Multi-agent coordination over local indexes via clique-based distributed assignment

This paper addresses a distributed coordination assignment problem of multi-agent systems via state measurements over local indexes. In this problem, no assignment of reference states is predetermined, but arbitrary assignments are allowed to agents. Due to this flexibility, assignment can be determined in a distributed way over local indexes, made by each agent under its own rule. First, we design a distributed controller using only state measurements over local indexes for the coordination assignment with the best performance. For this purpose, a new distributed decision-making method for assignment, called clique-based distributed assignment, is developed, which divides the assignment problem of all agents into small-size problems in terms of cliques (i.e. complete subgraphs). Next, we investigate the local/global achievement of the coordination assignment over a δ-proximity graph, where agents within distance of δ are connected. Finally, the effectiveness of the proposed method is demonstrated through numerical examples.