IEEE Trans Neural Netw Learn Syst. 2018 Dec;29(12):6015-6025. doi: 10.1109/TNNLS.2018.2817880. Epub 2018 Apr 12.
Practical time-varying formation tracking problems for second-order nonlinear multiagent systems with multiple leaders are investigated using adaptive neural networks (NNs), where the time-varying formation tracking error caused by time-varying external disturbances can be arbitrarily small. Different from the previous work, there exists a predefined time-varying formation formed by the states of the followers and the formation tracks the convex combination of the states of the leaders with unknown control inputs. Besides, the dynamics of each agent has both matched/mismatched heterogeneous nonlinearities and disturbances simultaneously. First, a practical time-varying formation tracking protocol using adaptive NNs is proposed, which is constructed using only local neighboring information. The proposed control protocol can process not only the matched/mismatched heterogeneous nonlinearities and disturbances, but also the unknown control inputs of the leaders. Second, an algorithm with three steps is introduced to design the practical formation tracking protocol, where the parameters of the protocol are determined, and the practical time-varying formation tracking feasibility condition is given. Third, the stability of the closed-loop multiagent system is proven by using the Lyapunov theory. Finally, a simulation example is showed to illustrate the effectiveness of the obtained theoretical results.
使用自适应神经网络（NNs）研究了二阶非线性多智能体系统的实用时变编队跟踪问题，其中由时变外部干扰引起的时变编队跟踪误差可以任意小。与以前的工作不同，存在由跟随器的状态形成的预定义时变编队，并且编队跟踪具有未知控制输入的领导者的状态的凸组合。此外，每个代理的动力学同时具有匹配/不匹配的异构非线性和干扰。首先，提出了一种使用自适应 NNs 的实用时变编队跟踪协议，该协议仅使用局部邻域信息构建。所提出的控制协议不仅可以处理匹配/不匹配的异构非线性和干扰，还可以处理领导者的未知控制输入。其次，介绍了一个具有三个步骤的算法来设计实用的编队跟踪协议，其中确定了协议的参数，并给出了实用的时变编队跟踪可行性条件。第三，通过 Lyapunov 理论证明了闭环多智能体系统的稳定性。最后，通过仿真示例说明了所得到的理论结果的有效性。
