Research On Robot Positioning Accuracy Compensation Technology Based On Error Model

The application of robot in modern industry mainly relies on off-line programming technology to plan robot movements,which puts higher requirements on the absolute positioning accuracy of robots.The general way to improve the absolute positioning accuracy of the robot is to calibrate the error of the robot.However,the error of the base coordinate system cannot be recognized by the traditional calibration technology of geometric parameter error.The accuracy of calibration parameter is not high,and the compensation effect is limited by the deformation of the joint under the weight of robot linkages.Aiming at the deficiency of the traditional geometric parameter error calibration technology,the error of the base coordinate system and the flexible deformation of joints are introduced respectively based on the geometric parameter error model,and the precision compensation technology of the robot is studied.The main research contents of this paper are as follows:The D-H model is used to describe the motion of the robot linkage.By analyzing the principle of parameter mutation at the parallel joint,the MD-H model is introduced at the parallel joint to model the robot kinematics.The forward and reverse kinematics analysis of the robot is then given.On this basis,the construction method of the robot Jacobian matrix is given based on the differential kinematics method,and the robot geometric parameter error model is constructed.Owing to eliminate the error of robot base coordinate system error,a robot distance error model is constructed,and an error correction model of robot coordinate system is proposed.A simplified model of flexible deformation of 2th and 3th joints under the weight of the linkages was constructed for joint deformation of the robot,and a method for joint flexible deformation identification based on residual error was proposed.The parameter redundancy of the error model is analyzed,and the redundant parameters in the model are then identified and eliminated.The mechanism of the error correction model of the base coordinate system to predict the pose error of the base coordinate system is also given.Based on the L-M algorithm,the parameter identification method and process are designed.Based on the MATLAB / GUI platform,a calibration program and a calibration visual interface are developed.Numerical simulation of various error models is then performed to verify the effectiveness of the error model and identification algorithm,and to give the application scope of the error model.A calibration test platform based on IRB 6700-150 / 320 industrial robot and Leica AT960-LR laser tracker was built,and error measurement and error compensation experiments were carried out to increase the average position error of 50 test points from 3.4355 mm to 0.1750 mm,and the compensation effect remains within 1% after the perturbation error of the base coordinate system,and the average position error is further optimized to 0.1122 mm after introducing joint deformation compensation.