Robust fault detection for the networked control system with model uncertainty

Robust fault detection for a networked control system (NCS) including model uncertainty and stochastic data packet dropout is investigated, the NCS model with stochastic variable is established. Based on the observer method, a robust fault detection filter can generate a residual signal, which has good robustness to the model uncertainty, external disturbance, data packet dropout, and sensitivity to the fault. Then, the sufficient condition of asymptotically mean-square stable for the residual dynamical system is achieved. The desired robust fault detection filter is obtained, which is constructed in terms of certain linear matrix inequality (LMI). In the residual evaluation part, a threshold involved false alarm rate is presented, and the problem of estimating false alarm rate is analyzed. Finally, two numerical examples illustrate the effectiveness of the proposed approach.