Precision robotic deburring based on force control for arbitrarily shaped workpiece using CAD model matching

Abstract

Surface finishing processes such as deburring are crucial for ensuring the quality of a workpiece and human safety by removing burrs. However, deburring involves excessive noise, dust, and vibration, which can be harmful to human workers. Thus, there has been extensive research into the use of robots to perform deburring instead of human workers. In robotic deburring, the precise tracking of the contour of an arbitrarily shaped workpiece is of major concern for precision deburring. In this study, to achieve precision deburring, a tool-path modification method based on a computer-aided design (CAD) model and direct teaching is proposed taking into consideration the position/orientation errors of the workpiece. In addition, based on this trajectory, impedance control is used to avoid applying an excessive contact force and a virtual wall is adopted to improve the force-control performance. Without knowing the position/orientation of the workpiece, the optimal deburring trajectory can be generated by matching the extracted tool path from the CAD model to the teaching points. From the simulations of the tool path modification method using an iterative closest point (ICP)-based contour matching algorithm and a series of experiments on robotic deburring, the performance of the proposed method was verified.