Research On Several Key Technologies Of Chinese Chess Robot

The current chess robots are mostly based on industrial robots for secondary development.Due to the large size and high cost of industrial robots,they do not conform to the positioning of chess robots as service and entertainment robots.Chess robots use special chessboard and chess pieces,so that the recognition algorithm of chessboard state is not universal.It doesn't have fun to pursue a winning goal and the human-machine game experience is poor for using the traditional computer chess game algorithm.To deal with these problems,this thesis studies the key technical points such as the recognition of chessboard and chess pieces,location and recognition of chess pieces in the chess area and path planning of putting chess pieces,the anthropomorphic game of the chess robot,the structure design of the chess robot and its software and hardware control.Finally,developed a chess robot system platform.The main research contents of this thesis are as follows:Chessboard and chess pieces recognition technology of chess robot is studied.The rough location of chessboard corners is realized Based on the background image of chessboard.Scanning chessboard corners sequentially,using corner neighborhood image based on HSV space to determine whether there is a chess piece and its color categories at corners.In order to solve the problem of low accuracy of chess recognition caused by uncertainty of direction,position and scale of chess image,a chess character classification and recognition model based on multi-layer feedforward network is constructed for red and black chess,which takes the character outline normalized Fourier descriptor feature as input.The accuracy of chess piece recognition is over 99.3%.Chess robot' s technology of location and recognition of chess pieces in the chess area and path planning of putting pieces is studied.Based on the circular contour matching method,the chess piece circle contour detection is realized to realize all the chess pieces positioning,and the chess pieces are recognized based on the chess piece recognition method of this thesis.Based on the greedy algorithm,path of putting pieces is planned,and the selection order of the pieces in the chess area and the placement position of the pieces in the chessboard area are determined,and the access position queue of the end of the chess robot is obtained.After experimental verification,Comparing with putting the chess pieces in order,putting chess pieces path planning based on the greedy algorithm shortened the total path length of putting chess pieces,and the running time of the algorithm meets the requirements of real-time planning.Anthropomorphic game decision of chess robot is studied.Put forward the concept of piece preference and game style of chess robot from the perspective of personification.Combining with the piece preference factor,Adjust the evaluation of the situation generated by the root node and the influence of the chess preference factor on the game decision is analysed.Combining with the game style factor,Adjust the evaluation of the situation generated by the root node.By adjusting the size of the game style factor,the game program has a radical,moderate and robust style.The influence of the game style factor on the game decision is analysed.Finally,it is proved through anthropomorphic game experiment that it is feasible to study chess anthropomorphic game from the perspective of chess preference and game style.Developed a chess robot system platform.Starting from the functional requirements of the chess robot system,the functional modules required by the system are analyzed,and the design and development of each module are completed.Finally,the chess robot system platform is developed.Finally,the feasibility of the chess robot system is verified by the game comprehensive experiment of the chess robot.