A Model-Free Detection Strategy for False Data Injection Attacks in Networked Control Systems

In this paper, a model-free detection strategy is proposed for stealthy false data injection (FDI) attacks targeted at setpoints of control systems. A redundant item is added to the transmitted data, and decoding is carried out at the receiver before a few detection rules are checked against. Furthermore, this model-free detection scheme is proved effective in detecting FDI attacks on setpoint signals while combined with watermarking method or the idea of random point. Finally, effectiveness of the proposed strategy is verified through numerical simulation.