Dynamic output feedback robust model predictive control with guaranteed quadratic boundedness

This paper proposes a novel synthesis approach to dynamic output feedback robust model predictive control for systems with both polytopic description and bounded disturbance. The notion of quadratic boundedness is utilized to characterize the stability properties of the augmented closed-loop system. An error signal is defined, which is a linear combination of the true state, controller state and output. The constraints related with the unmeasurable state are handled by applying the bounds on this error signal. The parameters for this error signal are on-line refreshed in an always feasible optimization problem which is separated from the main optimization problem. The augmented state of the closed-loop system converges to a neighborhood of the equilibrium provided that the main optimization is feasible at the initial time. A numerical example is given to illustrate the effectiveness of the controller.