부산대학교 도서관

This study presents a solution to the significant profile error in surface milling caused by low absolute positioning accuracy and weak robot stiffness. It proposes a predictive model for surface machining profile error that considers both robot positioning error and deformation error. A hierarchical compensation method, integrating offline and online approaches, is also introduced. The model employs the relevance vector machine to address the strong nonlinearity of positioning errors, while the deformation error prediction model is based on the robot stiffness model. The influence of robot error on the surface profile error is analyzed. To mitigate the accuracy loss resulting from the coupling relationship between positioning and deformation error, an error compensation method is proposed. This method utilizes an offline iterative algorithm to calculate the compensation value for positioning error and adjust the machining toolpath. Then, a method for calculating the compensation value for deformation error is presented, which combines online cutting force measurement and the cutting depth model to reduce the calculation time during online iterations. The effectiveness of these error compensation methods in improving the accuracy of robot surface milling is demonstrated through experimentation.