Output-Feedback Multi-Agent Consensus with Intermittent Communication on Directed Graphs

This note studies output-feedback consensus of linear multi-agent systems (MASs) on directed graphs. In view of limited communication resources in MAS tasks, the aim of this note is to reduce inter-agent communication. A fully distributed dynamic output-feedback protocol is proposed, where an event-triggered strategy is designed to determine when to exchange protocol states while a time-triggered one is designed to determine when to sample relative outputs. Design conditions are firstly established for strongly connected graphs and then refined for directed graphs with a spanning tree. In contrast to the existing works, the merits of the proposed protocol are threefold: it is applicable for directed graphs; it relies on relative outputs, rather than absolute outputs or absolute/relative states; it only requires intermittent communication but no continuous monitoring of neighboring agents.