Along with the development of automation in industrial field, the industrial robot, as a importantautomatic media, can be found various appliactions such as painting robot, transfer robot and weldingetc..Although the repositioning accuracy of the industrial robot is high, its absolut postioning accuracyis poor, which severely limits the spectrum of its application.Therefore, it is important and meaningfulto perform the research on the kinematic calibration.In this thesis, a robot designed in the lab is selected to be the research subject. To avoid thesingularity in D-H model; a modified five-parameter method is applied to establish the kinematicsmodel of the robot. To determine the relation between the coordinates of the base and the tool, thedirect and inverse kinematics is solved. Meanwhile, the coordinate of the tool is set up via thenoncontact nine-point method. Then, the error source about this robot is analyzed to conduct itsparameters indentification and error compensation. There are many error sources for industrial robot,we classified these resources into two error resources i.e., steady state error and dynamic error.Thesteady error includes the parameter error of connecting links and the revolution error of joints.Thedynamic error includes the elastic deformation of joints because of the gravity. Kinematics model andthe error model are established for the end efector. Through the analysis of the relationship betweenthe error sources and the posture error of the end effector,a parameters identification algorithm isproposed to obtain the actual parameters of the industrial robot by measuring series of actual positonand orientation. Relative rotation error model and distance error model are introduced to avoid theconversion between the base cooridinate of the robot and the measure coordination. The errorcompensation algorithm is studied by input motion programming method.Futher, the correctness and accuracy of the parameter identification algorithm and thecompensation algorithms are verified by simulations in Mathematica software. Then, the kinematiccalibration experiment of the industrial robot is conducted to measure the actual parameters of therobot. With the error compensation algorithm, the error of the end effector is compensated. Theexperimental method in this thesis is simple and reliable. The compensated terms is related to therevolutionary variables of the joints. After the compensation, the accuary of this robot is improved.Meanwhile, the requirments for on-line compensations are fulfilled. |