Anomaly Identification for Cyber-Physical Systems Subject to Replay Attacks and Sensor Faults

This brief investigates the identification of replay attacks and sensor faults in cyber-physical systems subject to periodic disturbances. For each type of anomaly (replay attack or sensor fault), an anomaly identification framework consisting of an adaptive anomaly isolation estimator (AIE) and an anomaly identification processor (AIP) is proposed. By dynamically configuring the blocking zeros within AIE and through co-design with the compensation signal inside AIP, decoupling between the watermarks generated by AIPs for anomaly identification and periodic disturbances can be achieved. Numerical simulation cases validate the effectiveness of the proposed framework.