Anti-Disturbance Secure State Estimation for Continuous-Time Cyber-Physical Systems Under Sparse Sensor Attacks

This article investigates the anti-disturbance secure state estimation problem of continuous-time linear cyber-physical systems under sparse sensor attacks and external disturbances. First, a Luenberger-like observer is designed to eliminate the impact of attacks by automatically truncating the compromised communication channels, while embedding a disturbance observer to compensate for the effects of disturbances on estimation performance. Then, a necessary and sufficient condition for $\text {strong}^{*}$ detectability of the observation error system, as well as the conditions for the existence of the designed observer, are provided. The feasibility of these conditions is also analyzed. Unlike the existing results, based on the $\text {strong}^{*}$ detectability property of the observation error system, the designed observer is proven to asymptotically converge to the true system state despite disturbances and attacks. Finally, the effectiveness and superiority of the designed observer are verified through a practical numerical example.