polynomial approximation, a suboptimal policy is developed to construct a state feedback in a sample-and-hold manner. The results show that the suboptimal policy succeeds in providing a stabilizing suboptimal state feedback law that drives the system relatively close to the optimal trajectories and respects the given constraints.
10:00-10:20	FrA5.3
Discretized Switching Time Optimization Problems, pp. 3179-3184
Flaßkamp, Kathrin	Univ. of Paderborn
Murphey, Todd	Northwestern Univ.
Ober-Blöbaum, Sina	Univ. of Paderborn
A switched system is defined by a family of vector fields together with a switching law which chooses the active vector field at any time. Thus, the switching law encoding the switching times and the sequence of modes may serve as a design parameter. Switching time optimization (STO) focuses on the optimization of the switching times in order to govern the system evolution to a desired behavior described by some cost function. However, it is rare that a STO problem can be solved analytically leading to the use of numerical approximation methods. In this contribution, we directly start with applying integration schemes to approximate the system's state and adjoint trajectories and study the effect of this discretization. It turns out that in contrast to the continuous time problem, the discretized problem loses differentiability with respect to the optimization variables. The isolated nondifferentiable points can be precisely identified though. Nevertheless, to solve the STO problem, nonsmooth optimization techniques have to be applied which we illustrate using a hybrid double pendulum.
10:20-10:40	FrA5.4
Stability of Switching Systems and Generalized Joint Spectral Radius, pp. 3185-3190
Ogura, Masaki	Texas Tech. Univ.
Martin, Clyde F.	Texas Tech. Univ.
This paper studies the mean stability of stochastic switching linear systems. We first show that the mean stability is characterized by an extended version of so called generalized joint spectral radius. Then it is shown that, under an invariance condition, the quantity can be computed as the spectral radius of a certain matrix associated with the given switching system. Also we show that the mean square stability is equivalent to the existence of a Lyapunov function. Our results are illustrated by numerical examples.
10:40-11:00	FrA5.5
Convex Optimization Methods for Computing the Lyapunov Exponent of Matrices, pp. 3191-3196
Protasov, Vladimir	Moscow State Univ.
Jungers, Raphaël	Univ. catholique de Louvain
We introduce a new approach to evaluate the largest Lyapunov exponent of a family of matrices, which describes the stability with probability one of a randomly switching linear system. For positive systems, of particular importance in systems and control, the rate of convergence of our approximation is estimated and the efficiency of the algorithm is demonstrated on particular switching systems of different dimensions. This is done by introducing new upper and lower bounds for the largest Lyapunov exponent of nonnegative matrices. We generalize this approach to arbitrary systems (not necessarily positive), derive a new universal upper bound for the Lyapunov exponent, and show that a similar lower bound, in general, does not exist.
11:00-11:20	FrA5.6
Generic Fractal Structure of the Optimal Synthesis in Problems with Affine Multi-Dimensional Control, pp. 3197-3202
Hildebrand, Roland	Univ. Joseph Fourier / CNRS
Lokutsievskiy, Lev	Moscow State Univ.
Zelikin, Michail	Moscow State (Lomonosov) Univ.
We consider a linear-quadratic deterministic optimal control problem where the control takes values in a two-dimensional simplex. The phase portrait of the optimal synthesis contains second-order singular extremals and exhibits modes of infinite accumulations of switchings in finite time, so-called chattering. We prove the presence of an entirely new phenomenon, namely a chaotic behaviour of the set of optimal trajectories. The set of optimal non-wandering trajectories has the structure of a Cantor set, and the dynamics of the system is described by a topological Markov chain. We compute the entropy and the Hausdorff dimension of the non-wandering set. This behaviour is generic for piece-wise smooth Hamiltonian systems in the vicinity of a junction of three discontinuity hypersurface strata.

FrA6	HG E1.2
Sliding Mode Control I (Regular Session)
Chair: Pisano, Alessandro	Univ. di Cagliari
Co-Chair: Boiko, Igor	Petroleum Inst.
09:20-09:40	FrA6.1
Improved Sliding Mode Control of a Class of Nonlinear Systems: Application to Quadruple Tanks System, pp. 3203-3208
Larguech, Samia	National School of Engineers of Sfax
Aloui, Sinda	Univ. of Picardie Jules Verne, France
Chaari, Abdessattar	National Engineering School of Sfax
El Hajjaji, Ahmed	Univ. de Picardie-Jules Verne
Koubaa, Yassine	National Engineering School of Sfax
In this paper, a stable adaptive sliding mode based on tracking control is developed for a class of nonlinear Multi Input Multi Output (MIMO) systems with external disturbances. In order to reduce the chattering phenomenon without deteriorating the tracking performance, the discontinuous term in the classical sliding mode technique is replaced by an adaptive Proportional Derivative (PD) term. The effect of the approximation error which arises from the PD term is reduced by adding a robust term in the proposed control law. All parameter, adaptive laws and the robust control term, are derived based on the Lyapunov stability analysis. The overall adaptive sliding mode scheme guarantees the asymptotic convergence to zero of tracking errors and the boundedness of all signals in the closed-loop system. The proposed approach is applied to a quadruple tanks system and achieves satisfactory simulation results.
09:40-10:00	FrA6.2
Dead-Beat Sliding Mode Control of Perishable Inventories with Transportation Losses and Multiple Suppliers, pp. 3209-3215
Bartoszewicz, Andrzej	Tech. Univ. of Lodz
Maciejewski, Michal	Inst. of Automatic Control, Tech. Univ. of Lodz
In this paper a new dead-beat sliding mode warehouse management strategy is proposed. The strategy ensures full customers' demand satisfaction and prevents from exceeding the warehouse capacity. Furthermore, it ensures that the ordered quantities of goods are always non-negative and upper bounded. The proposed strategy not only explicitly takes into account decay of goods stored in the warehouse (perishing inventories) but it also accounts for transportation losses which take place on the way from suppliers to the warehouse.
10:00-10:20	FrA6.3
Sliding-Mode Control of a Hydrostatic Drive Train with Uncertain Actuator Dynamics, pp. 3216-3221
Sun, Hao	Rostock Univ.
Aschemann, Harald	Univ. of Rostock
A sliding-mode approach with disturbance compensation is proposed in this paper for the tracking control of a hydrostatic drive train, which is commonly used in off-road vehicles. A control-oriented modelling results in a system of four nonlinear differential equations which are subject to uncertain actuator dynamics with saturation effects and unknown disturbances -- a leakage volume flow and a disturbance torque acting on the hydraulic motor. The disturbances are estimated by a nonlinear reduced-order disturbance observer and used for a disturbance compensation. Thereby, the switching part in the sliding-mode control law and chattering phenomena can be reduced. Simulation results point out that the proposed approach leads to an excellent tracking performance despite the given uncertainties in the actuator dynamics.
10:20-10:40	FrA6.4
Analysis of Performance of a Liquid Level Process Controlled by the Super-Twisting Algorithm, pp. 3222-3227
Boiko, Igor	Petroleum Inst.
Al Ameri, Osama	The Petroleum Inst.
Second-order sliding mode control algorithms are a popular research area. In the present research, a study of performance of the Super-Twisting (STW) controller is done, with application to a liquid level process. Comparison of its performance with the performance of an optimally tuned PID controller is done. The study is carried out through analytical approach, simulations and experiments. From the comparison of the performance, conclusions are made about suitability and perspectives of the use of the STW controller for the considered process.
10:40-11:00	FrA6.5
An Adaptive Version of a Second Order Sliding Mode Output Feedback Controller, pp. 3228-3233
Estrada, Antonio	Inria
Plestan, Franck	Ec. Centrale De Nantes-CNRS
Allouche, Benyamine	Ec.
This paper proposes a novel adaptive-gain second order sliding mode controller. The proposed controller is based on discontinous-impulsive action and only uses information of the output, no differentiation is needed. A practical second order sliding mode is ensured in spite of bounded perturbations with unknown bound. The gain adaptation is based on particular conditions attainable on sliding mode which can be verified with the only information of the output. Both, the convergence conditions and adaptation proposal are tested through simulations.
11:00-11:20	FrA6.6
A New Approach to Causal Output Tracking for Non-Minimum Phase Nonlinear Systems Via Combined First/second Order Sliding Mode Control, pp. 3234-3239
Pisano, Alessandro	Univ. di Cagliari
Baev, Simon	Department of Computer Science, Georgia Southwestern State Univ.
Salimbeni, Domenico	Univ. of Cagliari, Dept of Electrical and Electronic Engine
Shtessel, Yuri B.	Univ. of Alabama at Huntsville
Usai, Elio	Univ. degli Studi di Cagliari
This paper deals with the output-tracking control problem for a class of non minimum phase nonlinear systems. In order to provide the desired output tracking features while, at the same time, stabilizing the internal dynamics in finite time, a novel cascade-like control structure is devised, which mixes together first and second order sliding mode methodologies. The scheme provides asymptotic convergence, under full state feedback, bringing the main benefit, as compared with previous literature, that the sliding mode dynamics of the output tracking error variable is of lower dimension, which simplifies the tuning of the scheme and gives rise to improved transient features of the error variables. Theoretical analysis and simulation results support the effectiveness of the proposed solution

FrA7	HG E3
Applications of Robust Control (Regular Session)
Chair: Zenger, Kai	Aalto Univ. School of Electrical Engineering
Co-Chair: Richards, Arthur	Univ. of Bristol
09:20-09:40	FrA7.1
Nonlinear and Asymmetric Thermal-Aware DVFS Control, pp. 3240-3245
Durand, Sylvain	CNRS - CINVESTAV
Lesecq, Suzanne	CEA
Embedded devices using highly integrated circuits must cope with conflicting constraints. They have become more sensitive to variabilities with technology scaling while requiring computational efficiency under even more limited energy storage. Power management techniques, mainly based on Dynamic Voltage and Frequency Scaling (DVFS) algorithms, are hence of great interest. Also, temperature increase/decrease is directly related to the power consumption and platform characteristics highly depend on the temperature. As a result, temperature must be controlled, at least limited. In the present paper, the nonlinearity between power and temperature is analyzed and a thermal-aware DVFS control technique is discussed. The proposed strategy implements a chopped scheme on top of a robust DVFS approach in order to limit the temperature increase. The nonlinearity notably leads to an asymmetry in the temperature behavior, which makes the final attained temperature is higher than in the linear case.
09:40-10:00	FrA7.2
Disturbance Feedback for Handling Uncertainty in Air Traffic Flow Management, pp. 3246-3251
Clare, Gillian	Univ. of Bristol
Richards, Arthur	Univ. of Bristol
This paper presents the novel application of disturbance feedback optimization techniques for uncertainty management in Air Traffic Flow Management (ATFM) problems. The efficiency of ATFM optimizations in preventing local demand-capacity imbalances is reliant on predictions of future capacity states. However these predictions are inherently uncertain due to factors such as weather effects and unscheduled demand. A pre-existing ATFM flow based model is augmented to include feedback on the disturbances which perturb the weather scenario away from the nominal. Two formulations for modelling the feedback disturbance signal are explored. Results are presented demonstrating the benefits, in terms of reduced delays, of incorporating feedback on the problem solutions over single-solution approaches. Some initial studies of the relative computational scaling properties are also presented, demonstrating that taking advantage, within the formulation, of linearly related scenarios can yield computational advantages. Directions for further computational improvement are also discussed.
10:00-10:20	FrA7.3
A Semi-Active Controller Tuning and Application to Base Seismically-Isolated Structures, pp. 3252-3257
Teodorescu, Catalin-Stefan	CEA Saclay
Diop, Sette	CNRS
Politopoulos, Ioannis	CEA Saclay
Benidir, Messaoud	L2S-Supelec-UPS-11
This paper proposes a modified version of Leitmann and co-authors' classical result on the stabilization of uncertain nonlinear systems. In particular, for usual models of structure dynamics in earthquake engineering it is shown that applying a specific control law drives the state variables into a ball around the origin (arbitrarily chosen) in finite time as long as the radius of the ball is not lower than a limiting value. In addition estimates of this limiting ball radius and the time limit for arbitrary ball radius are provided. The semi-active control thus provides the control designer with interesting design parameters. It is also an attempt to explicitly use pseudoacceleration floor response spectrum as a performance criterion. Though not limited to two-degree-of-freedom structures, this semi-active control is applied to these plant models for simplicity and illustrated through simulations.
10:20-10:40	FrA7.4
Rotorcraft MAV Having an Onboard Manipulator: Longitudinal Modeling and Robust Control, pp. 3258-3263
Escareno Castro, Juan Antonio	FEMTO-ST Inst.
Rakotondrabe, Micky	FEMTO-ST Inst. UMR CNRS - UFC / ENSMM / UTBM,
Flores Colunga, Gerardo Ramon	Univ. of Tech. of Compiègne
Lozano, Rogelio	Univ. de Tech. de Compiegne
The paper presents results on the modeling and control of a multi-body air vehicle concept that incorporates a robotic arm to a miniatute rotorcraft robot. One of our goals was to obtain a detailed model in order to identify and qualify inherent constraints provided by the interaction of both mechanical subsystems.To do this, we used the Euler- Lagrange formulation. The knowledge of the model of the interactive aerial robot allowed us to propose a hierarchical control scheme, where a nonlinear Sliding Mode Controller (SMC) handle the translational mo- tion while the rotational motion is addressed by a H∞ - based controller. Numerical simulations were carried out to validate the overal control approach.
10:40-11:00	FrA7.5
An Instability Condition for Uncertain Systems Toward Robust Bifurcation Analysis, pp. 3264-3269
Inoue, Masaki	Tokyo Inst. of Tech.
Imura, Jun-ichi	Tokyo Inst. of Tech.
Kashima, Kenji	Osaka Univ.
Arai, Takayuki	Tokyo Inst. of Tech.
Aihara, Kazuyuki	The Univ. of Tokyo
In this paper, we consider instability analysis of uncertain feedback systems. First, we present an instability counterpart of the small gain stability theorem. Based on the instability theorem, we solve the problem of instability analysis for systems with dynamic uncertainties. Then, we propose a novel concept of robust bifurcation analysis. An illustrative example is presented for bifurcation analysis of uncertain genetic network models.
11:00-11:20	FrA7.6
Nonsmooth Optimization Based Multiple Robust Controller Design under Coupling Partitioned Uncertainty, pp. 3270-3275
Liu, Jiaqi	Aalto Univ.
Gao, Xiaozhi	Aalto Univ. of Automation and Systems Tech.
Zenger, Kai	Aalto Univ. School of Electrical Engineering
This paper proposes a nonsmooth optimization techniques to solve Hinf synthesis problems under weak coupling uncertainty. For some kind of systems, a single controller cannot be found in terms of given robust performance objective, whereas the multiple robust controller design can satisfy the desired performance by introducing extra controllers and by dividing the entire uncertainty into small partitioned sets. Such partitioned sets can be generated based on small gain theorem and u-analysis. Besides for solving weak-coupled parametric uncertainty, the partition with decoupled algorithm can achieves the aims of accelerating the computing speed and improving the computing efficiency. In that case, the multiple robust controller design has to be transformed into a nonsmooth optimization problem. This paper chooses the Clarke subdifferential and steepest descent method to compute the gradient of the nonsmooth function. The validity of this method is illustrated by a realistic controller design for a satellite attitude control.

FrA8	HG E5
Distributed and Coordinated Control (Regular Session)
Chair: Stursberg, Olaf	Univ. of Kassel
Co-Chair: Cantoni, Michael	Univ. of Melbourne
09:20-09:40	FrA8.1
Distributed Model Predictive Control Techniques Applied to an Irrigation Canal (I), pp. 3276-3281
Alvarez, Antonio	Univ. of Seville
Ridao, Miguel A.	Univ. de Sevilla
Ramirez, Daniel R.	Univ. of Sevilla
Sanchez, Laura	AECOM
This paper presents the application of a Distributed Model Predictive Controller (DMPC) to the control of an accurate model of an irrigation canal in Spain. The canal is modelled using the Saint-Venant equations and implemented using the well known modelling software for irrigation canals SIC. The DMPC algorithm has been implemented in Matlab and interfaced to SIC. In the distributed control algorithms, the local controllers exchange information so that their control policies are optimal in the sense of getting the best values of a performance index. The results show that the proposed distributed control algorithm obtains better control performance than a more conventional decentralized control scheme without information exchange. This better performance translates directly into money and resource savings.
09:40-10:00	FrA8.2
A Technique for Designing Stabilizing Distributed Controllers with Arbitrary Signal Structure Constraints, pp. 3282-3287
Rai, Anurag	Massachusetts Inst. of Tech.
Warnick, Sean	Brigham Young Univ.
This paper presents a new approach to distributed controller design that exploits a partial-structure representation of linear time invariant systems to characterize the structure of a system. This partial-structure representation, called the dynamical structure function, characterizes the signal structure, or open-loop causal dependencies among manifest variables, capturing a significantly richer notion of structure than the sparsity pattern of the transfer function. The design technique sequentially constructs each link in an arbitrary controller signal structure, and the main result proves that the resulting controller is either stabilizing or no controller with the desired structure can stabilize the system.
10:00-10:20	FrA8.3
Design of Distributed Controllers and Communication Topologies Considering Link Failures (I), pp. 3288-3294
Groß, Dominic	Univ. of Kassel
Jilg, Martin	Univ. of Kassel
Stursberg, Olaf	Univ. of Kassel
This paper presents results on optimization based design of distributed control laws for linear discrete-time systems interconnected through states, inputs, and a cost function. In the proposed scheme the subsystems may communicate state information, resulting in a trade-off between good closed-loop performance and high communication costs as well as robustness issues if communication links are unreliable. Previous results on optimization based synthesis of a communication topology and associated distributed control laws are extended to the case that (i) some communication links are prone to failure and that (ii) a controller may reconfigure itself if it is directly affected by a link failure. To include this scenario into the synthesis problem, a link failure is modeled as a change in the network topology and of the affected controllers. The problem is formulated as a mixed-integer semi-definite problem (MISDP) which combines the discrete optimization of the network topology subject to communication constraints and link failures with subproblems for structured controller synthesis.
10:20-10:40	FrA8.4
Distributed MPC Applied to Power Demand Side Control (I), pp. 3295-3300
Larsen, Gunn Kristine Holst	Univ. of Groningen
Pons, Jarno	Univ. of Groningen
Achterop, Sietse	Univ. of Groningen
Scherpen, Jacquelien M.A.	Univ. of Groningen
In the future, global energy balance of a smart grid system can be achieved by its agents deciding on their own power demand locally and the exchange of these decisions. In this paper, we model a network of households with washing machine programs that can be shifted in time so that the overall power demand is flattened. The network model describes how the information of power imbalance of individual agents can be exchanged in the system. Additionally, dynamics, washing machine constraints and power demand forecasts of each agent are included. Compared to existing smart grid models with hierarchical structures, our model, together with a market mechanism, achieves the power balance in the system in a completely distributed way. The market mechanism is a distributed MPC scheme based on dual decomposition and sub-gradient iterations. We provide results with a realistic power and washing machine demand pattern and we test scalability of the problem. Finally, we provide insights in the scalability of the algorithms.
10:40-11:00	FrA8.5
Formation Control Via Quasi-Time Optimal Protocol, pp. 3301-3305
Morozov, Yury	V.A. Trapeznikov Inst. of Control Sciences RAS
The problem of formation control is considered. The dynamics of an agent are described as a Lagrangian system without/with internal dynamic of controler. A quasi-time-optimal decentralized control law is proposed for synchronization of networked agents. Construction of the formation is to be completed in finite time using signum control protocols. During the motion, the formation follows a leader, real or virtual. Every agent determines relative position, velocity and acceleration of its neighbors, links between which are determinated by a time-stable directed tree. Theoretical results are illustrated by numerical examples.
11:00-11:20	FrA8.6
On Computing Quadratic Controls for Acyclic Networks of Heterogenous Systems, pp. 3306-3311
Shames, Iman	Univ. of Melbourne
Cantoni, Michael	Univ. of Melbourne
In this paper we consider the problem of computing finite-horizon quadratic optimal controls for a network of interconnected heterogeneous linear systems. We assume that the network is isomorphic to an acyclic directed graph and propose a solution to the problem via dynamic programming. This yields a finite sequence of smaller problems to be solved. Each smaller problem is associated with a node (or collection of such) in the network graph. It involves only local information and information from neighbours. As such, the solution algorithm lends itself to a distributed implementation. We numerically compare the performance of the proposed solution with computing the finite-horizon optimal controls by solving the global problem without exploiting structure.

FrA9	HG E7
Control of UAVs (Regular Session)
Chair: Aguiar, A. Pedro	Faculty of Engineering, Univ. of Porto (FEUP)
Co-Chair: Huck, Stephan Marc	ETH Zurich
09:20-09:40	FrA9.1
Health Aware Planning under Uncertainty for UAV Missions with Heterogeneous Teams, pp. 3312-3319
Ure, Nazim Kemal	Massachusetts Inst. of Tech.
Chowdhary, Girish	Massachusetts Inst. of Tech.
How, Jonathan P.	Massachusetts Inst. of Tech.
Vavrina, Matthew A.	Boeing Res. and Tech.
Vian, John	Boeing Res. and Tech.
In large-scale persistent missions, the vehicle capabilities and health often degrade over time. This paper presents a Health Aware Planning (HAP) Framework for long-duration complex UAV missions by establishing close feedback between the high-level planning based on Markov Decision Processes (MDP) and the execution level learning-focused adaptive controllers. This feedback enables the HAP framework to plan by anticipating the failures and reassessing vehicle capabilities after the failures. This proactive behavior allows for efficient replanning to account for changing capabilities. Simulations for a 4 UAV target tracking scenario is presented to demonstrate the effectiveness of the proactive replanning capability of the presented HAP framework.
09:40-10:00	FrA9.2
Moving Path Following for Autonomous Robotic Vehicles, pp. 3320-3325
Oliveira, Tiago	Acad. da Força Aérea Portuguesa
Encarnação, Pedro	Catholic Univ. of Portugal
Aguiar, A. Pedro	Faculty of Engineering, Univ. of Porto (FEUP)
This paper introduces the moving path following (MPF) problem for autonomous robotic vehicles, in which the vehicle is required to converge to and follow a desired geometric moving path, without a specific temporal specification. This case generalizes the classical path following problem, where the given path is stationary. Possible tasks that can be formulated as a MPF problem include terrain/air vehicles target tracking and gas clouds monitoring, where the velocity of the target/cloud specifies the motion of the path. Using the concept of parallel-transport frame associated to the geometric path, we derive the MPF kinematic-error dynamics for 3D paths with arbitrary motion specified by its linear and angular velocity. An application is made to the problem of tracking a target on the ground using an Unmanned Aerial Vehicle. The control law is derived using Lyapunov methods. Formal convergence results are provided, and numerical and hardware in the loop simulations demonstrate the effectiveness of the proposed method.
10:00-10:20	FrA9.3
A Multiplicative Filter for GLMAV Attitude Estimation, pp. 3326-3331
Grandvallet, Bertrand	CRAN / ISL
Zemouche, Ali	Nancy-Univ.
Boutayeb, M.	Lorraine Univ.
In this paper the authors investigate the problem of attitude estimation applied to the Gun Launched Micro Aerial Vehicle (GLMAV), using low cost sensors. The main contribution of the paper consists in the design of a sliding window filter for quaternion estimation. The proposed approach is considered in an EKF framework in order to ensure a better robustness to external perturbations. Finally, high performances of the filter are shown through numerical simulations.
10:20-10:40	FrA9.4
An Evaluation of UAV Path Following Algorithms, pp. 3332-3337
Pb, Sujit	Univ. of Porto
Saripalli, Srikanth	Arizona State Univ.
Sousa, Joao	Univ. do Porto - Faculdade Engenharia
Path following is the simplest desired autonomous navigation capability for UAVs. Several approaches have been developed in the literature for path following. However, there is no adequate information on the comparison analysis of these algorithms. In this paper, we compare five path following algorithms that are easy to implement, take less implementation time and are robust to disturbances. The comparison of these guidance laws is carried out through simulations using the total cross-track error and control effort performance metrics.
10:40-11:00	FrA9.5
On the Construction of Barrier in a Connectivity Maintenance Game, pp. 3338-3345
Bhattacharya, Sourabh	Iowa State Univ.
In this work, we address the problem of maintaining an operable communication link between two aerial vehicles in the presence of a jammer. In this work, we formulate the problem as a game of kind, and present the equations that characterize the barrier in the game using Isaacs' technique. We derive the optimal control of the vehicles on the barrier, and present simulation results for specific terminal conditions on the boundary of the usable part.
11:00-11:20	FrA9.6
Position Trajectory Tracking of a Quadrotor Helicopter Based on L1 Adaptive Control, pp. 3346-3353
De Monte, Paul	Tech. Univ. München
Lohmann, Boris	Tech. Univ. Muenchen
We present an adaptive backstepping controller for the position trajectory tracking of a quadrotor. The tracking controller is based on the L1 adaptive control approach and uses a typical nonlinear quadrotor model. We slightly modify the L1 adaptive control design for linear systems to comply with the time-varying nonlinear error dynamics that arise from the backstepping design. Our approach yields a stable adaptive system with verifiable bounds on the tracking error and input signals. The adaptive controller compensates for all model uncertainties and for all bounded disturbances within a particular frequency range, which we specify a priori. The design of this frequency range involves a trade-off between control performance and robustness, which can be managed transparently through the L1 adaptive control design. Simulation results show the powerful properties of the presented control application.

FrA10	HG D1.1
Stability of Linear Systems (Regular Session)
Chair: Knorn, Steffi	Univ. of Newcastle
Co-Chair: Charalambous, Themistoklis	Royal Inst. of Tech. (KTH)
09:20-09:40	FrA10.1
String Stability Analysis of a Vehicle Platoon with Communication Range 2 Using the Two-Dimensional Induced Operator Norm, pp. 3354-3359
Knorn, Steffi	Univ. of Newcastle
Middleton, Rick	The Univ. of Newcastle
Vehicle strings can be modelled as continuous-discrete 2D systems. However, every such 2D system includes a structural nonessential singularity on the stability boundary and special care is needed to analyse BIBO stability of the system. A unidirectional vehicle string with communication range 2 is analysed and simulated.
09:40-10:00	FrA10.2
Integral Inequality for Time-Varying Delay Systems, pp. 3360-3365
Seuret, Alexandre	LAAS
Gouaisbaut, Frederic	LAAS CNRS
This paper considers the stability analysis of time- varying delay systems. We develop a new integral inequality which is proved to encompasses the celebrated Jensen's inequal- ity. These technical tools allow to construct simple Lyapunov- Krasovskii functionals very efficient in practice. Notice that our procedure is coupled with the use of the reciprocal convexity result in order to reduce the conservatism induced by the LMIs optimisation setup. The effectiveness of the proposed results is illustrated by some classical examples from the literature.
10:00-10:20	FrA10.3
Stability and Persistence Analysis of Large Scale Interconnected Positive Systems, pp. 3366-3371
Ebihara, Yoshio	Kyoto Univ.
Peaucelle, Dimitri	LAAS-CNRS
Arzelier, Denis	LAAS-CNRS
This paper is concerned with the analysis of large-scale interconnected systems constructed from positive subsystems and a nonnegative interconnection matrix. We first show that the interconnected system is admissible and stable if and only if a Metzler matrix built from the coefficient matrices of the positive subsystems and the interconnection matrix is Hurwitz stable. By means of this key lemma, we further provide several results that characterize the admissibility and stability of interconnected systems in terms of the weighted L1-induced norm of each positive subsystem and the Frobenius eigenvalue of the interconnection matrix. Moreover, in the case where every subsystem is SISO, we provide explicit conditions under which the interconnected system has the property of persistence, i.e., the state of the interconnected system converges to a unique strictly positive vector (up to a strictly positive constant multiplicative factor) irrespective of nonnegative and nonzero initial states. We illustrate the effectiveness of the persistence results by formation control of multi-agent systems.
10:20-10:40	FrA10.4
On the Rate of Convergence of Continuous-Time Linear Positive Systems with Heterogeneous Time-Varying Delays, pp. 3372-3377
Feyzmahdavian, Hamid Reza	Royal Inst. of Tech. - KTH
Charalambous, Themistoklis	Royal Inst. of Tech. (KTH)
Johansson, Mikael	Royal Inst. of Tech.
In this work, a set of conditions are presented for establishing exponential stability and bounds on the convergence rates for both general and positive linear systems with heterogeneous time-varying delays. First, a sufficient condition for delay-independent exponential stability of general linear systems is derived. When the time delays have a known upper bound, we present an explicit expression that bounds the decay rate of the system. We demonstrate that the best decay rate that our bound can provide, can be easily found via convex optimization techniques. Finally, for positive linear systems, we show that the stability condition that we have developed is also necessary. The validity of the results is demonstrated via simple examples.
10:40-11:00	FrA10.5
Stability and Stabilization of Discrete-Time Descriptor Systems with Several Extensions, pp. 3378-3383
Masubuchi, Izumi	Kobe Univ.
Ohta, Yuzo	Kobe Univ.
This paper is concerned with stability criterion of discrete-time descriptor systems in terms of LMIs. Though there have been proposed several LMIs for discrete-time descriptor systems, their use is seemingly limited in comparison with LMIs for continuous-time descriptor systems. A remedy of those existing results is that they are not able to be linearized with respect to the coefficient matrices of the descriptor equation. This makes it difficult to apply them to synthesis problems, while LMIs for analysis of state-space systems or those for continuous-time descriptor systems enjoy change-of-variable methods to obtain LMIs for synthesis. In this paper, a new LMI is presented for admissibility of discrete-time descriptor systems that has such a useful feature of affineness of coefficient matrices in the inequality. After showing the main result with the new LMI, we apply it to several problems on discrete-time descriptor systems, namely, state-feedback synthesis, stability analysis of delta-operator systems and D-stability problems generalized to descriptor systems, and a generalization of quadratic stabilization. Some of these are illustrated via numerical examples.

