Predictive fuzzy control for a mobile welding robot seam tracking

Abstract

For the curved weld seam tracking problem of wheeled welding mobile robot used in shipbuilding and large spherical tank welding, predictive fuzzy control method was proposed in this paper. To overcome the disadvantage of rotational arc sensor can not detect the next seam position, the algorithm employed weighted least square fitting method to predict the weld seam next position and used linearization model to predict the robot next orientation and position. In addition, fuzzy control was adopted to deal with nonlinear characteristic of the system. The main advantages of this predictive fuzzy controller include small tracking error, small robot orientation delay angle, and the availability of a linearized predictive model that reduce the computing complexity, and the highly tracking quality. In experiments, we applied this control method to tracking curved fillet weld seams and compared with a traditional fuzzy controller. The experiment results demonstrated that the feasibility and advantages of this predictive fuzzy control on the curve weld seam tracking.