Event-triggered cooperative stabilization of multiagent systems with partially unknown interconnected dynamics

Abstract

We consider a stabilization problem in interconnected multiagent systems subject to both matched and unmatched modeling uncertainties. We propose a mixed graph theoretic and optimal control formulation, build a two-layer structure with separate agent and control layers, and capture the architectural aspect of cyber–physical systems. The control layer, to be designed, represents a linear cooperative control strategy and is built by control nodes, communication protocol, and information broadcast event-triggered strategies. We propose a systematic design framework to obtain two robust static gains such that the same control layer implementation can be used for both matched and unmatched modeling uncertainties. In the proposed event-triggered strategy, each agent uses only two decoupled impulsive approximators to determine its information broadcast events independently of other agents. We prove exponential convergence to the origin of all trajectories of the proposed two-layer interconnected multiagent system under the linear cooperative protocols; free of Zeno behavior in spite of the partially known nonlinearities over the agent layer and non-synchronous, non-periodic communication over the control layer. We also study practical stability with Zeno-freeness in order to reduce the noise sensitivity of the event-triggered strategy, or the number of the information broadcast events. We examine the feasibility of our ideas in simulation.