Distributed consensus tracking under variable communication network is an important research topic of multi-agent systems (MASs). Many results have been obtained for consensus of MASs with switching networks that are either jointly connected or periodically connected, but it remains challenging when the disconnected switching topologies are activated more frequently than the connected ones. This paper investigates both the consensus tracking and average consensus tracking problems for double-integrator MASs under switching topologies. By using the tools of multiple Lyapunov function and linear matrix inequality method, a neighbor-based distributed controller is designed for each following-agent so as to achieve (average) consensus tracking, even if the underlying switching communication network contains disconnected digraphs. Sufficient conditions, in terms of dwell time, total running time ratio of connected to disconnected topologies, and the exponentially convergent margin, are derived. Compared with existing research, the proposed method presents a quantified switching stability result allowing more disconnected digraphs to participate in switching, which can significantly reduce the communication pressure in applications. Simulation results are finally provided to verify the theoretical analysis.

可变通信网络下的分布式共识跟踪是多智能体系统（MAS）的一个重要研究课题。对于 MAS 与联合连接或周期性连接的交换网络的共识已经获得了许多结果，但是当断开连接的交换拓扑比连接的交换拓扑更频繁地激活时仍然具有挑战性。本文研究了切换拓扑下双积分器 MAS 的一致性跟踪和平均一致性跟踪问题。通过使用多重李雅普诺夫函数和线性矩阵不等式方法的工具，为每个跟随代理设计了一个基于邻居的分布式控制器，以实现（平均）共识跟踪，即使底层交换通信网络包含断开的有向图。足够的条件，就驻留时间而言，可以推导出连接拓扑与断开拓扑的总运行时间比以及指数收敛裕度。与现有研究相比，所提出的方法提出了一个量化的切换稳定性结果，允许更多断开的有向图参与切换，这可以显着降低应用中的通信压力。最后给出仿真结果来验证理论分析。