Research Of Kinematic Calibration And Error Analysis For Industrial Robot

Kinematic calibration of industrial robot is one of the key techniques of robot off-line programming, and is also a very important in Robotics. It has great significance in both theoretical research and engineering practice as the industrialization of Robot is current tread. Kinematic calibration of industrial robot is a process for identifying actual parameters based on kinemctic modeling, which is used in compensating the robot error.When industrial robot works as a teaching and playback robot, the main performance index is repeatability accuracy, and when works as an off-line programming robot, the main performance index is absolute accuracy. However,the repeatability accuracy of industrial robot is high enough but the absolute accuracy is low which cannot meet the requirement in off-line programming. So we have to calibrate the robot to improve its absolute accuracy. As the robot off-line programming system's developing, kinematic calibration of industrial robot is becoming more and more important.This paper detailedly discusses interrelated conceptions and theories, principles and approaches of kinematic calibration of industrial robot, which are the base of calibration and modeling.Firstly classical D-H model is established according the robot's structure and the solution to both forward kinematic problems and inverse kinematic problems are worked out. Then we point out the limitation of D-H model in calibration and establish a new model according to both the classical and a modified D-H modeling convention for calibration. Further the robot's error model is established according final actual kinematic model.After getting the solution to IKP (inverse kinematic problems) with algebraic method, a study of accurate solution to IKP based on genetic neural network is introduced. This method uses the forward kinematic results as training data set in the GA-RBF network and the nonlinear mapping from joint variable-space to operation variable-space was obtained with iterative training and learning method.After discussing two kinds of robot calibration approaches which both construct a closed kinematic chain to calibrate robots, a new approach called virtual closed kinematic chain method is described. Furthermore, the principle, system setup and details of this method are discussed and illuminated respectively. This method magnifies the end effector error on a observing plane using a laser line so as to get more accurate joint values which help identify more accurate parameters.To test the proposed method of Kinematic Calibration and facilitate further study, ABB IRB140 robot is calibrated in experiments and Simulations are conducted by taking certain data into the process of algorisms to test the robustness and validity.