Data-Driven Design for Parity Space-Based Replay Attack Detection

This article studies the data-driven design of parity space based replay attack detection method for cyber-physical systems. By utilizing the input and output data, the proposed data-driven detection can uncover stealthy replay attacks. First, the replay attack detector is formulated by following the parity space method. Subsequently, data-driven parity space identification, residual quantification, and parity matrix optimization are presented to support data-driven residual generation and evaluation for the detector. To enhance detection performance, an integration of attack defense model information and process data is presented, and the data-driven detector design is proposed. Finally, simulations are conducted to testify the performance of our schemes.