Intelligent Control Method Of The Table Tennis Robot—the Generation Of The Hitting Ball Trajectory

Table tennis robot is a hand-eye coordination system, which includes mechanical systems, vision systems and control system, respectively, equivalent to a human’s arm, eyes and brain. This paper focuses on the robot’s control system, that is to say how to control the robot’s wrist effectively to complete the desired strike action. Table tennis robot intelligent control of the racket is not only a focus but also a difficult issue; only solving this issue, the robot can achieve the control of the table tennis placement and control strategy when the robot play table tennis with people.In this paper, we use the imitation learning about table tennis robot to fulfill the intelligent control of table tennis, that is to say to make the robot’s arm mimic the action about the person’s arm hit the table tennis. Further said that the racket trajectory is the action sequences which should be imitated by robot when the people playing the table tennis. The first is to collect the video of people playing table tennis, and then to process the image which is the each frame in the video to get the table tennis in the image, and adopt appropriate processing algorithms, such as the Hough transform, Canny operator, the least squares method etc. to get the auxiliary lines which is marked on the racket. Then find the coordinates of the center of the racket and its pose according to the PnP algorithm and the auxiliary lines in the racket.After get the data of the center of racket, we will know the trajectory of the racket that is the hand trajectory when people to play table tennis. In this way, you can take advantage of these calculated data and make use of the BP neural network model to train and teach the robot to hit the table tennis using the humanoid posture. Through the imitation study, training a variety of batting trajectory, for example, to block the ball trajectory, spike trajectory and cut the ball trajectory etc, and then to implement the robot’s intelligent control of the racket.In this paper, we use the OPENGL technical to achieve the simulation system about the table tennis robot, to provide a simulation platform for the subject of future work.