Nonlinear Anti-windup Control Based on a Gain-Scheduling Algorithm

Abstract

We propose a design method of a nonlinear anti-windup controller for discrete-time systems subject to l2 disturbances. The proposed controller is composed of a gain-scheduled feedback control law and a standard observer. The gain-scheduled control law has a structure that a high-gain control law and a low-gain control law are interpolated by a single scheduling parameter. The scheduling parameter is computed on-line by solving a convex optimization problem based on the information of the state variables estimated by the observer. We show that, by using the proposed control law, both a large region of attraction and high distubance attenuation performance can be achieved.