The control method of tracking curved fillets based on wheeled mobile welding robot with rotational arc as sensor is researched. A predictive fuzzy controller is designed to coordinately control horizontal slider moving and robot body turning when tracking curved welding seam. Through detecting the horizontal deviation and inclination of welding gun, the angular deviation between the orientation of robot moving and the direction of welding seam trajectory is acquired. The proposed predictive fuzzy controller employs weighted least square fitting method to predict the next position of welding seam trajectory and uses linearization model to predict the next orientation and position of welding gun. In addition, fuzzy control is adopted to deal with nonlinear characteristic of the system. In experiments, the proposed controller is applied to tracking curved seams and compared with a traditional fuzzy controller. The experiment results demonstrate that the feasibility and advantages of this predictive fuzzy control on the curved weld seam tracking. © 2010 Journal of Mechanical Engineering.
