Path Generation Method for Robotic Blades Grinding Based on Predictive Control and Extended Kalman Filter

The surface quality of air engine blades influences their dynamic performance directly, and it depends on the grinding process. Consequently, robotic grinding has a vital role in the manufacturing of large-scale blades. In this paper, a control method for curve tracking is proposed to generate paths for grinding heavily curved blades, especially their edge areas. Cartesian positions of the robot are controlled by a visual servoing based algorithm, which is divided into the prior measurement stage and the performing task stage for safety reasons. Owing to the employment of predictive control with designed transition sequences and the extended Kalman filter, the path generated by the method demonstrates required accuracy as well as promoted stability. Experiments have been carried out to verify its tracking performance faced with a curved workpiece and environment noises. It has been confirmed that the method can be used to control grinding tasks.