Robust Controller Design Using Parameter Space

Abstract Robust controllers are designed using the mean and covariance matrix of the estimate of the plant parameters. The aim of the design Is to maximise the probability that the parameters of the closed loop characteristic equation remain within some desired region in the parameter space. The direct approach is only applicable to simple systems and thus alternative methods have been developed. An approach which considers the distance of the closed loop parameter vector from the boundary of the desired region, called the Minimum ratio design, is shown to be almost as good as the direct design on a wide range of simple examples.