Robotic Curve-Shaped Seam Tracking System Using Vision Sensor

This paper presents a vision-based curve-shaped seam tracking system. The measurement of the curve trajectory is based on a camera that is installed ahead of the end effector of the robot. Effective image processing method is designed to extract the feature points of the curve. Then a fuzzy controller is proposed to control the position of the robot to track the curve. To improve the performance of the controller, a self-tuning mechanism is added to the controller. Finally, experiments are fully conducted to verify the effectiveness of the proposed method.