基于滑模变结构的空间机器人神经网络跟踪控制

Neural-network tracking control of space robot based on sliding-mode variable structure

研究了在无需模型估计值的情况下不确定空间机器人轨迹跟踪问题,提出了滑模变结构的神经网络控制方案.首先基于Lyapunov理论设计了一种径向基函数(RBF)神经网络控制器来补偿系统中的未知非线性,该神经控制器能够保证闭环系统的稳定性,而通过利用饱和函数把神经网络和滑模控制结合起来的控制器来不仅可以进一步削弱滑模控制输入的抖振,且当神经网络控制器无效时仍能保证系统鲁棒性.仿真结果证明了该控制器能在初期及强干扰情况下均能达到较好的控制效果.

This paper investigates the tracking problem of space robot with uncertainties,without using the estimation values of a model,and puts forward a neural-network control scheme with sliding-mode variable structure.A radial-basis-function（RBF） neural-network controller based on Lyapunov theory is designed to compensate for the unknown nonlinearity in the system.The neural-network controller guarantees the stability of the closed-loop system.The controller that inte-grates the neutral network with the variable structure by saturation function not only effectively eliminates the chattering in sliding-mode input,but also maintains the robustness of the closed-loop system when the neutral-network controller fails.Simulation results show the desirable performances of the presented controller in the early phase of operation and in the strong disturbance situation.