Hybrid triggering control for average consensus of multi-agent systems

Abstract

This paper investigate the event-triggered consensus problem for multi-agent systems in the average consensus case. An event of each agent is triggered only when certain measurement error reaches some prescribed threshold. Then at each event time instant, each agent communicates with its neighbors, sends information to them and then updates its controller. In the event design in this work, we use discontinuous threshold and as a result, the continuous communication among agents is avoided. This may reduce energy consumption in practice. To exclude the Zeno behavior, we propose an event/time hybrid triggering scheme. Then asymptotic consensus can be guaranteed while the Zeno behavior can be avoided. Moreover, the proposed algorithm is then extended to a self-triggered case and thus continuous self-state monitoring can be further avoided. The effectiveness of the proposed approach is illustrated by simulation examples.