论文部分内容阅读
针对机械臂末端运动受约束位置/力混合控制问题,提出一种基于观测器的神经网络自适应控制算法.假定机械臂非线性动力学模型未知,并且仅有机械臂关节角位置和接触力可测量,通过非线性坐标变换对位置和力控制进行解耦,得到降阶的位置和力动力学模型.利用高通滤波器对关节角速度进行重构,用神经网络对降阶未知动力学模型进行逼近,同时添加滑模控制项用于补偿神经网络的逼近误差,以改善系统的跟踪性能.最后,基于李亚普诺夫稳定性理论给出系统的稳定条件,并通过数字仿真验证了该算法的有效性.
Aiming at the mixed position and force control problem of manipulator’s end movement, a neural network adaptive control algorithm based on observer is proposed. Assuming that the nonlinear dynamics model of manipulator is unknown, and only the position and contact force of manipulator’s joint angle And the position and force control are decoupled by nonlinear coordinate transformation to get the reduced-order position and force dynamics model.High-pass filter is used to reconstruct the joint angular velocity, and the neural network is used to approximate the reduced-order unknown dynamics model , And add the sliding mode control term to compensate the approximation error of the neural network to improve the tracking performance of the system.Finally, the stability conditions of the system are given based on the Lyapunov stability theory, and the effectiveness of the algorithm is verified by numerical simulation .