论文部分内容阅读
针对航空制孔机器人绝对定位精度补偿中存在的建模复杂及运算量大的问题,提出了一种基于极限学习机的绝对定位精度补偿方法。该方法通过将机器人视为一个黑箱系统,忽略机器人的几何因素和非几何因素的影响,通过高精度的激光跟踪仪测量获得机器人的末端运动误差,采用极限学习机建立机器人误差预测模型。由机器人误差预测模型获得机器人在期望位置的位置偏差,通过修正机器人位置坐标来实现机器人的绝对定位精度补偿。最后该方法在航空制孔机器人上进行了试验,试验结果显示机器人的绝对位置误差的平均值和最大值分别降低了75.69%和78.16%。
Aiming at the problem of complicated modeling and large amount of computation in the absolute positioning accuracy compensation of aeronautic hole robots, an absolute positioning accuracy compensation method based on extreme learning machine is proposed. The method considers the robot as a black box system, ignores the geometric factors and the non-geometric factors of the robot, obtains the robot’s end movement error by the high-precision laser tracker measurement, and establishes the robot error prediction model by the limit learning machine. The position error of the robot at the expected position is obtained from the robot error prediction model, and the absolute positioning accuracy of the robot is compensated by correcting the position coordinates of the robot. Finally, the method is tested on aerohole robot. The experimental results show that the average and maximum absolute error of the robot are reduced by 75.69% and 78.16% respectively.