Event-triggered consensus control for stochastic multi-agent systems under state-dependent topology

In this study, the mean square consensus problem for stochastic multi-agent systems with nonlinear protocols is concerned. In particular, the topological structure of the system is characterised by state-dependent directed graph, in which the edges are described by nonlinear function related to the states and possess much better performance than before. A novel proposition based on matrix theory is proposed to conquer the challenges of time-invariant asymmetric matrix. In view of the communication burden and practically, a time-dependent event-triggered strategies are explored, where the control input on each agent is updated only when a certain triggering condition is violated. Then, some sufficient conditions for mean square consensus of stochastic multi-agent with state-dependent topology are established. Furthermore, a positive lower bound for inter-event times can be found such that the Zeno phenomenon is eliminated. Finally, a simulation example is utilised to illustrate the validly of developed theoretical results and the effectiveness of the control protocol.

Automated summary

This study focuses on the mean square consensus problem for stochastic multi-agent systems with nonlinear protocols, proposing a novel proposition based on matrix theory to overcome challenges, and exploring time-dependent event-triggered strategies for communication efficiency. Sufficient conditions for mean square consensus with state-dependent topology are established, eliminating Zeno phenomenon with a positive lower bound for inter-event times. A simulation example demonstrates the validity of theoretical results and effectiveness of the control protocol.