Stealthy Unbounded False Data Injection Attack Strategy Based on Zero Dynamics Theory

This is the authors' abstract. We don't add key points for this paper.

This paper investigates stealthy unbounded dual-channel false data injection (FDI) attack strategy in cyber-physical systems (CPSs) from the perspective of an attacker. In response to the issue that many attack strategies cannot exert an unbounded impact on stable systems, a stealthy unbounded FDI attack strategy based on zero dynamics theory is proposed. The proposed strategy aims to ensure the stealthiness of the attack while maximizing system performance degradation. Firstly, an inherent coupling mechanism between the forward channel attack signal and the feedback channel attack signal is established. Secondly, the recursive expressions of the residual variations and the state estimation error variations under the attack in the form of a zero dynamics system are constructed, enabling unbounded stealthy attacks. Unlike existing approaches, the proposed attack strategy is not constrained by system stability, and the attack impact can be flexibly adjusted by the attacker, thereby offering superior generality. Finally, the effectiveness and feasibility of the proposed attack strategy are demonstrated through numerical simulations.