论文部分内容阅读
介绍一种能提高弧焊机器人焊缝跟踪精度的神经网络控制器．通过神经网络的补偿作用，弥补了由于无法知道机器人精确模型所造成的控制上的误差．不同于机器人控制中传统的神经网络控制器，本文提出并应用了基于笛卡尔空间轨迹控制的机器人焊缝跟踪神经网络，大大简化了控制算法．计算机模拟及实验结果表明，该控制器非常适用于机器人的实际焊接，对于现有的机器人，无须改变其控制器内部结构，即可应用该技术．与常用的机器人关节力矩控制法相比，有效地提高了跟踪精度并具有较强的鲁棒性．
A neural network controller that can improve the tracking accuracy of arc welding robots is introduced. Through the compensation function of neural network, it can make up for the control error caused by unable to know the accurate model of robot. Different from the traditional neural network controller in robot control, this paper proposes and applies robotic seam tracking neural network based on Cartesian space trajectory control, which greatly simplifies the control algorithm. Computer simulation and experimental results show that the controller is very suitable for the actual robot welding, for the existing robot, without changing the internal structure of the controller, you can apply the technology. Compared with the common robot joint torque control method, the tracking accuracy is effectively improved and has strong robustness.