Research of visual servoing control algorithms for industrial welding vehicles

Visual servoing control algorithms are proposed to improve tracking precision and robustness of welding vehicle with vision sensor. Firstly, a mathematical model of image features is established and then a detection algorithm for false data is designed. Secondly, an adaptive structure for visual servoing control is introduced to deal with position delay existing in the process of tracking welding seams shaped zigzag line. The adaptive control structure can detect variance of direction of weld seams and then adjust control parameters. In this control structure, an algorithm to detect variance of direction of weld seams based on Kalman filter, an algorithm to adjust reference pixel coordinates and a control algorithm which combines prediction control value with real-time adjusting value are given. Finally, the results of experiments on the proposed algorithms indicated the better robustness to disturbance and the higher tracking precision.