Memory-Based Adaptive Event-Triggered Filter Subject to Hybrid Cyber Attacks and Input Limitation

Abstract

This article focuses on the design of a secure memory-based adaptive event-triggered filter for networked systems subject to hybrid cyber attacks and input limitations. First, a hybrid attack model incorporating denial-of-service (DoS) attacks, deception attacks, and replay attacks is established for filter design. Second, a novel memory-based adaptive event-triggered strategy sensitive to cyber attacks is introduced into the filter design to save network resources, optimize network channel utilization, and prevent network congestion. Subsequently, a novel event-triggered filtering error model is established under hybrid cyber attacks and input limitations. Utilizing Lyapunov–Krasovskii functionals and linear matrix inequality (LMI) techniques, sufficient conditions can be concluded to prove the exponential mean-square stability of the filtering error model with a given $H\infty$ performance index. Finally, the effectiveness and the practicality of the obtained conclusions are demonstrated by a numerical simulation and tunnel diode circuit.