Co-design of dynamic event-triggered mechanism and observer-based control for networked systems under hybrid cyber attacks: A GA-LMI-based approach.

This paper presents a co-design framework for an observer-based controller and a dynamic event-triggered mechanism (DETM) in networked control systems subject to hybrid cyber attacks. The hybrid attacks, including false data injection and replay attacks, are modeled using two independent Bernoulli processes. To reduce the network load a DETM is introduced between the sensor and controller with limited communication bandwidth. The closed-loop system's stability and H∞ performance are ensured through the derivation of nonlinear matrix inequalities. To address the nonlinear coupling in the co-design conditions, a variable transformation and matrix decomposition approach are employed to decouple the controller gains. However, this method relies on predefined DETM parameters, which may limit design flexibility. To overcome this issue a GA-LMI-based co-design algorithm is proposed to jointly optimize the controller and DETM parameters. Simulation results demonstrate that the proposed method achieves better H∞ performance compared to some existing approaches. The effectiveness and low conservatism of the method are further validated through its application to a buck DC-DC converter system under hybrid cyber attacks.