Harmonious visual servoing control for tracking weld seams shaped curve

A visual serving control method which harmonizes the torch motion with visual tracking is proposed for robot tracking weld seams shaped curve. Firstly, a mathematical model for feature points is established, and then rotation axis is computed. A fuzzy visual serving controler with two-layers structure is designed to keep image features among vision field by computing valid range of control outputs. In order to get precise valid range of control output, a state estimator based on Kalman filter with dynamic model compensation is designed. The proposed method is verified by tracking experiments.