Reaserch On The Key Technologies Of Robotic Automatic Polishing Sysytem

With the development of industrial automation technology, robots are increasingly applied to the automated assembly lines. Grinding and polishing the surface of the ware is a complicated process, therefore, manual operation is not only difficult to ensure product quality, also poor working environment for workers results in great health hazard. In order to improve working efficiency and product quality, developing an automated robotic polishing system has become the focus of ware processing industries. Currently robotic grinding and polishing systems on the market have two common problems:the quality of processed ware is poor; the process of artificial teach programming is complex but with low efficiency. To address these problems, this work conduct a series of research work as follows:l.The belt grinding was used for ware polishing. Specifically, this work studied the belt grinding process to identify the grinding parameters which influenced the ware polished quality. Moreover the relationship between them was discussed.2.Based on SVR(Support Vector Regression), a modeling method was put forward aiming at predicting and controlling the surface roughness of ware during the robotic grinding process. By using belt linear velocity, robot feed rate and curvature as input while the roughness as output, a regression model of the grinding process was built, which can help programmers to set appropriate grinding parameters according to the desired roughness.3.The forward and inverse kinematics of robot was also analyzed. Using the D-H method, the kinematics model of ABB IRB4400 robot was established, and the forward and inverse kinematics equations were obtained. Besides, the Matlab simulation of robot motion provided the variation of each joint and verified the correctness of the kinematics analysis, which laid the foundation for robot off-line programming.4.Robot off-line programming techniques were studied. In the off-line programming software (Robotstudio), path planning was based on the CAD model of the ware. Firstly, the grinding curve was obtained by cross-sectional method, and then the grinding path points depended on the curvature were set up to get the correct grinding path. Secondly, robot program was automatically generated based on the grinding path. After the simulation, robot program was finally loaded to the robot controller and completed the automation process.