Event-triggered fault detection and isolation for discrete-time linear systems

Abstract

In this study, the problem of event-triggered fault detection and isolation (FDI) for discrete-time linear time invariant systems is considered. Using a Leunberger observer as the residual generator, a multi-objective formulation of the FDI problem is presented based on l1, H- and H? performance indices. For each performance index, sufficient conditions for the design of FDI observer are presented based on linear matrix inequalities (LMIs). To reduce the conservativeness of the multi-objective problem, extended LMIs are used to eliminate the couplings of Lyapunov matrices with the system state space matrices. It is shown that through an event-triggered data transmission mechanism, the amount of data that is sent to the FDI module is significantly decreased. Simulation results corresponding to a remotely operated vehicle demonstrate the effectiveness and capabilities of the proposed design methodology. The Institution of Engineering and Technology 2016.