Using complex zonotopes for stability verification

This paper is concerned with computer control systems. In particular, we consider stability of quasi-periodic linear impulsive systems which can be used to model sampled-data systems and networked control systems. In this paper, we introduce a class of sets complex zonotopes and use them to verify global exponential stability by deriving a sufficient condition for their contraction. Their advantage is that the eigenstructure of the system can be exploited to choose good candidate complex zonotopes for verifying stability. Using this approach we could show, compared to the previous approaches, either larger or competitive bounds for the maximum uncertainty in stabilizing sampling times on a set of benchmark systems.