Distributed Matrix Pencil Formulations for Prescribed-Time Leader-Following Consensus of MASs With Unknown Sensor Sensitivity
研究了异构多智能体系统在未知传感器灵敏度下的规定时间领导跟随一致性问题，提出一种基于最差情况传感器的分布式矩阵束公式，设计出鲁棒且保守性低的控制参数，并通过时变反馈方案将协议扩展至无限时间域。
This article investigates the prescribed-time leader-following consensus problem for heterogeneous multiagent systems (MASs) with unknown sensor sensitivity. Considering a connected undirected topology, we introduce a time-varying dual observer/controller design framework that leverages both regular local and inaccurate feedback to achieve consensus tracking within a prescribed time. The proposed analytical framework applies to MASs equipped with sensors exhibiting uncertain sensitivities. A key innovation of our design is the framework of a distributed matrix pencil formulation based on the worst case sensor, leading to control parameters that exhibit sufficient robustness and relatively low conservativeness. Additionally, we establish a bounded time-varying feedback (TVF) scheme that extends the prescribed-time distributed protocol to an infinite time domain without compromising final control accuracy. This includes a detailed discussion of the analytical relationship between switching time and the upper bound of the time-varying gain. In particular, we employ the proportional coefficient obtained from several matrix pencil formulations along with a monotonically increasing time-varying (blow-up) function to derive the feedback gain, simplifying the complexity of control design. Simulations validate the effectiveness of the methodology through a series of electromechanical systems and single-link robot manipulators.