Distributed Command Governor strategies for constrained coordination of multi-agent networked systems

In this paper a novel distributed FeedBack Command Governor (FB-CG) supervision strategy is presented for multi-agent networked systems subject to pointwise-in-time coordination constraints on the overall network evolutions. Specifically, a sequential distributed strategy where only one agent at a time is allowed to manipulate its own reference signal is fully detailed and its stability, feasibility and viability (liveness) properties rigorously proved. It is shown that, unlike the centralized case, liveness (the absence of deadlock situations) can be ensured only if the constraints satisfy a Constraints Qualification (CQ) condition. Several results and open problems in the liveness analysis are discussed and a counterexample is also provided for showing that for multi-input agents case the issue requires further investigations.