Observer-Based Dynamic Event-Triggered Resilient Control for Heterogeneous Multi-Agent Systems Under DoS Attacks

摘要

This paper studies the issue of dynamic event-triggered (ET) resilient control for heterogeneous multi-agent systems (MASs) under DoS attacks. Most of the existing work only considers ideal linear models and undirected graph communications. However, in practice, both disturbance and noise exist in the system model, and directed graph communication is more common. To solve this issue, a fully distributed dynamic ET control strategy with a prediction-based dynamic compensation algorithm is designed to deal with the difficulty of communication jamming. By using the Lyapunov method, it is proved that bounded consensus can be achievable in heterogeneous MASs under DoS attacks, and Zeno behavior is excluded. Furthermore, the tolerable attack intensity is quantified by attack frequency and duration. Finally, a numerical simulation is conducted to validate the efficacy of the proposed method.Note to Practitioners—In real-world scenarios, numerous complex tasks necessitate collaborative endeavors among heterogeneous intelligent agents, such as robots swarms, multiple vehicles and smart grids. Due to physical or geographical limitations, direct communication between all agents and a leader is not always feasible, thereby restricting individual agents to indirectly accessing the leader’s information. Furthermore, the practical constraints on network resources, coupled with the susceptibility to malicious DoS attacks, impede the completion of tasks by multiple agents. This paper presents a resilient dynamic ET control strategy for the consensus task of heterogeneous agents, aiming to improve system resilience while conserving network resources.