Interval estimation-based attack detection framework for CPSs: an improved dynamic event-triggered mechanism

This paper proposes an improved dynamic event-triggered mechanism (IDETM) and attack detection framework based on an interval estimation method for discrete-time cyber-physical systems (CPSs) subject to deception attacks and uncertainties. First, by integrating a relative threshold with a time-decaying component, the IDETM is designed to schedule data transmission, significantly reducing communication redundancy while ensuring bounded transmission errors. Second, a quadratic Lyapunov analysis yields a dissipation-based small-gain condition for the feedback interconnection between the augmented closed-loop dynamics and the IDETM-induced transmission error, establishing input-to-state practical stability (ISpS) of the attack-free nominal closed-loop system without presupposing closed-loop state boundedness. A zonotopic residual evaluation scheme constructs dynamic safe residual sets through estimation-error reachable-set propagation, and a sufficient condition is derived for detecting coordinated actuator and sensor deception attacks at their onset. Numerical simulations using a vehicle lateral dynamics model illustrate the communication efficiency and attack detection performance of the proposed framework.