Predictive Control of Constrained Linear Systems with Multiple Missing Measurements

Abstract

This paper investigates the problem of predictive control for constrained control systems, in which the measurement signal may be multiply missing. An augmented stochastic model is firstly introduced to describe and compensate these missing measurements. By means of an infinite horizon quadratic performance objective, a state-feedback predictive control law involving missing probability is designed by minimizing the upper bound on performance objective at each sampling instant. It is shown that the on-line optimization problem subject to input and state constraints can effectively be solved in terms of linear matrix inequalities. The designed predictive controller can achieve the desired control performance and also guarantee the closed-loop stability in mean square sense. Finally, an example is given to illustrate the proposed results.