Secure state estimation with disturbance rejection against switched sparse sensor attacks

Abstract

This paper studies the issue of secure state estimation for cyber-physical systems vulnerable to sparse sensor attacks, where attackers might change their attack targets at any time to deteriorate the estimation performance. To address this issue, a Luenberger-like observer is designed to estimate the system state under the attack interference, which employs an improved projection operator and the least square method to handle the switched attacks and introduces a disturbance estimator to enhance the anti-disturbance capability. It is proven that the observation error system can achieve asymptotic stability and dissipative performance, even under the influence of switched sensor attacks and disturbances. Different from the existing results, this method can not only defend against more erratic attacks but also effectively reject disturbances. The conclusion presented is validated through simulation experiments.