Improved LMI representations for bounded real criterion of systems with polytopic type uncertainty

An improved bounded real criterion is derived by employing a descriptor system approach. When used in robust performance analysis of systems with convex polytopic type uncertainty, it can reduce conservatism inherent in the classical quadratic method by providing a parameter-dependent Lyapunov function. Compared with the existing similar results, our approach shows some favorable features in computation. Two numerical examples are included to illustrate the effectiveness of the proposed representations.