Improved Prediction Dynamics for Robust MPC

The objectives of this paper are twofold. The first is to present a particular choice of the parameters of dynamic feedback laws for polytopic uncertain and/or time-varying systems with state and input constraints. We show that it has the same desired property as that of algorithms in [1], [2], i.e., the domain of attraction of the controlled system under a linear/saturated dynamic feedback law is identical to the domain of attraction under any static linear/saturated state feedback law. The second objective is to propose new procedures for robust constrained prediction dynamics based MPC that do not require the assumption of quadratic stability. With respect to other well known techniques, the main advantages of this new approach is the reduced conservativeness.