Design Of Chess Game System Between Human And Robot For Parallel Robot

As one of the ten fields in China’s manufacturing power strategy "Made in China 2025",intelligent robot has high value to be researched.And game robot,which can quickly verify the effectiveness of robot intelligence algorithm,has always been a researching hotspot.Parallel robots,with the characteristics of large stiffness,strong load-carrying capacity and good dynamic performance,and can be applied to most working scenarios only with a few degrees of freedom,is attracting the attention of many researchers.In this paper,the delta robot is studied,the 3-RRR delta robot model is analyzed,the vision system and game decision system are designed,and the robot platform is comprehensively developed.The main work is as follows:(1)Robot model analysis and simulation.The 3d model of the robot is obtained according to the real robot,and the model analysis is carried out;Robot workspace visualization;The kinematics simulation is carried out and controller is designed for the robot.(2)Game-decision-system design.The common methods in the field of game is introduced,and the system-overall-framework for gobang game is put forward.The intelligent decision-making system was designed and tested according to alpha Zero algorithm,and the shortcomings of the algorithm were analyzed and the improvement scheme was proposed.(3)Research and design of robot vision system.Introduce the key technology of vision system;The overall vision system architecture of autonomous chess robot is proposed.Focus on the design of chessboard image acquisition and processing,chess classification and chess situation generation;Two solutions are proposed to solve the chess piece classification problem.One is the traditional chess piece classification model based on HSV space,which is suitable for distinguishing the colors of different chess pieces.The other is the pawn classification model based on convolutional neural network(CNN).The visual system was built and its performance was tested.(4)Robot autonomous chess test.The overall structure of the autonomous chessplaying robot is proposed,the platform-development of the robot system is completed and tested,and the overall performance was verified.This system realizes the function of playing chess by robot perception,which is of great significance to promote the diversification of robot application and improve the quality of robot teaching.