Moving horizon control of linear systems with input saturation and plant uncertainty Part 1. Robustness

Abstract

We present a moving horizon feedback system, based on constrained optimal control algorithms, for linear plants with input saturation. The system is a non-conventional sampled-data system: its sampling periods vary from sampling instant to sampling instant, and the control during the sampling time is not constant but determined by the solution of an open-loop optimal control problem. This is a two-part paper. In this part, we show that the proposed moving horizon control system is robustly stable, whether the state of the plant is measurable or not. In the second part, we show that the proposed moving horizon control system is capable of following a class of reference inputs and suppressing a class of disturbances. Experimental results show that the behaviour of the moving horizon control system is superior to that resulting from some alternative control laws.