Multi-Agent Consensus with Event-Triggered Communication and Relative Output Sensing

This paper investigates the consensus problem of linear multi-agent systems with event-triggered (ET) communication on undirected graphs. More practically, it is assumed that each agent can only access the relative output about its neighbors. Hence, the existing event-triggered (ET) protocols based on relative state sensing or absolute output sensing are not applicable. Inspired by this, an ET output-feedback protocol based on relative output sensing is developed. By a new Lyapunov function with more augmented states, it is proved that the resulting closed-loop system can reach consensus and the protocol states converge to zero. It is also shown that Zeno behavior is excluded from the triggering process. Furthermore, a method for reconstructing the continuous-time relative output from sampled data is formulated, based on which, the protocol can be implemented without continuous monitoring of neighbors. Finally, the effectiveness of the theoretic developments is illustrated by a numerical example.