Distributed Asymptotic Secure Estimation for Cyber-Physical Systems Under Stealthy Attacks: An Interval Observer-Based Method

This article focuses on the problem of distributed asymptotic secure estimation for cyber-physical systems (CPSs) under stealthy attacks and unknown but bounded process disturbance. First, interval observers based on two-point observers are constructed to produce upper and lower bounds for the attacks. Then, augmented systems with a resort to auxiliary variables are introduced, and algebraic relationships among augmented states, disturbance and attacks are obtained by interval estimations of the augmented system outputs. Further, from the algebraic relationships, distributed secure estimation, including distributed state estimation and stealthy attack reconstruction, is designed and achieves omniscience asymptotically. Compared with the existing results, a novel interval observer-based distributed estimation approach is proposed, which can realize simultaneous state estimation and attack reconstruction. Finally, a simulation example is presented to illustrate the effectiveness of the proposed distributed estimation strategy.