False data injection attacks for a class of output tracking control systems

With the development of cyber-physical systems (CPSs), the security becomes an important and challenging problem. Attackers can launch various attacks to destroy the control system performance. In this paper, a class of linear discrete-time time-invariant control systems is considered, which is open-loop critically stable and only has one critical eigenvalue. By including the output tracking error as an additional state, a Kalman filter-based augmented state feedback control strategy is designed to solve its output tracking problem. Then a stealthy false data attack is injected into the measurement output, which can completely destroy the output tracking control systems without being detected. Simulation results on a numerical example show that the proposed false data injection attack is effective.