Research On Path Planning Of Robocup Small Size League

With the development of science and technology and the progress of the society, more and more new events caused the people's interests. RoboCup Soccer is one of them. The football game environment is complex and changes all the time; it is a common multi-agent system. The setup of this system created the extremely good experimental platform for the artificial intelligence research, become the focus of research, and its application prospect is also very extensive.In the soccer match, who win the game is closely linked with the number of goals, and therefore, the decision-making part in the whole game plays an important position. The path planning is the core of the decision-making, is also a very challenging subject, this topic is the main research method of path planning and its implementation. There are many path planning methods and more commonly are lattice method, genetic algorithms, artificial potential field method, neural network algorithms, etc, these methods each has its advantages and disadvantages, in real-time and dynamic environment it is difficult to achieve the desired requirements, need to be further improved and optimized.This paper focuses on the small size league robot soccer of RoboCup competitions as the research object, according to one of the most key decision-making subsystem and the path planning method is studied, and finding out a highly effective route to realize the soccer robot action. The specific content of this article include the following aspects:First, an overview of robot soccer games were described in detail, the importance of the path planning problem in the whole system was analyzed,the importance of robot research background and research actuality were discussed in detail, and summarizes the purpose and significance of this paper.Secondly, four subsystems were discussed, namely, the visual subsystem, decision-making subsystem, wireless communication subsystem and the structure of robot car subsystems. And then explain the position and role of various components in the whole system, the decision-making sub-system has been emphasis discussed at length. Then several traditional soccer robot paths planning method was analyzed, and compared the advantages and disadvantages for each method.Finally, based on the research of the model of the robot car, aimed at selecting the optimal path problem, this paper presents a rapidly exploring random trees method to path planning for the car, and then through simulation software and material to verify the feasibility and effectiveness of this algorithm.The path planning algorithm as small size league of decision-making system, Botnia team participated in 2009 China's open RoboCup competition and has been achieved good result.