Event-Based Fault Detection for Networked Switched Systems Subject to Hybrid Attacks

This paper focuses on the issue of event-based fault detection (FD) for networked switched systems (NSSs) subject to hybrid attacks. Compared with the existing fault detection literature, both stochastic deception attacks and aperiodic denial-of-service (DoS) attacks are considered to disturb the information transmission in the network channel simultaneously. In this context, a new attack-resilient fault detection filter (ARFDF) is devised as a residual generator while resisting the effects of hybrid attacks. Since the DoS attacks affect the operation of the event-triggering mechanism (ETM), a resilient ETM is adopted to save the limited network resources and adapt to DoS attacks. The dynamics of the initial system combined with fault detection filters, the effects of the ETM and hybrid attacks are formulated as a new switched residual model. By employing Lyapunov stability theory and average dwell time (ADT) method, the new residual system can be mean-square exponential stable when the parameters of attacks meet certain conditions. And the quantitative relationship between DoS attacks and the switching law is formulated. Moreover, a co-design method of ETM parameters and filter gains is proposed. Finally, simulation results are provided to verify the feasibility and effectiveness of the proposed method.