基于神经网络的机械手鲁棒输出跟踪控制

针对一类参数未知的机械手输出跟踪鲁棒控制问题，应用输入／输出反馈线性化方法和Lyapunov稳定理论，提出了一种基于神经网络建模的机械手输出跟踪自适应鲁棒控制器，采用三层前向神经网络来逼近未知非线性函数，网络的权值依据Lyapunov稳定性进行实时修正，保证了相应闭环系统的稳定性，所提出的控制器保证了跟踪误差及相应闭环系统的状态－致最终有界，且不需预知不确定性的上界，以两自由度移动关节刚性机械手的跟踪控制问题的为例进行了仿真，结果表明所提出方法是行之有效的。

Tracking control of robot manipulator robust output based on neural network

For the tracking control of robust output of a class of robot manipulators with unknown parameters, an adaptive robust control based on neural network for robot manipulator was developed by applying the input/output feedback linearization method and the Lyapunov stability theory. The unknown non linear functions were modeled with a three layer forward neural network. The network weights were updated by the Lyapunov stability in real time, and the stability of the closed loop systems was then guaranteed by this method. Moreover, the proposed control guaranteed that the output tracking error and the states of the established closed loop system were uniformly ultimately bounded. The up boundary of the uncertainties was not known in advance. The tracking control of the two DOF rigid robot manipulator with motive joint was used as example for simlation to show that the proposed method is effective.