Optimal Linear FDI Attacks with Side Information: A Comparative Study

摘要

This paper studies the problem of false data injection attacks against remote state estimation. We consider the scenario where malicious attackers can intercept the original innovation and also obtain some side information of system states with an extra sensor. A linear attack strategy that can successfully bypass anomaly detectors is proposed based on the combined information. The recursion of a posteriori error covariance under attacks is derived; then the optimal attack strategy is obtained by solving a convex optimization problem. An explicit solution of attack coefficients for scalar systems is derived. Simulation results show that, compared with existing work, the proposed attack can successfully bypass the anomaly detector and cause more severe performance degradation in remote state estimation.