Quantised scaled consensus of linear multiagent systems on faulty networks

In this paper, the practical mean-square scaled-consensus problem of multiagent systems is studied. Higher-order general linear dynamics is investigated, where all agents are subject to faults or denial of service attacks (DoS). The considered network comprises agents taking quantised states under undirected communication channels. Different from the existing works, no strict conditions have been imposed on the eigenvalues of the agents' state matrix. First, we propose a distributed scaled-consensus protocol that utilises a probabilistic quantisation algorithm. Second, scaled-disagreement analysis is provided via mean-square stability methodology. Finally, sufficient conditions in terms of linear matrix inequalities are presented in order to obtain scaled-consensus gain for both leaderless and leader-follower cases. The presented theoretical results are examined through two simulation examples.

Automated summary

This paper investigates the practical mean-square scaled-consensus problem of multiagent systems, considering higher-order general linear dynamics and potential faults or denial of service attacks. A distributed scaled-consensus protocol utilizing a probabilistic quantisation algorithm is proposed, and scaled-disagreement analysis is provided through mean-square stability methodology. Sufficient conditions in terms of linear matrix inequalities are presented to obtain scaled-consensus gain for both leaderless and leader-follower cases, which are validated through simulation examples.