Yu Ziquan, Liu Zhixiang, Zhang Youmin, Qu Yaohong, Su Chun-Yi
IEEE Trans Neural Netw Learn Syst. 2020 Jun;31(6):2077-2091. doi: 10.1109/TNNLS.2019.2927887. Epub 2019 Aug 9.
This paper investigates the distributed finite-time fault-tolerant containment control problem for multiple unmanned aerial vehicles (multi-UAVs) in the presence of actuator faults and input saturation. The distributed finite-time sliding-mode observer (SMO) is first developed to estimate the reference for each follower UAV. Then, based on the estimated knowledge, the distributed finite-time fault-tolerant controller is recursively designed to guide all follower UAVs into the convex hull spanned by the trajectories of leader UAVs with the help of a new set of error variables. Moreover, the unknown nonlinearities inherent in the multi-UAVs system, computational burden, and input saturation are simultaneously handled by utilizing neural network (NN), minimum parameter learning of NN (MPLNN), first-order sliding-mode differentiator (FOSMD) techniques, and a group of auxiliary systems. Furthermore, the graph theory and Lyapunov stability analysis methods are adopted to guarantee that all follower UAVs can converge to the convex hull spanned by the leader UAVs even in the event of actuator faults. Finally, extensive comparative simulations have been conducted to demonstrate the effectiveness of the proposed control scheme.
本文研究了存在执行器故障和输入饱和情况下多架无人机（multi-UAVs）的分布式有限时间容错包容控制问题。首先开发了分布式有限时间滑模观测器（SMO）来估计每个跟随无人机的参考值。然后，基于估计的知识，借助一组新的误差变量，递归设计分布式有限时间容错控制器，以引导所有跟随无人机进入由领航无人机轨迹所跨越的凸包内。此外，利用神经网络（NN）、神经网络的最小参数学习（MPLNN）、一阶滑模微分器（FOSMD）技术和一组辅助系统，同时处理多无人机系统中固有的未知非线性、计算负担和输入饱和问题。此外，采用图论和李雅普诺夫稳定性分析方法，以确保即使在执行器出现故障的情况下，所有跟随无人机也能收敛到由领航无人机所跨越的凸包内。最后，进行了广泛的对比仿真，以证明所提出控制方案的有效性。