FrA11	HG D1.2
Chemical Process Control (Regular Session)
Chair: Fikar, Miroslav	Slovak Univ. of Tech.
Co-Chair: Kinnaert, Michel	Univ. Libre de Bruxelles
09:20-09:40	FrA11.1
Multi-Objective Optimal Control of Ultrafiltration/diafiltration Processes, pp. 3384-3389
Jelemensky, Martin	STU in Bratislava
Paulen, Radoslav	Faculty of Chemical and Food Tech. Slovak Univ. Tec
Fikar, Miroslav	Slovak Univ. of Tech.
Kovacs, Zoltan	Corvinus Univ. of Budapest
This paper deals with multi-objective optimal operation of a general batch ultrafiltration/diafiltration (UF/DF) process. We consider the optimization criterion consisting of two parts as combination of two classical optimization problems, namely the minimum time problem and the problem of minimum diluant consumption. We apply Pontryagin's minimum principle for obtaining the necessary conditions for optimality. Based on these, we show how to obtain a Pareto front and how to derive multi-objective optimal operation for considered type of processes. The presented methodology will be applied to the most commonly used model in membrane filtration - the limiting flux case and to another experimentally derived membrane model.
09:40-10:00	FrA11.2
NCO-Tracking Based Control of Semi-Batch Antisolvent Crystallization Processes in the Presence of Uncertainties, pp. 3390-3395
Kamaraju, Vamsi Krishna	National Univ. of Singapore
Srinivasan, B.	Ec. Pol. Montreal
Chiu, Min-Sen	National Univ. of Singapore
Control of crystallization processes in order to obtain desired product specifications is very critical in pharmaceutical industries, as it influences not only the efficiency of the other downstream processes but also the bioavailability of the drug. Hence, motivated to counter the pragmatic limitations of implementing the optimal control policies in the presence of plant-model mismatch, measurement-based optimization (MBO) schemes for real-time optimal operation of (semi-)batch pharmaceutical antisolvent crystallization processes has been presented in this paper. MBO schemes are based on tracking the Necessary Conditions of Optimality (NCO), usually a sequence of boundary and (or) interior arcs, using measurement feedback. The current study explores this approach by designing neighboring extremal (NE) controller to follow the optimal trajectory in the interior arcs during the real time optimal control. Furthermore, performance assessment of the NE controller is done by comparing to different control strategies for various scenarios of plant model mismatch.
10:00-10:20	FrA11.3
Temperature-Temperature Cascade Control of Binary Batch Distillation Columns, pp. 3396-3401
Castellanos Sahagun, Eduardo	PARCAR DESARROLLOS Y SERVICIOS SA DE CV
Alvarez, Jesus	Univ. Autónoma Metropolitana, Unidad Iztapalapa
In this paper it is addressed the control design for binary batch distillation columns, based on temperature measurements. The combination of nonlinear constructive control theory with passivation and observability notions and existing batch distillation concepts yields: (a) a simple methodology for temperature sensor locations, (b) the on-line generation of temperature policies that ensure constant distillate purity, and (c) a temperature-to-temperature cascade tracking controller to force the prescribed constant distillate product purity policy. The methodology is tested with representative binary systems.
10:20-10:40	FrA11.4
Application of First-Principles Based Techniques for Compounds Prediction in a Gasification Plant, pp. 3402-3407
Zanoli, Silvia Maria	Univ. Pol. delle Marche
Astolfi, Giacomo	Univ. Pol. marche
In this work a first-principles based model for the prediction of the Syngas composition produced in a gasification plant is described. The model is formulated on the basis of chemical-physical principles of the gasification process and is based on nonlinear static equations. For the formulation of the equations the Gibbs energy minimization of the compound used in the gasification process is taken into account together with computations of material balance. The model has been validated with real data gathered from a gasification plant. From the analysis of the results it can be stated that good estimation performances have been achieved. This chemical-physical model overcomes the limitation introduced by the use of a linear model presented by the authors in a previous work. In fact the new model provides a good estimation even when the range of the stoichiometric ratio is extended outside the boundaries of validity of a linear model. Given the results the model can be used in a Model-Based Predictive Control (MPC) strategy for process optimization and/or for the development of a Fault Diagnosis system.
10:40-11:00	FrA11.5
POD-Based State Estimation of Simulated Moving Bed Chromatographic Processes, pp. 3408-3414
Boulkroune, Boulaid	HEI
Kinnaert, Michel	Univ. Libre de Bruxelles
Zemouche, Ali	Nancy-Univ.
Chromatographic separation is a powerful technique based on the differential adsorption of the components of a mixture. In simulated moving bed (SMB) chromatographic processes, a counter-current movement of the liquid and the solid phases is achieved by periodically switching the inlet and the outlet ports in a closed loop of chromatographic columns. This paper addresses the problem of state observation of simulated moving bed (SMB) chromatographic processes. The proposed approach is based on a reduced model obtained via the proper orthogonal decomposition (POD) method. The state estimation is performed using a nonlinear observer designed for Lipschitz LPV (linear parameter varying) systems. Numerical simulations are provided to validate the proposed approach.
11:00-11:20	FrA11.6
Static and Dynamic Set Point Optimization Techniques for Optimal Operation of Wastewater Treatment Plants, pp. 3415-3420
Vega, Pastora	Univ. of Salamanca
Revollar, Silvana	Univ. de Salamanca
Martin Noguerol, Jose Manuel	Univ. de Salamanca
Francisco, Mario	Univ. of Salamanca
Optimization and control strategies are necessary to keep wastewater treatment plants operating in the best possible conditions with the minimum consumption of energy. In this paper static and dynamic set point optimization strategies to improve the operation of the plant in order to achieve the economically optimal conditions are studied. Different objective function formulations are compared and the overall performance of the plant operating in a rain weather scenario is evaluated to determine the most adequate operation strategy.

FrA12	HG D3.2
Reduced Order Systems (Regular Session)
Chair: de Carvalho, Martins J .L.	Faculdade de Engenharia da Univ. do Porto
Co-Chair: Belur, Madhu N.	Indian Inst. of Tech. Bombay
09:20-09:40	FrA12.1
Continuous-Time IO Systems Identification through Downsampled Models, pp. 3421-3426
Lopes dos Santos, P.	Univ. do Porto
Azevedo Perdicoulis, Teresa Paula	ISR-Coimbra & UTAD
Ramos, Jose A.	Nova Southeastern Univ.
Jank, Gerhard	RWTH
de Carvalho, Martins J .L.	Faculdade de Engenharia da Univ. do Porto
An indirect downsampling approach for continuous-time input/output system identification is proposed. This {it modus operandi} was introduced to system identification through a subspace algorithm, where the input/output data set is partitioned into lower rate $m$ subsets. Then, a state-space discrete-time model is identified by fusing the data subsets into a single one. In the present work the identification of the input/output downsampled model is performed by a least squares and a simplified refined instrumental variables (IV) procedures. In this approach, the inter-sample behaviour is preserved by the addition of fictitious inputs, leading to an increase of excitation requirements of the input signal. This over requirement is removed by directly estimating from the data the parameters of the transfer function numerator. The performance of the method is illustrated using the Rao-Garnier test system.
09:40-10:00	FrA12.2
H2 Pseudo-Optimality in Model Order Reduction by Krylov Subspace Methods, pp. 3427-3432
Wolf, Thomas	Tech. Univ. München
Panzer, Heiko K. F.	Tech. Univ. München
Lohmann, Boris	Tech. Univ. Muenchen
New sufficient conditions for H2 pseudo-optimality in Krylov-based model reduction of linear dynamical systems are presented. The conditions are easy to evaluate and permit first applications: a new algorithm to generate H2 pseudo-optimal reduced models with respect to the projecting subspace and a procedure to generate superior local optima in iterative methods. Numerical examples illustrate the contributions.
10:00-10:20	FrA12.3
On Parametric Model Order Reduction by Matrix Interpolation, pp. 3433-3438
Geuss, Matthias	TU München
Panzer, Heiko K. F.	Tech. Univ. München
Lohmann, Boris	Tech. Univ. Muenchen
A general framework for model order reduction is proposed for high-order parameter-dependent, linear time-invariant systems. The procedure is based on matrix interpolation and consists of six steps. At first a set of high-order nonparametric systems is computed for different parameter vectors. The resulting local high-order systems are then reduced by a projection-based reduction method. Thereby, proper right and left subspaces for the reduced systems are calculated. Next the bases of the right subspaces of the reduced systems are adapted and the bases of the left subspaces are adjusted. For that the concept of duality is introduced. Finally, the precomputed matrices of the local systems are interpolated in a matrix manifold with an interpolation method. In this paper the six steps of the algorithm and the degrees of freedom which arise therein are presented. Furthermore, advantages and difficulties in the selection of the degrees of freedom are pointed out. It is additionally shown that two existing methods for parametric model order reduction by matrix interpolation are special cases of the proposed general procedure as they -- often implicitly -- determine a limiting selection of the degrees of freedom.
10:20-10:40	FrA12.4
Distributed Proper Orthogonal Decomposition for Large-Scale Networked Dynamical Systems, pp. 3439-3445
Kojima, Chiaki	Univ. of Tokyo
Kawasaki, Issei	The Univ. of Tokyo
Moriyama, Satoshi	The Univ. of Tokyo
Wada, Jun	The Univ. of Tokyo
Recently, dynamical systems in engineering and science problems become drastically larger and too complex. One of the ways to solve the difficulty is to model systems with hierarchical network structures. Proper orthogonal decomposition (POD) is a model reduction method using available data and its singular value decomposition. In this paper, we apply the POD to a networked linear dynamical system and propose the distributed POD which can specify the degree of approximate of each subsystem and preserve the network structure. We also characterize an upper bound of the approximation error and give a criterion to determine the optimal degrees of reduced-order subsystems. As an application to the distributed POD, the l₁-norm minimizing POD is also proposed for a construction of a simple network structure which approximates the behavior of the entire system appropriately. A numerical example is provided to demonstrate the distributed and l₁-norm minimizing PODs and to show their efficiency.
10:40-11:00	FrA12.5
Minimal Controller Structure for Generic Pole Placement, pp. 3446-3451
Kalaimani, Rachel Kalpana	Indian Inst. of Tech. Bombay
Belur, Madhu N.	Indian Inst. of Tech. Bombay
In this paper we address the generic pole placement problem for a system represented by differential algebraic equations. The genericity aspect is relevant when dealing with large dynamical systems where the plant equations are sparse. We capture the sparsity structure of the plant equations into an edge weighted and undirected bipartite graph. We propose an algorithm that furnishes a `minimal' controller structure for achieving generic arbitrary pole placement: minimality in the sense of the sparsity within controller equations. More precisely, we introduce a procedure to come up with a set of controller equations such that, in addition to generically achieving arbitrary pole placement, the bipartite graph constructed for this controller has the minimum number of edges amongst all controllers that generically achieve arbitrary pole placement. The algorithm we propose involves finding a minimum number of paths that cover a given set of vertices corresponding to plant equations. We introduce an integer that captures the extent of MIMO features inside the plant equations, since this turns out to crucially decide the minimum number of required edges. This paper's minimal controller structure problem and the proposed solution turn out to also solve the problem of generically completing a given rectangular polynomial matrix into a unimodular matrix using the minimum number of nonzero entries.

FrA13	HG D5.2
Fractional Dynamical Systems and Signals - Part 1 – Fractional Differentiation in Modeling and Systems Analysis (Invited Session)
Chair: Sabatier, Jocelyn	LAPS - Bordeaux 1 Univ.
Co-Chair: Sierociuk, Dominik	Warsaw Univ. of Tech. (ISEP)
Organizer: Sabatier, Jocelyn	LAPS - Bordeaux 1 Univ.
Organizer: Farges, Christophe	IMS-LAPS
09:20-09:40	FrA13.1
Fractional Order Identification of Human Arm Dynamics: Preliminary Results (I), pp. 3452-3457
Tejado, Ines	Univ. Técnica de Lisboa, Inst. Superior Técnico
Valério, Duarte Pedro Mata de Oliveira	IDMEC/IST, TULisbon
Pires, Pedro	IST, TULisbon
Martins, Jorge M.	Inst. Superior Tecnico
Recently, control approaches for human-like behaviour have been attracting considerable attention in the field of surgery robotics. To this respect, precise dynamic models of the human arm are required to give the surgeon the physical feeling of working with a human assistant rather than a machine, which will result in a safer physical interaction. Musculo-skeletal systems of several species, including human muscles, have been successfully modelled by fractional differential equations. This study presents fractional order closed-loop identification for estimating the dynamics of the human arm. In particular, fractional and integer order models are identified in the frequency domain from real experiments with human subjects, using continuous random force as input and position in angle as output. The results show that more general dynamic models, i.e., fractional order models, allow adequate frequency responses to be attained but with a smaller number of parameters. A comparison with different dynamic models of the human arm reported in the literature is also given to demonstrate the validity of the proposed models.
09:40-10:00	FrA13.2
Fractional Models for Lithium-Ion Batterie (I), pp. 3458-3463
Sabatier, Jocelyn	LAPS - Bordeaux 1 Univ.
Francisco, Mbala Junior	PSA Peugeot Citroën
Guillemard, Franck	PSA Peugeot Citroën
With hybrid and electric vehicles development, automobile battery monitoring systems (BMS) have to meet the new requirements. These systems have to give information, on state of health, state of charge, available power…To get this information, BMS often implement battery models. Accuracy of the information manipulated by the BMS thus depends on the model accuracy. This paper is within this framework and addresses lithium-ion battery modeling. The proposed fractional model is based on simplifications of an electrochemical model and on resolution of some partial differential equations used in its description. Such an approach permits to get a simple model in which electrochemical variables and parameters still appear.
10:00-10:20	FrA13.3
Equivalent Switching Strategy and Analog Validation of the Fractional Variable Order Derivative Definition (I), pp. 3464-3469
Sierociuk, Dominik	Warsaw Univ. of Tech. (ISEP)
Malesza, Wiktor	Warsaw Univ. of Tech.
Macias, Michal	Warsaw Univ. of Tech.
The paper presents switching strategy that is equivalent to one type of variable order derivative definitions. The numerical scheme, based on matrix approach, for this type of the definition is also introduced. Using this approach the identity of the switching strategy and considered definition is derived. The switching scheme can be used as an interpretation of this type of definition. Paper presents also numerical examples for introduced methods. Finally, the idea and results of analog (electrical) realization of the switching fractional order integrator (of orders 0.5 and 1) are presented and compared with numerical approach.
10:20-10:40	FrA13.4
Model Order Identification for Fractional Models, pp. 3470-3475
Victor, Stéphane	Univ. de Bordeaux
Malti, Rachid	Univ. de Bordeaux
This paper deals with continuous-time system identification using fractional differentiation models. So far, no algorithm exists concerning model order identification of fractional models. The simplified refined instrumental variable method is proposed to estimate the parameters of the fractional differential equation model when all the fractional orders are assumed known. Then an optimization approach based on the use of the developed instrumental variable estimator is presented. This two-stage algorithm, called oosrivcf is then used for model order identification by using simultaneously two criteria: the Young Information Criterion (YIC) and the RT2 criterion.
10:40-11:00	FrA13.5
Analysis of Fractional-Order Telegraph Model for Neutron Transport in Nuclear Reactor with Slab Geometry (I), pp. 3476-3481
Vyawahare, Vishwesh	Indian Inst. of Tech. Bombay
Nataraj, P.S.V.	Indian Inst. of Tech.
This paper deals with the analysis of fractional-order (FO) telegraph equation which models the transport of thermal neutrons inside a nuclear reactor with slab geometry. The FO model, previously developed by the authors, models the neutron transport as anomalous diffusion, precisely a subdiffusion. This model removes the lacunae of the conventional integer-order model of neutron movements. Here, the FO model is solved using the well known technique of separation of variables and the spatial distribution and time evolution of the neutron flux in the slab reactor are computed. This exercise is probably being performed for the first time for the fractional-order model of neutron transport. It clearly depicts the wavelike nature of the neutron flux. Also, the convergence of neutron flux for FO telegraph and subdiffusion models for asymptotic time establishes the long-time subdiffusive behaviour of the FO telegraph equation. The analysis carried out in this paper is thus forms a crucial step in the process of development of fractional-order model for a nuclear reactor.
11:00-11:20	FrA13.6
A Non-Square MIMO Fractional Robust Control for the Airpath of a Diesel Engine, pp. 3482-3487
Lamara, Abderrahim	Univ. of Bordeaux1
Chamaillard, Yann	Univ. of Orléans
Charlet, Alain	Univ. d'Orléans
Lanusse, Patrick	Univ. of Bordeaux / Bordeaux Pol. Inst. /ENSEIRB
Colin, Guillaume	Univ. of Orléans
In order to minimize pollutant emissions and more specifically nitrogen oxides, a decentralized multivariable robust control has been designed for the airpath of a Diesel engine with a turbocharger, an intake throttle and Exhaust Gas Recirculation. The control system is designed with a fractionalorder approach called CRONE (Commande Robust d'Ordre Non Entier) control, via a MIMO open-loop transfer function optimization, which maintains performance and robust stability for a wide set of operating points. The experimental results from a test-bench show the relevance of the proposed approach.

FrA14	HG D7.1
Traffic Control (Regular Session)
Chair: De Schutter, Bart	Delft Univ. of Tech.
Co-Chair: Kulcsar, Balazs	Chalmers Univ. of Tech.
09:20-09:40	FrA14.1
Model Based Tracking Control Using Jerky Behavior in Platoon of Vehicles, pp. 3488-3493
Merzouki, Rochdi	Ec. Pol. de Lille
Conrard, Blaise	LAGIS - Univ. de Lille 1
Kumar, Pushpendra	Univ. Lille1, Pol. LAGIS
Coelen, Vincent	Univ. Lille1, Pol. LAGIS
In this work, a reconstruction of the sensor based tracking control in a platoon of autonomous vehicles is developed. This control is reproduced with accuracy in the absence of data from the laser rangefinder. Only the velocity of each involved vehicle in the platoon is considered, as well as the inter-vehicle communication. This reconstruction occurs when the laser rangefinder is disturbed or failed. It helps maintaining a fault tolerant control when the measurement is unavailable. The control strategy is validated on a real-time simulator of vehicles dynamic, based on experimental measurements, acquired from autonomous vehicles.
09:40-10:00	FrA14.2
Decentralized Optimal Control of a Car Platoon with Guaranteed String Stability, pp. 3494-3499
Morbidi, Fabio	Inria, Grenoble - Rhone-Alpes
Colaneri, Patrizio	Pol. di Milano
Stanger, Thomas	Johannes Kepler Univ. Linz
This paper presents new decentralized optimal strategies for "Cooperative Adaptive Cruise Control" (CACC) of a car platoon under string-stability constraints. Two related scenarios are explored in the article: in the first one, a linear-quadratic regulator in the presence of measurable disturbances is synthesized, and the string-stability of the platoon is enforced over the controller's feedback and feedforward gains. In the second scenario, H2- and Hinf-performance criteria, respectively accounting for the desired group behavior and the string stability of the platoon, are simultaneously achieved using the recently-proposed compensator blending method. An analytical study of the impact of actuation/communication delays and uncertain model parameters on the stability of the multi-vehicle system, is also conducted. The theory is illustrated via numerical simulations.
10:00-10:20	FrA14.3
Optimal Hybrid Macroscopic Traffic Control for Urban Regions: Perimeter and Switching Signal Plans Controllers, pp. 3500-3505
Hajiahmadi, Mohammadreza	Delft Univ. of Tech.
Haddad, Jack	Tech. - Israel Inst. of Tech.
De Schutter, Bart	Delft Univ. of Tech.
Geroliminis, Nikolas	Ec. Pol. Fédérale de Lausanne (EPFL), Urban Transport
The dynamics of a heterogeneous large-scale urban network is modeled as R homogeneous regions with the macroscopic fundamental diagrams (MFDs) representations. The MFD provides for homogeneous network regions a unimodal, low-scatter relationship between network vehicle density and network space-mean flow. In this paper, the optimal hybrid control problem for an R-region MFD network is formulated as a mixed integer nonlinear optimization problem, where two types of controllers are introduced: (i) perimeter controllers, and (ii) switching signal timing plans controllers. The perimeter controllers are located on the border between the regions, as they manipulate the transfer flows between them, while the switching controllers control the dynamics of the urban regions, as they define the shape of the MFDs, and as a result affect the internal flows within each region. Moreover, to decrease the computational complexity due to the nonlinear and nonconvex nature of the formulated optimization problem, we rewrite the problem as a mixed integer linear problem utilizing a piecewise affine approximation technique. The performance of the two problems is evaluated and compared for different traffic scenarios for a two-region urban case study.
10:20-10:40	FrA14.4
Feedback Perimeter Control for Multi-Region Large-Scale Congested Networks, pp. 3506-3511
Aboudolas, Konstantinos	Ec. Pol. Fédérale de Lausanne (EPFL), Urban Transport
Geroliminis, Nikolas	Ec. Pol. Fédérale de Lausanne (EPFL), Urban Transport
It was recently observed from empirical traffic data that by aggregating the highly scattered plots of flow versus density from individual loop detectors for city regions with homogeneous spatial distribution of congestion, the scatter significantly decreases and a well-defined Macroscopic Fundamental Diagram (MFD) exists between space-mean flow and density. This result can be of great importance to unveil simple perimeter control policies in such a way that maximizes the network outflow (trip endings). Single-region perimeter control might be sub-optimal if there is a significant number of destinations outside the region of analysis or if the network is heterogeneously loaded. This paper integrates an MFD modeling to perimeter and boundary control optimization for large-scale networks with multiple centers of congestion, if these networks can be partitioned into a small number of homogeneous regions. Perimeter control actions may be computed in real-time through a linear multivariable feedback regulator. The impact of the perimeter control actions to a three-region real urban network is demonstrated via micro-simulation. A key advantage of the proposed approach is that it does not require high computational effort and future demand data if the current state of each region can be observed.
10:40-11:00	FrA14.5
Synthesis of Intelligent Control of Traffic Flows in Urban Roads Based on the Logical Network Operator Method, pp. 3512-3517
Diveev, Askhat	Inst. of Russian Acad. of Sci.
Sofronova, Elena	Peoples' Friendship Univ. of Russia
A problem of optimal control of traffic flows in the urban network is considered. A derivation of the mathematical model of control by the traffic lights at intersections is given. The mathematical model of control object is obtained using the controlled networks theory. This model is a system of nonlinear finite-difference equations. To present a large scale road networks the model contains the connection matrices. The connection matrices describe connections between input and output roads in subnetworks. These matrices allow considering the influence of traffic flows in one subnetwork the ones in other subnetwork. The traffic flow control is done by the coordination of active phases of traffic lights. The optimal control problem is to minimize the difference between the total input flow and total output flow for all subnetworks. To solve the optimal control synthesis problem we use a logical network operator method. An example of model for the network of three subnetworks is given.
11:00-11:20	FrA14.6
Incident Parameter Estimation, pp. 3518-3523
Dabiri, Azita	Chalmers Univ.
Kulcsar, Balazs	Chalmers Univ. of Tech.
The paper proposes a sampled data based estimation methodology to reconstruct local incident parameter of the macroscopic Incident Traffic Flow (ITF) models. The key idea in ITF models is to dynamically relax the traffic mean speed to the traffic equilibrium one based of an time and space varying incident term. First, the analysis of incident corrupted traffic flow models, described as an inhomogenuous nonlinear Partial Differential Equation (PDE), is presented in continuous time. Second, space and time discretization techniques are applied to derive traffic management oriented ITF models. Online parameter estimation is suggested to capture the severity of incident throughout the proposed parameter, i.e. to estimate the incident parameter. Numerical example is carried out to show the viability of macroscopic incident parameter estimation technique using data obtained from a high-fidelity microscopic simulation.

FrA15	HG D7.2
Control of Power Systems (Regular Session)
Chair: Dimarogonas, Dimos V.	Royal Inst. of Tech. (KTH)
Co-Chair: Rantzer, Anders	Lund Univ.
09:20-09:40	FrA15.1
Distributed vs. Centralized Power Systems Frequency Control, pp. 3524-3529
Andreasson, Martin	Royal Inst. of Tech. (KTH)
Dimarogonas, Dimos V.	Royal Inst. of Tech. (KTH)
Johansson, Karl Henrik	Royal Inst. of Tech.
Sandberg, Henrik	Royal Inst. of Tech. (KTH)
This paper considers a distributed control algorithm for frequency control of electrical power systems. We propose a distributed controller which retains the reference frequency of the buses under unknown load changes, while asymptotically minimizing a quadratic cost of power generation. For comparison, we also propose a centralized controller which also retains the reference frequency while minimizing the same cost of power generation. We derive sufficient stability criteria for the parameters of both controllers. The controllers are evaluated by simulation on the IEEE 30 bus test network, where their performance is compared.
09:40-10:00	FrA15.2
A Detailed Study on a DC-Voltage-Based Control Scheme Using a Multi-Terminal HVDC System for Frequency Control, pp. 3530-3535
Chen, Yijing	Lab. des Signaux et Systemes (LSS), CNRS
Dai, Jing	Lab. des Signaux et Systemes, CNRS
Damm, Gilney	Lab. IBISC - CNRS/Evry Univ.
Lamnabhi-Lagarrigue, Francoise	CNRS-EECI
This paper investigates a control strategy that uses a multi-terminal high voltage direct current (HVDC) system for frequency control. It first describes a DC voltage control law which, based only on local information, makes the interconnected non-synchronous systems share their primary reserves. Then, a sufficient condition for choosing the controller gains is developed based on Lyapunov theory and LMI techniques with rigorous mathematical stability proofs. Simulation results show that the control strategy has achieved the objective of sharing primary reserves, guarantees the whole system's ultimate boundedness and improves the transient and steady-state performance.
10:00-10:20	FrA15.3
Nonlinear Control Design for a Multi-Terminal VSC-HVDC System, pp. 3536-3541
Chen, Yijing	Lab. des Signaux et Systemes (LSS), CNRS
Dai, Jing	Lab. des Signaux et Systemes, CNRS
Damm, Gilney	Lab. IBISC - CNRS/Evry Univ.
Lamnabhi-Lagarrigue, Francoise	CNRS-EECI
This paper presents a nonlinear control strategy based on dynamic feedback linearization control theory and a backstepping-like procedure for a multi-terminal voltage-source converter based high-voltage direct current (multi-terminal VSC-HVDC) system. The controller is able to provide asymptotic stability for the power transmission system with multiple terminals. In particular, it is shown that the zero dynamics (mostly representing the DC network) is exponentially stable. These results are obtained by a stability proof for the whole system under the proposed controller, and its performance is illustrated by computer simulations.
10:20-10:40	FrA15.4
Non-Stationary Harmonic Tracking Using Piecewise-Overlapped Group-Harmonic Algorithm, pp. 3542-3547
Lin, Hsiung Cheng	National Chin-Yi Univ. of Tech.
Huang, Guo-Shing	National Chin-Yi Univ. of Tech.
The Discrete Fourier Transform (DFT) is still a widely-used tool for analyzing and measuring both stationary and transient signals in power systems. However, the misapplications of DFT can lead to incorrect results caused by some problems such as aliasing effect, spectral leakage and picket-fence effect. A strategy of Piecewise-overlapped Group-harmonic Energy Restoring (PGER) algorithm is developed for both stationary and non-stationary harmonic evaluation in power systems. The proposed algorithm can restore dispersing spectral leakage energy caused by the DFT, and thus retrieve the respective real harmonic value. Based on the piecewise-overlapped method, especially every time-varying harmonic can be extracted fast and accurately. The numerical examples are presented to verify the proposed algorithm in term of robust, fast and high-precision performance.
10:40-11:00	FrA15.5
On Optimal Sensor Placement for Mitigation of Vulnerabilities to Cyber Attacks in Large-Scale Networks, pp. 3548-3553
Vaidya, Umesh	Iowa State Univ.
Fardad, Makan	Syracuse Univ.
We propose a system theoretic approach to the identification and mitigation of vulnerabilities to cyber attacks, in networks of dynamical systems. Using the controllability and observability gramians, we define a network's vulnerability in terms of the impact of an attack input and the degree of difficulty with which this impact can be detected. In this framework, a network is deemed as vulnerable if it is easy for an attacker to steer it to a certain state and yet such a state is hard to observe through the network's sensing mechanisms. We propose strategies for finding the optimal location of a small number of sensors that minimize the network's vulnerability. Such strategies are obtained as the solution of convex optimization problems, formulated so as to strike a balance between maximal reduction of the system's vulnerability and employing a minimal number of sensors. The utility of the developed framework is demonstrated on a standard IEEE nine bus power system network model.
11:00-11:20	FrA15.6
Dynamic Power Coordination for Load Reduction in Dispatchable Wind Power Plants, pp. 3554-3559
Madjidian, Daria	Lund Univ.
Kristalny, Maxim	Tech. - IIT
Rantzer, Anders	Lund Univ.
In a dispatchable wind power plant, turbines are free to continuously vary their power production as long as the sum of their productions meets the total power demand. Recent research has shown that this freedom can be used to reduce structural loads by allowing turbines in the plant to coordinate their power. This paper explains the mechanisms that make power coordination useful for reducing structural loads on the turbine tower and the low speed shaft. In addition, it assesses the benefits of coordination at different operating points.

