15th Triennial World Congress of the International Federation of Automatic Control
  Barcelona, 21–26 July 2002 
RECEDING HORIZON LINEAR QUADRATIC CONTROL WITH FINITE INPUT CONSTRAINT SETS
Daniel E. Quevedo* José A. De Doná*
Graham C. Goodwin*
* Department of Electrical & Computer Engineering, The
University of Newcastle, Callaghan, NSW 2308, Australia

By exploring the geometry of the underlying constrained optimization, a finitely parameterized solution to the discrete time receding horizon linear quadratic control problem with a finite input constraint set is obtained. The resulting controller gives rise to a closed loop system which is piece-wise affine in the plant state. The switching regions are polytopes and are related to those obtained when dealing with ℓ* (saturation-like) constraint sets.
Keywords: predictive control, quadratic control, constraints, binary control, finite fields, multilevel controllers, quantized signals, switching values
Session slot T-We-M17: Predictive Control/Area code 2d : Optimal Control