We build a two-layer multiagent system with separate agent and control layers. The control layer represents a distributed stabilization protocol to be designed, and the agent layer represents the dynamics of an interconnected multiagent system subject to agent- and multiagent system-level modeling uncertainties. We propose a dynamic event triggered strategy to schedule the controller updates and information broadcast in a nonperiodic and nonsynchronous manner develop a step-by-step procedure to systematically design the required stabilization gain, and characterize the condition under which the candidate stabilization gain and control layer topology represent a valid control layer. We theoretically prove that all trajectories of the proposed two-layer interconnected MAS exponentially converge to the origin in the presence of modeling uncertainties over the agent layer, and find a strictly positive lower bound on the inter-event time length to rule out the Zeno phenomenon. We further prove that the proposed dynamic event-triggered strategy will not result in more information broadcast events than those of the associated static event-triggered mechanism. We validate the proposed ideas in simulation.
展开▼
