Integration of Visual Information and Robot Offline Programming System for Improving Automatic Deburring Process

Abstract

The automatic deburring process of casting parts has been commonly investigated by industrial robot manipulators in recent decade. Majority of solutions are dependent on human teaching or robot Off-Line Programming system (OLP), which assume that deburring paths are completely predefined offline and have a perfect relative calibration between the workpiece and robot. Those limiting assumptions lack practical applications in uncertainty as diverse as modeling error, robot positioning error and environment. In this paper, we consider the method of automatic deburring process with visual assistance for synchronously detecting the deburring path and compensating calibration error. To do so, a deburring system which consists of two interconnected modules is presented, including the detection and extraction of deburring paths using a industrial CCD camera, and a self-developed OLP system (RobSim) is applied to generate robot executable trajectories with the presence of limited visual data. Our deburring system that has the automatic deburring capability depends on the ability to have a good insight of the workpiece and environment. An industrial application is considered to verify the feasibleness and effectiveness of our system. Experimental results show that the improvement of deburring performance by the seamless integration of visual system and the advantages of an OLP system in the field of automatic deburring process.