FrST1	HG F1, F3
Semi-Plenary 5: Information Processing and Control in Biological Systems; Some Fundamental Limits (Semi-Plenary Session)
Chair: Ferrari Trecate, Giancarlo	Univ. degli Studi di Pavia
11:30-12:15	FrST1.1
Information Processing and Control in Biological Systems; Some Fundamental Limits*
Vinnicombe, Glenn	Univ. of Cambridge

FrST2	HG F7, F5
Semi-Plenary 6: Legged Dynamics and Control: Basic Models, Neuromuscular Interpretation, and Robotic Application (Semi-Plenary Session)
Chair: Smith, Roy S.	ETH Zurich
11:30-12:15	FrST2.1
Legged Dynamics and Control: Basic Models, Neuromuscular Interpretation, and Robotic Application*
Geyer, Hartmut	Carnegie Mellon Univ.

FrB1	HG F1
Emerging Control Applications I (Regular Session)
Chair: Horn, Joachim	Helmut-Schmidt-Univ. / Univ. of the Federal Armed Forces Hamburg
Co-Chair: Martinez-Salamero, Luis	Escola Tecnica Superior d'EnginyeriaUniversitat Rovira i Virgili
13:40-14:00	FrB1.1
State Estimation with Time Delay and State Feedback Control of Cathode Exhaust Gas Mass Flow for PEM Fuel Cell Systems, pp. 3560-3565
Schultze, Martin	Helmut-Schmidt-Univ. Hamburg
Horn, Joachim	Helmut-Schmidt-Univ. / Univ. of the Federal Armed Forc
Polymer electrolyte membrane (PEM) fuel cells are highly efficient energy converters and provide electrical energy, cathode exhaust gas with low oxygen content and water. On aircraft fuel cells have the potential of replacing the auxiliary power unit (APU) that is currently used for electrical power generation during ground operations. An APU is a significant source of noise and greenhouse gases. PEM fuel cells, however, can significantly reduce these pollutants. In this study a PEM fuel cell system is investigated for generation of oxygen depleted cathode exhaust air (ODA). This gas is intended for inerting on aircraft and must have a low oxygen concentration. A nonlinear simulation model comprising the fuel cell stack, stack cooling system and cathode exhaust gas dehumidifying system has been developed and a reduced order model for controlling ODA-gas mass flow was derived. An observer for state estimation combined with a predictor to compensate for the large time delay occurring has been designed. ODA mass flow is controlled for by a linear state feedback controller. Simulation results performed at the nonlinear simulation model in Matlab/Simulink® and experimental results are shown.
14:00-14:20	FrB1.2
A Fixed-Structure Automaton for Load Management of Electric Vehicles, pp. 3566-3571
Stuedli, Sonja	The Univ. of Newcastle
Middleton, Rick	The Univ. of Newcastle
Braslavsky, Julio H.	Commonwealth Scientific and Industrial Res.
The uncontrolled charging of electric vehicles (EVs) imposes additional stresses on the grid. These stresses are set to increase due to the predicted increase in penetration of EVs. However, EV charging loads offer opportunities for controlling the actual demand to limit the peak demand or track variable generation to support the grid. This also can facilitate the integration of intermittent renewable energy generation in the near future. In this paper a learning automaton is proposed to manage EVs capable of on/off charging. We propose a new algorithm for distributed control of charging based on the broadcast of a congestion signal to regulate the aggregated demand. We show that the proposed algorithm converges to steady state operation, and analyse its implications on the distribution of power demand amongst the EVs. The potential of the proposed algorithm is illustrated by simulations for capping the aggregate demand of the EVs, and for tracking of slowly varying power generation signals.
14:20-14:40	FrB1.3
Temperature Control of the FTU Liquid Lithium Limiter, pp. 3572-3577
Boncagni, Luca	ENEA
Carnevale, Daniele	Univ. of Rome
Cristaldi, Alessia	Univ. of Rome, Tor Vergata
De Maio, Stefano	Univ. of Rome, Tor Vergata
Mazzitelli, Giuseppe	enea
Sassano, Mario	Univ. of Rome, Tor Vergata
Vitale, Vincenzo	ENEA
Vitelli, Riccardo	Univ. of Rome, Tor Vergata
Zaccarian, Luca	CNRS
Several experiments on the Tokamak devices have shown serious difficulties in the development of appropriate technical solutions concerning the structure and material composition of the plasma-facing elements. For instance the first wall (i.e. the innermost surface of the Tokamak vacuum chamber), the limiters and the divertors, are a main concern due to the neutron flux and the thermal loads. A possible solution is to abandon solid plasma facing materials, using liquid metals for those surfaces. In the FTU (Frascati Tokamak Upgrade) an experimental Liquid Lithium Limiter (LLL) has been installed and tested, by empirically tuning the plasma position in order to prevent the modules of the limiter from reaching critical temperatures. In this paper, we propose two alternative nonlinear control strategies for the temperature regulation of the surfaces of the LLL modules. The paper is concluded with numerical simulations that validate the performances of the proposed approaches.
14:40-15:00	FrB1.4
Model Identification of a Proton-Exchange Membrane Fuel-Cell from an Input-Output Experiment : The Diffusive Representation Approach (I), pp. 3578-3583
Restrepo, Carlos	Delft Univ. of Tech.
Garcia, Germain	LAAS-CNRS
Calvente, Javier	Univ. Rovira i Virgili
Giral, Roberto	Univ. Rovira i Virgili
Martinez-Salamero, Luis	Escola Tecnica Superior d'EnginyeriaUniversitat Rovira iVirgili
This paper proposes a diffusive Proton-Exchange Membrane (PEM) Fuel-Cell (FC) model that reproduces the static and low-frequency dynamic characteristic of a PEMFC. Due to the nature of physical phenomena involved in a PEMFC, a fractional operator is a good candidate for describing its input-output behavior. In this paper, this operator is replaced by a well suited representation called "diffusive model" which simplifies its analysis and which leads to numerical realizations suitable for identification purposes. The diffusive model is approximated by a finite dimensional model that allows to formulate an optimization problem of least square error type to estimate the model distribution. The diffusive model is identified by means of experimental measurements of current (input) and voltage (output) in the FC. The obtained model is simple and can be used in systems that require real-time emulators or complex long-time simulations. The experimental results using the Ballard NEXA 1.2 kW fuel cell validate the advantages of the diffusive model.
15:00-15:20	FrB1.5
Pricing Long-Term Permits and Scheduling of Electric Vehicle Charging in Parking Lots with Shared Resources, pp. 3584-3589
Deshpande, Ajay	IBM T J Watson Res. Center
Murali, Pavankumar	IBM T J Watson Res. Center
With the interest from car owners in going green being at an all-time high, electric vehicles (EVs) are flooding the automobile market. One of the primary concerns in owning an EV is the availability of charging infrastructure while away from home. There has been a renewed interest in managing and pricing the usage of shared commercial EV chargers, while maximizing the operator's profits. Towards this end, we propose a combined pricing-scheduling quadratic integer programming (QIP) model that iteratively prices and schedules EV charging. A pricing module is used to accept/reject charging requests and control the right number and types (arrival-departure times, charge demand etc.) of EVs to charge. The scheduling module ensures that the demand can be met subject to price-demand sensitivity and other scheduling constraints. Once the EVs to be accepted have been finalized and their permit prices determined, the scheduling module can be run every night once the day-to-day arrival and departure times of each EV is revealed to the operator.

FrB2	HG F3
Embedded Optimization for Control and Estimation II – Hardware Implementation (Invited Session)
Chair: Zeilinger, Melanie N.	Ec. Pol. Fédéral de Lausanne (EPFL)
Co-Chair: Domahidi, Alexander	Automatic Control Lab. ETH Zurich
Organizer: Zeilinger, Melanie N.	UC Berkeley
Organizer: Domahidi, Alexander	Automatic Control Lab. ETH Zurich
13:40-14:00	FrB2.1
A Predictive Control Solver for Low-Precision Data Representation (I), pp. 3590-3595
Longo, Stefano	Cranfield Univ.
Kerrigan, Eric C.	Imperial Coll. London
Constantinides, George Anthony	Imperial Coll. London
We propose a method to efficiently exploit the non-standard number representation of some embedded computer architectures for the solution of constrained LQR problems. The resulting quadratic programming problem is formulated to include auxiliary decision variables as well as the inputs and states. The new formulation introduces smaller roundoff errors in the optimization solver, hence allowing one to trade off the number of bits used for data representation against speed and/or hardware resources. Interestingly, because of the data dependencies of the operations, the algorithm complexity (in terms of computation time and hardware resources) does not increase despite the larger number of decision variables.
14:00-14:20	FrB2.2
A Computationally Efficient Parallel Coordinate Descent Algorithm for MPC: Implementation on a PLC (I), pp. 3596-3601
Necoara, Ion	Univ. Pol. Bucharest
Clipici, Dragos	Pol. Univ. of Bucharest
In this paper we propose a parallel coordinate descent algorithm for solving smooth convex optimization problems with separable constraints that may arise e.g. in model predictive control (MPC) for linear network systems. The basis for the new algorithm are block coordinate descent updates, that are computed in parallel, have low iteration complexity and use only local information. As a result, the algorithm is suitable for implementation on autonomous hardware with low computational power. In the case of a strongly convex objective function, we prove that the algorithm has linear rate of convergence. An MPC scheme based on this algorithm is derived, such that the computations of feasible and stabilizing inputs are distributed and cheap, and can be done by each subsystem in part. Some implementation issues of the new algorithm for MPC problems are discussed, as well as it being tested on a 4 tank laboratory setup.
14:20-14:40	FrB2.3
Fixed-Point Dual Gradient Projection for Embedded Model Predictive Control (I), pp. 3602-3607
Patrinos, Panagiotis	IMT Inst. for Advanced Studies Lucca
Guiggiani, Alberto	IMT Inst. for Advanced Studies Lucca
Bemporad, Alberto	IMT Inst. for Advanced Studies Lucca
Although linear Model Predictive Control has gained increasing popularity for controlling dynamical systems subject to constraints, the main barrier that prevents its widespread use in embedded applications is the need to solve a Quadratic Program (QP) in real-time. This paper proposes a dual gradient projection (DGP) algorithm specifically tailored for implementation on fixed-point hardware. A detailed convergence rate analysis is presented in the presence of round-off errors due to fixed-point arithmetic. Based on these results, concrete guidelines are provided for selecting the minimum number of fractional and integer bits that guarantee convergence to a suboptimal solution within a prespecified tolerance, therefore reducing the cost and power consumption of the hardware device.
14:40-15:00	FrB2.4
Energy-Aware MPC Co-Design for DC-DC Converters (I), pp. 3608-3613
Suardi, Andrea	Imperial Coll. London
Longo, Stefano	Cranfield Univ.
Kerrigan, Eric C.	Imperial Coll. London
Constantinides, George Anthony	Imperial Coll. London
In this paper, we propose an integrated controller design methodology for the implementation of an energy-aware explicit model predictive control (MPC) algorithms, illustrating the method on a DC-DC converter model. The power consumption of control algorithms is becoming increasingly important for low-power embedded systems, especially where complex digital control techniques, like MPC, are used. For DC-DC converters, digital control provides better regulation, but also higher energy consumption compared to standard analog methods. To overcome the limitation in energy efficiency, instead of addressing the problem by implementing sub-optimal MPC schemes, the closed-loop performance and the control algorithm power consumption are minimized in a joint cost function, allowing us to keep the controller power efficiency closer to an analog approach while maintaining closed-loop optimality. A case study for an implementation in reconfigurable hardware shows how a designer can optimally trade closed-loop performance with hardware implementation performance.
15:00-15:20	FrB2.5
Embedded Predictive Control on an FPGA Using the Fast Gradient Method (I), pp. 3614-3620
Jerez, Juan Luis	Imperial Coll. London
Goulart, Paul J.	ETH Zurich
Richter, Stefan	ETH Zurich
Constantinides, George Anthony	Imperial Coll. London
Kerrigan, Eric C.	Imperial Coll. London
Morari, Manfred	ETH Zurich
Model predictive control (MPC) in resource-constrained embedded platforms requires faster, cheaper and more power-efficient solvers for convex programs than is currently offered by software-based solutions. In this paper we present the first field programmable gate array (FPGA) implementation of a fast gradient solver for linear-quadratic MPC problems with input constraints. We use fixed-point arithmetic to exploit the characteristics of the computing platform and provide analytical guarantees ensuring no overflow errors occur during operation. We further prove that the arithmetic errors due to round-off can lead only to reduced accuracy, but not instability, of the fast gradient method. The results are demonstrated on a model of an industrial atomic force microscope (AFM) where we show that, on a low-end FPGA, satisfactory control performance at a sample rate beyond 1 MHz is achievable, opening up new possibilities for the application of MPC.
15:20-15:40	FrB2.6
Monitoring Control Updating Period in Fast Gradient Based NMPC, pp. 3621-3626
Alamir, Mazen	CNRS-Gipsa-Lab.
In this paper, a method is proposed for on-line monitoring of the control updating period in fast-gradient-based Model Predictive Control (MPC) schemes. Such schemes are currently under intense investigation as a way to accommodate for real-time requirements when dealing with systems showing fast dynamics. The method needs cheap computations that use the algorithm on-line behavior in order to recover the optimal updating period in terms of cost function decrease. A simple example of constrained triple integrator is used to illustrate the proposed method and to assess its efficiency.

FrB3	HG F5
Genetic Regulatory Systems (Regular Session)
Chair: Goncalves, Jorge M.	Univ. of Cambridge
Co-Chair: Stan, Guy-Bart Vincent	Imperial Coll. London
13:40-14:00	FrB3.1
Network Reconstruction Using Knock-Out and Over-Expression Data, pp. 3627-3632
Hayden, David P.	Univ. of Cambridge
Yuan, Ye	Univ. of Cambridge
Goncalves, Jorge M.	Univ. of Cambridge
This paper outlines necessary and sufficient conditions for network reconstruction of linear, time-invariant systems using data from either knock-out or over-expression experiments. These structural system perturbations, which are common in biological experiments, can be formulated as unknown system inputs, allowing the network topology and dynamics to be found. We assume that only partial state measurements are available and propose an algorithm that can reconstruct the network at the level of the measured states using either time-series or steady-state data. A simulated example illustrates how the algorithm successfully reconstructs a network from data.
14:00-14:20	FrB3.2
Control of Protein Concentrations in Heterogeneous Cell Populations, pp. 3633-3639
Vignoni, Alejandro	Univ. Pol. de Valencia
Oyarzún, Diego A.	Imperial Coll. London
Picó, Jesús	Univ. Pol. de Valencia
Stan, Guy-Bart Vincent	Imperial Coll. London
In this work we propose a synthetic gene circuit for controlling the variability in protein concentration at a population level. The circuit, based on the use of an intracellular nonlinear controller coupled to a cell-to-cell communication mechanism, allows for independent control of the mean and variance of a signalling molecule across cell population. Via a piecewise affine approximation of the nonlinearity, we provide set invariance results that imply the stability of the closed loop system. We also obtain closed-form expressions for the mean and variance as a function of the tuneable parameters of the controller. The predictions offered by the theoretical analysis are in agreement with numerical simulations performed with physiologically realistic parameters in Escherichia coli.
14:20-14:40	FrB3.3
Multistability Equivalence between Gene Regulatory Networks of Different Dimensionality, pp. 3640-3645
Schittler, Daniella	Univ. of Stuttgart
Allgower, Frank	Univ. of Stuttgart
Waldherr, Steffen	Univ. Stuttgart
This paper formulates and addresses the problem of equivalence in terms of multistability properties between gene regulatory systems of different dimensionality. Given a dynamical system of a gene regulatory network, and the structure of another, higher-dimensional gene regulatory network, the aim is to find a dynamical system for the latter which has the same stability properties as the former. We propose construction rules for the dynamics of a high-dimensional system, given the low-dimensional system and the high-dimensional network structure. These construction rules yield a multistability-equivalent system, as we prove in this work. The usefulness of our method is illustrated with an exemplary gene regulatory network, for which we construct a dynamical system which has the same multistability properties as a given, low-dimensional model of gene regulation.
14:40-15:00	FrB3.4
Stochastic Simulation of Enzymatic Reactions under Transcriptional Feedback Regulation, pp. 3646-3651
Lugagne, Jean-Baptiste	INRIA Paris-Rocquencourt, 78153 Le Chesnay
Oyarzún, Diego A.	Imperial Coll. London
Stan, Guy-Bart Vincent	Imperial Coll. London
The interaction between gene expression and metabolism is a form of feedback control that allows cells to up- or downregulate specific reactions according to the environmental conditions. Although gene expression is an inherently stochastic process, the effect of genetic feedback on the propagation of noise to the metabolic layer remains largely unexplored. These systems operate in two timescales, and a major challenge is to devise stochastic simulation techniques that can cope with this stiffness in reasonable computational time. We propose a simulation technique, based on the slow-scale Stochastic Simulation Algorithm, to rapidly compute realizations of the Markov process associated to an enzymatic reaction under genetic feedback regulation. We show that in the case of constant substrate, the enzyme-substrate complexes have a binomial stationary distribution. With this result we can avoid the explicit simulation of the binding/dissociation of the enzyme and substrate, leading to a significant improvement in computational speed. We discuss the extension of the algorithm to networks of enzymatic reactions. The proposed method can be used to systematically compute the stationary distributions of the species for different combinations of model parameters, thus opening the way for the identification of the cellular processes that can modulate the amplification or attenuation of genetic noise in enzymatic reactions.
15:00-15:20	FrB3.5
Identification of Biological Models from Single-Cell Data: A Comparison between Mixed-Effects and Moment-Based Inference, pp. 3652-3657
Gonzalez Vargas, Andres Mauricio	Univ. degli studi di Pavia
Uhlendorf, Jannis	INRIA, Paris-Rocquencourt, France
Schaul, Joé	INRIA, Paris-Rocquencourt, France
Cinquemani, Eugenio	INRIA Grenoble - Rhone-Alpes
Batt, Gregory	INRIA Paris Rocquencourt
Ferrari Trecate, Giancarlo	Univ. degli Studi di Pavia
Experimental techniques in biology such as microfluidic devices and time-lapse microscopy allow tracking of the gene expression in single cells over time. So far, few attempts have been made to fully exploit these data for modeling the dynamics of biological networks in cell populations. In this paper we compare two modeling approaches capable to describe cell-to-cell variability: Mixed-Effects (ME) models and the Chemical Master Equation (CME). We discuss how network parameters can be identified from experimental data and use real data of the HOG pathway in yeast to assess model quality. For CME we rely on the identification approach proposed by Zechner <em>et al.</em> (PNAS, 2012), based on moments of the probability distribution involved in the CME. ME and moment-based (MB) inference will be also contrasted in terms of general features and possible uses in biology.
15:20-15:40	FrB3.6
State Estimation for Gene Networks with Intrinsic and Extrinsic Noise: A Case Study on E.coli Arabinose Uptake Dynamics, pp. 3658-3663
Carta, Alfonso	Inria Sophia Antipolis - Méditerranée
Cinquemani, Eugenio	INRIA Grenoble - Rhone-Alpes
We address state estimation for gene regulatory networks at the level of single cells. We consider models that include both intrinsic noise, in terms of stochastic dynamics, and extrinsic noise, in terms of random parameter values. We take the Chemical Master Equation (CME) with random parameters as a reference modeling approach, and investigate the use of stochastic differential model approximations for the construction of practical real-time filters. To this aim we consider a Square-Root Unscented Kalman Filter (SRUKF) built on a Chemical Langevin Equation (CLE) approximation of the CME. Using arabinose uptake regulation in Escherichia coli bacteria as a case study, we show that performance is comparable to that of a (computationally heavier) particle filter built directly on the CME, and that the use of information about parameter uncertainty allows one to improve state estimation performance.

FrB4	HG F7
Computation and Control (Regular Session)
Chair: Varga, Andreas	German Aerospace Center
Co-Chair: Svaricek, Ferdinand	Univ. of the German Armed Forces, Munich
13:40-14:00	FrB4.1
On Computing Inner-Outer Factorizations of Periodic Systems, pp. 3664-3669
Varga, Andreas	German Aerospace Center
A numerically reliable state space algorithm is proposed for computing inner-outer factorizations of causal periodic descriptor systems. The main computational ingredients are the computation of a special condensed Kronecker-like form of periodic pairs using orthogonal reduction algorithms and the solution of periodic Riccati equations. The proposed approach is completely general, being applicable to arbitrary causal periodic systems with time-varying state dimensions.
14:00-14:20	FrB4.2
Computing the Distance to Instability for Large-Scale Nonlinear Eigenvalue Problems, pp. 3670-3675
Michiels, Wim	K.U. Leuven
Guglielmi, Nicola	Univ. of L'Aquila
A quadratically converging algorithm for the computation of the distance to instability of a broad class of nonlinear eigenvalue problems is presented, including the polynomial eigenvalue problem and the delay eigenvalue problem. The algorithm is grounded in a recently presented approach for computing the pseudospectral abscissa. The application of the algorithm only relies on the availability of a method to compute the rightmost eigenvalue of perturbed problems obtained by adding rank one perturbations to the coefficient matrices, for which, in case of large and sparse matrices, efficient iterative algorithms can be used
14:20-14:40	FrB4.3
A Comparison of Methods for Higher-Order Numerical Differentiation, pp. 3676-3681
Listmann, Kim Daniel	ABB AG
Zhao, Zheng H.	ABB AG
This article compares three different methods for the online computation of higher-order derivatives from a measurement signal. In general such measurements are noise corrupted and the application of finite difference schemes is inappropriate. Thus, all methods are compared w.r.t. their noise suppression capabilities for different noise types and levels, their computational complexity to compute a derivative and their tuning effort for proper commissioning. Finally, a recommendation is provided which of the differentiators is best used when.
14:40-15:00	FrB4.4
Fast Jacobi-Type Algorithm for Computing Distances between Linear Dynamical Systems, pp. 3682-3687
Jimenez, Nicolas	Center for Imaging Science, Johns Hopkins Univ.
Vidal, Rene	Johns Hopkins Univ.
Afsari, Bijan	Johns Hopkins Univ.
The alignment distance is a novel metric between linear dynamical systems that has been shown to be very useful in many applications in computer vision. However, since the computation of the alignment distance requires solving a minimization problem on the orthogonal group, it is important to develop computationally efficient algorithms for solving this problem. In this paper, we present a fast and accurate Jacobi-type algorithm that solves this problem. Each step of the algorithm is equivalent to finding the roots of a quartic polynomial. We show that this rooting may be done efficiently and accurately using a careful implementation of Ferrari's classical closed-form solution for quartic polynomials. For linear systems with orders that commonly arise in computer vision scenarios, our algorithm is roughly twenty times faster than a fast Riemannian gradient descent algorithm implementation and has comparable accuracy.
15:00-15:20	FrB4.5
Computing the L-Infinity[0, H)-Induced Norm of a Compression Operator, pp. 3688-3693
Kim, Jung Hoon	Kyoto Univ.
Hagiwara, Tomomichi	Kyoto Univ.
This paper is concerned with the computation of the induced norm of a compression operator defined on the Banach space L-infinity[0,h), which is a difficult problem because it is an infinite-rank operator. This paper provides two methods for this problem, each of which can compute an upper bound and a lower bound of the induced norm by using an idea of staircase or piecewise linear approximation. Another key idea in both methods is to apply fast-lifting by which the interval [0,h) is divided into M subintervals with equal width, and the computation errors in these methods are ensured to be reciprocally proportional to M or M^2. The effectiveness of the proposed methods is demonstrated through a numerical example.
15:20-15:40	FrB4.6
Fractional Order PID Controller (FOPID)-Toolbox, pp. 3694-3699
Lachhab, Nabil	Federal Armed Forces Univ. Munich, Germany
Svaricek, Ferdinand	Univ. of the German Armed Forces, Munich
Wobbe, Frank	IAV GmbH
Rabba, Heiko	IAV GmbH
This paper presents a fractional order PID controller (FOPID)-Toolbox to design robust fractional PID controllers achieving a desired crossover frequency and a desired phase margin. A novel approach based on nonsmooth optimization techniques is used. The requirements to be fulfilled by the controller are expressed in terms of a desired open-loop response. Loop shaping configuration is used to synthesize the controller. To optimize the fractional orders an optimization algorithm based on the steepest descent method is used. Simulation results show the benefit of our method.

