Resilient Control against Multiplicative Attacks in Cyber-Physical Systems

In this article, a resilient control scheme against multiplicative attacks is studied within the unified control and detection framework. Firstly, multiplicative attacks are modeled as outputs of a dynamic system, the inputs of which are real-time input and output data of a CPS. Using factorization techniques and Bezout’s identity, the closed-loop dynamics of the CPS with and without parameter uncertainties under multiplicative attacks is analyzed. Subsequently, in virtue of the small gain theorem, necessary and sufficient conditions to guarantee the stability of the CPS are presented. On this basis, it is proved that reducing the H-infinity norm of stable kernel representation of the controller by tuning the Youla parameter can enhance the CPS’s stability margin. Thus, the resilient controller exhibits better performance against potential multiplicative attacks. Finally, the validity of the presented approach is demonstrated on a boost circuit model.