Kinematic Parameters And Dynamic Behavior Identification Technology Of Industrial Robot

The repeated and absolute accurac ies of an industrial robot are two important indexes to characterize the performance of the robot.Nowadays, the repeated accuracy of the robot has been able to reach a high level,generally being guaranteed within 0.1mm.While the level of the robot's absolute accuracy is still very low, and different robots have different levels, their differences are very large.some of them can reach millimeter,but others may even remain at centimeter level.With the popularization of the robot's off- line programming technology, improving the robot's absolute accuracy has become an urgnet problem to be solved in the field of robot manufacturing and application technology.In this paper, the Comau SMART5 NJ 220-2.7 industrial robot is used as the research object, a kinematic calibration method based on a single telescoping ballbar is presented. Based on the kinematics error model, the errors of the kinematic parameters are identified for each joint of the robot. At the same time, the preliminary study on dynamic behavior identification of the robot is completed. The paper 's main research contents and results are as following:1. Based on the classical DH modeling theory, the nominal kinematic parameters of each joint are obtained combing with the mechanical structure of the robot,and the kinematics model of the robot is established.Then the accuracy and rationality of the established model is verified by using Matlab. Finally,a method is proposed to optimize the operation space of the robot based on the kinematics model.2. Based on the differential motion theory and DH kinematics model, the kinematics error model of the robot is established. It is verified by using Matlab.And then based on it,this paper presents an error pre-compensation method on joint space.3. Based on a single telescoping ballbar, a position and attitude measurement method is proposed,and the measurement device is designed.Then the Levenberg-Marquardt algorithm is used to identify the error according to the error model.4. According to the calibration theory described in the above, the actual calibration experiment is conducted, and the error identification is carried out, and the results of the identification are compensated and verified. The results of the experiments show that after compensation, the maximum error value of the position error decreases 68.88%, the average error decreases 74.02%, the root mean square error decreases 72.72%;Meanwhile,the maximum error value of the attitude error decreases 77.53%,the average error decreases 87.69%, the root mean square error decreases 86.90%.5. The dynamic behavior identification of the robot in the milling process was designed and completed using the impact hammer.And then,the milling experiment is designed to verify the identification results.The processes and results of the experiment show that the identification results are reasonable.