| Abstract: | This paper studies the problem of cooperative attack strategy design for cyber‐physical systems modeled by discrete‐time linear systems. A function that reflects the attacker's target is established, and a linear attack strategy is designed based on the objective function, in which the attack signal is unknown but energy bounded, while the statistical information of the existing ones is required to be known. Unlike the existing analysis methods for attack strategy, new Lyapunov functions that depend on the system states at different instants are constructed in the attack strategy analysis, and the sensitivity to the state estimator and the robustness against the detector are both enhanced on attack signals through using the H?/H∞ scheme. Finally, two simulation examples are given to show that the designed attack strategy achieves larger estimation errors and smaller detection probability simultaneously than the existing design methods. |
