Design And Implementation Of Intelligent Automatic Table Tennis Ball Picking Robot

With the continuous development of research and application of service robot technology,the application requirements of service robot in the field of sports have begun to be presented and gradually met.The intelligent table tennis ball picking robot designed in this article meets these demand.It is believed that there will be broad development and application prospects.In this paper,Kinect v2 and RPLIDAR A1 radars are used as the main body sensors.Adopt Kobuki as the robot chassis.Under the Linux system,software libraries such as ROS,Open CV,Kinect SDK and the C++ programming language are used to design and implement an intelligent table tennis automatic ball picking robot.The main work of this article includes:(1)Analyze the needs and determine the target function to be achieved by the robot,and design the system hardware structure and software modules according to the functional requirements.(2)Aiming at the problem of table tennis detection,a fast and stable table tennis detection algorithm was designed using the color and shape characteristics of table tennis,and the image recognition rate of table tennis in complex scenes was improved.A method for filtering invalid circles based on the relationship between distance and size is proposed to reduce the false detection rate of table tennis;a table tennis positioning algorithm is designed based on the camera pinhole model,which realizes the real-time global positioning function of table tennis.(3)Based on the Ant Colony algorithm,a ball picking sequence planning algorithm module is designed,and the optimal parameter setting of the ant colony algorithm is studied;In order to improve the global search capability,an optimized ant colony algorithm based on cross-operation is proposed.This algorithm introduces the ‘cross-operation' during the iterative process.,which reduced probability of getting stuck in a local optimal solution;Based on the differential wheel motion model,another improved Ant Colony algorithm is proposed,which considers the effects of robot steering to optimize the path selection mode in actual operation,and improves the operation efficiency of the intelligent ball picking robot.(4)The navigation and obstacle avoidance module is designed from the aspects of global path planning and local obstacle avoidance.Aiming at the global path planning algorithm,an improved A* algorithm based on JPS+ algorithm is designed,which makes up for the shortcomings of the low calculation efficiency of the A* algorithm and implements a stable and real-time path planning function.Aiming at the local obstacle avoidance algorithm,this paper presents an improved Dynamic Window Approach(DWA)based on the Velocity Obstacle(VO)algorithm which realizes the obstacle avoidance function for moving dynamic obstacles and improves the running safety of the ball picking robot.Integrate each module into a system,and conduct experimental tests on each module and the overall system.The experimental results show that the performance of the the intelligent table tennis automatic ball picking robot designed in this paper has reached the expected design goal,which provides a reference design model for the product development of table tennis ball picking service robot.