Interval threshold-based finite-frequency sensor attack detection for interconnected cyber-physical systems

This study addresses the problem of observer design and its applications to attack detection for an interconnected cyber-physical system subjected to unknown disturbances and sensor attacks. First, an overall system is constructed using the information on the interconnection of several subsystems, and problems related to the observer design are then solved. Second, to facilitate the residual's sensitivity to attacks and its robustness against unknown disturbances, an observer is provided. The gain matrix is developed using both the H₋ performance and mixed L₂−L_∞/H_∞ performance. Third, for the suppression of the effect of unknown disturbances and attack signals on the errors in the state estimation and on attack detection in a finite-frequency domain, the generalized Kalman-Yakubovich-Popov lemma is applied to set the conditions under which the observer can operate effectively. In light of the detection scheme, we suggest a detection logic that prevents false alarms by comparing the evaluation function to the corresponding threshold. Additionally, an interval threshold-based method is introduced to establish the threshold bounds, and numerical simulation results are provided to demonstrate the method's validity in this paper.