Industrial Robot Positioning Error Compensation Based On The Motion Model

With the change of car body size in automobile manufacturing industry, the laser detection system in production line is widely used. The detection precision is closely related to the quality, while the positioning accuracy of the robot in the detection system will influence the overall detection precision. It is important to study how to improve the positioning accuracy of the robot.In the paper, the most common KUKA robot in the laser detection system is taken as an example to study the kinematics problems. The corresponding motion model is established based on the actual structure of the KUKA robot, and the calculating method for both the forward and inverse kinematics solutions. Simulation is applied to compare and analyze the common inverse kinematics method and neural network method.Based on the established kinematics model, body coordinate system is introduced to transform the car body positioning error to the differential in the system, considering the coupling influence to the robot positioning accuracy from the car body positioning precision. According to the geometric error, the model of KUKA robot positioning error is established and verified with simulations.Considering the real-time compensation for the KUKA robot positioning error and the influence of temperature, RBF Neural Network and the Spatial Interpolation algorithm is applied to illustrate that the precision of comprehensive compensation method is more satisfied compared with the NewTon-Raphson Iteration method.