Robust predictive control of perturbed switched linear systems with stable and unstable subsystems in the presence of uncertainties

Abstract

In this article, robust model predictive control is proposed for discrete-time switched linear systems in the presence of external disturbances, model uncertainties and constraints on the states and input. To solve the problem, a mode-dependent average dwell time switching strategy is constructed in controller design procedure, with less limitations than the existing average dwell time switching. Using some mathematical tools, the model predictive control problem is turned into a feasible linear matrix inequality. The control algorithm is presented via two theorems, and the stability analysis is analytically presented, based on multiple Lyapunov function method. Compared with the previous studies, the underlying switched system may consist of stable and unstable subsystems, perturbed by model uncertainties and external disturbances. The simulation results are also given to study the effectiveness of the proposed robust model predictive control algorithm, from a comparison viewpoint.