基于神经网络滑模的机械臂轨迹跟踪控制方法

针对机械臂轨迹跟踪控制中存在建模误差以及外界干扰造成的控制性能下降问题,提出一种改进的自适应神经滑模控制方法。分别选取状态反馈和改进的神经网络滑模方法来控制系统的确定部分和不确定部分。利用神经网络的非线性映射能力自适应地学习系统不确定性的未知上界,其输出作为滑模控制器的动态补偿项,Lyapunov函数法推导得到神经网络权值更新律。为保证神经网络映射的有效性,提高收敛速度,采用遗传算法对神经网络结构参数进行优化。双关节机械臂系统的仿真结果表明了该方案的有效性。

Trajectory tracking control method of manipulators based on neural network sliding mode

Aiming at the modeling errors existing in trajectory tracking control of manipulator and the degradation of control performance caused by external disturbances,an improved adaptive neural sliding mode control method was proposed. The state feedback and the improved neural network sliding mode were respectively selected to control the deterministic part and the indefinite part of the system. The unknown upper bound of system uncertainty was learnt adaptively utilizing the nonlinear mapping ability of neural network,and the output was used as the dynamic compensation of the sliding mode controller. To ensure the global stability of the system,Lyapunov function was used to determine the neural network weight update rule. To ensure the validity of neural network mapping and improve the convergence rate,the genetic algorithm was used to optimize the structural parameters of neural network. The simulation of the dual joint manipulator system shows the effectiveness of the proposed method .