Research On Design And Key Issues For Paint Robot With Hollow Non-spherical Wrist

Paint Robot is one of the key systems for automobile manufacturing, and therefore its structural design, related theories and methodologies have significant value in the academic researches and the industrial circles. Several key issues relevant to the development of a Paint Robot system are extensively investigated in this dissertation, such as the robot kinematics, dynamics, mechanism performance evaluation, dimension optimization, and geometry parameter calibration, et al. A prototype has been developed and calibrated. The following contributions have been made.The space motion and pose of the robot are described using the dual quaternions. The mapping between the traditional D-H parameter method and the dual quaternions is established. According to the characteristics of this mapping, the projective geometry method is introduced for the kinematics analysis of the robot. The elements of the rigid motion group, which expresses the space pose and motion of the robot, can be mapped to a point of the Study quadratic surface in the seven-dimension projective space utilizing the projective mapping. Then, the kinematics analysis method is established via geometry frame. The inverse kinematics algorithm of the non-spherical wrist 6R robot is realized utilizing the intersection of the constraint manifolds of two 3R sub-mechanism with the Study quadratic surface. This algorithm serves as the basis of the offline programming and trajectory planning of the non-spherical wrist robot.The dynamics model of the robot is established based on Kane Equation, and the body Jacobian matrix of the robot based on screw theory is introduced to solve the difficult problem of the partial rate in the traditional Kane Equation. The consistency in the mathematical expression of the body Jacobian matrix and the partial velocity is demonstrated. Utilizing screw theory, the body Jacobian matrix in screw form makes the Kane Equation simple and clear in geometric meaning. The velocity and acceleration of the links can be derived from the recursive formula based on Newton-Euler equation. The dynamic model of the robot is then established. The dynamic model is validated by comparing the computational analysis results with the software simulation. The peak torque and the moment of inertia on the motor shaft in the acceleration and deceleration phases are calculated based on the peak torque and the inertia matching principle. The servo motors are selected according to these two sets of parameters.In order to evaluate the velocity and acceleration performances of the robot, the global performance index of the velocity and acceleration is established according to the condition number of Jacobian matrix and Hessian matrix. The solving equations for the elements of Hessian matrix are derived based on the exponential product formula. The solution process is simplified by the application of screw theory. Moreover, setting the global performance index of the velocity and acceleration as the optimization objective, the relationship between the size of the links and the performance index of the robot is investigated. The size combinations of the main links are: 0.8m and 1.5m under the optimal performance of global velocity; 1.1m and 1.2m under the optimal performance of global acceleration. The index system for the synthetical performance evaluation of different robots with the same functions is also established, where the formation of the performance level is derived from the system evolution's perspective, and the self-organization feature map algorithm is utilized to simulate the process of evolution. The effectiveness of the proposed index system is validated through a specific case study.The detailed design of the robot with non-spherical hollow wrist is completed according to the previous design theory. The manufacturing and assembly has been finished; and the control system is built upon YASKAWA servo system. The MDH kinematics model is established for the geometry error calibration of the robot. The redundant items of the geometry error are derived using singular value decomposition of the extended Jacobian matrix. And a simplified calibration model is obtained after excluding the redundant geometric parameters. Computational analysis is performed, where the maximum position error is reduced from 4.978mm to 0.056mm after the compensation. The results fully satisfy the accuracy requirements of the Paint Robot. Furthermore, the calibration experiment has been performed, where a laser tracker tool is utilized to measure the space position of the robot; and the real geometric error is calculated to compensate the parameter error. The maximum position error is reduced from 14.46mm to 0.12mm after the calibration. The experimental results validate the calibration model and further reveal the fact that the developed robot could meet the accuracy requirements of painting works after the error compensation.