Robust fault detection of nonlinear uncertain systems with event-triggered communication

The robust fault detection problem for a class of nonlinear uncertain systems with event-triggered measurement communication is addressed in this study. The proposed event-triggered robust fault detection scheme has two steps. First, with the event-triggered measurement, an adaptive approximator is proposed to learn the unknown modeling uncertainty online. Second, after the learning procedure, an event-triggered residual generator is designed by integrating the adaptive approximator with the residual signal for fault detection. The adaptive threshold for fault detection decision is derived by taking the event-triggered scheme into consideration. The performance of the event-triggered fault detection system is rigorously analyzed by characterizing the event-triggered sampling error, including the stability of the adaptive approximation and the fault detectability.