Robot Soccer game is an international and high-tech oppositional activity t hat develops rapidly in recent years. It's a basic research project in the domain of artificial intelligence and robot, and also a great challenging and high-tech intensive project. Aim at the decision-making subsystem in a soccer robot system, the valid intelligent reasoning methods are discussed in this thesis. Soccer robot system is a typical multi-agent system. It is involved in robotics, computer vision, pattern recognition, multi-agent system, path planning, self-organization and self-learning theory etc.The Robot Soccer system is composed of 4-subsystem: vision subsystem, command subsystem, communication subsystem, and robots. Command subsystem is the core of the whole system. It should be able to pick up the knowledge on its own, and decide robots-cooperation-task, which make the whole system smart. Soccer robot system is a dynamic, uncertain andReal-time platform, and in such an adversarial and competitive environment, how to realize the real-time robot path-planning is a challenge problem. The main contents of this thesis are as follows:1. Investigated the current research status and main contents of soccer robots, according to the domestic and foreign relevant references. At the same time, gave the detailed introduction of current research status of soccer robot system, path planning and formation control.2. Introduced some kinds of theories and algorithms of path planning, many researchers apply various technologies to the soccer robot path-planning, such as artificial potential field, grads modeling method, evolutionary algorithms and artificial neural network. However, the application of these theories is not perfect in high dynamic and real-time environment, and some more effective algorithms should be investigated. The author discusses their feasibility in real-system. Finally, I choose the artificial potential field method as my study mission. 3. Through the study of formation control of soccer robots, explained the problem of coordination and cooperation in soccer robots path planning. Presented an observer based formation control algorithm for soccer robots.4. Built the simulated environment for soccer robots formation control with the robotsoccer software and compiled lingo program to control the simulated soccer robots.In this paper, we focus on the model of the decision-making subsystem, and analyze the structure of some system models. A path-planning method in dynamic surroundings which was based on imp roved artificial potential field was proposed in this paper. Because the exact movement information was hard to acquire, it is difficult to finish path planning of soccer robots in dynamic world. Searching the moving goal by gradient approximation method, the unreachable problem is solved. The results of simulation have verified that the method can solve the problem effectively, even though there are stochastic moving obstacles in the surroundings. |