Fault Detection and Isolation for a Kind of NCSs with Markov Delays

Abstract A new approach for fault detection and isolation (FDI) of Networked Control Systems (NCSs) is proposed in this paper. The paper first regards NCSs with unknown network-induced delay as a Jump Markov linear system (JMLS) with unknown input, then gives a Bayesian estimation of the JMLS based on Unknown Input Kalman filter (UIKF), and finally uses the estimation results to detect and isolate faults based on Log-likelihood ratio (LLR) approach. Simulation example shows that the method is robust to random networked induced-delay and unknown input.