Optimal model based anti-windup design

In this paper we discuss an optimal tuning strategy for a model based anti-windup (AW) compensator which is straightforward to use and can be tuned empirically with respect to a general cost-function. The presented cost-function is representative for the performance during saturation of the combination of a linear-controller and an AW-compensator. The simulation results showed that adding the tuned anti-windup compensator to the existing control structure can improve the performance of the closed loop system with respect to the minimized control-objectives of the cost-function. Simulation tests of a position control on a damped mass-spring system showed that the optimized anti-windup controller performance is closer to a MPC than a full-order anti-windup compensator derived by a linear matrix inequality.