FrB5	HG E1.1
Switched Systems II (Regular Session)
Chair: Di Bernardo, Mario	Univ. of Naples Federico II
Co-Chair: Bolzern, Paolo	Pol. di Milano
13:40-14:00	FrB5.1
Hybrid Controllers for Mode-Observability of Switching Linear Systems: Existence and Genericity, pp. 3700-3705
Baglietto, Marco	Univ. of Genova
Battistelli, Giorgio	Univ. of Firenze
Tesi, Pietro	Univ. of Genoa
This paper presents recent developments in the study of non-autonomous switching linear systems. For such systems, we address the issue of how to systematically design linear controllers allowing the active process mode to be observable from closed-loop data. The results are stated formally by introducing the notions of feedback distinguishability and discerning controllers. Both existence and genericity problems are discussed. It is finally shown how a given family of discerning controllers can be implemented as a single hybrid controller which preserves the discerning capability of the 6original controllers.
14:00-14:20	FrB5.2
Incremental Stability of Planar Filippov Systems, pp. 3706-3711
Di Bernardo, Mario	Univ. of Naples Federico II
Liuzza, Davide	Univ. of Naples Federico II
We study the problem of proving incremental stability of a planar Filippov system. In particular, referring to systems that present an attractive sliding region on their discontinuity boundary, we will give a differential condition on such region able to guarantee incremental exponential stability of sliding mode trajectories. We will then derive conditions for the incremental stability of the whole system. The approach is based on using tools from contraction theory, extending their applicability to include discontinuous dynamical systems.
14:20-14:40	FrB5.3
Adaptive Control for State Dependent Switched Systems in Brunovsky Form, pp. 3712-3717
Angulo García, Fabiola	Univ. Nacional de Colombia
Di Bernardo, Mario	Univ. of Naples Federico II
Montanaro, Umberto	Italian National Res. Council
Rincon, Alejandro	Univ. Católica de Manizales
Santini, Stefania	Univ. di Napoli Federico II
In this work an adaptive strategy for state dependent switching systems in Brunovsky form is presented. Lyapunov direct method is used to derive the adaptive control law to track a given reference signal. The resulting controller is equipped with a set of adaptive gains able to tackle plant commutations. Control gains are adapted through a feedback law depending on the state space region visited by the system trajectories. A consistent proof of convergence of the tracking error is given and a numerical example is used to illustrate the analytical result.
14:40-15:00	FrB5.4
A Reduced Model of Three Ways Catalyst Converter and Stored Oxygen Rate Estimation Using Switched Observer, pp. 3718-3723
Ngo, Caroline	INP Grenoble
Koenig, Damien	Grenoble - Inp
Sename, Olivier	INPG
Bechart, Hubert	RENAULT
Pollutant emissions limitations ruled by the latest standards force car manufacturers to improve after treatment systems. This improvement is achieved by new systems on one hand, and by more efficient control strategies, on the other. This paper describes a 0-D model of Three Ways Catalyst converter (TWC) which is able to reproduce the transient dynamics of the converter although simple. A switched observer of stored oxygen rate is derived from this reduced model, and its convergence is proved. An estimation of the oxygen capacity of TWC provides precious information for on-board diagnosis or monitoring purpose. First, a description of the TWC dynamics and its reduced model will be described, then a discrete time switched observer of the stored oxygen rate and the free site capacity estimation by least squares method are presented.
15:00-15:20	FrB5.5
Mean Square Stability of Non-Homogeneous Markov Jump Linear Systems Using Interval Analysis, pp. 3724-3729
Chitraganti, Shaikshavali	Univ. de Lorraine
Aberkane, Samir	UHP, NANCY 1
Aubrun, Christophe	Univ. Henri Poincaré
This paper deals with a discrete-time Markov jump linear system with a non-homogeneous Markov Chain. In particular, we consider the situation when the transition probability matrix of the non-homogeneous Markov Chain is varying in an interval, and obtained a sufficient condition for the mean square stability of the proposed system.
15:20-15:40	FrB5.6
On the Interplay between Periodic Switches and Uncontrolled Jumps in Linear Discrete-Time Systems, pp. 3730-3735
Bolzern, Paolo	Pol. di Milano
Colaneri, Patrizio	Pol. di Milano
This paper deals with the analysis of stability and the characterization of input-output norms for dual switching linear systems in discrete-time. These systems are subject to two independent switching signals, taking values in finite sets, coming from different sources. The first switching signal is periodic while the second is uncontrolled and generate parameter jumps which are either completely arbitrary or satisfy a prescribed dwell-time constraint. The overall system is a linear time-varying system exhibiting a complex dynamic behavior due to the interplay between the periodic switches and the uncontrolled jumps. LMI conditions for stability and guaranteed H_2 norm input/output performances are provided. The results are obtained by merging the theory of linear periodic systems with recent developments on switched linear time-invariant systems.

FrB6	HG E1.2
Sliding Mode Control II (Regular Session)
Chair: Bandyopadhyay, Bijnan	IIT Bombay
Co-Chair: Fischer, Claudia	ETH Zurich
13:40-14:00	FrB6.1
Discrete-Time Sliding Mode Control of GMAW Systems Using Infrequent Output Measurements, pp. 3736-3741
Bera, Manas Kumar	Indian Inst. of Tech. Bombay
Priya, P.S Lal	Indian Inst. of Tech. Bombay
Bandyopadhyay, Bijnan	IIT Bombay
Paul, A.K	Welding Industry,Mumbai
Gas metal arc welding (GMAW) is the most popular industrial welding process today, preferred for its versatility, speed and relative ease of adapting to robotic automation. This process has been used extensively by many industrial environments especially, the automobile industry. The advantages of automated welding process with feedback controllers include increased productivity, consistency in welding quality, as well as health and safety benefits for the welder. This paper proposes a discrete time sliding mode controller, a highly robust controller, to control the welding current and arc voltage of a GMAW system using output feedback. The concept of multirate output feedback technique, with infrequent output measurements has been employed to design the robust controller. A linear multi input multi output (MIMO) system model has been considered here for the design. Since the proposed algorithm is based on output feedback, it is more practical in comparison to any state feedback based control algorithms. The performance of the controller is analyzed in the presence of model parameter uncertainties and the simulation results prove the efficiency of the controller proposed.
14:00-14:20	FrB6.2
Closed Loop NOx Control by Discrete Time Sliding Mode, pp. 3742-3747
Rafetzeder, Christoph	Johannes Kepler Univ. Linz
Stadlbauer, Stephan	Johannes Kepler Univ. Linz
Waschl, Harald	Johannes Kepler Univ. Linz
Del Re, Luigi	Johannes Kepler Univ. Linz
Nitrogen oxides (NOx) are besides particulate matter (PM), specially for diesel engines, the most relevant emission components and, for a given engine, reduction of one pollutant lead to increases emissions of the other. As passenger car emission legislation is permanently becoming stricter, there is a growing interest in fixing the instantaneous target of one variable while providing the requested torque and minimizing the other main pollutant. As engines are nonlinear highly dynamic systems, the design of a suitable controller is a complex task. In this paper we address the NOx emissions and propose a discrete time sliding mode controller (DSMC) developed to control the exhaust gas recirculation (EGR) valve in order to provide tracking of a time varying NOx reference while rejecting disturbances due to changes e.g. in engine speed and torque. The method presented in this paper avoids using intermediate variables like intake fresh air and inlet manifold pressure and is shown to work properly both in simulation and in measurements with a production Euro 5 engine.
14:20-14:40	FrB6.3
Terminal Sliding Mode Impedance Control for Bilateral Teleoperation under Unknown Constant Time Delay and Uncertainties, pp. 3748-3753
Vafaei, Alaleh	Univ. of Tehran
Yazdanpanah, M. J.	Tehran Univ.
Finite time control of a teleoperation system under unknown constant time delay and parametric uncertainties is considered using terminal sliding mode and impedance control. Terminal sliding mode controller is a finite time control scheme that provides finite time convergence to the equilibrium point as well as fast response and high tracking precision in comparison with conventional sliding mode controllers, which are widely used for control of teleoperators. The proposed control scheme for bilateral teleoperation under unknown constant time delay is a sliding mode-impedance control for master robot and a terminal sliding mode control for slave robot. These two control laws can provide perfect tracking and transparency that are main goals in bilateral teleoperations. Simulation results verify the effectiveness of the proposed scheme for bilateral teleoperation under time delay.
14:40-15:00	FrB6.4
Performance Analysis of Relay Feedback Position Regulators for Manipulators with Coulomb Friction, pp. 3754-3759
Aguilar, Luis T.	Inst. Pol. Nacional
Freidovich, Leonid	Umeå Univ.
Orlov, Yury	CICESE
Mérida, Jován O.	Inst. Pol. Nacional
The purpose of the paper is to analyze the performance of several global position regulators for robot manipulators with Coulomb friction. All the controllers include a proportional-differential part and a switched part whereas the difference between the controllers is in the way of compensation of the gravitational forces. Stability analysis is also revisited within the nonsmooth Lyapunov function framework for the controllers with and without gravity pre-compensation. Performance issues of the proposed controllers are evaluated in an experimental study of a five degrees-of-freedom robot manipulator. In the experiments, we choose two criteria for performance analysis. In the first set of experiments, we set the same gains to all the controllers. In the second set of experiments, the gains of the controller were chosen such that the work done by the manipulator is similar.
15:00-15:20	FrB6.5
Integral Higher Order Sliding Mode in Singular Optimal Stabilization, pp. 3760-3763
Fridman, Leonid M.	National Autonomous Univ. of Mexico
Jimenez-Lizarraga, Manuel	Univ. Autonoma de Nuevo Leon
Ibarra, Efrain	UANL
This paper explores the possibilities of the socalled Integral Higher-Order Sliding-Mode (IHOSM) control framework to solve singular optimal stabilization (SOS) problems of arbitrary order for uncertain systems. We connect the order of singularity of SOS with the co-dimension of the set of singular optimal manifold. IHOSM is designed allowing to reach SOM in prescribing time moment and ensures SOS. Moreover, this controller provides the insensitivity of trajectory w.r.t. matched bounded uncertainties.
15:20-15:40	FrB6.6
Time Convergence Estimation of a Perturbed Double Integrator: Family of Continuous Sliding Mode Based Output Feedback Synthesis, pp. 3764-3769
Raul Santiesteban, Raulcos	Inst. Tecnológico de Culiacán
In this paper mechanical systems of relative degree two, textit{i.e.} a perturbed double integrator is under study. A sliding mode based algorithm is under study using strict non smooth Lyapunov functions, such that compensation of growing perturbations together with state variables is shown. Indeed, the well known twisting algorithm and a generalized smooth family of this algorithm are considered. A strict non-smooth Lyapunov function is proposed allowing to design tuning rules for the gains of a family of controllers such that global exact finite time stability of the origin is shown. The proposed methodology estimate an upper bound for convergence time of the closed loop system in spite of growing perturbation with respect to the state. To illustrate performance and robustness properties a numerical experiment is presented, using one-link pendulum as a test bed.

FrB7	HG E3
Autonomous Maintenance of Model Based Control Systems (Invited Session)
Chair: Van den Hof, Paul M.J.	Eindhoven Univ. of Tech.
Co-Chair: Ozkan, Leyla	Eindhoven Univ. of Tech.
Organizer: Ozkan, Leyla	Eindhoven Univ. of Tech.
Organizer: Van den Hof, Paul M.J.	Eindhoven Univ. of Tech.
13:40-14:00	FrB7.1
Performance Monitoring for Model Predictive Control Maintenance (I), pp. 3770-3775
Modén, Per Erik	ABB
Lundh, Michael	ABB
An economically motivated performance measure is proposed for use with model predictive control applications. The typical case handled is that you have a trade-off situation between two conflicting goals: (i) the need to be close to a constraint, since this will give good production economy, and (ii) the need to avoid constraint violation, since that would mean producing an inferior quality, exceeding environmental constraints (thereby requiring a fee), or having some other costly drawback. The proposed performance measure uses consumption and production rates in the process and the corresponding cost and benefit factors, combined with the risks of constraint violations and their associated costs. There is a discussion on the hard task of choosing threshold values, and the approach is illustrated by applying it to a couple of examples. One of the examples is the Autoprofit benchmark pulp digester model.
14:00-14:20	FrB7.2
Perspectives of Data-Driven LPV Modeling of High-Purity Distillation Columns (I), pp. 3776-3783
Bachnas, Ahmad Alrianes	Eindhoven Univ. of Tech.
Tóth, Roland	Eindhoven Univ. of Tech.
Mesbah, Ali	Massachusetts Inst. of Tech.
Ludlage, Jobert	Delft Univ. of Tech.
This paper investigates data-driven, Linear-Parameter-Varying (LPV) modeling of a high-purity distillation column. Two LPV modeling approaches are studied: a local approach, corresponding to the interpolation of Linear Time-Invariant (LTI) models identified at steady-state purity levels, and a global Least-Square Support Vector Machine (LS-SVM) approach which offers non-parametric estimation of the system w.r.t. data with varying operating conditions. In an extensive simulation study, it is observed that the global LS-SVM approach outperforms the local methodology in capturing the dynamics of the high-purity distillation column under study. The simulation results suggest that the global LS-SVM approach provides a reliable modeling tool under realistic noise conditions.
14:20-14:40	FrB7.3
Asymptotic Behaviour of Toeplitz Matrix in Multi-Input Multi-Output Model Predictive Control (I), pp. 3784-3789
Tran, Quang N.	Eindhoven Univ. of Tech.
Ozkan, Leyla	Eindhoven Univ. of Tech.
Ludlage, Jobert	Delft Univ. of Tech.
Backx, Ton	Eindhoven Univ. of Tech.
The singular value decomposition (SVD) of the Toeplitz matrix in the quadratic performance index of Model Predictive Control (MPC) is studied. The underlying goal is to find connection between the frequency domain information and the finite time optimal control and use this connection as a basis for stability, robust performance analysis and tuning of the dynamic MPC criterion. In a recent work by the authors, it was shown that the singular value decomposition of the Toeplitz matrix provides gain and phase information of the associated system for sufficiently long prediction and control horizons. This work is extended to MIMO case and is shown that singular value decomposition of the Toeplitz matrix can be used for stability analysis of closed loop system.
14:40-15:00	FrB7.4
Model Predictive Control with Integrated Experiment Design for Output Error Systems (I), pp. 3790-3795
Larsson, Christian A.	KTH Royal Inst. of Tech.
Annergren, Mariette	KTH Royal Inst. of Tech.
Hjalmarsson, Håkan	Royal Inst. of Tech.
Rojas, Cristian R.	KTH - Royal Inst. of Tech.
Bombois, Xavier	Delft Univ. of Tech.
Mesbah, Ali	Massachusetts Inst. of Tech.
Modén, Per Erik	ABB
Model predictive control has become an increasingly popular control strategy thanks to the ability to handle constrained systems. Obtaining the required models through system identification is often a time consuming and costly process. Applications oriented experiment design is a means of reducing this effort but is often formulated in terms of the input's spectral properties. Therefore, time domain constraints are difficult to enforce. In this contribution we combine MPC with experiment design to formulate a control problem where excitation constraints are included. The benefits are that time domain constraints are respected while the experiment design criteria are fulfilled. The method is evaluated on a numerical example.
15:00-15:20	FrB7.5
Fault-Tolerant Servo Systems against Sensor Failures Using Limited Integrators, pp. 3796-3802
Suyama, Koichi	Tokyo Univ. of Marine Science and Tech.
Sebe, Noboru	Kyushu Inst. of Tech.
No effective controller design for achieving the tolerance against sensor failures has been proposed for servo systems until now because it is difficult to deal with ramp signals generated by integrators in the servo performance. In this paper, we obtain the condition for the limiting function in limited integrators to behave effectively as the measure against ramp signals for the first time. Then, the limiting function and a controller can be designed in a unified framework of simultaneous optimization with consideration given to the servo performance in the normal case and in faulty cases. The obtained result succeeds in establishing the potential applicability of limited integrators to fault-tolerant servo systems.
15:20-15:40	FrB7.6
Applicative Fault Tolerant Control for Semi-Active Suspension System : Preliminary Results, pp. 3803-3808
Varrier, Sébastien	Gipsa-Lab.
Vivas-Lopez, Carlos	Tecnologico de Monterrey, Campus Monterrey
Lozoya-Santos, Jorge de-J	Tecnologico de Monterrey, Campus Monterrey
Koenig, Damien	Grenoble - Inp
Martinez, John-Jairo	Gipsa-Lab. INP-Grenoble
Morales-Menendez, Ruben	Tecnológico de Monterrey, Campus Monterrey
Tudon-Martinez, Juan Carlos	Tecnológico de Monterrey
This paper presents applicative results about fault detection and control for semi-active suspension system. The control is extended from the “Acceleration Driven Damping” (ADD) controller combined with a robust fault detection scheme in order to estimate the fault. The approach, based on a quarter of vehicle model includes the nonlinearities of the shock absorber. The robust fault estimation module is based on H1 background providing robustness. It aims to estimate a sensor fault on the system, especially a bias on accelerometers. Combination of both strategies allows to attenuate the effect the fault on the system. This approved structure aims to be implemented in real time. Validations has been made in a F-Class vehicle in CarsimTM software which provides realistic non-linear vehicle behavior. Results show the effectiveness of the fault tolerant strategy for semi-active dampers.

FrB8	HG E5
Synchronization in Networks (Regular Session)
Chair: Lunze, Jan	Ruhr-Univ. Bochum
Co-Chair: Cao, Ming	Univ. of Groningen
13:40-14:00	FrB8.1
Graph Comparison and Its Application in Network Synchronization, pp. 3809-3814
Liu, Hui	Univ. of Groningen
Cao, Ming	Univ. of Groningen
Wu, Chai Wah	IBM T. J. Watson Res. Center
Using the tool of graph comparison from spectral graph theory, we propose new methodologies to guarantee complete synchronization in complex networks. The main idea is to utilize flexibly topological features of a given network so that the eigenvalues of the Laplacian matrix associated with the network can be estimated. The proposed methodologies enable the construction of different coupling-strength combinations in response to different knowledge about sub-networks. The obtained bounds of the network graphs' eigenvalues can be further used to study the robust synchronization problem in face of link failures in networks. Examples are utilized to demonstrate how to apply the methodologies to networks.
14:00-14:20	FrB8.2
Optimal Synchronization of Circulant Networked Multi-Agent Systems, pp. 3815-3820
Mosebach, Andrej	Ruhr-Univ. Bochum
Lunze, Jan	Ruhr-Univ. Bochum
This paper investigates the synchronization problem of identical agents while minimizing an LQ-based performance index that reflects the behavior of the synchronization errors. For an objective function which penalizes the synchronization errors with a circulant structured weighting matrix, an optimal networked controller is proposed which has the same structure. The resulting communication graph is a linear combination of circulant subgraphs which couple all agents with all other agents.
14:20-14:40	FrB8.3
Robustness of Synchronization in Heterogeneous Multi-Agent Systems, pp. 3821-3826
Kim, Jaeyong	Seoul National Univ.
Yang, Jongwook	Seoul National Univ.
Shim, Hyungbo	Seoul National Univ.
Kim, Jung-Su	Seoul National Univ. of Science and Tech.
This paper studies robustness of synchronization in heterogeneous multi-agent systems, which is gained by interactions with other agents through the network. In order to effectively deal with the heterogeneous cases, we introduce the concept of the averaged dynamics which is the average of all agents' dynamics, and then claim that two sources enhance the robustness of the group behavior against differences among agents. First, we show that strong coupling among heterogeneous agents makes the trajectories of all agents remain in an arbitrarily small neighborhood of the trajectory of the averaged dynamics. Second, we observe that the amount of individual variation of each agent, that contributes to the averaged dynamics, gets smaller as the number of agents increases, and thus, the averaged dynamics becomes more robust to the differences among agents. Simulation results confirm our claim that a large number of agents with strong couplings have robust synchronization.
14:40-15:00	FrB8.4
Regulated Output Synchronization for Heterogeneous Networks of Non-Introspective, Minimum-Phase SISO Agents without Exchange of Controller States, pp. 3827-3832
Grip, Håvard Fjær	NTNU
Saberi, Ali	Washington State Univ.
Stoorvogel, Anton A.	Univ. of Twente
In this paper we study the problem of achieving regulated output synchronization in a network of minimum-phase SISO agents. Our problem formulation is characterized by the combination of three different challenges: the network is heterogeneous (meaning that the agents are governed by non-identical models); the agents are non-introspective (meaning that they do not have access to information about their own state or output); and the agents are not allowed to exchange internal controller states via the network. To handle these challenges, we present an observer-based control methodology that combines elements of low-gain and high-gain design techniques.
15:00-15:20	FrB8.5
Attitude Synchronization of Spacecraft Formation with Adaptation of Consensus Penalty Terms, pp. 3833-3838
Zhang, Kewen	Worcester Pol. Inst.
Demetriou, Michael A.	Worcester Pol. Inst.
The main thrust of this work is on the time variation of the consensus weights used to enforce agreement (synchronization) amongst a network of spacecraft. The main idea is to allow a flexibility in the consensus gains to adapt in proportion to the Euclidean distance of the spacecraft states. A prelude to the time adaptation of the consensus controllers is also provided and takes the form of considering different gains in the synchronization signal thereby arriving in two cases: node-dependent and edge-dependent synchronization gains. A way to choose these fixed gains is in terms of the initial mismatch between the pairwise difference of the spacecraft states. Further, an adaptation of the synchronizing controllers is proposed and compared to the fixed gains. Extensive numerical studies are provided to further support the theoretical predictions and provide insights on the choice of consensus gains in synchronization control used for spacecraft formation.
15:20-15:40	FrB8.6
Synchronization of Dynamical Networks under Sampling, pp. 3839-3844
Giraldo, Jairo	Univ. de los Andes
Mojica-Nava, Eduardo	National Univ. of Colombia
Quijano, Nicanor	Univ. de Los Andes
Motivated by the increasing interest in networked multi-agent systems and the wide number of applications in decentralized distributed control of smart grids, we address the problem of synchronization of microgrids (MGs), when each node (microgrid)is modeled as a linear-time continuous system whose output measurements are sent through communication links. However, the inclusion of a communication infrastructure adds new challenges to control strategies and some problems may arise such as time delays, packet losses, sampling period, just to name a few. In this work, we consider that data is sampled with homogeneous sampling periods. Then, using the novel concept of average passivity, we define the conditions for synchronizability when all nodes are identical and unstable dynamics are present. Additionally, results are extended to the case of non-uniform agents, and some simulations of synchronization in smart grids are introduced.

FrB9	HG E7
Nonlinear Control of UAVs (Regular Session)
Chair: Lozano, Rogelio	Univ. de Tech. de Compiegne
Co-Chair: Nikolakopoulos, George	Luleå Univ. of Tech. Sweden
13:40-14:00	FrB9.1
Geometric Nonlinear PID Control of a Quadrotor UAV on SE(3), pp. 3845-3850
Abootorabi Goodarzi, Farhad	The George Washington Univ.
Lee, Daewon	The George Washington Univ.
Lee, Taeyoung	George Washington Univ.
Nonlinear PID control systems for a quadrotor UAV are proposed to follow an attitude tracking command and a position tracking command. The control systems are developed directly on the special Euclidean group to avoid singularities of minimal attitude representations or ambiguity of quaternions. A new form of integral control terms is proposed to guarantee almost global asymptotic stability when there exist uncertainties in the quadrotor dynamics. A rigorous mathematical proof is given. Numerical example illustrating a complex maneuver, and a preliminary experimental result are provided.
14:00-14:20	FrB9.2
Fast and Saturating Attitude Control for a Quadrotor Helicopter, pp. 3851-3857
Fritsch, Oliver	Tech. Univ. München
Henze, Bernd	Tech. Univ. München
Lohmann, Boris	Tech. Univ. Muenchen
In this paper a continuous, nonlinear state feedback attitude controller for a quadrotor helicopter is presented. The design is based on an energy shaping approach and prioritizes the alignment of the thrust direction due to its prominent role for the translational dynamics. Since the controller tends to saturate the control torques, it achieves short settling times. Almost global asymptotic stability of the desired attitude is proven for arbitrary unknown moment of inertia matrices and the performance of the closed loop system is analyzed by simulations.
14:20-14:40	FrB9.3
Quadrotors Flight Formation Control Using a Leader-Follower Approach, pp. 3858-3863
Mercado Ravell, Diego Alberto	Univ. de Tech. Compiegne
Castro-Linares, Rafael	CINVESTAV-IPN
Lozano, Rogelio	Univ. de Tech. de Compiegne
In this paper it is presented a control strategy to solve the trajectory tracking and flight formation problem, in horizontal plane, of multiple unmanned aerial vehicles (UAVs) kind quadrotor, by means of a leader-follower scheme. Time scale separation of the translational and rotational quadrotor dynamics is used to achieve trajectory tracking. A sliding mode controller is proposed for the translational dynamic and provides the desired orientation for the UAV, which is controlled by a linear PD control. Finally, from the formation error dynamics, a sliding mode control is used by the follower to preserve the formation with respect to a leader. Experimental results, using a virtual leader and a follower in formation, are shown to evaluate the proposed control law.
14:40-15:00	FrB9.4
Full Quaternion Based Attitude Control for a Quadrotor, pp. 3864-3869
Fresk, Emil	Luleå Univ. of Tech.
Nikolakopoulos, George	Luleå Univ. of Tech. Sweden
The aim of this article is to present a novel quaternion based control scheme for the attitude control problem of a quadrotor. A quaternion is a hyper complex number of rank 4 that can be utilized to avoid the inherent geometrical singularity when representing rigid body dynamics with Euler angles or the complexity of having coupled differential equations with the Direction Cosine Matrix (DCM). In the presented approach both the quadrotor's attitude model and the proposed non-linear Proportional squared (P^2) control algorithm have been implemented in the quaternion space, without any transformations and calculations in the Euler's angle space or DCM. Throughout the article, the merits of the proposed novel approach are being analyzed and discussed, while the efficacy of the suggested novel quaternion based controller are being evaluated by extended simulation results.
15:00-15:20	FrB9.5
Nonlinear Control of the Longitudinal Dynamics and the Rotor Dynamics of a Small Scale Helicopter, pp. 3870-3875
Benítez-Morales, José Gerardo	Centro de Investigación y Estudios Avanzados
Rodriguez-Cortes, Hugo	Northeastern Univ.
Castro-Linares, Rafael	CINVESTAV-IPN
A nonlinear control scheme is proposed for the trajectory tracking problem of a small scale helicopter's longitudinal dynamics. The control scheme is based on a control design procedure that constructs static stabilizers for nonlinear systems which are linearizable by dynamic feedback. The flatness properties of the helicopter's longitudinal dynamics are exploited to design the desired trajectory.Sufficient conditions are given to guarantee asymptotic convergence to zero of the tracking error assuming that all the helicopter's parameters are known and that all states are measured. Numerical simulations are provided to illustrate the properties of the proposed controller.
15:20-15:40	FrB9.6
Time-Parametrization Control of Quadrotors with a Robust Quaternion-Based Sliding Mode Controller for Aggressive Maneuvering, pp. 3876-3881
Sanchez, Anand	CINVESTAV
Parra Vega, Vicente	Centro de Investigación y de Estudios Avanzados del IPN(Unidad S
Garcia Salazar, Octavio	Aerospace Engineering Res. and Innovation Center (CIIIA) - U
Ruiz Sanchez, Francisco Jose	CINVESTAV Saltillo
Ramos Velasco, Luis Enrique	Univ. Pol. de Pachuca
In this paper, we introduce the concept of time parametrization for Quadrotor flight maneuvers establishing a novel, robust and model free feedback controller based on a quaternion representation without singularities. Three prominent features are remarked: firstly, the control algorithm assures exponential stability of the full position/attitude dynamics of the system with smooth control efforts. Secondly, the closed-loop system is robust in presence of external forces and induced moments generated during the flight maneuvers. Finally, the controlled Quadrotor offers capabilities for aggressive maneuvers. Additionally, based on time-based generators, the sliding mode gives rise to well-posed terminal stability, parametrized by a desired convergence time defined by the user, independently of initial conditions and feedback gains. This allows tasks, such as cooperative Quadrotor, or force and interaction task, to be solved on a nonlinear and simplified setting. Simulations are presented in several scenarios in order to carry out a wide idea of our approach.

