Secure State Estimation: A Novel Retransmission and Roll-Back Mechanism

As the scale of networked control systems gradually expands, an increasing number of sensors are installed to monitor the operational status of the systems. This paper focuses on the problem of secure state estimation for multi-sensor networked control systems under cyber attacks, utilizing available retransmitted information. Firstly, we establish retransmission rules for data considered suspicious due to additive external attacks, based on which a novel state estimation method called distributed retransmission and roll-back estimator (DRR-Estimator) is proposed. More specifically, this designed estimator incorporates a term for historical discarded data, which is able to enhance the estimation performance. Using linear matrix inequalities, we prove that the estimation error converges to a finite bound exponentially in mean square, even in the presence of attacks. It is further demonstrated that the proposed estimator works effectively under time delays. Finally, the proposed method is tested effectively in a regional temperature monitoring system.