Fuzzy approximation based fault detection of nonlinear discrete delay stochastic systems via observers

Abstract

This paper deals with fault detection for nonlinear discrete delay stochastic systems. The nonlinear mappings in the systems are approximated by fuzzy systems. Real-valued parameters of fuzzy systems are determined by error back propagation(EBP) training algorithm. Then, observer is proposed for nonlinear discrete delay stochastic systems and Linear matrix inequality(LMI) approach is developed to establish sufficient conditions for error system to be exponentially ultimately bounded. Furthermore, threshold for fault detection is described in corollary. Finally, simulation results indicate the validity of the proposed method.