FrB10	HG D1.1
Linear Matrix Inequalities (Regular Session)
Chair: Scherer, Carsten W.	Univ. of Stuttgart
Co-Chair: Szabo, Zoltan	MTA SZTAKI
13:40-14:00	FrB10.1
Iterated Approximate Value Functions, pp. 3882-3888
O'Donoghue, Brendan	Stanford Univ.
Wang, Yang	Stanford Univ.
Boyd, Stephen P.	Stanford Univ.
In this paper we introduce a control policy which we refer to as the iterated approximate value function policy. The generation of this policy requires two stages, the first one carried out off-line, and the second stage carried out on-line. In the first stage we simultaneously compute a trajectory of moments of the state and action and a sequence of approximate value functions optimized to that trajectory. The next stage is to perform control using the generated sequence of approximate value functions. This yields a time-varying policy, even in the case where the optimal policy is time-invariant. We restrict our attention to the case with linear dynamics and quadratically representable stage cost function. In this case the pre-computation stage requires the solution of a semidefinite program (SDP). Finding the control action at each time-period requires solving a small convex optimization problem which can be carried out quickly. We conclude with some examples.
14:00-14:20	FrB10.2
Homogeneous Polynomially Parameter-Dependent State Feedback Controllers for Finite Time Stabilization of Linear Time-Varying Systems, pp. 3889-3894
Borges, Renato Alves	Univ. of Brasilia
Ishihara, João Yoshiyuki	Univ. of Brasília
Kussaba, Hugo Tadashi	Univ. of Brasília
Silva, Laríssa de Paiva	Univ. of Brasilia
This paper investigates the problem of parameter-dependent state feedback control of continuous-time systems in the context of finite time stability. The controller is designed in order to guarantee that the closed-loop system is finite time stable. The system is considered time varying with the parameters modeled within a unit simplex. The design conditions obtained by means of Lyapunov functions are expressed as linear matrix inequalities. The finite time stability is assessed by using homogeneous polynomially parameter-dependent state feedback gains with arbitrary degree g. LMI relaxations are proposed based on Pólya's theorem. A controller is obtained by the solution of a factibility problem and the effect of the relaxation procedures analysed by an optimization problem. Numerical examples are provided.
14:20-14:40	FrB10.3
LMI & BMI Technics for the Design of a PI Control for Irrigation Channels, pp. 3895-3900
Rodrigues, Mickael	Univ. CLAUDE BERNARD LYON 1
Wu, Yongxin	Univ. Claude Bernard Lyon 1
Aberkane, Samir	UHP, NANCY 1
Dos Santos Martins, Valérie Sylvie	Univ. Claude Bernard Lyon 1
This paper considers the problem of control design for a nonlinear distributed parameter system in infinite dimension which is described by hyperbolic Partial Differential Equations (PDEs) of de Saint-Venant. For describing the dynamic of this nonlinear system over a wide operating range, the Multi-Models approach, which takes into account Linear Time Invariant (LTI) models defined around a set of operating points, has been used. By means of an Internal Model Boundary Control (IMBC), a new design of Proportional Integral (PI) feedback is performed through Bilinear Matrix Inequality (BMI) and Linear Matrix Inequality (LMI) technics. The new results have been simulated and also compared to the previous results, illustrating the new theoretical contribution.
14:40-15:00	FrB10.4
Gain-Scheduled Synthesis with Dynamic Stable Strictly Positive Real Multipliers: A Complete Solution, pp. 3901-3906
Scherer, Carsten W.	Univ. of Stuttgart
In this paper we continue our efforts to arrive at a solution of the gain-scheduling controller design problem with dynamic multipliers. Although a solution for D-scalings is available, this scenario leads to conservatism if scheduling on real parameters. This motivates to consider the very same problem with D/G-scalings, for which we recently proposed a design framework with generalized positive real multipliers. The present paper closes a gap in this previous approach and provides necessary and sufficient convex conditions for the existence of gain-scheduling controllers if the multipliers are stable and positive real. Examples illustrate the possibility for a substantial reduction of conservatism with the novel solution.
15:00-15:20	FrB10.5
Multivariate S-Procedure, pp. 3907-3912
Szabo, Zoltan	MTA SZTAKI
Biro, Zsolt	MTA SZTAKI
Bokor, Jozsef	Hungarian Acad. of Sciences
Different variants of the S-procedure provides a very important tool in robust stability and robust performance analysis. Concerning performance assessment this paper shows that the design framework based on the full block S-procedure (extended KYP lemma) contains an inherent conservativeness. The main result of the paper is a multivariate version of the classical S-procedure, stated for negative graph subspaces. A new solvability condition for the Elimination lemma is also provided.
15:20-15:40	FrB10.6
Generalizing the KYP Lemma to the Union of Intervals, pp. 3913-3918
Pipeleers, Goele	KU Leuven
Iwasaki, Tetsuya	UCLA
Hara, Shinji	The Univ. of Tokyo
A recent generalization of the Kalman-Yakubovich-Popov (KYP) lemma establishes the equivalence between a semi-infinite inequality on a segment of a circle or straight line in the complex plane and a linear matrix inequality. In this paper we further generalize the KYP lemma to particular curves in the complex plane, which include the union of segments of a circle or line as a special case.

FrB11	HG D1.2
Process Control (Regular Session)
Chair: Vilanova, Ramon	UAB
Co-Chair: Hahn, Juergen	Rensselaer Pol. Inst.
13:40-14:00	FrB11.1
Automatic Feedforward Tuning for PID Control Loops, pp. 3919-3924
Veronesi, Massimiliano	Yokogawa Italy
Visioli, Antonio	Univ. of Brescia
In this paper we propose a method for the automatic design of the feedforward compensator for PID control loop where the process is affected by a disturbance. The parameters of the compensator are automatically computed after the disturbance is rejected for the first time by a simple PID controller which can be roughly tuned. Further, a technique to assess the performance of the compensator is also proposed. Simulation and experimental results show the effectiveness of the methodology.
14:00-14:20	FrB11.2
Paradigms for Unified Runtime Systems in Industrial Automation, pp. 3925-3930
Gruener, Sten	RWTH Aachen Univ.
Epple, Ulrich	RWTH Aachen Univ.
Borders between different levels of the industrial automation pyramid become more and more blurry because of the increasing production flexibility and plant complexity. The arising challenges can be overcome by the use of the same runtime systems and programming paradigms for distributed automation at the manufacturing execution system, the process control and the field device levels. Paradigms for unified runtime system are proposed involving message-based communication, co-existence of cyclic/event-based execution semantics of function block diagrams, a flexible hierarchical tasking concept, development workflow and non-functional system properties.
14:20-14:40	FrB11.3
Model Reference Based Robust Tuning of Five-Parameter 2DoF PID Controllers for First-Order Plus Dead-Time Models, pp. 3931-3936
Alfaro, Victor M.	Univ. of Costa Rica
Vilanova, Ramon	UAB
The aim of this paper is to present a robust tuning method for two-degree-of-freedom proportional integral derivative controllers for first-order plus dead time controlled processes. A two-degree-of-freedom Ideal PID with filter control algorithm is selected and its complete set of parameters (included the filter time constant) is obtained. This is based on the use of a model reference optimization procedure with servo and regulatory target closed-loop transfer functions. The designer is allowed to select a maximum sensitivity MS from 2.0 to 1.4.
14:40-15:00	FrB11.4
A Tdof Pid Control System Design by Referring to the Md-Pid Control System and Its Sensitivities, pp. 3937-3942
Shigemasa, Takashi	TOSHIBA MITSUBISHI-ELECTRIC INDUSTRIAL SYSTEMS Corp.
Negishi, Yasunori	TOSHIBA MITSUBISHI-ELECTRIC INDUSTRIAL SYSTEMS Corp.
Baba, Yasushi	Toshiba IT & Control Systems Corp.
PID controls are widely used as a basic control technology in the industrial control system today. However, tuning of PID control systems is not always easy, because of their simple control structures for wide classes of industrial control processes. The tuning of a two-degree-of-freedom (TDOF) PID control system which is increased configuration parameters, is more difficult. In order to obtain good control performances for wide classes of process control systems, we developed a Model-Driven PID(MD-PID) control system about ten years ago, by extending the "Model Driven control" concept proposed by Kimura. The MD-PID control system has a cascade control structure that is an Internal Model Control(IMC) by Morari as the main controller and a PD control loop as the lower loop including the control process. In this paper, after reviewing the design method for a MD-PID Control system with new PD feedback design method, we propose a new design method for a TDOF PID control system by referring to the Model-Driven PID control system and its sensitivities.
15:00-15:20	FrB11.5
Computing Optimal Operating Condition Profiles for Fed-Batch Fermentation of Fuel-Grade Ethanol, pp. 3943-3948
Dai, Wei	Rensselaer Pol. Inst.
Hahn, Juergen	Rensselaer Pol. Inst.
This paper investigates optimization of operational strategies of an industrial ethanol fermentation process. One of the challenges associated with this type of process is that most of the measurements are only taken sporadically, thereby, complicating process monitoring and optimization. The one exception to this rule involves temperature measurements, which are readily available. However, existing models used in industry do not include an energy balance and, accordingly, the temperature measurements cannot be used to estimate model parameters. This paper addresses these deficiencies and proposes modifications to an existing ethanol fermentation model. The proposed changes include the derivation of an energy balance, modification of the reaction kinetics to include additional inhibition terms, and also estimation of model parameters from industrial data. The new model is validated against plant data and then used for optimization of the process operations. It is shown that modifications of the input profiles for the cooling rate and the glucoamylase addition can lead to an approximately 10% increase in ethanol yield. These are promising results, even though these findings will ultimately need to be validated during real plant operations.
15:20-15:40	FrB11.6
ANOVA Model Based Moving Window Approach for RtR Control in High-Mix Semiconductor Manufacturing Industry, pp. 3949-3954
Ling, Dan	Huazhong Univ. of Science and Tech.
Zheng, Ying	Huazhong Univ. of Science and Tech.
Fang, Huajing	Huazhong Univ. of Science and Tech.
Fan, Huijin	Huazhong Univ. of Sci. & Tech.
Zhao, Jin	Huazhong Univ. of Science and Tech.
Abstract—In this work, a new state estimation method based on run-to-run control for high-mix semiconductor manufacturing industry is proposed. This algorithm exerts moving window approach on the analysis of variance (ANOVA) model, MW-ANOVA approach, and employs the right pseudo-inverse of the context matrix to provide better estimate of the relative states including disturbances of each product and tool at next run. This approach takes the commonly encountered non-stationary disturbances, i.e., an integrated moving average with first order IMA(1,1) disturbance, into account. Comparing to the traditional t-EWMA algorithm, it provides better control performance, which can be verified by simulation examples.

FrB12	HG D3.2
System Modeling (Regular Session)
Chair: Abel, Dirk	RWTH Aachen Univ.
Co-Chair: Privara, Samuel	CzechTechnicalUniversityinPrague,FacultyofElectricalEngineering
13:40-14:00	FrB12.1
Building Semi-Physical Modeling: On Selection of the Model Complexity, pp. 3955-3960
Vana, Zdenek	Department of Control Engineering, FacultyofElectricalEngineerin
Privara, Samuel	CzechTechnicalUniversityinPrague,FacultyofElectricalEngineering
Zacekova, Eva	Department of Control Engineering, FacultyofElectricalEngineerin
Cigler, Jiri	Czech Tech. Univ. in Prague, Faculty ofElectricalEngine
Buildings account for significant amount of final energy consumption and therefore there is an intensive research aimed at its optimization. Predictive control has become a very popular approach in many industries buildings included. The main bottleneck of this method is a need for a good model. There are many different identification frameworks, plenty of methods and approaches, some of them more or less suitable for the building modeling. A common situation is that there are number of models at hand (often of different complexities), and there is a need for selection of the ``best'' model for predictive control. A logical choice is to start testing the statistical significance of an additional complexity (in a sense of the structure and number of disturbance inputs) of the model. This paper proposes a systematic way of building-up the model, starting from a simple structure. Then, more complex models are considered in an iterative manner. In each iteration, the statistical significance of the additional information due to the more complex model is checked. The procedure stops when selecting more complex models brings no quality improvements. In this paper, a semi-physical modeling using CTSM and a model selection based on statistical tests are presented. Finally, the properties of the proposed algorithm are investigated.
14:00-14:20	FrB12.2
Greybox Modeling of the Diesel Combustion by Use of the Scalar Dissipation Rate, pp. 3961-3966
Zweigel, Rene	Inst. of automatic control, RWTH Aachen Univ.
Albin, Thivaharan	RWTH Aachen Univ.
Hesseler, Frank-Joseph	Inst. of automatic control, RWTH Aachen Univ.
Abel, Dirk	RWTH Aachen Univ.
This paper deals with the greybox modeling of a diesel combustion process for the purpose of model-based combustion control. For this a real-time capable model with a high prediction quality and a high interpolation and extrapolation capability is required. An existing neural network model of the process will be introduced. The model inputs are the quantity of injected fuel, start of injection and the intake manifold fraction of recirculated exhaust gas. The variables to be predicted are the position of the combustion average, the indicated mean effective pressure and the maximum cylinder pressure gradient which correlates with the combustion noise of the engine. To enhance the existing model an analytic description of the scalar dissipation rate is used as a new input into the neural network. The aim here is to decrease the model error, improve the model robustness and reduce the complexity of the neural network. The scalar dissipation rate descripes the diffusivity in a combustion mixture and represents a combustion characteristic value. For a couple of engine working points the scalar dissipation rate is calculated by a 3D CFD simulation. Afterwards the calculation results are disposed to find an analytic description. To use this new model information a restructuring of the neural network is necessary. The modeling results will be shown and compared to the existing blackbox model.
14:20-14:40	FrB12.3
A High-Fidelity Simulation Model for an Asymmetric Nonlinear Electrothermal Actuator, pp. 3967-3972
Muffato, Leonardo Angelo	SAUTER AG
Kunz, Dominique-Stephan	SAUTER AG
In this paper a high-fidelity single-input and single-output (SISO) simulation model for an asymmetric nonlinear electrothermal actuator is presented. The model was mainly developed using grey-box system identification and considers energy conversion and energy flow between electrical, thermal and mechanical domains. It is particularly useful for investigation of different control strategies (control performance, control stability, actuator life cycle and energy consumption) for plants employing SAUTER AXT211 thermal actuator. Some insights are given in the process of system identification, model optimization and validation using measured data. Correlation between simulation results and measured data is investigated, as well as simulation model confidence.
14:40-15:00	FrB12.4
Identifying Second-Order Models of Mechanical Structures in Physical Coordinates: An Orthogonal Complement Approach, pp. 3973-3978
Ramos, Jose A.	Nova Southeastern Univ.
Mercère, Guillaume	Poitiers Univ.
Prot, Olivier	Univ. de Limoges
The problem of identifying the mass, damping, and stiffness matrices of a mechanical structure is a well known constrained system identification problem in the literature. The constraints come from the symmetry of the mass, damping, and stiffness matrices, as well as the number of sensors and actuators placed on the structure. Here we present two solutions to this problem, one based on a structured system identification approach and the other based on a similarity transformation approach. The latter approach takes advantage of the non-uniqueness of the problem to force the solution to a particular basis. Examples of both approaches show the feasibility of the methods, and it is expected to shed light on solving the most restrictive of the structural identification class of problems.
15:00-15:20	FrB12.5
Modeling and Online-Identification of Electrically Stimulated Antagonistic Muscles for Horizontal Shoulder Abduction and Adduction, pp. 3979-3984
Spagnol, Pierfrancesco	Pol. di Milano
Klauer, Christian	Tech. Univ. Berlin
Previdi, Fabio	Univ. degli Studi di Bergamo
Raisch, Joerg	Tech. Univ. Berlin
Schauer, Thomas	Tech. Univ. Berlin
A comparison of two different models of a pair of antagonistic muscles for horizontal shoulder abduction and adduction is presented. The proposed models are based on the so called Hill-model: a mechanical framework (inertia, dampers and springs) is used for their development. The models consider as inputs the estimates of the activation states of the muscles based on digital filtering of the evoked electromyogram (eEMG). Model outputs are angular velocity and position. Both models show good results in terms of performance, but they are different in terms of number of parameters that need to be identified and in terms of physical interpretation. One of the models, in fact, describes the muscle as a spring that generates torque by changing its stiffness parameter depending on its activation level. In order to enable adaptive model-based feed-forward and feedback control strategies for angular position/velocity control, an online-identification method based on an Extended Kalman Filter (EKF) is introduced for one of the two models. Simulation and experimental results show the good performance in terms of convergence time and accuracy of the estimation.

FrB13	HG D5.2
Fractional Dynamical Systems and Signals - Part 2 – Fractional Differentiation in Automatic Control (Invited Session)
Chair: Sabatier, Jocelyn	LAPS - Bordeaux 1 Univ.
Co-Chair: Mohajerin Esfahani, Peyman	Swiss Federal Inst. of Tech.
Organizer: Sabatier, Jocelyn	LAPS - Bordeaux 1 Univ.
Organizer: Farges, Christophe	IMS-LAPS
13:40-14:00	FrB13.1
Free Time Fractional Optimal Control Problems (I), pp. 3985-3990
Pooseh, Shakoor	Univ. of Aveiro
Almeida, Ricardo	Univ. of Aveiro
Torres, Delfim F. M.	Univ. of Aveiro
We present necessary optimality conditions for a class of optimal control problems. The dynamic constraints involve fractional-order and integer-order derivatives and the final time is free. Optimality conditions are obtained using variational principles and some typical problems are solved by approximating the fractional derivatives in terms of integer ones.
14:00-14:20	FrB13.2
A Comparison of Fractional Smith Predictors (I), pp. 3991-3996
Maamri, Nezha	Ec. Nationale Supérieure d'Ingénieurs de Poitiers
Tenoutit, Mammar	Univ. of Poitiers.
Trigeassou, Jean-claude	Univ. of Poitiers
A new class of fractional Smith predictors is proposed in this paper, associating Internal Model Control for the design of time delay controllers to the CRONE template in order to improve robustness to static gain variations. Numerical simulations illustrate intrinsic and structural robust performances of the fractional PI and PID predictors. Comparisons of the two controllers show that the Smith PID presents the greatest ability to reject gain variations and system structure uncertainty.
14:20-14:40	FrB13.3
Fractional Order Control of the Injection System in a CNG Engine (I), pp. 3997-4002
Lino, Paolo	Tech. Univ. of Bari
Maione, Guido	Pol. di Bari
Common rail injection systems of compressed natural gas (CNG) engines require robust control, because of large variations in their parameters and working conditions and because of their highly nonlinear behavior. This paper shows that a noninteger (fractional) order PI (or FOPI) controller, designed by a loop-shaping approach, improves robustness to parameter variations and guarantees an optimal performance. Efficacy of the FOC is tested by simulation based on a CNG linearized model and on the approximation of the FOC by a rational transfer function.
14:40-15:00	FrB13.4
Inversion-Based Feedforward Design for Constrained Fractional Control Systems (I), pp. 4003-4008
Visioli, Antonio	Univ. of Brescia
Padula, Fabrizio	Univ. of Brescia
In this paper we propose an input-output inversion-based methodology for the synthesis of the feedforward action for a fractional control system in order to achieve a predefined process variable transition from a steady-state value to another. In particular, the feedforward action is implemented either as a signal to be added to the feedback control variable or as a command signal to be applied (instead of the typical step signal) to the closed-loop system. The method allows the minimization of the transition time by taking explicitly into account constraints on the process input an output and their derivatives. Simulation results show the effectiveness of the technique.
15:00-15:20	FrB13.5
Computer Algorithms for Solving Optimization Problems for Discrete-Time Fractional Systems (I), pp. 4009-4014
Czyronis, Przemysław, Marcin	Warsaw Univ. of Tech.
Dzielinski, Andrzej	Warsaw Univ. of Tech.
Dynamic programming and discrete-time calculus of variations optimization problems for fractional discrete-time systems with quadratic performance index have been formulated and solved. A new methods for numerical computation of optimization problems have been presented. The efficiency of the methods have been demonstrated on numerical example and illustrated by graphs. Graphs also show the differences between the fractional and classical (standard) systems theory. Results for both methods have been obtained through a computer algorithms written for this purpose.
15:20-15:40	FrB13.6
Fractional Order PD Control of a Visual Servoing Manipulator System (I), pp. 4015-4020
Copot, Cosmin	Ghent Univ.
Ionescu, Clara	Ghent Univ.
Lazar, Corneliu	"Gh. Asachi" Tech. Univ. of Iasi
De Keyser, Robin M.C.	Univ. of Gent
In this paper, a fractional order PD controller based on visual features is presented. A manipulator robot with 6 degrees of freedom and an eye-in-hand camera configuration is employed to design a visual servoing control architecture. The image based control law was designed using point features. A Matlab simulator which implements the visual control architecture was developed and the results were compared with a classical integer order controller. The validity of the visual based controllers is shown by the simulation results which demonstrate that the proposed approach based on a fractional order controller has a stable and convergent behavior when dealing with visual servoing applications. The fractional order PD controller provides better performances in comparison with an integer order PD controller.

FrB14	HG D7.1
Prediction and Control of Freeway Traffic (Invited Session)
Chair: Canudas de Wit, Carlos	CNRS-LAG-Grenoble
Co-Chair: Bianchi, Domenico	Univ. of L'Aquila
Organizer: Pisarski, Dominik	Inria Rhone-Alpes
Organizer: Ferrara, Antonella	Univ. of Pavia
Organizer: Sacone, Simona	Univ. of Genova
13:40-14:00	FrB14.1
Optimal Balancing of Freeway Traffic Density: Application to the Grenoble South Ring (I), pp. 4021-4026
Pisarski, Dominik	Inria Rhone-Alpes
Canudas de Wit, Carlos	CNRS-LAG-Grenoble
This paper presents the application of the idea of optimal balancing of traffic density distribution. The idea was previously studied by the authors, and here it is implemented to the Grenoble South Ring in the context of the Grenoble Traffic Lab. The traffic on the ring is represented by the Cell Transmission Model that was tuned by using real data and Aimsun micro-simulator. A special attention is paid to the calibration of a flow merging model. A large-scale optimization problem is solved by using decomposision methods and implementing combinatorial procedures. The main difficulties in the implementation as well as the limitations of the designed software are highlighted. Finally, the results of different traffic scenarios on the Grenoble South Ring are presented.
14:00-14:20	FrB14.2
Case-Study Based Performance Assessment of an Event-Triggered MPC Scheme for Freeway Systems (I), pp. 4027-4032
Ferrara, Antonella	Univ. of Pavia
Nai Oleari, Alberto	Univ. of Pavia
Sacone, Simona	Univ. of Genova
Siri, Silvia	Univ. of Genova
The paper is aimed at the definition of an efficient control framework for freeway systems. To this end, a Model Predictive Control scheme based on the cell transmission model is adopted in which the considered control action refers to ramp metering. Two major aspects characterize the paper. First, the prediction model is formulated as a mixed logical dynamical system (i.e. it is described by linear dynamic equations and linear inequalities in which both real and integer variables are involved) in order to obtain a finite-horizon optimal control problem with the structure of a mixed-integer quadratic programming problem. In such a problem the objective function quadratically penalizes the deviation of the state variables from a specific equilibrium point. Secondly, an event-triggered scheme is used to further reduce the computational load of the control scheme making it more adequate for real applications. Numerical results based on data coming from a real case study relevant to Grenoble South Ring are also presented and discussed.
14:20-14:40	FrB14.3
Model Predictive Control for Freeway Traffic Using Discrete Speed Limit Signals (I), pp. 4033-4038
D. Frejo, José Ramón	Univ. de Sevilla
Nuñez, Alfredo	Delft Univ. of Tech.
De Schutter, Bart	Delft Univ. of Tech.
Camacho, Eduardo F.	Univ. of Sevilla
A model predictive control (MPC) approach for freeway traffic control using discrete signals for the activation of variable speed limit panels is proposed. The discrete characteristics of the panels and some necessary constraints for their real operation are usually underestimated in the literature, so first we propose a way to include them using the macroscopic traffic model METANET within an MPC framework. For obtaining practical discrete signals, the MPC controller has to solve a highly non-linear optimization problem, including mixed-integer variables. Since solving such a problem is complex and difficult to execute in real-time, we propose some methods to obtain reasonable control actions in a limited computation time. The methods consist of first relaxing the discrete constraints for the inputs; and then, based on this continuous solution and together with different search methods, to find discrete speed limit signals that provide the best performance, keeping the number of simulations reduced. The proposed methods are tested by simulation, showing not only a good performance but also keeping the computational time reduced.
14:40-15:00	FrB14.4
Networked Model Predictive Traffic Control with Time Varying Optimization Horizon: The Grenoble South Ring Case Study (I), pp. 4039-4044
Bianchi, Domenico	Univ. of L'Aquila
Ferrara, Antonella	Univ. of Pavia
Di Benedetto, M. Domenica	Univ. of L'Aquila
This work discusses the design of a networked traffic control scheme. We refer to the Grenoble South Ring traffic system, as case study, in which the control actions are computed in a control centre far from the traffic system and then send, through a wireless communication channel, to the actuators placed along the road, i.e. on-ramp traffic lights in the considered case of ramp metering control. The communication channel is affected by delays and packet loss. In order to counteract the effects due to the transmission of the variable over the communication channel, we suggest to adopt a model predictive control (MPC) strategy based on the use of a buffer. Moreover, in order to limit the computational burden and improve the effectiveness of the proposal, the length of the optimization horizon of the predictive control algorithm, and, consequently, the buffer length, is updated relying on a delay estimation. The performances of the proposed approach are assessed in simulation relying on a traffic model the parameters of which are identified using data produced by a commercial microscopic simulator of the Grenoble South Ring traffic system. The capabilities of the considered control scheme when the system parameters vary and the transmitted signals are affected by time delays are highlighted.
15:00-15:20	FrB14.5
Online Dynamic Travel Time Prediction Using Speed and Flow Measurements (I), pp. 4045-4050
Leon ojeda, Luis Ramon	INRIA, Grenoble. Univ. de Grenoble
Kibangou, Alain Yuwa	GIPSA-Lab. Joseph Fourier/CNRS
Canudas de Wit, Carlos	CNRS-LAG-Grenoble
Traffic forecasting is considered nowadays as one of the most important traffic management techniques on road networks. To provide suitable control strategies and advanced traveler information, which improve traffic performance, a continuous short-term prediction is a significant requirement. In this paper, we propose a new approach for travel time forecasting between two points of interest of a given highway divided in nodes and links. Since nodes and links have distinct characteristics, two different prediction methods are proposed. The resulting predicted travel time is then computed as the sum of predicted travel times in nodes with those in links. An adaptive Kalman filtering approach is considered for predicting sojourn time in nodes and flows at boundaries of links. Inside links, divided in cells for improving resolution, a deterministic observer is used for computing unmeasured densities. The performance of the proposed method is evaluated by using data of the Grenoble south ring, a case study of the NoE Hycon2.

FrB15	HG D7.2
Optimization for Energy Systems (Regular Session)
Chair: Sandberg, Henrik	Royal Inst. of Tech. (KTH)
Co-Chair: Borrelli, Francesco	Univ. of California, Berkeley
13:40-14:00	FrB15.1
Energy and CO2 Efficient Scheduling of Smart Home Appliances, pp. 4051-4058
Sou, Kin Cheong	Chalmers Univ. of Tech.
Kordel, Mikael	KTH Royal Inst. of Tech.
Wu, Jonas	KTH Royal Inst. of Tech.
Sandberg, Henrik	Royal Inst. of Tech. (KTH)
Johansson, Karl Henrik	Royal Inst. of Tech.
A major goal of smart grid technology (e.g., smart meters) is to provide consumers with demand response signals such as electricity tariff and CO$_2$ footprint so that the consumers can consciously control their electricity consumption patterns. These demand response signals provide incentives for the consumers to help reduce peak energy demand by load balancing, as this is particularly relevant in a situation with high level of renewable energy penetration. However, the volume of information can be overwhelming for the consumers. Further, in some situation minimization of electricity bill and CO$_2$ emission can be conflicting goals and a trade-off analysis is required. To enable the consumers to participate in smart grid effort this paper proposes a decision aiding framework for optimal household appliances scheduling and trade-off analysis through Pareto frontier exploration. To compute the optimal schedules associated with Pareto optimal points, linear optimization problems with SOS2 (special ordered set of type 2) constraints are solved using CPLEX, in the case where the demand response signals are assumed to be piecewise constant. For arbitrary demand response signals, a corresponding dynamic programming solution is proposed. A numerical study demonstrates that in a realistic test case the Pareto frontier analysis can provide valuable information leading to schedules with drastically different electricity and CO$_2$ emission patterns. In addition, the case study verifies that the Pareto frontier can be computed in real-time in a realistic residential computing environment.
14:00-14:20	FrB15.2
Load Sharing Optimization of Parallel Compressors, pp. 4059-4064
Paparella, Francesco	Pol. di Milano
Domínguez, Luis F.	ABB Schweiz
Cortinovis, Andrea	ABB Corp. Res.
Mercangöz, Mehmet	ABB Corp. Res.
Pareschi, Diego	ABB
Bittanti, Sergio	Pol. di Milano
This paper deals with the problem of optimizing the load distribution and the on/off switching sequences of parallel compressor units in natural gas pipeline compression stations. Natural gas pipelines are used to deliver gas from production sources to customers. Compression stations on these pipelines are generally composed of the interconnection of several compressors units and the aim of the load sharing optimization is to operate these units in an energy efficient way while continuously satisfying the varying demand of gas flow. As the gas demand changes start-up and shut-down of compressor units might be required and the impact of these switching events on the expected lifetime of the compressors also needs to be taken into consideration. Two types of operational constraints are involved: continuous constraints concerning the conservation equations of mass and flow and combinatorial constraints concerning the possibility of changing the number of active compressors. Overall the optimization problem can be formulated as a mixed integer nonlinear programming (MINLP) problem. In this article the generic optimization problem setup and appropriate techniques for the solution are presented. Compared to other traditional strategies such as equal load balancing or equal distance operation from surge, the MINLP approach has shown considerable improvement in terms of energy savings.
14:20-14:40	FrB15.3
Performance Evaluation of Battery Balancing Hardware, pp. 4065-4070
Preindl, Matthias	Univ. of Padua
Danielson, Claus	Univ. of California, Berkeley
Borrelli, Francesco	Univ. of California, Berkeley
In this paper we evaluate the performance of seven proposed hardware topologies for balancing the cells in a battery pack. We consider four classes of battery balancing hardware; shunting, cell-to-stack, storage element, and dissipative hardware. We present models of these hardware topologies that capture the steady-state behavior of the balancing hardware dynamics. We evaluate the hardware topologies based on time required to balance the cells and energy dissipated during balancing. A linear programming based method for calculating the worst-case time to balance and energy dissipated during balance is provided. The number of linear programs required to compute both metrics grows exponentially with the number of cells. We show how to use symmetries to efficiently compute both metrics. Our approach scales well for large-scale battery packs and provides non intuitive solutions.
14:40-15:00	FrB15.4
MPC for Wind Power Gradients - Utilizing Forecasts, Rotor Inertia, and Central Energy Storage, pp. 4071-4076
Hovgaard, Tobias Gybel	Vestas Tech. R&D
Larsen, Lars Finn Sloth	Vestas A/S
Jørgensen, John Bagterp	Tech. Univ. of Denmark
Boyd, Stephen P.	Stanford Univ.
We consider the control of a wind power plant, possibly consisting of many individual wind turbines. The goal is to maximize the energy delivered to the power grid under very strict grid requirements to power quality. We define an extremely low power output gradient and demonstrate how decentralized energy storage in the turbines' inertia combined with a central storage unit or deferrable consumers can be utilized to achieve this goal at a minimum cost. We propose a variation on model predictive control to incorporate predictions of wind speed. Due to the aerodynamics of the turbines the model contains nonconvex terms. To handle this nonconvexity, we propose a sequential convex optimization method, which typically converges in fewer than 10 iterations. We demonstrate our method in simulations with various wind scenarios and prices for energy storage. These simulations show substantial improvements in terms of limiting the power ramp rates (disturbance rejection) at the cost of very little power. This capability is critical to help balance and stabilize the future power grid with a large penetration of intermittent renewable energy sources.
15:00-15:20	FrB15.5
Dynamic Optimization of a Campus Cooling System with Thermal Storage, pp. 4077-4082
Powell, Kody	Univ. of Texas at Austin
Cole, Wesley	Univ. of Texas at Austin
Ekarika, Udememfon	Univ. of Texas at Austin
Edgar, Thomas F.	Univ. of Texas at Austin
Thermal energy storage gives a system enhanced operational flexibility because thermal loads can be shifted, not only spatially—from one piece of equipment to another—but also temporally, using storage, from one point in time to another. The resulting optimization problems become non-convex and difficult to solve. This paper illustrates how to take advantage of the enhanced flexibility that storage provides, while simplifying the dynamic optimization problems involved. The optimization strategy presented decouples the static and dynamic pieces of the problem using a hierarchical structure, where a static sub-problem is solved for each trial decision variable in the dynamic problem. Energy savings as high as 9.4 % are observed with cost savings as high as 17.4%.

