Adaptive Predefined-Time Neural Tracking Control for High-Order Nonlinear Systems: A Switching Event-Triggered Approach
本文针对高阶非线性系统，提出一种切换事件触发策略，实现预定时间内的神经跟踪控制，能提前设定收敛时间上限并节省通信资源。
This article concerns the problem of an event-triggered predefined-time neural tracking control for high-order nonlinear systems (HONSs) through a switching event-triggered strategy (ETS). In contrast to the previous research on HONSs, one prominent virtue of this work is that the upper bound of settling time can be determined a priori by a separate controller parameter. First, by employing the addition of a power integrator technique and a switching event-triggered rule, an adaptive event-triggered controller is developed under the backstepping framework. Therein, the unknown nonlinearities can be approximated effectively through neural networks (NNs). Unlike many existing approaches, the proposed control strategy is not only proficient in coping with the coupling term caused by the event-triggered rule and the nonlinear function, but also has excellent effects in saving communication resources. In terms of the predefined time stability theory, the boundedness of all variables and the fast convergence characteristics of the tracking error within the predetermined time in the system are guaranteed. Finally, the simulations are provided to validate the effectiveness of the suggested control scheme.