Design of compensators for actuator saturation

Abstract

Actuator saturation is a common non-linear control problem; if it is not being compensated properly, the system can become unstable. When the actuator saturates, the control that cannot be implemented can be interpreted as a non-linear disturbance being injected into the closed-loop system. After transformation, the applied set-point is altered by the disturbance, giving the effective set-point. If the effective instead of the applied set-point is used to calculate the control, no actuator saturation occurs. Since the effective set-point always replaces the applied set-point whenever the actuator saturates, a compensator can be designed aiming to produce a more acceptable effective set-point. The conditions for its implementation are given, followed by the properties of the effective set-point. A procedure for selecting the parameter of the compensator is also described. Examples are presented to illustrate the design procedure and to compare the performance of the proposed and the existing compensators.