Fault detection for uncertain nonlinear networked systems with Markov data transmission pattern

Abstract

This paper is concerned with robust fault detection problem of nonlinear uncertain networked control systems (NCSs). Two irrelevant Markov chains are introduced to describe Markov data transmission pattern in both from sensors to controller (S/C) part and from controller to actuators (C/A) part, then a Markovian jump system model is established. By employing mode-dependent fault detection filter as residual generator, fault detection of nonlinear NCSs is formulated as nonlinear H ∞ filtering problem. Furthermore, Lyapunov function approach is used to derive the sufficient condition which satisfies the stochastic stability and prescribed H ∞ attenuation level simultaneously. Especially, the desired mode-dependent fault detection filters are constructed in terms of certain linear matrix inequalities (LMIs), and explicit parameters are characterized if these LMIs are feasible. The effectiveness of the proposed method is demonstrated by simulation example.