Research On Painting Robot Based On Multi-sensor Fusion

Robot is an intelligent product combined of machinery and electronic equipment developed in recent years. With the development and progress of materials, machining, electronics, controlling, sensors and other related technology, the robot has been widely used in various fields. Especially with the moving forward of urbanization process, the demand for painting robot is dramatically increased. In this background, the research on the painting robot is particularly important.Multi-sensor data fusion technology combines controlling theory, signal processing, artificial intelligence and other disciplines' advanced theory and technology, through the integrated treatment of the data collected by a number of sensors, the information collected can be more accurate, and have the redundancy and complementarity. Using the Multi-sensor fusion technology in the robot, the accuracy in controlling of the robot can be significantly improved.In this paper, Firstly, combined with the working environment and design requirements, presenting the overall structure and material of the painting robot, designing the structure of major joints and selecting the drive motor, completing the parts'three-dimensional designation of in solidworks, assembling the overall structure of the painting robot. Then, processing the kinematics analysis and inversing kinematics analysis of the painting robot, establishing the D-H parameters, linking coordinate system and kinematic equations of the painting robot, using the Robotics Toolbox in MATLAB analyzed the simulation of the manipulator model. Secondly, processing the dynamic analysis of the painting robot, educing the models created in solidworks into ADAMS, setting he equations of motion of each joint in ADAMS, then simulating and analyzing the painting robot with the ADAMS solver, generating the data curve. After these, researching the multi-sensor fusion technique in robotics application. Through colligating the data collected from a number of sensors, The multi-sensor fusion technique can achieve the unified description of the information outside. The information obtained will be more accurate. So that the multi-sensor fusion technique can improve the robustness and stability of the system, expand the system's space-time coverage. The paper focuses on the analysis of system model and algorithm processes of Kalman filter, researches the application of data fusion in speed sensor and acceleration sensor, uses MATLAB to simulate the process of Kalman filter estimates the acceleration's deviation. Finally, discussing the application of RBF neural network in adaptive control of robot, introducing the RBF neural network, researching the transform of Cartesian coordinates and joints'position of manipulator in working space, modeling the manipulator in RBF neural network. On this basis, processing the simulation of two degrees of freedom manipulator.