基于最小二乘法的机械手偏移距离预测

Prediction of Offset Distance of Manipulator Based on Least Square Method

机械手在各行各业的应用越来越多,随着带有机械手的设备工作时间的增加,一些用来固定机械手的零件逐渐老化、松动,就会引起机械手因偏离正常允许偏移范围而不能准确抓取、放置产品。本文采用最小二乘法对前一个月机械手横向、纵向和摆动偏移距离为样本建立模型,从而获取回归系数。此外,分别以上个月三个偏移量的最大值来预测下个月的三个方向偏移量。经大量采集某实际生产车间的数据及分析表明,该方法能够在一定程度上解决机械手偏移距离的随意性,验证了该方法的有效性。

The manipulator is increasingly used in different industries. With its working hours increasing,some of the parts used to fix the robot are gradually aging and getting loose,which leads the robot to deviate from the normal offset range and fail to accurately grasp,place the products. In this paper,the least squares method is used to establish the model for the horizantal,vertical and swing offset distance,and to obtain the regression coefficients. In addition,the three maximum offsets for the previous month is used to predict the those of the following month. The analysis and the data collected in production show that the method can effectively solve the arbitrariness of the manipulator offset distance.