Research On Man-machine Game Parallel Robot System

In recent years,the entertainment robot industry has developed rapidly,especially the game robot with man-machine game function.Many countries have made efforts to deepen the research in this direction,and extended the game theory to various fields,achieving remarkable results.However,the domestic research on man-machine game is still in the stage of continuous improvement and development,far behind the developed countries,in view of this situation,it is urgent to carry out the game robot research with man-machine interaction function,and promote the development of the domestic game robot industry.Delta parallel robot has beautiful appearance,compact structure and good dynamic performance,which is very suitable for the hardware structure of game robot.Therefore,a Delta parallel human-machine chess robot system is designed.Combining the advantages of traditional game software and new intelligent robot,this system can complete the whole process of chess seeing,calculating and picking according to the rules of gobang,and realize the face-to-face gobang game between robot and human.The Delta parallel human-machine chess robot system is composed of three parts: visual recognition subsystem,board executive subsystem and game algorithm subsystem.It can simulate the whole process of human game and realize the game between human and computer on the physical board.The vision recognition subsystem consists of four parts: the main control computer,the camera,the video acquisition card and the image recognition algorithm.The checkerboard executive mechanism subsystem is composed of two parts: the checkerboard main mechanism and the Delta parallel mechanism,which can realize the functions of automatic child fetching,child moving and down son,etc.The subsystem of game algorithm is composed of game search strategy,game database and man-machine interaction module to realize man-machine chess function.In terms of the design of visual recognition subsystem,a whole set of recognition system is constructed for chess,checkerboard calibration,placement error compensation and chess color recognition.In terms of visual control technology,in order to realize the judgment of the chess game and provide the next game operation carried out by the executive organ for the chess game,the vision control system scheme of look-and-move structure based on position is constructed:(1)aiming at the problem that the camera may be installed on the side of the checkerboard,which causes the distortion of the acquired image,the fitting polynomial method is designed to locate the checkerboard grid points.(2)after a series of gray level processing,smoothing denoising,thinning and segmentation of the collected chess images,the contour tracking method is adopted to identify the chess target in the board,and the maximum area recognition method is used to compensate the position error of the opponent's outposts to achieve accurate identification and positioning.(3)to design a set of modular visual control system,the visual recognition and visual control is divided into two modules,subdivision of the work content,to ensure the checking accuracy of each module,at the same time,parallel working mode,improve the overall control efficiency of the system.In terms of the subsystem design of the actuator,a set of execution system with Delta parallel robot as the main body was designed.According to the requirements of the gozi game,the kinematics model was analyzed to design and select the parts and components,cooperate with the visual control system component selection,design the control system software,and finally complete the platform construction.In terms of the design of game algorithm subsystem,based on the four-game rule of continuous punches of five pieces of chess,the paper optimized the maximum-minimum alpha-beta pruning search algorithm,and designed the C4 alpha-beta pruning search algorithm,which significantly improved the computational efficiency and game level.