This paper considers observer–predictor feedback for delay compensation in consensus control under undirected switching graphs. Each node involved in the network is described by a linear system with an actuator delay of uncertain length, and the connection graph is allowed to be unknown and switch arbitrarily. An observer is proposed to estimate the disagreement state using locally available information, and a delay-adaptive consensus protocol is proposed via observer–predictor feedback. With the employment of a “backstepping” state transformation involving the observer state and the estimation delay state, an explicit Lyapunov function is constructed to prove the consensus and stability properties of the resultant closed-loop system. Numerical simulations are presented to demonstrate the effectiveness of the theoretic results.

本文考虑了无向切换图下共识控制中延迟补偿的观察者 - 预测器反馈。网络中涉及的每个节点都由一个线性系统描述，其执行器延迟长度不确定，连接图允许未知并任意切换。建议观察者使用本地可用信息估计分歧状态，并通过观察者 - 预测者反馈提出延迟自适应共识协议。通过使用涉及观察者状态和估计延迟状态的“后步”状态变换，构造了一个显式的 Lyapunov 函数来证明合成闭环系统的一致性和稳定性。数值模拟被提出来证明理论结果的有效性。