FrC1	HG F1
Emerging Control Applications II (Regular Session)
Chair: Kugi, Andreas	Vienna Univ. of Tech. (VUT)
Co-Chair: Haidar, Ihab	L2S, Supelec-CNRS-Univ. Paris Sud
16:00-16:20	FrC1.1
Basal Ganglia Oscillations: The Role of Delays and External Excitatory Nuclei, pp. 4083-4088
Haidar, Ihab	L2S, Supelec-CNRS-Univ. Paris Sud
Pasillas Lepine, William	CNRS
Panteley, Elena V.	CNRS
Chaillet, Antoine	Univ. Paris Sud
Basal ganglia are interconnected deep brain structures involved in movement generation. Their beta-band oscillations (13-30Hz) are known to be linked to Parkinson's disease motor symptoms. In this paper, we provide conditions under which these oscillations may occur, by explicitly considering the role of the pedunculopontine nucleus (PPN). We analyse the existence of equilibria in the associated firing-rate dynamics and study their stability by relying on a delayed MIMO frequency analysis. These results are illustrated with simulations that confirm numerically the analytic predictions of our two main theorems.
16:20-16:40	FrC1.2
Control of Temperature to Suppress the Population of Rhyzopertha Dominica (F.) (Coleoptera, Bostrichidae) in a Grain Silo Prototype, pp. 4089-4093
W. F. De Souza, Denis	PID BRASIL Automacao Industrial Ltda
Vargas, Alessandro N	Univ. Tec Federal do Parana, UTFPR
Do Val, Joao B.R.	UNICAMP - FEEC
M. Freitas, Adriana	Embrapa Soja
Lorini, Irineu	Embrapa Soja
This note presents some results from laboratory experiments that were conducted to characterize the influence of temperature in the mortality of adults of the insect known as lesser grain borer, Rhyzopertha dominica (F.). The insects were separated into strains and were appropriately immersed into a mass of wheat, and the infested wheat was stored in a silo bin of small dimensions with control of temperature. Our experiments indicates that the control of temperature can be used as a successful tool to increase the mortality of {it R. dominica} in grain silos. The paper also describes the construction of the electrical device that implements the control of temperature in the proposed grain silo prototype.
16:40-17:00	FrC1.3
Subsea Solution for Anti-Slug Control of Multiphase Risers, pp. 4094-4099
Jahanshahi, Esmaeil	Norwegian Univ. of Science & Tech.
Skogestad, Sigurd	Norwegian Univ. of Science & Tech.
Lieungh, Mats	Siemens Energy
A top-side choke valve is usually used as the manipulated variable for anti-slug control of multi-phase risers at offshore oil-fields. With new advances in the subsea technology, it is now possible to move top-side facilities to the sea floor. The two main contributions in this paper are to consider an alternative location for the control valve and to consider how to deal with nonlinearity. This research involved controllability analysis based on a simplified model fitted to experiments, simulations using the OLGA simulator, as well as an experimental study. It was concluded that a control valve close to the riser-base is very suitable for anti-slug control, and its operation range is the same as the top-side valve. However, a subsea choke valve placed at the well-head can not be used for preventing the riser-slugging.
17:00-17:20	FrC1.4
Radar Resource Management: Dynamic Programming and Dynamic Finite State Machines, pp. 4100-4105
Seok, Jinwoo	Univ. of Michigan
Zhao, Jinxin	Univ. of Michigan
Selvakumar, Jhanani	Univ. of Michigan
Sanjaya, Edwin	Univ. of Michigan
Kabamba, Pierre T.	Univ. of Michigan
Girard, Anouck	Univ. of Michigan at Ann Arbor
Finite state machines are a standard tool to model event-based control logic, and dynamic programming is a staple of optimal decision-making. We combine these approaches in the context of radar resource management for Naval surface warfare. There is a friendly (Blue) force in the open sea, equipped with one multi-function radar and multiple ships. The enemy (Red) force consists of missiles that target the Blue force's radar. The mission of the Blue force is to foil the enemy's threat by careful allocation of radar resources. Dynamically composed ﬁnite state machines are used to formalize the model of the battle space and dynamic programming is applied to our dynamic state machine model to generate an optimal policy. To achieve this in near-real-time and a changing environment, we use approximate dynamic programming methods. Example scenario illustrating the model and simulation results are presented.
17:20-17:40	FrC1.5
Control of Radiant Tubes in an Indirect-Fired Strip Annealing Furnace for Improved Efficiency, pp. 4106-4111
Niederer, Martin	Vienna Univ. of Tech.
Steinboeck, Andreas	Vienna Univ. of Tech.
Strommer, Stephan	Vienna Univ. of Tech.
Kugi, Andreas	Vienna Univ. of Tech. (VUT)
The possibilities for improving the efficiency of an indirect-fired strip annealing furnace with radiant tubes are explored. First, the influence of the fuel/air-ratio on the efficiency is analyzed by means of linearized relationships. Second, the optimum operating range of an individual radiant tube and a new control strategy for the fuel supply are derived. Both analyses are based on the balance model of a single radiant tube. The input data were taken from an experimental measurement campaign of an industrial furnace.
17:40-18:00	FrC1.6
On the Use of the Inclinometers in the PnP Problem, pp. 4112-4117
D'Alfonso, Luigi	Univ. della Calabria
Garone, Emanuele	Univ. degli Studi della Calabria
Muraca, Pietro	Univ. della Calabria
Pugliese, Paolo	Univ. della Calabria
This paper deals with the problem of estimating the relative pose between a camera and an object. It is assumed that both the camera and the object are equipped with an Inertial Measurement Unit (IMU) which measures their inclinations with respect to the gravity vector. Moreover, it is assumed that the object contains a feature of $n geq 2$ points, the position of which in the object reference frame is known {it a priori}. The resulting pose estimation problem can be seen as a P$n$P problem with inclination information. Early results of the authors on the subject showed that in this case the P$2$P problem always gives 2 solutions, except for a few singular configurations where the number of solution is infinite. Moreover, to avoid singular configurations and to resolve ambiguities a very simple solution for the P$3$P problem, based on the idea of re-projection, was proposed. In this paper, it will be shown that, thanks to a simple test based on geometrical considerations, very often one of two solutions of the P$2$P problem can be discarded. Moreover to solve the remaining ambiguities and to ameliorate the pose estimation, a novel and more robust algorithm for the general P$n$P problem will be proposed. The results will be validated through numerical and experimental tests.

FrC2	HG F3
Embedded Optimization for Control and Estimation III – Applications (Invited Session)
Chair: Zeilinger, Melanie N.	Ec. Pol. Fédéral de Lausanne (EPFL)
Co-Chair: Lucia, Sergio	TU Dortmund
Organizer: Zeilinger, Melanie N.	UC Berkeley
16:00-16:20	FrC2.1
Mixed-Integer NMPC for Predictive Cruise Control of Heavy-Duty Trucks (I), pp. 4118-4123
Kirches, Christian	Heidelberg Univ.
Bock, Hans Georg	Univ. of Heidelberg
Schlöder, Johannes	Ruprecht-Karls-Univ. Heidelberg
Sager, Sebastian	OVGU Magdeburg
We present a novel numerical method for nonlinear model-predictive control of heavy-duty trucks. This method realizes a predictive online cruise controller, and includes the opportunity for multiple predictive gear choices. The combination of nonlinear dynamics, constraints, and objective with the hybrid nature of the gear choice makes for a challenging combinatorial prediction problem. The numerical algorithm presented to attack this problem is based on a direct and simultaneous method for the efficient solution of optimal control problems in ordinary differential equations and differential-algebraic equations. Partial outer convexification is applied to treat the mixed-integer aspect of the problem, maintaining certificates of $varepsilon$-optimality. A vanishing constraint formulation for engine speed constraints ensures feasibility of predictive gear choices. Suitable exploitation of direct multiple shooting structures allows to achieve encouraging real-time capable mixed-integer control feedback rates. Numerical results for a heavy-duty truck cruise controller including predictive gear shifts are presented for several artificial and real-world scenarios.
16:20-16:40	FrC2.2
Robust Nonlinear Model Predictive Control of a Batch Bioreactor Using Multi-Stage Stochastic Programming (I), pp. 4124-4129
Lucia, Sergio	TU Dortmund
Engell, Sebastian	TU Dortmund
This paper presents a robust nonlinear model predictive control scheme and its application to a batch bioreactor. The approach is based on the description of the uncertainty evolution as a scenario tree. This makes it possible to take explicitly into account the future disturbances and control inputs leading to a non-conservative approach that is not based on the tracking of a nominal solution. The main challenge of the approach is that the size of the resulting optimization problem grows exponentially with the prediction horizon and with the number of uncertainties. The potential of the approach is demonstrated by simulation examples of a nonlinear penicillin fermentation process where the proposed scheme can fulfill the state and the input constraints for all the possible values of several uncertain parameters, improving the performance of existing robust approaches such as tracking of the necessary conditions of optimality.
16:40-17:00	FrC2.3
Nonlinear Moving Horizon Estimation for Combined State and Friction Coefficient Estimation in Autonomous Driving (I), pp. 4130-4135
Zanon, Mario	KU Leuven
Frasch, Janick	Univ. of Heidelberg,Germany
Diehl, Moritz	KU Leuven
Real-time autonomous driving requires a precise knowledge of the state and of the ground parameters, especially in dangerous situations. In this paper an accurate nonlinear multi-body vehicle model is developed, featuring a detailed Pacejka tire model and a Moving Horizon Estimation (MHE) scheme is formulated. To meet the real-time requirements, an efficient algorithm based on the Real Time Iteration (RTI) scheme for the Direct Multiple Shooting method is exported through automatic C code generation. The exported plain C-code is tailored to the model dynamics, resulting in computational times in the range of few milliseconds. In addition to state estimates, MHE provides friction coefficient estimates, allowing the controller to adapt to varying road conditions. Simulation results from an obstacle avoidance scenario on a low friction road are presented.
17:00-17:20	FrC2.4
An Auto-Generated Nonlinear MPC Algorithm for Real-Time Obstacle Avoidance of Ground Vehicles (I), pp. 4136-4141
Frasch, Janick	Univ. of Heidelberg,Germany
Gray, Andrew	Univ. of California, Berkeley
Zanon, Mario	KU Leuven
Ferreau, Hans Joachim	ABB Schweiz AG
Sager, Sebastian	OVGU Magdeburg
Borrelli, Francesco	Univ. of California, Berkeley
Diehl, Moritz	KU Leuven
We address the problem of real-time obstacle avoidance on low-friction road surfaces using spatial Nonlinear Model Predictive Control (NMPC). We use a nonlinear four-wheel vehicle dynamics model that includes load transfer. To overcome the computational difficulties we propose to use the ACADO Code Generation tool which generates NMPC algorithms based on the real-time iteration scheme for dynamic optimization. The exported plain C code is tailored to the model dynamics, resulting in faster run-times in effort for real-time feasibility. The advantages of the proposed method are shown through simulation.
17:20-17:40	FrC2.5
Moving Horizon for Friction State and Parameter Estimation (I), pp. 4142-4147
Boegli, Max	KU Leuven
De Laet, Tinne	KU Leuven
De Schutter, Joris	KU Leuven
Swevers, Jan	KU Leuven
Efficient on-line state and parameter estimation is essential for model-based friction compensation in order to track changes of friction characteristics in time and space. This paper presents a moving horizon estimation (MHE) algorithm for on-line friction state and parameter estimation using a smoothed (analytic) version of the Generalized Maxwell-Slip (GMS) model, a multi-state friction model known to describe all essential friction characteristics in presliding and sliding motion. In contrast to the GMS model, which consists of a switching structure to accommodate for its hybrid nature, the Smoothed GMS (S-GMS) model consists of an analytic set of differential equations well suited for gradient-based state and parameter estimation, as in MHE or in extended Kalman filtering (EKF). Moreover, MHE is known to better handle model nonlinearities, disturbances and constraints than EKF. This paper discusses the implementation of an MHE algorithm for the S-GMS friction model and experimentally compares its performance to an EKF implementation for joint state and parameter estimation.
17:40-18:00	FrC2.6
Moving Horizon Observation for Autonomous Operation of Agricultural Vehicles (I), pp. 4148-4153
Frasch, Janick	Univ. of Heidelberg,Germany
Saeys, Wouter	KULeuven
Diehl, Moritz	KU Leuven
Kraus, Tom	KU Leuven
Varying terrain conditions influencing the ground friction challenge model-based control methods for precise autonomous driving of agricultural vehicles on off-road ter- rains. We apply moving horizon estimation (MHE) to cope with these uncertainties in a predictive control framework for autonomous driving of a sensor-equipped tractor using a (nonlinear) rigid-body dynamic model featuring a simple tire slip model. Using the ACADO Code Generation tool feedback times in the range of few miliseconds are achieved. Simulation results and estimation results on recorded experimental data are presented.

FrC3	HG F5
Biological Systems (Regular Session)
Chair: Khammash, Mustafa H.	Univ. of California at Sta. Barbara
Co-Chair: Bernard, Olivier	Inria
16:00-16:20	FrC3.1
Efficient Parameter Identification for Stochastic Biochemical Networks Using a Reduced-Order Realization, pp. 4154-4159
Hori, Yutaka	The Univ. of Tokyo
Khammash, Mustafa H.	Univ. of California at Sta. Barbara
Hara, Shinji	The Univ. of Tokyo
In this paper, we propose a parameter identification method for stochastic biochemical reaction networks using flow cytometry data. A distinctive feature of the proposed method is that it is computationally efficient compared to existing works, thus it is applicable to complex biochemical networks. To this end, we first show that it is possible to construct a significantly small-order realization of the stochastic biochemical system using flow cytometry measurements. Then, the small-order realization is utilized for the development of the efficient identification method. Finally, the proposed method is demonstrated with an existing biological example.
16:20-16:40	FrC3.2
A Mathematical Model of Fractone-Controlled Morphogenesis, pp. 4160-4165
Chyba, Monique	Univ. of Hawai'i at Manoa
Mercier, Frederic	Univ. of Hawaii
Tamura-sato, Aaron	Univ. of Hawai'i at Manoa
Zou, Rong	Kyushu Univ.
It has been hypothesized that the generation of new neural cells (neurogenesis) resulting from stem and progenitor cell proliferation and differentiation in the developing and adut brain is guided by the extracellular matrix. The extracellular matrix of the neurogenic niches comprises basement membranes, which are organized in meningeal and vascular sheaths beneath subsets of stem/progenitor cells, and/or specialized structures termed fractones, which are scattered in between stem/progenitor cells. Growing evidence indicates that fractones of the adult brain bind and activate growth factors at the surface of stem/progenitor cells to influence their proliferation. Neural stem/progenitor cells during early brain development are named neuroepithelial cells. It has been shown that neuroepithelial cell proliferation is also associated with fractones during early brain development, although the functional links between fractones and neuroepithelial cells during development have not been elucidated. We hypothesize that the distribution of fractones, which is constantly changing during development, is correlated with the decision of where and when the neuroepithalial cells proliferate, therefore reflects morphogenesis. We present a mathematical model that considers the role of fractones as captors and activators of growth factors that influence the rate of proliferation and the location of the newly generated neuroepithelial cells in the forming brain. This model allows for the dynamic placement and removal of fractones into the evolving cell mass, giving us control over its developing shape. Using this model, we simulate the observed process of post-neurulation, focusing on the formation of the lateral ventricle walls from the anterior portion of the neural tube.
16:40-17:00	FrC3.3
Single Event Molecular Signalling for Estimation and Control, pp. 4166-4171
Parag, Kris	Univ. of Cambridge
Vinnicombe, Glenn	Univ. of Cambridge
Cell biology is characterised by low molecule numbers and coupled stochastic chemical reactions with intrinsic noise permeating and dominating the interactions between molecules. Recent work of Lestas, Vinnicombe and Paulsson has shown that in such environments there are hard limits on the accuracy with which molecular populations can be controlled and estimated. These limits are predicated on a continuous diffusion approximation of the target molecule (although the remainder of the system is non-linear and discrete). The principal result of Lestas, Vinnicombe and Paulsson's work assumes that the birth rate of the signalling species is linearly dependent on the target molecule population size. In this paper, we investigate the situation when the entire system is kept discrete, and arbitrary non-linear coupling is allowed between the target molecule and downstream signalling molecules. In this case it is possible, by relying solely on the event triggered nature of control and signalling reactions, to define non-linear reaction rate modulation schemes that achieve improved performance in certain parameter regimes. These schemes would not appear to be biologically relevant, raising the question of what are an appropriate set of assumptions for obtaining biologically meaningful results.
17:00-17:20	FrC3.4
Approximate Model Reductions for Combinatorial Reaction Systems, pp. 4172-4177
Petrov, Tatjana	ETHZ
Koeppl, Heinz	ETH Zurich
The paper considers a model reduction technique that is well-suited for biochemical reaction systems giving rise to the assembly of a large number of different molecular species. The reduction is performed by grouping species with common properties, directly from the model specification in terms of a rule-based language. In recent works, general algorithms for the exact reductions of rule-based models were established, but the state space often remains combinatorial. We extend this line of research by introducing approximate reductions, and an error measure which allows us to quantitatively study the effect of approximate model reductions.
17:20-17:40	FrC3.5
A Dynamic Model Coupling Photoacclimation and Photoinhibition in Microalgae, pp. 4178-4183
Hartmann, Philipp	INRIA
Nikolaou, Andreas	Imperial Coll. London
Chachuat, Benoît	Imperial Coll. London
Bernard, Olivier	Inria
Microalgae are often considered a promising alternative for production of renewable energy, particularly as a potential producer of biodiesel. In order to improve largescale, industrial culturing systems, the development of mathematical models that are capable of predicting photosynthetic productivity under dynamic conditions is crucial. Especially important are the processes of growth inhibition due to excess light, known as photoinhibition, and of adjustment of the light harvesting capacity to photon flux, known as photoacclimation. In this paper, we develop a dynamic model that accounts for the processes of photoinhibition, photoacclimation and growth in microalgae, thereby spanning multiple time scales. The properties of the model are investigated under quasi steady state conditions and the model is validated against several experimental data sets from the literature. We also discuss how the model can provide new insights into the mechanisms underlying photoacclimation.
17:40-18:00	FrC3.6
Driving Microalgal Production in Raceway Systems to Near Optimal Productivities, pp. 4184-4189
Muñoz-Tamayo, Rafael	INRIA
Mairet, Francis	Univ. Técnica Federico Santa María
Bernard, Olivier	Inria
We propose a simple operational criterion for raceway systems that when integrated in a closed-loop controller allows to attain microalgal productivities very near to the maximal theoretical producitivities. The strategy developed was tested numerically by using a mathematical model of microalgae growth in raceways.

FrC4	HG F7
Applications of State Estimation (Regular Session)
Chair: Cenedese, Angelo	Univ. of Padova
Co-Chair: Lemos, Joao M.	Inesc-id
16:00-16:20	FrC4.1
Robust Nonlinear Estimation of Varying Optical Phase, pp. 4190-4195
Rehman, Obaid Ur	UNSW at ADFA
Petersen, Ian R.	Australian Defence Force Acad.
Song, Hongbin	The Univ. of New South Wales Canberra
Huntington, Elanor H.	Univ. of New South Wales, Canberra
This work presents a robust approach to estimate the continuously varying phase of an optical system. Here we extend the adaptive homodyne estimator scheme by using a nonlinear guaranteed cost robust filter in the feedback loop to improve the estimator performance. The nonlinear robust filter is designed by using a copy of the sector bounded nonlinear uncertainty present in the measurement of the adaptive homodyne estimator. Finally, closed loop simulations considering both linear and nonlinear uncertainty are performed and the results are compared with the standard Kalman filter results. Simulation results show a significant improvement using the nonlinear filter which provides a better choice for the time varying phase estimation problem.
16:20-16:40	FrC4.2
On the Estimation of Atmospheric Turbulence Layers for AO Systems, pp. 4196-4201
Beghi, Alessandro	Univ. di Padova
Cenedese, Angelo	Univ. of Padova
Masiero, Andrea	Univ. di Padova
In current and next generation of ground telescopes, Adaptive Optics (AO) are employed to overcome the detrimental effects induced by the presence of atmospheric turbulence, that strongly affects the quality of data transmission and therefore limits the actual resolution of the overall system. The analysis as well as the prediction of the turbulent phase affecting the light wavefront is therefore of paramount importance to guarantee the effective performance of the AO solution. In this work, a layered model of turbulence is proposed, based on the definition of a Markov-Random-Field whose parameters are determined according to the turbulence statistics. The problem of turbulence estimation is formalized within the stochastic framework and conditions for the identifiability of the turbulence structure (numbers of layers, energies and velocities) are stated. Numerical simulations are used to assess the proposed procedure and validate the results.
16:40-17:00	FrC4.3
Multiple-Model Adaptive State Estimation of the HIV-1 Infection Using a Moving Horizon Approach, pp. 4202-4207
Casal, Filipe R.	ISR-IST
Aguiar, A. Pedro	Faculty of Engineering, Univ. of Porto (FEUP)
Lemos, Joao M.	Inesc-id
This paper addresses the problem of state estimation under parametric uncertainty of discrete lumped nonlinear systems with application to the HIV-1 infection. We present an estimation algorithm using a multiple-model adaptive estimation approach with a bank of moving horizon estimators with decimated observations. This is motivated by its possible applications to the HIV-1 infection where, in practice, we are unable to observe the patient on a regular basis (non-periodic measurements) and because the HIV-1 dynamics depends on parameters unique to each patient (parameter uncertainty). We show that under reasonable assumptions, the proposed estimation algorithm is robust to parametric uncertainty and the estimation error converges to a small neighborhood of zero. The robustness and performance of the algorithm are illustrated through computer simulations.
17:00-17:20	FrC4.4
Position Estimation Approach by Complementary Filter-Aided IMU for Indoor Environment, pp. 4208-4213
Fourati, Hassen	Univ. Joseph Fourrier, GIPSA-Lab.
Manamanni, Noureddine	Univ. of Reims- Champagne
Afilal, Lissan-eddine	CReSTIC, Reims
Handrich, Yves	Univ. de Strasbourg, DEPE/IPHC
This paper proposes a foot-mounted Zero Velocity Update (ZVU) aided Inertial Measurement Unit (IMU) filtering algorithm for pedestrian tracking in indoor environment. The algorithm outputs are the foot kinematic parameters, which include foot orientation, position, velocity, acceleration, and gait phase. The foot motion filtering algorithm incorporates methods for orientation estimation, gait detection, and position estimation. A novel Complementary Filter (CF) is introduced to better pre-process the sensor data from a foot-mounted IMU containing tri-axial angular rate sensors, accelerometers, and magnetometers and to estimate the foot orientation without resorting to GPS data. A gait detection is accomplished using a simple states detector that transitions between states based on acceleration measurements. Once foot orientation is computed, position estimates are obtained by using integrating acceleration and velocity data, which has been corrected at step stance phase for drift using an implemented ZVU algorithm, leading to a position accuracy improvement. We illustrate our findings experimentally by using of a commercial IMU during regular human walking trial in a typical public building. Experiment results show that the positioning approach achieves approximately a position accuracy less than 1 m and improves the performance regarding a previous work of literature.
17:20-17:40	FrC4.5
3-D Inertial Trajectory and Map Online Estimation: Building on a GAS Sensor-Based SLAM Filter, pp. 4214-4219
Lourenço, Pedro	Inst. for Systems and Robotics / Inst. Superior Técnico
Guerreiro, Bruno J. N.	Inst. Superior Tecnico
Batista, Pedro	Inst. Superior Técnico
Oliveira, Paulo Jorge	Inst. Superior Técnico
Silvestre, Carlos	Univ. of Macau
This paper addresses the problem of obtaining an inertial trajectory and map, with the associated uncertainty, using the sensor-based map provided by a globally asymptotically stable SLAM filter. An optimization problem with a closed-form solution is formulated, and its uncertainty description is derived resorting to perturbation theory. The combination of the algorithm proposed in this paper with the sensor-based SLAM filter results in a complete SLAM methodology, which can be directly applied to unmanned aerial vehicles (UAVs). Both simulation and preliminary experimental results, using an instrumented quadrotor equipped with a RGB-D camera, are included in this work to illustrate the performance of the proposed algorithm under realistic conditions.
17:40-18:00	FrC4.6
Robustness Analysis of Feedback Linearisation with Robust State Estimation for a Nonlinear Missile Model, pp. 4220-4225
Norton, Peter	Univ. of Leicester
Prempain, Emmanuel	Univ. Leicester
A nonlinear missile model with time-varying uncertain parameters is controlled with a simple feedback linearisation and time-scale separation design, with synthesis based on the nominal model and full state feedback. The closed-loop system is then represented as a linear fractional transformation (LFT). A robust H-infinity filter is designed for the controlled plant, to estimate unknown states. Robust stability of the closed-loop system is then verified by using a scaled linear differential inclusion (LDI) technique.

