ABSTRACT
Virtually all control actuation devices are subject to magnitude and/or rate limits and this typically leads to degradation of the nominal performance and even to instability. Historically, this phenomenon has been called "windup" and it has been addressed since the 1950's (see, e.g., [25]). To deal with the "windup" phenomenon, "anti-windup" constructions correspond to introducing control modifications when the system saturates, aiming to prevent instability and performance degradations. Early developments of anti-windup employed ad-hoc methods (see, e.g., [6,9,18] and surveys in [1,19,30]). In the late 1980's, the increasing complexity of control systems led to the necessity for more rigorous solutions to the anti-windup problem (see, e.g., [8]) and in the last decade new approaches have been proposed with the aim of allowing for general designs with stability and performance guarantees [12,17,29,31,36,39,42,47].
