An RBF-PD Control Method for Robot Grasping of Moving Object

In order to solve the uncertainty of robot’s grabbing position of moving objects, a control method based on RBF (radial basis function) neural network and PD (proportional-derivative) for crawling dynamic targets is proposed. The Kalman filter algorithm is used to estimate the pose of the moving target. The information of the pose estimator is used as the input of the adaptive neural network controller. An adaptive robust control scheme based on RBF neural network and PD is proposed. It ensures that the trajectories are accurately tracked even in the presence of external disturbances and uncertainties. The machine learning method is implemented into a vision-based control scheme to compensate for the uncertainty of the estimated grasping position and improve the success rate of the robot’s accurate grasping. Finally, the experiment was carried out to verify the effectiveness of the proposed method.