FrC5	HG E1.1
Formal Verification and Design of Hybrid Systems (Invited Session)
Chair: Prabhakar, Pavithra	IMDEA
Co-Chair: Liu, Jun	Univ. of Sheffield
Organizer: Prabhakar, Pavithra	IMDEA
Organizer: Liu, Jun	Univ. of Sheffield
16:00-16:20	FrC5.1
An Approximate Abstraction Approach to Safety Control of Differentially Flat Systems (I), pp. 4226-4231
Colombo, Alessandro	Pol. di Milano
Girard, Antoine	Univ. Joseph Fourier
Control for safety specifications of large nonlinear systems is a challenging task. By reducing the system to a discrete abstraction the computational demands of the controller can be greatly reduced. We propose a supervisor for differentially flat systems, based on an approximate abstraction of the flat output. By defining the abstraction on the flat output space, we simplify the design of the abstraction and further reduce the computational complexity of the resulting supervisor, and by exploiting approximate simulation techniques we obtain a controller that is simpler and more flexible than previously proposed solutions. The resulting algorithm is tested on an eight-dimensional nonlinear planar crane model.
16:20-16:40	FrC5.2
Aggregation of Thermostatically Controlled Loads by Formal Abstractions (I), pp. 4232-4237
Esmaeil Zadeh Soudjani, Sadegh	TU Delft - Delft Univ. of Tech.
Abate, Alessandro	TU Delft - Delft Univ. of Tech.
This work discusses a two-step procedure based on formal abstractions, which generates a finite stochastic dynamical model as an aggregation of the continuous temperature dynamics of a population of Thermostatically Controlled Loads (TCL). The temperature of each single TCL is described by a stochastic difference equation and the TCL status (ON, OFF) by a deterministic switching mechanism. The procedure is formal as it allows the exact quantification of the error introduced by the abstraction. The work discusses extensions to the case of controlled TCL. The procedure is tested on a case study and benchmarked against an alternative approach in the literature.
16:40-17:00	FrC5.3
Modeling and Verification of a Robotic Surgical System Using Hybrid Input/Output Automata, pp. 4238-4243
Capiluppi, Marta	Univ. of Verona
Schreiter, Luzie	Karlsruhe Inst. of Tech. (KIT)
Fiorini, Paolo	Univ. of Verona
Woern, Heinz	Karlsruhe Inst. for Tech. (KIT)
Raczkowsky, Jörg	Karlsruhe Inst. of Tech. - KIT
The area of robotic surgical systems has to deal with several important safety aspects to ensure that the patient and the Operating Room staff are safe. A robotic surgical system has to fulfill specific safety requirements and to ensure that the system reacts like its specification. To this end, a verification process is necessary. In this paper an architecture for robotic surgery is modelled using the framework of Hybrid Input/Output Automata (HIOAs). A case study based on a surgical robotic operation scenario is presented and modelled using HIOAs. Exploiting the modularity and compositionality theory of HIOAs, the verification of the system is performed.
17:00-17:20	FrC5.4
Modular Modeling of Control Systems in SpaceEx (I), pp. 4244-4251
Donze, Alexandre	UC Berkeley, EECS Department
Frehse, Goran	Univ. Joseph Fourier Grenoble 1
The Hybrid I/O-automaton (HIOA) is a rigorous formal model designed for the analysis of complex hybrid (discrete-continuous) dynamical systems. The use of the HIOA formalism renders compositional reasoning possible, in the sense that once a property has been established for an automaton, it still holds if the automaton is composed with other automata. In this paper, it is shown how control systems can be modeled and verified in SpaceEx as HIOA in a modular fashion. Formally, HIOA models distinguish between controlled and uncontrolled variables. With examples and usage guidelines, they are related to the concepts of input/output and state/algebraic variables most control designers are familiar with. Additionally, the applicability of HIOA is enlarged by allowing variables to be controlled in more than one automaton. While this invalidates compositionality, it gives users who do not intend to use compositional reasoning more freedom in their modeling choices. Finally, it is shown how control systems given by semi-explicit differential algebraic equations can be algorithmically brought to the form understood by the SpaceEx reachability algorithm.
17:20-17:40	FrC5.5
Verification of Embedded Control Programs (I), pp. 4252-4256
Dang, Thao	VERIMAG
Jeannet, Bertrand	IRISA/INRIA
Testylier, Romain	VERIMAG
In this paper we are concerned with the problem of verifying embedded control programs. The approach we use combines the logico-numerical techniques developped for the verification of Lustre programs and the set-based image computation for continuous systems. The practical interest of this approach lies in the fact that there exists a tool to generate Lustre code for controllers described in Simulink. We also illustrate the approach with some experimental results obtained for a robotic controller for LEGO Mindstorm.
17:40-18:00	FrC5.6
Symbolic Models for Stochastic Control Systems without Stability Assumptions (I), pp. 4257-4262
Zamani, Majid	Delft Univ. of Tech.
Mohajerin Esfahani, Peyman	Automatic Control Lab. Swiss Federal Inst. ofTechnolo
Abate, Alessandro	TU Delft - Delft Univ. of Tech.
Lygeros, John	ETH Zurich
Symbolic approaches provide a mechanism to construct discrete and possibly finite abstractions of continuous control systems. Discrete abstractions are in turn amenable to automata-theoretic techniques targeted to the construction of controllers satisfying complex specifications, difficult to enforce over continuous models with conventional control design methods. Although construction of discrete abstractions has been extensively studied for non-probabilistic continuous-time control systems, it has received scant attention on stochastic continuous-time control systems. In this paper, we propose an abstraction technique that is applicable to any stochastic continuous-time control system, as long as we are only interested in its behavior over a compact set. The effectiveness of the proposed results is illustrated with the synthesis of a controller for a jet engine model, which is not stable, affected by noise, and subject to a schedulability constraint expressed by a finite automaton.

FrC6	HG E1.2
Oscillations and Synchronization (Regular Session)
Chair: Boiko, Igor	Petroleum Inst.
Co-Chair: Zechner, Christoph	ETH Zurich
16:00-16:20	FrC6.1
Synchronization in a Heterogeneous Network of Discrete-Time Introspective Right-Invertible Agents, pp. 4263-4268
Wang, Xu	New York Univ.
Saberi, Ali	Washington State Univ.
Yang, Tao	Royal Inst. of Tech.
This paper studies output synchronization problem for a heterogenous network of discrete-time introspective right-invertible agents. We propose a decentralized control scheme to solve the output synchronization problem for a set of communication topologies. Moreover, if the synchronization trajectories are assumed to be bounded, a universal controller can be constructed for all communication topologies which contain a directed spanning tree.
16:20-16:40	FrC6.2
Marginal Dynamics of Stochastic Biochemical Networks in Random Environments, pp. 4269-4274
Zechner, Christoph	ETH Zurich
Deb, Soudeep	ETH Zurich
Koeppl, Heinz	ETH Zurich
Stochastic simulation algorithms provide a powerful means to understand complex biochemical processes as well as to solve the inverse problem of reconstructing hidden states and parameters from experimental single-cell data. At presence, a repertoire of efficient algorithms for simulating and calibrating stochastic reaction networks is available. However, most of these approaches do not account for the fact that each cell of a clonal population is exposed to a random extrinsic environment, i.e., the agglomerate of so-called extrinsic factors such as cell size, shape or cell cycle stage. We recently proposed a dynamic description of stochastic chemical kinetics in random but unknown extrinsic environments, reflected by a stochastic process where uncertain parameters are marginalized out. In this work we further investigate that process and provide additional analytical results. We demonstrate the marginalization using several biologically relevant parameter distributions and derive exact waiting-time distributions. We further show that the marginalized process model can achieve a variance reduction in the context of parameter inference.
16:40-17:00	FrC6.3
Synchronization Behaviors in Goodwin Oscillator Networks Driven by External Periodic Signals, pp. 4275-4280
Nguyen, Dinh Hoa	Graduate School of Information Science and Tech. Univ.
Hara, Shinji	The Univ. of Tokyo
In this paper, we present a systematic approach based on harmonic balance method to study the induced oscillations in a class of Goodwin oscillator networks forced by external periodic signals. Analytical expressions on the dependence of the phases and amplitudes of network oscillations to those of forcing inputs are revealed. Based on those expressions, we further show that under some specific conditions, the amplitude and phase shift of synchronized oscillations in networks of Goodwin oscillators monotonically depend on the amplitude of exciting inputs. The theoretical results are then illustrated through some examples.
17:00-17:20	FrC6.4
Observer-Based Approach for Fractional-Order Chaotic Synchronization and Communication, pp. 4281-4286
N'Doye, Ibrahima	Univ. of Luxembourg
Darouach, Mohamed	CRAN CNRS UMR 7039, Nancy Univ.
Voos, Holger	Univ. of Luxembourg
This paper presents a method based on the linear state observer design for constructing a chaotically synchronized systems. Fractional-order direct Lyapunov theorem is used to derive the closed-loop asymptotic stability in terms of linear matrix inequalities (LMIs) formulation. The gains of the observer and observer-based controller are obtained in terms of linear matrix inequalities. The proposed approach is then applied to suggest a means to secure communications. The method combines chaotic masking and chaotic modulation, where the information signal is injected into the transmitter and simultaneously transmitted to the receiver. Chaotic synchronization and chaotic communication are achieved simultaneously via a linear state observer design technique. An numerical fractional-order chaotic Lorenz system is given to demonstrate the applicability of the proposed approach.
17:20-17:40	FrC6.5
Frequency-Domain Analysis of Self-Excited Oscillations for a Class of Multivariable Relay Systems, pp. 4287-4292
Boiko, Igor	Petroleum Inst.
Pisano, Alessandro	Univ. di Cagliari
Usai, Elio	Univ. degli Studi di Cagliari
This paper investigates certain complex oscillatory phenomena taking place in multivariable relay control systems. A frequency based approach is applied, by making use of such tools of frequency-domain analysis as the Locus of a Perturbed Relay System (LPRS) and transfer properties analysis via the equivalent gain concept. As the novelty, we provide mathematical models and methods of analysis applicable to complex oscillatory modes not dealt with in the existing literature. A thoroughly discussed analysis and simulation example are outlined to validate the presented results.
17:40-18:00	FrC6.6
A Non-Quadratic Criterion for Stability of Forced Oscillations and Its Application to Flight Control, pp. 4293-4298
Pogromsky, A. Yu.	Eindhoven Univ. of Tech.
Matveev, Alexey S.	St.Petersburg Univ.
Andrievsky, Boris R.	Inst. for Problems of Mechanical Engin.
Leonov, Gennady	Saint Peterburg satate Univ.
Kuznetsov, Nikolay	Saint Petersburg State Univ. Univ. of Jyvaskyla
A new test for stability of forced oscillations in nonlinear systems is applied to a marginally stable plant of the second order with a saturated PID-control and a harmonic reference signal. The range of its frequencies and amplitudes that guarantee stability of the forced oscillations is found. This result is used for stability analysis of an aircraft roll angle control system.

FrC7	HG E3
Fuzzy Control (Regular Session)
Chair: Oliveira, Vilma A.	Univ. de Sao Paulo
Co-Chair: Chadli, Mohammed	Univ. de Picardie-Jules Verne
16:00-16:20	FrC7.1
Actuator Faults Reconstruction Using Reduced-Order Fuzzy Observer Structures, pp. 4299-4304
Krokavec, Dusan	Tech. Univ. of Kosice
Filasova, Anna	Tech. Univ. of Kosice
This paper focuses on the principle for designing reduced-order fuzzy-observer-based actuator fault reconstruction for nonlinear systems. The problem addressed can be indicated as the generalized approach for a kind of reduced-order fuzzy observer design with special gain matrix structure that depends on the given matching conditions. Using the Lyapunov theory, design conditions are obtained and expressed in terms of linear matrix inequalities and the terms for uniform estimation of the actuator faults are given.
16:20-16:40	FrC7.2
Stabilizing Switched T-S Fuzzy Systems Using a Fuzzy Lyapunov Function Approach, pp. 4305-4310
Faria, Flavio Andrade	Univ. de Sao Paulo
Valentino, Michele Cristina	Univ. de Sao Paulo
Oliveira, Vilma A.	Univ. de Sao Paulo
In this work, sufficient conditions for the existence of switching laws for stabilizing switched T-S fuzzy systems via a fuzzy Lyapunov function are proposed. The conditions are found by exploring properties of the membership functions and are formulated in terms of linear matrix inequalities (LMIs). A state feedback design is also used to extend the applicability of the results. Numerical examples illustrate the effectiveness of the proposed design methods.
16:40-17:00	FrC7.3
Synthesis of a Unknown Inputs Proportional Integral Observer for TS Fuzzy Models, pp. 4311-4315
Youssef, Tewfik	M'hamed Bougara Univ. of Boumerdes (UMBB)
Chadli, Mohammed	Univ. de Picardie-Jules Verne
Zelmat, Mimoun	M'hamed Bougara Univ. of Boumerdes
In this paper, synthesis conditions of a proportional integral observer for a Takagi-Sugeno (TS) fuzzy models subject to unknown inputs and unmeasurable decision variables are established. These unknown inputs affect both state and output of the system. The synthesis of this observer is based on hypothesis that the unknown inputs are under the polynomials form with bounded norm of their kth derivatives. The Lyapunov theory and L2–gain technique are used to develop the stability conditions of such observers in LMIs formulation. A numerical example is proposed to validate the proposed design conditions.
17:00-17:20	FrC7.4
Advanced Control Solutions to Increase Efficiency of a Furnace Combustion Process, pp. 4316-4321
Zanoli, Silvia Maria	Univ. Pol. delle Marche
Barchiesi, David	Univ. Pol. delle Marche
Astolfi, Giacomo	Univ. Pol. marche
Barboni, Luca	api raffineria di Ancona
In the present work the problem of the furnaces combustion optimization in petrochemical environment is presented. In particular, the paper is focused on the combustion efficiency that directly affects the operating costs of the plant. A preliminary study of the combustion process has been performed. A model of the system has been obtained by a black-box approach and limitations of the existing control architecture have been analyzed. A new control architecture, based on advanced PID control architecture, coupled “cross-limiting” control logics and Fuzzy logic has been developed and implemented in a Distributed Control System (DCS). The major benefits introduced by the new control system can be found in its reliability and in its robustness to compensate the measurable disturbances that affect the furnace. Moreover, the proposed control scheme has been proven to be effective in the reduction of the O2 content in the exhaust of furnace gas as well as in the reduction of the fuel consumption. As a consequence of the O2 reduction a reduction of the exhaust gas temperature has been achieved thus further increasing the furnace efficiency. The total efficiency increase has been estimated of about 2.2% with a significant energy saving of about 500 k€/year. Finally, the reduction of nitrogen oxide and carbon monoxide concentrations in the exhaust gases achieved by the new control strategy, allows minimizing the pollution emissions satisfying the actual national environmental requirements.
17:20-17:40	FrC7.5
An Extension of the Invariance Principle for Switched T-S Fuzzy Systems, pp. 4322-4327
Valentino, Michele Cristina	Univ. de Sao Paulo
Faria, Flavio Andrade	Univ. de Sao Paulo
Oliveira, Vilma A.	Univ. de Sao Paulo
In this work, we propose an extension of the invariance principle for switched T-S fuzzy systems using an auxiliary function V for the convex combination of the subsystems which applies to unstable subsystems. The results are given in terms of LMIs. Numerical examples illustrate the effectiveness of the proposed design methods.
17:40-18:00	FrC7.6
Model Predictive Control of a HVAC System Based on the LoLiMoT Algorithm, pp. 4328-4333
Schwingshackl, Daniel	Alpen-Adria-Univ. Klagenfurt
Rehrl, Jakob	Alpen-Adria-Univ. Klagenfurt
Horn, Martin	Klagenfurt Univ.
In this paper a multi-variable control for the simultaneous control of the air temperature and the air humidity of an industrial heating ventilating and air conditioning (HVAC) system is presented. For the multi-input-multi-output (MIMO) control the model predictive control (MPC) method is applied. To model the system local linear neuro fuzzy models (LLNFM) are used and computed by the so-called Local Linear Model Tree (LOLIMOT) algorithm. The main focus of this work is the communication between the MPC and the LLNFMs and the application on a real world test plant. The proposed method is compared to a linear MPC scheme and a conventional PI control strategy.

FrC8	HG E5
Distributed Control of Large-Scale Systems (Regular Session)
Chair: Farina, Marcello	Pol. di Milano
Co-Chair: Stursberg, Olaf	Univ. of Kassel
16:00-16:20	FrC8.1
Building Temperature Distributed Control Via Explicit MPC and “Trim and Respond” Methods, pp. 4334-4339
Koehler, Sarah	Univ. of California, Berkeley
Borrelli, Francesco	Univ. of California, Berkeley
We study the distributed control of heating, ventilation, and air conditioning systems in buildings. We present and compare two control design techniques. In the first part, a one-step distributed model predictive controller (DMPC) is introduced. We compute the explicit state-feedback solution and show how to implement it on existing networked buildings control platforms. In the second part, we discuss a control logic currently used in the buildings industry called “Trim and Respond”. We show that “Trim and Respond” controllers can be seen as a special class of one-step DMPC algorithms. We conclude the paper with a simulation study which shows the DMPC control design and its equivalence with “Trim and Respond” heuristics.
16:20-16:40	FrC8.2
Optimized Distributed Control and Topology Design for Hierarchically Interconnected Systems, pp. 4340-4346
Jilg, Martin	Univ. of Kassel
Stursberg, Olaf	Univ. of Kassel
A method to reduce the computational complexity for the simultaneous design of a communication topology and feedback control laws for large scale systems is proposed. In general, such combined procedures may be posed as Mixed- Integer Programs (MIPs), which suffer from high combinatorial complexity when the number of possible communication links grows large. Although some explicit solutions for MIP formulations exist, these are either based on very restrictive assumptions or yield an iterative LMI procedure, which is computationally expensive. The presented scheme tackles the problem by pre-analyzing the coupling structure of the plant and dividing it into hierarchically coupled, distinct groups (clusters). This enables one to decompose the global MIP into a set of smaller MIPs, which can then be solved independently. In contrast to existing approaches, a global LMI optimization has to be solved only once, not repeatedly.
16:40-17:00	FrC8.3
A Solution to the Tracking Problem Using Distributed Predictive Control, pp. 4347-4352
Farina, Marcello	Pol. di Milano
Scattolini, Riccardo	Pol. di Milano
Betti, Giulio	Pol. di Milano
A Distributed Predictive Control (DPC) algorithm for tracking (piecewise) constant output reference signals is presented in this paper. The overall system is assumed to be composed by a number of discrete-time linear subsystems interconnected through the states and/or the inputs. The algorithm is non-cooperative, based on neighbor-to-neighbor communication, and does not require an iterative exchange of information among neighbors. Unfeasible reference signals, that cannot be reached due to state and control constraints, can also be considered by computing the nearest feasible reference value. A simulation example is reported.
17:00-17:20	FrC8.4
Distributed Estimation and Control for Large Population Stochastic Multi-Agent Systems with Coupling in the Measurements, pp. 4353-4358
Abedinpour Fallah, Mehdi	École Pol. de Montréal
Malhame, Roland P.	Ec. Pol. de Montreal
Martinelli, Francesco	Univ. di Roma Tor Vergata
In this paper, we investigate a class of large population stochastic multi-agent systems where the agents have linear stochastic dynamics and are coupled via their measurement equations. Using the state aggregation technique, we propose a distributed estimation and control algorithm that combines the Kalman filtering for state estimation and the linear-quadratic-Gaussian (LQG) feedback controller. Moreover, the stability analysis in terms of exponential boundedness in the mean square is given for the proposed algorithm.
17:20-17:40	FrC8.5
Stability Analysis of Networked Systems with Similar Dynamics, pp. 4359-4364
Schuh, René	Ruhr-Univ. Bochum
Lunze, Jan	Ruhr-Univ. Bochum
The paper concerns systems that are composed of subsystems with similar dynamics. A method for analyzing the stability of such systems is developed, where the subsystems are decomposed into nominal models with identical dynamics and individual models describing the deviation of the subsystems from the common nominal models. The analysis is divided into three steps. First, the stability of the network controlled and physically interconnected nominal subsystems is checked by a state transformation based on the structure of the interconnection. The second step is to find an upper bound for the dynamics of the deviation model. The third step is to prove the stability of the overall system by considering the stabilized nominal system and the error bound. A sufficient condition for the stability of the overall system is derived and the method is demonstrated by proving the stability of platooning vehicles.
17:40-18:00	FrC8.6
A Class of Mean-Risk Decisions for Noncooperative Games and Distributed Controls, pp. 4365-4370
Pham, Khanh	U. S. Air Force Res. Lab.
Gubar, Elena	St. Petersburg State Univ.
The paper gives a comprehensive presentation of the broad and still developing area of risk-averse decision making in the stochastic linear-quadratic class of multiperson noncooperative differential games and/or distributed controls. The fundamental set of design principles is developed for self-directed decision makers, who are imitatively cautious and privately capable of incorporating summary performance-measure statistics into their distributed output-feedback Nash decision policies for competitive interactions and engagements.

FrC9	HG E7
Autonomous Robots and Vehicles (Regular Session)
Chair: Kyriakopoulos, Kostas J.	National Tech. Univ. of Athens
Co-Chair: Conte, Christian	ETH Zurich
16:00-16:20	FrC9.1
Reactive Navigation of a Mobile Robot to the Moving Extremum of a Dynamic Unknown Environmental Field without Derivative Estimation, pp. 4371-4376
Matveev, Alexey S.	St.Petersburg Univ.
Hoy, Michael C.	Univ. of New South Whales
Savkin, Andrey V.	Univ. of New South Wales
We consider a single kinematically controlled mobile robot traveling in a planar region supporting an unknown and unsteady field distribution. A single sensor provides the distribution value at the current robot location. We present a reactive navigation strategy that drives the robot to the time-varying location where the field distribution attains its spatial maximum and then keeps the robot in the pre-specified vicinity of the maximizer. The proposed control algorithm employs estimation of neither the entire field gradient nor derivative-dependent quantities, like the rate at which the available measurement evolves over time, and is non-demanding with respect to both computation and motion. Its mathematically rigorous analysis and justification are provided. Simulation results confirm the applicability and performance of the proposed guidance approach.
16:20-16:40	FrC9.2
Region Reaching Controller for Autonomous Underwater Vehicle without Velocity Measurement, pp. 4377-4381
Mukherjee, Koena	IIT Delhi
Kar, Indra Narayan	Indian Inst. of Tech. Delhi
Bhatt, R. K. P.	Indian Inst. of Tech. Delhi
This paper addresses the problem of region reaching by using the concept of potential energy of the specified region, for an Autonomous Underwater Vehicle (AUV). The six degrees of freedom AUV model has been exploited with Euler angle representation. The control law incorporates high filtered version of position error to regulate the system without velocity measurement. The proposed control law ensures stability in Lyapunov sense and simulation results shows the convergence of positional and orientation error.
16:40-17:00	FrC9.3
A Collision Avoidance Strategy for Safe Autonomous Navigation of an Intelligent Electric-Powered Wheelchair in Dynamic Uncertain Environments with Moving Obstacles, pp. 4382-4387
Wang, Chao	Univ. of New South Wales
Matveev, Alexey S.	St.Petersburg Univ.
Savkin, Andrey V.	Univ. of New South Wales
Nguyen, Tuan Nghia	Univ. of Tech. Sydney, Australia
Nguyen, Hung T.	Univ. of Tech. Sydney
We present a reactive navigation algorithm that guarantees the safety of automated intelligent wheelchairs for people with mobility impairments in dynamic uncertain environments. The proposed navigation algorithm restricts neither the natures nor the motions of the obstacles, the shapes of the obstacles can be time-varying (deforming obstacles). Furthermore, the proposed navigation algorithm does not require prior information about the positions and velocities of the obstacles to accomplish obstacle avoidance. Performance of the proposed algorithm is highlighted via theoretical study. Simulation and experimental results show that intelligent electric-powered wheelchairs are able to successfully avoid collisions with moving obstacles such as pedestrians or vehicles
17:00-17:20	FrC9.4
Robust Prescribed Performance Tracking Control for Unknown Underactuated Torpedo-Like AUVs, pp. 4388-4393
Bechlioulis, Charalampos	Aristotle Univ. of Thessaloniki
Kyriakopoulos, Kostas J.	National Tech. Univ. of Athens
This paper addresses the tracking control problem of 3D trajectories for underactuated torpedo-like Autonomous Underwater Vehicles (AUVs). A smooth control scheme is designed, without any a priori knowledge of the AUV model parameters, guaranteeing prescribed performance tracking despite the presence of external disturbances representing ocean currents and waves. This work is the first approach on underactuated systems employing the Prescribed Performance notion, originally proposed in cite{2008Bechlioulis_Rovithakis, 2009Bechlioulis_Rovithakis,2010Bechlioulis_Rovithakis}. Moreover, only the desired trajectory and none of its higher order derivatives is employed in the control scheme. Furthermore, the stability of the unactuated degrees of freedom is secured without incorporating the corresponding measurements in the control signal, thus simplifying implementation and increasing robustness against measurement noises. The proposed control scheme: i) is of low complexity, ii) avoids singularities that arise owing to the error definition in the kinematic model and iii) guarantees, through the appropriate selection of certain performance functions, configuration constraints that may arise owing to the limited field of view of cameras in case a vision based sensor system with on board cameras is adopted. Finally, simulation results clarify and verify the approach.
17:20-17:40	FrC9.5
AUV Range-Only Localization and Mapping: Observer Design and Experimental Results, pp. 4394-4399
Bayat, Behzad (Mohammadreza)	IST (Inst. Superior Tecnico)
Aguiar, A. Pedro	Faculty of Engineering, Univ. of Porto (FEUP)
This paper addresses the localization problem of an Autonomous Underwater Vehicle using relative range measurements to stationary beacons whose locations are also unknown. We propose an observer that under observability-like assumptions drives the estimation errors to a small neighborhood of zero (whose size depends on the noise and disturbances). The observer is designed by combining the concepts of minimum energy estimators applied to continuous processes with discrete measurements (not necessarily with a fixed sampling time), adaptive multiple models estimators, and simultaneous localization and mapping techniques. We also combine the proposed solution with a projection filter that significantly improves the performance. Experimental results with the Medusa robotic vehicle are presented and discussed.
17:40-18:00	FrC9.6
Capturing a Target with Range Only Measurement, pp. 4400-4405
Chaudhary, Gaurav	KPIT Cummins Infosystems Limited
Sinha, Arpita	Indian Inst. of Tech. Bombay
This paper addresses the problem of guiding a mobile robot towards a target using only range sensors. The bearing information is not available. The target can be stationary or moving. It can be the source of some gas leakage or nuclear radiation or it can be some landmark or beacon or any maneuvering vehicle. The mobile robot can be ground or aerial vehicle. In literature, many different strategies are proposed which use the range only measurement but they involve estimation of different parameters or have switching control strategy which makes them difficult to implement. We have proposed a strategy that can bring the robot arbitrarily close to the moving target. It is shown analytically that this strategies works for any initial condition of the robot with respect to the target. The results are validated through simulation.

FrC10	HG D1.1
Delay Systems (Regular Session)
Chair: Efimov, Denis	INRIA - LNE
Co-Chair: Sipahi, Rifat	Northeastern Univ.
16:00-16:20	FrC10.1
On Controller Design for Delay-Independent Stability of Linear Time-Invariant Systems with Multiple Delays, pp. 4406-4411
Nia, Payam M.	Northeastern Univ.
Sipahi, Rifat	Northeastern Univ.
Shafai, Bahram	Northeastern Univ.
In this paper, we present a new control-design procedure for linear time-invariant (LTI) systems with ``multiple'' delays. These procedures are based on algebraic tools and allow designing controllers that can stabilize such systems regardless of how large/small the delays are. That is, with these controllers, the system at hand can be rendered delay-independent stable (DIS). The essence of the control design is based on the Rekasius transformation, algebraic tools, elimination techniques, and Sturm sequences. The advantages of the design procedure are that it simplifies the control design to managing the roots of some single-variable polynomials while also preserving the controller structure and complying with the necessary and sufficient conditions of stability.
16:20-16:40	FrC10.2
Interval Observer Approach to Output Stabilization of Time-Varying Input Delay System, pp. 4412-4417
Polyakov, Andrey	INRIA Lille Nord-Europe
Efimov, Denis	INRIA - LNE
Perruquetti, Wilfrid	Ec. Centrale de Lille
Richard, Jean-Pierre	Ec. Centrale de Lille
The output stabilization problem for a linear system with an unknown time-varying bounded input delay is considered. The interval observer technique is applied in order to obtain guaranteed interval estimates of the system state. The procedure of the interval observer design, which is based on resolving of the Sylvester's equation, is discussed. Interval predictor is introduced and applied for a linear output stabilizing feedback design. The design procedure is based on Linear Matrix Inequalities (LMI). The theoretical results are supported by numerical simulations.
16:40-17:00	FrC10.3
Inverse Signal Shapers in Effective Feedback Architecture, pp. 4418-4423
Hromcik, Martin	Czech Tech. Univ.
Vyhlidal, Tomas	Czech Tech. Univ. in Prague
Kucera, Vladimir	Czech Tech. Univ. in Prague, Faculty of Mechanical Engi
New feedback control architecture is proposed in this paper suitable for manipulation of weakly damped flexible structures. Delay-based signal shapers are considered since they are very popular in such applications for their simple implementation and very good performance. Unlike the classical approach where the shapers are used just to filter the reference command, we consider placing it inside the closed loop. Compared to previous attempts by other authors, we propose to include the inverse of the signal shaper's dynamics in the feedback path, and we justify this architecture by thorough analysis of important feedback loop channels, namely the feedback response from input disturbance to the output. As results from the spectral properties, classical ZV shapers with lumped delays are very inconvenient for the shaper inverse operation, leading to neutrality of the dynamics with the associated risks of stability loss due to high frequency unstable modes. Therefore, a recently introduced DZV shaper with a distributed delay element is utilized in the feedback loop. Results are verified by simulation example.
17:00-17:20	FrC10.4
Stability Analysis of Time-Delay Systems Based on a Power of the Monodromy Operator, pp. 4424-4429
Saito, Yuki	Kyoto Univ.
Hagiwara, Tomomichi	Kyoto Univ.
This paper studies stability analysis of linear time-invariant time-delay feedback systems consisting of a finite-dimensional system and a pure delay. The continuous-time state transition is described by the monodromy operator in discrete-time through the lifting technique, and a stability condition in terms of its spectral radius is first restated in terms of the norm of a power of the operator. Quasi-finite-rank approximation is then applied on the power of the operator, and by evaluating the effect of the approximation error on the norm, a new necessary and sufficient stability condition is given finally.
17:20-17:40	FrC10.5
Stochastic $H_infty$ Control and Estimation of State-Multiplicative Discrete-Time Systems with Delay, pp. 4430-4435
Gershon, Eli	Tel Aviv Univ.
Shaked, Uri	Tel-Aviv Univ.
Linear, state delayed, discrete-time systems with stochastic uncertainties in their state-space model are considered. The problems of both $H_infty$ state-feedback control and filtering are solved, for the stationary case, via an input-output approach by which the system is replaced by a nonretarded system with deterministic norm-bounded uncertainties. In both problems, a cost function is defined which is the expected value of the standard $H_infty$ performance index with respect to the uncertain parameters.
17:40-18:00	FrC10.6
State Feedback H∞ Control of Time-Delay Switched Linear Systems: A Descriptor Approach, pp. 4436-4441
Galbusera, Luca	CNR
Bolzern, Paolo	Pol. di Milano
In this paper we propose contributions on the control of switched linear systems subject to time-delays. Our methods allow to synthesize state feedback switching laws able to ensure the stabilization and the fulfillment of a prescribed H-infinity disturbance attenuation property. The proposed criteria are delay-dependent and based on an application of the descriptor approach, previously introduced in the literature to study standard time-delay systems. The core of our procedure is based on the solution of sets of matrix inequalities, which allow simultaneously to assign a suitable switching law and to determine a feasible piecewise quadratic Lyapunov-Krasovskii functional associated to it.

