Robust fault detection filter for nonlinear networked control system

A study of robust fault detection filter for the nonlinear networked control system (NCS) with data dropouts is investigated. From controller to actuator, a variable is used to describe random data dropouts in communication link, which follow the Bernoulli distribution. In this study, the residual generator is designed by an observer-based robust fault detection filter, and a novel stochastic nonlinear NCS is established as the model of residual dynamical system(RDS). An optimal performance index is presented to manage the issue of uncertainties, which is to enhance the robustness of the residual generator against the random data dropouts and nonlinearity, without significant loss of the sensitivity to faults. Then, the sufficient condition of asymptotical mean-square stability for the RDS is proposed, and the suitable robust fault detection filter is designed by LMIs. An adaptive threshold is presented in residual evaluation. At last, an example is used to demonstrate the validity of the proposed fault-detection approach.