Robust Fault Detection for Fuzzy-Model-Based Networked Control Systems Under Denial-of-Service Attacks and Sensor Faults Within a Stochastic Communication Protocol

Abstract

This paper investigates the robust fault detection problem in non-linear networked systems, characterized by the Takagi-Sugeno (T-S) fuzzy model. The analysis is complicated by Denial-of-Service (DoS) attacks and sensor faults. In such systems, DoS attacks disrupt network communications, threatening both system stability and fault detection accuracy. Simultaneously, sensor faults compromise the reliability of the system's measurement outputs. To address these challenges, the stochastic communication protocol (SCP) is employed to optimize the transmission order between multiple sensors and the fault detector, effectively managing the limitations of communication capacity. Consequently, a fault detection scheme based on SCP is developed to generate residual signals, supported by an auxiliary residual system for stability analysis. Using the Lyapunov-Krasovskii (L-K) functional approach, sufficient conditions are derived to ensure the asymptotic stability of the residual system with a prescribed $H_{\infty }$ performance. Finally, the effectiveness of the proposed fault detection scheme is validated through simulation examples.