Wang Yujuan, Song Yongduan, Hill David J
IEEE Trans Cybern. 2021 Mar;51(3):1300-1310. doi: 10.1109/TCYB.2019.2893461. Epub 2021 Feb 17.
In this paper, we investigate the consensus tracking control problem for networked multiagent systems (MASs) with unknown nonaffine dynamics. Our goal is to achieve asymptotic (rather than ultimately uniformly bounded) consensus tracking, which is quite challenging especially if nonvanishing/nonparametric uncertainties are involved and at the same time the control protocol is required to be fully distributed and continuous everywhere. Here, we present a conceptually new and structurally simple solution with distributed and continuous control action. The developed method is capable of ensuring zero-error tracking with a unique converging feature in that the consensus tracking error first converges to a small adjustable residual set around zero within a prescribed finite time, and then further shrinks to zero exponentially. The key technique lies in the utilization of a state transformation based on certain scaling function. Our method also prevents the restrictive requirement that all subsystems have access to the linearly parameterized information as imposed in most existing consensus tracking results for nonlinear MAS.
在本文中，我们研究了具有未知非仿射动力学的网络化多智能体系统（MAS）的一致性跟踪控制问题。我们的目标是实现渐近（而非最终一致有界）一致性跟踪，这极具挑战性，特别是当涉及非零/非参数不确定性且同时要求控制协议完全分布式且在各处连续时。在此，我们提出一种概念上新颖且结构简单的具有分布式和连续控制作用的解决方案。所提出的方法能够确保零误差跟踪，并具有独特的收敛特性，即一致性跟踪误差首先在规定的有限时间内收敛到围绕零的一个小的可调节残差集，然后进一步指数收敛到零。关键技术在于利用基于特定缩放函数的状态变换。我们的方法还避免了大多数现有非线性MAS一致性跟踪结果中所施加的所有子系统都必须能够获取线性参数化信息这一限制性要求。
