Research On Kinematics Parameters Calibration Method Of Industrial Robot Based On Spherical Center Distance

The repetitive positioning accuracy of today's industrial robots is generally high,while the absolute positioning accuracy is very low,which hinders its development and popularization in the industry,so improving absolute positioning accuracy is of vital practical significance.The existing research methods have the disadvantages of complicated measurement process,expensive measurement equipment,and low calibration accuracy.In view of the above problems,a calibration method for kinematic parameters of industrial robots based on spherical center distance constraints is proposed,which can effectively improve the absolute positioning accuracy of the robot,and can simplify the calibration process and reduce the calibration cost.This article focuses on the calibration of operational parameters with six-axis series industrial robots as the research object.The main contents and conclusions are as follows:First,based on the principle of homogeneous coordinate transformation of industrial robots,two kinematic modeling methods are studied.Aiming at the singularity of the D-H model in parallel joints,combined with the MDH modeling method,a kinematics model is established for the FANUC LR Mate 200 i D industrial robot.The expression about the relationship between the joint angle of the robot and the movement of the terminal posture was obtained and the established model was verified by simulation.Secondly,the forward and inverse kinematics of the robot are derived based on the kinematics model,and provide a theoretical basis for the subsequent parameter identification and error compensation.The cause of the end error of the industrial robot is analyzed,a distance error model based on two standard spheres is proposed,and a calibration error equation is derived.Then,the redundant kinematic parameters existing in the error equation are studied,and the redundant parameters that affect the parameter identification are removed to obtain 18 identifiable kinematic parameters.The least square method and Levenberg-Marquardt algorithm are used to simulate and analyze the influence of redundant parameters on the calibration results.Comparing the simulation results,we can see that the calibration effect is better after removing the redundant parameters.Finally,an experimental system for calibration is built,and an error model for removing redundant parameters is used to perform calibration experiments on the above two identification algorithms.By comparing the evaluation indexes of the ball center distance errors before and after the calibration by the two algorithms,the analysis shows that the maximum deviation of the ball center distance using the Levenberg-Marquardt identification algorithm is reduced by 71.28%,the average deviation is reduced by 73.58%,the standard deviation is reduced by 75.76%Compared with the least squares identification algorithm,this algorithm is more effective for improving the absolute positioning accuracy of the robot.