Research On Improving The Accuracy Of Industrial Robot Based On Online Pose Measurement System

In modern manufacturing technology,industrial robots are the important part of intelligent manufacturing system.With the German Industrial 4.0 manufacturing technology and Made in China 2025 national strategic plan proposed,the importance of the flexible digital processing system based on industrial robots rises rapidly.However,the present absolute positioning accuracy of robot cannot meet the requirement of highly precise industrial production.At present,the method of modifying the kinematic parameters of robot could improve the absolute positioning accuracy of industrial robots to some extent.But the current kinematics model cannot accurately describe the actual error of industrial robot.In addition,the calibration process often off-line,which cannot reflect the affection of the environmental or load factors during working.Therefore,this paper researches an external real-time pose measurement system for the robot to improve the absolute positioning accuracy.Through the real-time measurement and compensation to the robot pose by the external measurement system,it can simplify the complex process of robot parameter error modeling,adapting the change of actual situation,and also describing the the actual robot errors.With the close-range photogrammetry system,dual inclinometer and PC,this paper combines a six degree of freedom(DOF)pose measurement system,which can real-time obtain the 6 DOF information of the robot manipulator by the measurement approaches of vision,electricity and inertial.This paper develops a calibration method based on the combination of the off-line kinematic parameter calibration and the external measurement.Based on the Kalman filter and Weighted Matrix Linear Minimum Variance Optimal Information Fusion Criterion,this paper presents a method of dynamic prediction and redundant data fusion.Through combined the high precision encoder of industrial robot with the biaxial inclinometer,this paper develops a three-dimensional attitude measurement method.The optimal operating range and working conditions of the online pose measurement system are studied and analyzed.This paper also presents a coordinate transformation method based on the coincidence of the characteristic lines of point clouds.It can accurately and quickly unify the coordinate systems.By implemented on the KR KUKA 5 arc industrial robot,a real-time control and compensation software algorithm is developed.And the aim that the real-time communication and feedback to the robot by the on-line pose measurement system is achieved in lab.