Resilient Event-Triggered Flocking Control of Multiagent Systems With Malicious Nodes

Résumé fourni par la source

This paper studies resilient flocking control for second-order multi-agent systems in the presence of non-cooperative agents and limited communication. Each cooperative agent only measures its own state, exchanges information with neighbors at asynchronously triggered instants, and runs a local estimator of the algebraic connectivity and its gradient. A switched distributed control architecture is proposed: when the estimated connectivity is low, a connectivity-enhancement mode is activated to increase the Fiedler eigenvalue of the cooperative subgraph. Once a connectivity margin is recovered, the system switches to a flocking mode that combines a repulsive-attractive potential with a WMSR-based velocity alignment term. An event-triggering mechanism with time-varying thresholds is designed to reduce communication while avoiding Zeno behavior. Under mild robustness and time-scale separation assumptions, it is shown that cooperative velocities remain uniformly ultimately bounded, the system switches to the flocking mode in finite time, and all cooperative agents asymptotically achieve resilient flocking: their velocities converge to a common value and all inter-agent distances stay within a prescribed safety interval. Numerical simulations validate the effectiveness of the proposed approach and demonstrate significant communication savings compared with continuous-feedback schemes.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.