Control design methods with anti-windup mechanism for disturbance rejection and reference following

There are many methods for designing feedback control systems with the objectives of disturbance rejection and reference following. Two well-known methods are to directly embed a disturbance model in the controller or to compensate for the effect of disturbance using its estimation from an observer. It is demonstrated in this paper that the first method has inherited a two-degree-of-freedom control system configuration, but has an integrator windup problem in the presence of actuator saturation, whilst the second approach has embedded an anti-windup mechanism, but suffers from poor performance in reference following. This paper examines these two mainstream control design methods for their applications in the presence of actuator saturation. It is shown with a simple example that the classical control system by directly embedding a disturbance model leads to integrator windup, and an anti-windup mechanism is hence proposed to overcome this drawback. The frequency response analysis presented in this paper leads to the conclusion that both methods offer similar closed-loop performance in terms of disturbance rejection and noise attenuation, however, the first approach that embeds a disturbance model in the controller provides a two-degree-of-freedom control system configuration, and hence gives a better performance in terms of reference following. Comparative simulation studies of both Single-Input and Single-Output (SISO) and Multi-input and Multi-Output (MIMO) systems are used to support the conclusions.