Consensus Control for Resilient Multi-Agent Systems Considering Leader Failure
Takuya Murakami, Toru Namerikawa · 2025
In this paper, we study a consensus control problem for resilient multi-agent systems considering leader failure. We present cooperative control to solve the formation problem, enabling the agents to maintain relative positions even if any agent, including the leader, malfunctions. To accomplish this goal, we designed an algorithm that implements distributed observers for fault detection to assess each agent's reliability. The algorithm reduces the effect of faults by limiting information exchange with low-reliability agents. Next, we extended the adjacency matrix so that any follower can replace the leader in response to the leader's failure. Finally, we verify the proposed method through simulations.