Cognition for robot scanner based remote welding

Abstract

The effort for reduced cycle times in manufacturing has supported the development of remote welding systems which use a combination of scanners for beam delivery and robots for scanner positioning. Herein, close coupling of both motions requires a precise command of the robot trajectory and the scanner positioning to end up with a combined beam delivery. Especially the path precision of the robot plays a vital role in this kinematic chain. In this paper, a sensor system is being presented which allows tracking the motion of the laser beam against the work piece. It is based on a camera system which is coaxially connected to the scanner thus observing the relative motion of the laser beam relative to the work piece. The acquired images are processed with computer vision algorithms from the field of motion detection. The suitability of the algorithms is being demonstrated with a motion tracking tool which visualizes the homogeneity of the tracking result. The reported solution adds cognitive capabilities to manufacturing systems for robot scanner based materials processing. It allows evaluation of the relative motion between work piece and the laser beam. Moreover, the system can be used to adapt system programming during set-up of a manufacturing task or to evaluate the functionality of a manufacturing system during production. The presented sensor system will assist in optimizing manufacturing processes