Attack Detection From the Perspectives of Control Center and Controlled Systems

This paper investigates the problem of attack detection for cyber-physical systems with input constraints in the presence of command and sensor attacks. For detecting a wider range of attacks, two attack detection strategies are designed, corresponding to the control center side and the controlled system side, respectively. First, from the perspective of the control center which sends commands and collects measurements, a novel attack detector is designed to detect irregularities by analyzing residual signals, tracking performance, and prediction performance. Second, for detecting the stealthy attacks bypassing the detector in the control center side, an attack detector in the controlled system side is also constructed by analyzing the reliability of the received command which is generated by a command governor. Compared to solely relying on residual signals, it is shown that combining the designed two detectors where the collected information is fully utilized ensures that more general attacks can be detected. Finally, a numerical simulation is provided to demonstrate the correctness and effectiveness of the proposed methods, showing that the proposed detectors equipped with multiple defense mechanisms can detect various attacks.