Practical approaches to low-order anti-windup compensator design: a flight control comparison

This paper considers three different methods for low-order anti-windup (AW) compensator design and compares their application to a realistic flight control problem, where the dominant actuator nonlinearity is aileron rate saturation. The compensator design methods all rigorously enforce at least local exponential stability via absolute stability results, but differ in both the construction of the AW compensator itself and the performance requirements used in the design. In particular, the paper compares a low order AW ``optimal'' design method in which the compensator poles and zeros are chosen by the designer and the accompanying gains synthesised optimally; a new low order method in which the optimisation procedure optimally chooses both the zeros and the gains; and a recently introduced classical design method where loopshaping is used to completely determine the gains and dynamics of the compensator. The methods are compared using a high-order nonlinear model of the lateral dynamics of an experimental aircraft, on which similar compensators have recently been flight tested.