基于神经网络的自由漂浮空间机械臂 自适应鲁棒控制

针对自由漂浮空间机械臂动力学模型难以精确获得，且无法表达为关于未知参数的线性形式问题，提出基于自适应神经网络的鲁棒控制方法.对于不确定性空间机械臂系统模型中存在的未知不确定部分，利用神经网络的万能逼近特性，设计神经网络控制器来补偿未知模型，避免传统控制中的保守上界估计；采用泰勒线性化技术将神经网络隐含层中的高斯函数线性化，设计包括网络权值、高斯中心及宽度在内的网络全参数自适应学习律，实现在线实时调整，提高控制精度；设计鲁棒自适应控制器来抑制外界扰动，并补偿逼近误差，提高系统鲁棒性；基于Lyapunov理论证明闭环系统的一致最终有界（UUB）.仿真试验表明所提控制方法能够获得较好控制效果，对空间机械臂控制具有一定工程应用价值.

Adaptive Robust Control Method of Free Floating Space Manipulators Based on Neural Network

Trajectory tracking control problems of free floating space manipulators with uncertainty are considered, and a robust adaptive control method is put forward based on neural network. The dynamics model of uncertain space manipulators is established, and the unknown nonlinear parts of system model are considered. A neural network controller is designed to approach the unknown uncertain parts, by the way the traditional upper bound of conservative is be avoided estimating. Taylor technology is used to linearize the Gaussian function of neural network hidden layer. Then the adaptive learning laws of the network weight parameters, the hidden layer center and width of Gaussian function are designed to guarantee the online real time adjustment. The robust adaptive controller is designed to compensate the amount of high order item and the approximation errors for better control accuracy. The uniformly ultimately bounded (UUB) of the closed loop system is proved based on the Lyapunov theory. Simulation results show that the proposed control method can get better control accuracy, and the controller has important engineering value for the space manipulator.