Robot Soccer Game In Real-time Intelligence Technology Research

Robot Soccer game is a high-tech oppositional activity that develops rapidly in recent years, and is the combination of the physical culture and high-tech. It is a milestone after the "Deep Blue" which won the international chess master. In the soccer robot system, several robots cooperate with each other and complete the aim of shooting in the real-time, noisy, oppositional and complex environment. At present, many researchers apply the real-time technologies, such as ANN(Artificial Neural Network), EA(Evolutionary Algorithms), MAS(Multi-Agent system), ES(Expert system), decision-making, path-planning and fuzzy logic to the normal plat of Robot Soccer, and make some progress. However, the development of Robot Soccer is underway, some theories is not so perfect, and some arithmetic should be improved.The paper focuses on the problem of path-planning in the system and effective real-time technologies in the system based on the design and development of the soccer robot system.Firstly, the paper introduces the background and actuality of the exploitation, tells the whole system based on the vision has four component: the decision-making subsystem, the vision subsystem, the wireless communication subsystem and the robot vehicle subsystem. Then, the function, structure, controlling model and using method of the robot vehicle subsystem are introduced. The structure and familiar problems of the wireless communication subsystem are analyzed. The working flow and operation method of Mirosot vision subsystem are expatiated. And some design methods of the decision-making subsystem are introduced.Secondly, as the important component of the soccer robot system, the decision-making subsystem, the design method of it is discussed. From bottom to top, the paper gives the denotation means of the entity, the design method of the action, the role of the robot and the order that all robots behave, and describe the process of the design.Thirdly, the real-time intelligent technologies that can be used in the system are discussed. The paper gives the methods of the action design using ANN, avoiding the obstacles using EA and making decision using decision-tree.At last, several methods applied to the problem of path-planning is provided. They are simulated and analyzed. The best one is selected and applied the de domestic and international games in 2003.