Dynamic Periodic Event‐Triggered Control for Multiagent Systems Under Hybrid DoS Attacks

ABSTRACT This article investigates the consensus problem of multiagent systems (MASs) based on state estimation under hybrid denial‐of‐service (DoS) attacks. Comparing with most existing articles that predominantly focuses on attacks targeting only the information transfer between subsystems (topology graph network of multiagent systems), this article aims to solve the consensus problem of MASs where both the network topology and the subsystem sensors are attacked. First, a novel identification method is proposed to quickly identify whether the transmission channel is attacked or not, and a class state estimators for the agents and the leader with corresponding parameter selection criteria are established. Second, advanced dynamic periodic event‐triggered control approaches, where the feedback information is fixed by the state estimator, are designed to achieve consensus under known and unknown connectivity‐maintained attacks (CMA), respectively. The provided control schemes have less conservative triggered conditions and lower controller update frequency. Third, the comprehensive control strategies based on dynamic periodic event‐based condition, state estimator, and dynamic compensator are provided to settle consensus issues of the systems under connectivity‐broken attacks (CBA) and hybrid DoS attacks, respectively. It can be shown that the provided control approaches can address consensus problems of the systems with attacked sensor and different kinds of attacks and can have the fewer triggered times than traditional periodic event‐triggered control schemes. Finally, numerical examples are provided to verify the conclusions.