Research On The Absolute Positioning Error Analysis And Compensation Of SIASUN SR10C Robot

Absolute positioning accuracy is an important factor to measure the performance of industrial robots.With the development of the intelligent flexible manufacturing technology nowadays,robot off-line programming will be used on more and more occasions,but the low absolute positioning accuracy will cause the problem that the offline programming can not achieve the desired results.The paper took SIASUN SR10 C robot as the research object to explore the methods of improving the absolute positioning accuracy of the industrial robot.The main research contents were as follows:According to the D-H modeling rule,combining with the actual institutional characteristics of the SR10 C robot,the kinematics model was established.The kinematics simulation proved that the established model was the same as the model of the robot controller.According to the kinematics model,the differential method was used to derive the geometric error model of robot.The laser tracker was used to mersure the position and orientation error of robot.According to the experimental requirement ed,a set of end executor was designed,and a tool coordinate system calibration method was put forward.According to the geometric error model and the position and orientation error,the circumference of a circle method and least squares identification method was adopted to get the actual D-H parameter to update the model parameters of robot controller.The mean square error of position from the method was 1.004 mm,accounting for 12.5% that before calibration.The mean square error of euler Angle were smaller than the one before the calibration.The paper studied the influence of two non-geometric parameters such as load and joint speed on the positioning accuracy of robot.The relative position errors of SR10 C robot with different speed and load were obtained through the experiment.The quadratic regression equation about the relationship of the position error with the speed and load was obtained by analyzing the experimental data.The paper used BP neural network to compensate the absolute positioning error of the robot.The absolute position error of the robot with 100% load and 99% joint speed was reduced from 1.429 mm to 0.329 mm from the compensation.