Dynamic Output Controller MPC via LMI

Authors:	Bernussou Jacques, LAAS-CNRS, France Granado Ernesto, Universidad Simón Bolívar, Venezuela Colmenares William, Universidad Simón Bolívar, Venezuela García Germain, LAAS-CNRS, France
Topic:	2.1 Control Design
Session:	Advances in Predictive Control
Keywords:	Predictive control; Output-feedback; Linear matrix inequalities; Parametric uncertainty; Stability; Performance.

Abstract

This paper extends some previous works on state feedback Model Predictive Control (MPC) to dynamic output-feedback in MPC methodology. The controller is updated at each sampling time by minimizing an upper bound of an infinite horizon quadratic cost. The optimization problem and its associated constraints, on input and output, are expressed by LMI (Linear Matrix Inequality). Only measurable output and a priori fixed range values of the non measurable states, are used to determine the controller. Stability of the closed loop system is proven. An extension to the robust case is discussed.