FrC11	HG D1.2
Optimization for Process Control (Regular Session)
Chair: Celikovsky, Sergej	Acad. of Sci. of Czech Republic
Co-Chair: Baramov, Lubomir	Honeywell Prague Lab.
16:00-16:20	FrC11.1
On Nonlinear Continuous-Time Optimal Control of Penicillin Cultivation, pp. 4442-4447
Pcolka, Matej	Czech Tech. Univ. in Prague
Celikovsky, Sergej	Acad. of Sci. of Czech Republic
This paper presents continuous-time optimal control of penicillin production maximizing penicillin concentration at chosen final time. First, control input parametrization is performed and the input is expressed as a piece-wise affine function of continuous time reducing the number of optimizable parameters. Then, iterative numerical gradient optimization with the initial conditions corresponding to the optimal input obtained by projected gradient optimization is used to find the optimal values of the optimized parameters. The proposed approach is compared to both the original gradient method and the traditionally used classical nonlinear feedback controller. All these strategies are tested on a set of numerical experiments for various cultivation lengths and the results are evaluated and discussed. The comparison reveals significant superiority of the proposed algorithm. Together with impressive reduction of memory space needed to store the solution, the results improvement makes the proposed algorithm very attractive from the industrial point of view.
16:20-16:40	FrC11.2
Robust Optimization on Receding Horizon of Processes with Storages and Periodic Production and Consumption Contracts, pp. 4448-4453
Baramov, Lubomir	Honeywell Prague Lab.
Havlena, Vladimir	Honeywell ACS AT Lab.
This paper presents a real-time optimizer for processes with material storages and periodic targets on integrals of optimized variables. A problem of robust target attainment with respect to uncertainties in future disturbance trajectory estimates was formulated. This robustness is achieved by means of soft constraints forcing future trajectories to a time-varying set from which reaching the target is feasible. An algorithm for computing this set is outlined, which is tractable for problems of medium size. The application area for this optimizer is assumed to be in (but not limited to) heat and power plant optimization with multiple fuels with periodic contracts on electric energy delivered and fuel gas consumed during the contract period.
16:40-17:00	FrC11.3
Optimized Pacing of Continuous Reheating Furnaces, pp. 4454-4459
Steinboeck, Andreas	Vienna Univ. of Tech.
Kugi, Andreas	Vienna Univ. of Tech. (VUT)
A pacing algorithm for optimizing process times in continuous reheating furnaces of a rolling mill is presented. The problem is solved by an iterative algorithm that uses quadratic programming in each iteration loop. The method takes into account relevant time constraints and the variable local reheating power based on a mathematical furnace model. An example problem taken from an industrial application demonstrates the feasibility of the proposed method.
17:00-17:20	FrC11.4
PID Design by Convex-Concave Optimization, pp. 4460-4465
Hast, Martin	LTH, Lund Univ.
Astrom, Karl J	Lund Inst. of Tech.
Bernhardsson, Bo M.	Lund Inst. of Tech.
Boyd, Stephen P.	Stanford Univ.
This paper describes how PID controllers can be designed by optimizing performance subject to robustness constraints. The optimization problem is solved using convex-concave programming. The method admits general process descriptions in terms of frequency response data and it can cope with many different constraints. Examples are presented and some pitfalls in optimization are discussed.
17:20-17:40	FrC11.5
On the Links between Real-Time Optimization, Neighboring-Extremal Control, and Self-Optimizing Control, pp. 4466-4471
Marchetti, Alejandro Gabriel	CIFASIS-CONICET
Zumoffen, David Alejandro	French-Argentine International Center for Information and System
In the context of process optimization, several techniques have been proposed in order to operate the plant near the economically optimal steady-state operating point. This paper discusses the links and differences between three of these techniques: (i) real-time optimization using the classical two-step approach, (ii) neighboring-extremal control with output feedback, and (iii) self-optimizing control using the null space method. It is argued that, under certain assumptions, the technique (ii) can be viewed as a first-order approximation of (i). Also, while (ii) and (iii) enforce similar optimal operations, their dynamic implementation and design features are very different.
17:40-18:00	FrC11.6
Optimal Selection of Sensor Network and Backed-Off Operating Point Based on Economics, pp. 4472-4477
Magbool Jan, Nabil	Indian Inst. of Tech. Madras
Skogestad, Sigurd	Norwegian Univ. of Science & Tech.
Narasimhan, Sridharakumar	IIT Madras
This paper discusses the simultaneous selection of measurements and economic backed off operating point when the nominal optimal operating point is constrained. However, operation at this point becomes infeasible due to uncertainties. In this regard, we recently proposed an optimization formulation that determines the backed-off point to ensure feasibility assuming accurate measurement of the states is available and disturbance as the only source of uncertainty. In the present work, we extend the formulation to partial state information case and also determine the optimal set of measurements for economical operation. The formulation also finds a suitable multivariable controller to achieve economic benefits. The final formulation is a mixed integer non-linear program. Here, we present a branch and bound type solution such that a two stage iterative problem is solved at each branching step. Finally, the proposed approach is demonstrated in an evaporator system.

FrC12	HG D3.2
Model Reduction (Regular Session)
Chair: Poussot-Vassal, Charles	Onera
Co-Chair: Jayawardhana, Bayu	Univ. of Groningen
16:00-16:20	FrC12.1
Model Reduction in Bio-Chemical Reaction Networks with Michaelis-Menten Kinetics, pp. 4478-4483
Hangos, Katalin M.	Computer & Automation Rsrch. Inst. of the HungarianAcademyof Sci
Gábor, Attila	IIM-CSIC
Szederkényi, Gábor	Computer and Automation Res. Inst.
In this paper, a model reduction procedure is proposed for the simplification of biochemical reaction network models. The approach is capable of reducing ODE models where the right hand side of the equations contains polynomial and/or rational function terms. The method is based on a finite number of mixed integer quadratic programming (MIQP) steps where the objective function effectively measures the fit between the time functions of the selected concentrations of the original and the reduced models, and the integer variables keep track of the presence of individual reactions. The procedure also contains the re-estimation of rate coefficients in the reduced model to minimize the defined model error. Two examples taken from the literature illustrate the operation of the method.
16:20-16:40	FrC12.2
H_2 and H_infinity Error Bounds for Model Order Reduction of Second Order Systems by Krylov Subspace Methods, pp. 4484-4489
Panzer, Heiko K. F.	Tech. Univ. München
Wolf, Thomas	Tech. Univ. München
Lohmann, Boris	Tech. Univ. Muenchen
We present rigorous bounds on the~$mathcal{H}_2$ and $mathcal{H}_infty$ norm of the error resulting from model order reduction of second order systems by textsc{Krylov} subspace methods. To this end, we use a strictly dissipative state space realization of the model and perform a factorization of the error system. The derived error expressions are easy to compute and can therefore be applied to models of very high order, as is demonstrated in numerical examples. In fact, all results hold true for arbitrary state space models in strictly dissipative realization that do not necessarily have to originate from second order systems.
16:40-17:00	FrC12.3
An Iterative Eigenvector Tangential Interpolation Algorithm for Large-Scale LTI and a Class of LPV Model Approximation, pp. 4490-4495
Poussot-Vassal, Charles	Onera
Vuillemin, Pierre	Onera - The French Aerospace Lab.
In this paper, a new Iterative Eigenvector Tangential Interpolation Algorithm (IETIA) is proposed to approximate large-scale LTI models, allowing to achieve H2 optimality-like conditions while preserving some user-defined eigenvalues. The proposed algorithm is also shown to be applicable to a certain class of LPV models. Numerical applications on standard model reduction benchmark assess the effectiveness of the approach.
17:00-17:20	FrC12.4
Method of Characteristics Based Model Reduction for Control of a Counter-Current Reactor Using Approximate Dynamic Programming, pp. 4496-4501
M, Sudhakar	INDIAN Inst. OF Tech. MADRAS
Narasimhan, Sridharakumar	IIT Madras
Kaisare, Niket	ABB Global Industries and Services Ltd.
The aim of this paper is to use the method of characteristics (MOC) as a technique to obtain reduced order model for control of a counter-current reactor. The reactor is described by a system of quasi-linear hyperbolic partial differential equations (PDEs). Due to the counter-current operation, the system exhibits characteristic lines with double slopes; the slope of the characteristics have opposing signs. A reduced order model is obtained by using only a limited number of characteristic lines in the time-space plane. An approximation technique is proposed to obtain state variables at each sampling instance. This reduced-order MOC model is used for control in the approximate dynamic programming (ADP) framework for set-point tracking.
17:20-17:40	FrC12.5
Model-Order Reduction of Biochemical Reaction Networks, pp. 4502-4507
Rao, Shodhan	Univ. of Groningen
van der Schaft, Arjan J.	Univ. of Groningen
van Eunen, Karen	Univ. Medical Center Groningen
Bakker, Barbara M.	Univ. Medical Center Groningen
Jayawardhana, Bayu	Univ. of Groningen
In this paper we propose a model-order reduction method for chemical reaction networks governed by general enzyme kinetics, including the mass-action and Michaelis-Menten kinetics. The model-order reduction method is based on the Kron reduction of the weighted Laplacian matrix which describes the graph structure of complexes in the chemical reaction network. We apply our method to a yeast glycolysis model, where the simulation result shows that the transient behaviour of a number of key metabolites of the reduced-order model is in good agreement with those of the full-order model.
17:40-18:00	FrC12.6
Singular Perturbation Approximation of Semistable Linear Systems, pp. 4508-4513
Ishizaki, Takayuki	Tokyo Inst. of Tech.
Sandberg, Henrik	Royal Inst. of Tech. (KTH)
Johansson, Karl Henrik	Royal Inst. of Tech.
Kashima, Kenji	Osaka Univ.
Imura, Jun-ichi	Tokyo Inst. of Tech.
Aihara, Kazuyuki	The Univ. of Tokyo
This paper proposes a singular perturbation approximation for semistable linear systems. In particular, we derive a novel expression of error systems in the Laplace domain. As a result, we obtain an H2-error bound in terms of the sum of eigenvalues of an index matrix, which coincides with a controllability gramian of the state-derivative. Furthermore, we show that the singular perturbation model appropriately preserves the semistability of the original system and also guarantees the stability of the error system. The efficiency of the proposed method is shown through a numerical example of a Markov chain model approximation.

FrC13	HG D5.2
Robust Control and Estimation (Regular Session)
Chair: Rovithakis, George A.	Aristotle Univ. of Thessaloniki
Co-Chair: Fikar, Miroslav	Slovak Univ. of Tech.
16:00-16:20	FrC13.1
Guaranteed Parameter Estimation in Nonlinear Dynamic Systems Using Improved Bounding Techniques, pp. 4514-4519
Paulen, Radoslav	Slovak Univ. of Tech.
Villanueva, Mario	Imperial Coll. London
Fikar, Miroslav	Slovak Univ. of Tech.
Chachuat, Benoît	Imperial Coll. London
This paper is concerned with guaranteed parameter estimation in nonlinear dynamic systems in a context of bounded measurement error. The problem consists of finding---or approximating as closely as possible---the set of all possible parameter values such that the predicted outputs match the corresponding measurements within prescribed error bounds. An exhaustive search procedure is applied, whereby the parameter set is successively partitioned into smaller boxes and exclusion tests are performed to eliminate some of these boxes, until a prespecified threshold on the approximation level is met. Exclusion tests rely on the ability to bound the solution set of the dynamic system for a given parameter subset and the tightness of these bounds is therefore paramount. Equally important is the time required to compute the bounds, thereby defining a trade-off. It is the objective of this paper to investigate this trade-off by comparing various bounding techniques based on interval arithmetic, Taylor model arithmetic and ellipsoidal calculus. When applied to a simple case study, ellipsoidal and Taylor model approaches are found to reduce the number of iterations significantly compared to interval analysis, yet the overall computational time is only reduced for tight approximation levels due to the computational overhead.
16:20-16:40	FrC13.2
Robust Filtering and Fixed-Lag Smoothing for Uncertain Discrete-Time Systems, pp. 4520-4525
Neveux, Philippe	Univ. d'Avignon et des Pays de Vaucluse
Blanco, Eric	Ec. Centrale de Lyon, Ampere Lab.
Robust filtering and fixed-lag smoothing problems for uncertain discrete-time systems are treated in this paper. The problem is set in a $lambda$-contaminated framework that does not required a model for the uncertainties. The results are presented in term of transfer function and in polynomials. The latter representation exhibits two spectral factors and a Diophantine equation in order to obtain the estimator. An exemple shows the effectiveness of the approach.
16:40-17:00	FrC13.3
An Intuitive Design for the Dual Mode Adaptive Robust Controller Based on Indirect Control, pp. 4526-4531
Teixeira, Leonardo	Federal Inst. of Education, Science and Tech. of Rio Gr
Barbosa, Oliveira, Josenalde	Agricultural School of Jundiaí
Araujo, Aldayr Dantas de	Federal Univ. of Rio Grande do Norte
In this paper it is proposed an indirect approach to the Dual Mode Adaptive Robust Controller (DMARC), which combines the typical transient and robustness properties of Variable Structure Systems, with a smooth control signal in steady-state, typical of conventional Adaptive Controllers, as Model Reference Adaptive Controller (MRAC). The aim of this indirect version, here named Indirect Dual Mode Adaptive Robust Controller (IDMARC), is to provide a more intuitive controller design, based on physical plant parameters, as resistances, inertia moments, capacitances, etc maintaining DMARC properties. In this paper, it will be presented a stability analysis for the proposed controller and simulations to an unstable second order plant.
17:00-17:20	FrC13.4
On the Numerical Optimization Design of Continuous-Time Quantizer: A Matrix Uncertainty Approah, pp. 4532-4537
Sawada, Kenji	The Univ. of Electro-Communications
Shin, Seiichi	the Univ. of Electro-Communications
For the networked control systems, the quantized control problem is one of the challenging problems since the continuous-valued signals are compressed and quantized to the discrete-valued signals via the communication channel and such the quantization often degrades the control performance. In terms of the broadbandization and the robustness of the networked control systems, this paper considers the continuous-time quantized control. In the quantized control, it is important to design a quantizer that minimizes the output difference between before and after the quantizer implementation. This paper describes a numerical optimization method of a continuous-time quantizer considering the switching speed. Using a matrix uncertainty approach of sampled-data control, we clarify that both of the temporal and spatial resolution constraints can be considered in analysis and synthesis, simultaneously.
17:20-17:40	FrC13.5
H-Infinity Output Feedback Control of Commensurate Fractional Order Systems, pp. 4538-4543
Fadiga, Lamine	Univ. de Bordeaux 1
Farges, Christophe	IMS-LAPS
Sabatier, Jocelyn	LAPS - Bordeaux 1 Univ.
Santugini, Kevin	Univ. de Bordeaux 1, Inst. de Mathématiques de Bordeaux
This paper addresses the problem of H-infinity output feedback control of commensurate Fractional Order Systems (FOS)of order 1<v<2. Along the lines of recent work on H-infinity norm computation for FOS, an extension of Bounded Real Lemma for FOS is proposed. This lemma is used to derive a method to design H-infinity output feedback control laws. The efficiency of this method is evaluated on a numerical example.
17:40-18:00	FrC13.6
Approximation-Free Prescribed Performance Control for Unknown SISO Pure Feedback Systems, pp. 4544-4549
Bechlioulis, Charalampos	Aristotle Univ. of Thessaloniki
Rovithakis, George A.	Aristotle Univ. of Thessaloniki
A universal control scheme is designed for unknown pure feedback systems, capable of guaranteeing, for any initial system condition, output tracking with prescribed performance and bounded signals in the closed loop. In this paper, by prescribed performance, it is meant that the output error converges to a predened arbitrarily small residual set, with convergence rate no less than a certain prespecied value. The proposed approximation-free and low-complexity control scheme isolates the output performance from the control gains selection and exhibits strong robustness against model uncertainties. In fact, any system in pure feedback form obeying certain controllability assumptions can be controlled by the proposed scheme without altering either the controller structure or the control gain values. Finally, a simulation study claries and veries the approach.

FrC14	HG D7.1
Railway Control (Regular Session)
Chair: Pendharkar, Ishan	Bombardier Transportation Switzerland
Co-Chair: Aschemann, Harald	Univ. of Rostock
16:00-16:20	FrC14.1
Optimal Trajectory Planning for Standard and Hybrid Railway Vehicles with a Hydro-Mechanic Transmission, pp. 4550-4555
Leska, Maik	Univ. of Rostock
Grüning, Tobias	Univ. of Rostock
Aschemann, Harald	Univ. of Rostock
Rauh, Andreas	Univ. of Rostock
This paper presents numerical results for the optimization of trajectories for standard diesel and hybrid railway vehicles with respect to the fuel consumption between two successive stops. First, a nonlinear model of the complete power train of a standard railway vehicle, purely driven by an internal combustion engine, is described. Based on the four different operating regimes – acceleration, cruising, coasting, and braking – fuel-optimal trajectories are computed by determining the optimal regime sequence. For this purpose, the golden ratio search method is used. Second, the presented approach is extended towards the trajectory optimization for parallel hybrid railway vehicles.
16:20-16:40	FrC14.2
Optimal Trajectory Planning for Trains under a Moving Block Signaling System, pp. 4556-4561
Wang, Yihui	Delft Univ. of Tech.
De Schutter, Bart	Delft Univ. of Tech.
van den Boom, Ton J. J.	Delft Univ. of Tech.
Ning, Bin	Beijing Jiaotong Univ.
The optimal trajectory planning problem for trains under a moving block signaling system is considered. This optimal trajectory planning problem is significant for punctuality, energy consumption, passenger comfort, etc. In a moving block signaling system, the minimum distance between two successive trains is the instantaneous braking distance required by the following train plus a safety margin. The constraints caused by the moving block signaling system are described as nonlinear inequalities, which can be transformed into linear inequalities using piecewise affine approximations. The optimal trajectory planning problem is subsequently recast as a mixed integer linear programming problem, which can be solved efficiently by existing solvers. A case study is used to demonstrate the performance of the proposed approach.
16:40-17:00	FrC14.3
Performance Analysis of MILP Based Model Predictive Control Algorithms for Dynamic Railway Scheduling, pp. 4562-4567
Rudan, János	Péter Pázmány Catholic Univ.
Kersbergen, Bart	Tech. Univ. Delft
van den Boom, Ton J. J.	Delft Univ. of Tech.
Hangos, Katalin M.	Computer & Automation Rsrch. Inst. of the HungarianAcademyof Sci
In this paper we analyze the performance of solvers for Mixed Integer Linear Programming (MILP) problems that appear from the model predictive control of railway networks. Our aim is to study techniques that reduce the amount of delay using dynamic traffic management by the rescheduling of trains. Due to the size of the emerging MILP problem and the given constraints on solution time, a thorough analysis of different MILP solution techniques was necessary. It has been proven that a significant speedup in the solution time can be achieved by the proper restructuring of the matrices of the MILP problem. The simulation results also confirm the effectiveness of the proposed control technique and the ability of this setup to analyze the most delay-sensitive trains in the network.
17:00-17:20	FrC14.4
A Robust Control of Contact Force of Pantograph-Catenary for the High-Speed Train, pp. 4568-4573
Mokrani, Nassim	Univ. of Picardie Jules Verne
Rachid, Ahmed	Univ. de Picardie Jules Verne
In this paper, a simplified model to three degree of freedom (3-DOF) of the pantograph-catenary (PAC) system is represented. A robust control of contact force between the pantograph and the catenary for the high-speed trains using the fuzzy sliding mode controller and the additional compensator is presented. A PID outer loop in the control law is use then the gains of the sliding term and PID term are tuned on-line by a fuzzy system. The result of the simulation is fit for the actual phenomenon. Thus, the method of this paper is valid
17:20-17:40	FrC14.5
Resonance Stability in Electrical Railway Systems – a Dissipativity Approach, pp. 4574-4579
Pendharkar, Ishan	Bombardier Transportation Switzerland
We propose a theory for stability analysis of railway systems based on energy exchange principles between the infrastructure and rolling stock. We show that if the rolling stock and infrastructure is dissipative with respect to specially constructed power functions (more general than voltage x current) stability is guaranteed. We thereby provide a proof of the Input Admittance Criterion (IAC) -- a widely used criterion in the railway industry -- as special case of our results. Application of the theory is demonstrated with an example from a metro train system.

FrC15	HG D7.2
Control of Electrical Motors (Regular Session)
Chair: Maciejowski, Jan M.	Univ. of Cambridge
Co-Chair: Marinkov, Sava	Eindhoven Univ. of Tech.
16:00-16:20	FrC15.1
Reference Frame Re-Alignment for Vector Control of the Brushless Doubly-Fed Machine, pp. 4580-4585
Broekhof, Alexander	Univ. of Cambridge
McMahon, Richard	Cambridge Univ. Engineering Department
Maciejowski, Jan M.	Univ. of Cambridge
Vector control provides stability and performance when applied to the brushless doubly-fed machine, however cross- coupling effects can arise between inputs and outputs. To address these effects, a procedure is proposed to both visualize and minimize the cross-coupling by means of steady-state mapping and a re-alignment of the dq reference frame. With this method implemented, gain-response tests show improved decoupling across the operating region.
16:20-16:40	FrC15.2
Robust Adaptive Control of Switched-Reluctance Motors without Velocity Measurements, pp. 4586-4591
Chumacero, Erik	Univ. Paris Sud
Loria, Antonio	CNRS
Espinosa-Perez, Gerardo	Univ. Nacional Autonoma de Mexico
We present a speed-sensorless tracking controller of switched reluctance motors with unknown parameters. Our approach relies on the design of two control loops: an outer control-loop for the rotor dynamics which is driven by a PID-type controller where the stator currents are viewed as virtual control inputs, and an inner tracking control-loop for the stator currents. We assume that the parameters of the rotor (inertia and the load torque) are unknown and we establish uniform global exponential stability. In the case that also the stator parameters are unknown, we add an adaptation law and we establish convergence of the tracking errors.
16:40-17:00	FrC15.3
Model Predictive Control of a High Speed Switched Reluctance Generator System, pp. 4592-4597
Marinkov, Sava	Eindhoven Univ. of Tech.
de Jager, Bram	Tech. Univ. Eindhoven
Steinbuch, Maarten	Eindhoven Univ. of Tech.
This paper presents a novel voltage control strategy for the high-speed operation of a Switched Reluctance Generator. It uses a linear Model Predictive Control law based on the average system model. The controller computes the DC-link current needed to achieve the tracking of a desired voltage reference in the presence of an unknown electrical load current. Its output is converted into a pair of excitation angles (turn-on/turn-off) by means of optimized scaling and mapping, while minimizing the dominant high-speed power loss. The constraints on the average DC-link voltage and current signals are directly handled. A numerical example is provided to validate the effectiveness and the robustness of the proposed control scheme.
17:00-17:20	FrC15.4
Robust Observer with Higher-Order Sliding Mode for Sensorless Speed Estimation of a PMSM, pp. 4598-4603
Kommuri, Suneel Kumar	Kyungpook National Univ. South Korea
Veluvolu, Kalyana C	Kyungpook National Univ.
Defoort, Michael	Valenciennes Univ.
This paper addresses the problem of sensorless speed estimation for the permanent magnet synchronous motor (PMSM). A higher-order sliding mode (HSM) observer is developed to provide estimation of back EMFs that are treated as unknown inputs without the use of lowpass filter. A finite-time smooth estimation is obtained and the chattering phenomenon is eliminated. An accurate speed estimate of PMSM can be algebraically computed with the estimated back EMFs. Experimental results show the effectiveness of the proposed method.
17:20-17:40	FrC15.5
Global Identification of Mechanical and Electrical Parameters of Synchronous Motor Driven Joint with a Fast CLOE Method, pp. 4604-4609
Robet, Pierre-Philippe	Univ. of Nantes
Gautier, Maxime	Univ. of Nantes
In many cases of new actuation of compliant controlled or bio-inspired joint driven robot, a global identification of electrical and mechanical coupled dynamics is required. This paper proposes a technique which mixes a closed loop output error method with the inverse dynamic identification model method which allows using linear least-squares technique to estimate the parameters. A first approach which has been validated on a DC motor allows a decoupled identification of the electrical and mechanical dynamics but fails to make a simultaneous identification. A major improvement of that method is proposed to carry out the coupled identification of both mechanical and electrical parameters. A validation on a synchronous motor driven joint shows the effectiveness of the new procedure..
17:40-18:00	FrC15.6
Online Identification for Auto-Tuning PID Based on Wavelet Neural Networks: An Experimental Validation on an AC Motor, pp. 4610-4615
Ramos Velasco, Luis Enrique	Univ. Pol. de Pachuca
Julio Cesar, Ramos Fernández	Univ. Pol. de Pachuca
Domínguez Ramírez, Omar Arturo	Univ. Autónoma del Estado de Hidalgo
Parra Vega, Vicente	Centro de Investigación y de Estudios Avanzados del IPN(Unidad S
Marquez Vera, Marco Antonio	Univ. Poitécnica de Pachuca
Carrillo Santos, Carlos Alberto	Univ. Autónoma del Estado de Hidalgo
Sanchez, Anand	CINVESTAV
Based on adaptive wavelet neural networks, a PID discrete control scheme for induction motor drives is presented. An auto-tuning wavenet scheme is synthesized for the PID feedback gains by means of gradient-descendent algorithm through online plant identification. The latter uses input and output data for a radial basis neural network (RBNN) with different daughter wavelet activation functions. Infinite impulse response filter in cascade with the output of the RBNN is introduced to prune nodes that contribute poorly for an efficient identification even though gradient-descendent is used. A comparative experimental study under real noisy conditions and no load perturbation shows results for different wavelets (Morlet, RASP1, POLYWOG1 and Shannon); better performance of Morlet and RASP1 wavenet-based tuning are obtained due to its better identification properties.

TuW1	HG D1.1
Workshop 1: Verification and Control of Stochastic Hybrid Systems (Workshop)
09:00-17:00	TuW1.1
Verification and Control of Stochastic Hybrid Systems*
Kamgarpour, Maryam	ETH Zurich
Abate, Alessandro	TU Delft - Delft Univ. of Tech.

TuW2	HG D1.2
Workshop 2: Distributed Optimization in Large Networks and Its Applications (Workshop)
09:00-17:00	TuW2.1
Distributed Optimization in Large Networks and Its Applications*
Johansson, Mikael	Royal Inst. of Tech.
Schenato, Luca	Univ. of Padova

Book of Abstracts of European Control Conference 2013