Fault-tolerant security control for switched systems via learning observer and DoS attacks

Abstract

This paper addresses the critical issue of security control for switched systems in the presence of denial-of-service (DoS) attacks and actuator faults. Traditional methods for handling such systems often face challenges in accurately determining transition probabilities and maintaining system stability under non-periodic DoS attacks. To tackle these issues, this paper proposes a novel switching rule that integrates dwell time and sojourn probabilities, significantly simplifying the determination of transition probabilities and reducing computational complexity. Additionally, an adaptive event-triggering mechanism is adopted to mitigate communication overhead while maintaining system control efficiency. A learning-based observer is implemented to accurately estimate non-differentiable fault functions, thereby enhancing the system’s resilience against external disturbances and component failures. Comprehensive simulations validate the effectiveness of the proposed methods, demonstrating improved system stability and reduced communication burdens under adverse conditions.