I. Introduction
As result of physical limitations, the output of actuators is always limited in amplitude and rate, such as maximum or minimum torque in an engine or the maximum safe pitch rate in an aircraft. Such limits must be taken in account in the control design, otherwise the controller output will be different from the plant input, leading to wrong update of the controller states and to consequences ranging from performance deterioration over large overshoots and sometimes even to limit cycles or stability loss. Therefore, this phenomenon – usually called “controller windup” – has a paramount practical relevance and therefore many existing techniques address this problem of actuator constraints, e.g. the “Model Predictive Control” (MPC) [1] or even the traditional “Anti-Windup (AW) Compensator” [2].