Abstract：
This paper investigates the consensus problem for discrete-time leader-following multi-agent systems subject to large time delays. Building upon two assumptions, a novel fully distributed protocol is devised by utilizing a normalized weighting matrix, depending solely on the relative output measurement. It is shown that, for arbitrarily large yet bounded input and communication delays that are constant and exactly known, the consensus problem can be effectively addressed by both the proposed protocol and its truncated version. Assuming further that followers incorporate solely input delays, then the permitted delays can be time-varying and different. The proposed protocols do not rely on global information of the directed communication topology, thus ensuring robustness against alterations in the communication topology. A numerical example is employed to validate the effectiveness of the suggested approach.
摘要：
本文研究了具有大时滞的离散时间领导者跟随多智能体系统的共识问题。基于两个假设，通过利用归一化加权矩阵设计了一种新颖的完全分布式协议，完全取决于相对输出测量。它表明,对于任意大但有界的输入和通信延迟，它们是恒定的，并且是完全已知的，所提出的协议及其截断版本都可以有效地解决共识问题。进一步假设追随者只包含输入延迟，那么允许的延迟可以是时变的和不同的。所提出的协议不依赖于有向通信拓扑的全局信息，从而确保针对通信拓扑中的改变的鲁棒性。数值算例验证了该方法的有效性。
