Distributed Control of Leader-follower Stochastic Multi-agent Systems: A Hierarchical and Dynamic Event-triggered Approach

This study introduces a hierarchical control framework for leader-follower stochastic multi-agent systems (MASs) with mismatched uncertainties in a dynamic event-triggered context. Existing event-triggered control approaches encounter difficulties in ensuring zero tracking/stabilization error while avoiding Zeno behavior. The underlying problem becomes even more intricate if such systems are structured as leader-follower networks with directed communication in the presence of stochastic disturbances and mismatched uncertainties. To address these challenges, we design a hierarchical control framework comprising two layers: Layer I for the reference signal generator and Layer II for the dynamic event-triggered controller. In Layer I, we develop a reference signal genera tor that enables each agent to asymptotically reconstruct the leader state. In Layer II, we design a dynamic event triggered controller for agents to track the reference signals generated by Layer I without relying on cooperation errors between neighboring agents, thereby separating inter agent coordination from local tracking control. This strategy formulates a clock-based dynamic triggering mechanismby introducing an auxiliaryvariable that evolves based on plant state values, while utilizing a clock variable to guarantee the existence of a positive lower bound on inter execution times. We demonstrate that all the closed-loop signals are globally uniformly bounded in probability, with the output tracking error of each agent converging to the equilibrium at the origin in probability. Numerical simulation verifies the benefits and efficiency of the algorithm.