Research And Design Of Table Tennis Robot System With Seven Degrees Of Freedom

In recent years,with the development of robotics and artificial intelligence technology,intelligent robots have gradually entered people's vision,and intelligent sports robots have also begun to move from experiments to applications.Table tennis is a very popular sport in China,the development of table tennis robot research has broad application prospects in China.The research of table tennis robots involves machine vision,deep learning,robot control and other technologies.This thesis takes the 7-DOF table tennis robot as the research object,focuses on the research and design of the vision system and control system in the table tennis robot system,and implements the entire system.This thesis first analyzes the vision system of the table tennis robot in detail,and then completes the design of the following three modules of the vision system according to their functions.The image acquisition and camera calibration module realizes the acquisition of ping-pong ball motion images and camera calibration.In the object detection and coordinate calculation module,after research and comparison,this thesis chooses color segmentation for target detection of ping-pong ball.Aiming at the speed of the conventional color segmentation method is slow,a parallel object detection method based on CUDA is designed to improve speed of target detection.Then the threedimensional position of the table tennis is obtained by the binocular camera target positioning principle.In the ping pong trajectory prediction module,a flight model of ping pong is established by means of force analysis,a rebound model of ping pong is established by means of support vector regression,and then the prediction of ping pong trajectory is completed using these two models.In this thesis,a 7-DOF robot is used as the execution mechanism,and the control system of the table tennis robot is established by detailed design of the 7-DOF robot kinematics solving module,the robot's obstacle avoidance path planning module and the batting action path planning module.The 7-DOF robot kinematics module is to establish the correspondence between the position of the hitting point in the three-dimensional space and the angle of each joint of the robot.In the robot obstacle avoidance path planning module,this thesis designs a method for the robot to avoid obstacles on the table tennis table by geometric analysis of the relative position of the robot and the table tennis table when the target point is determined.Then the method of analyzing the rotation characteristics of the robot's joint designed a method for the robot to avoid obstacles on the pingpong table in the course of hitting the ball.Finally,by designing the robot's ball striking path,the robot can be guaranteed to have a certain speed and direction.In addition,this thesis designed a three-dimensional model of a 7-DOF table tennis robot based on MATLAB,and verified the effectiveness of each module of the control system.In order to implement the entire system,a communication module between the vision system and the control system is designed to complete the data communication between the two sub-system.Then,in order to solve the problem that the robot cannot reach the hit point with a fixed attitude,this thesis designs a method of selecting the hit ball attitude based on the ball position.Finally,a 7-DOF table tennis robot system based on the KUKA LBR iiwa 7 R800 robot is built in this article,and the feasibility of each module design is verified.