【作者】 徐昶；

【导师】 郭庆平；

【作者基本信息】 武汉理工大学 ， 计算机应用技术， 2007， 硕士

【摘要】 机器人足球是近几年发展起来的机器人研究和竞赛项目，融合了人工智能、智能控制、实时视觉、无线通讯、机电一体化等多学科知识。机器人足球本身是一个典型的智能体系统，同时也为智能体系统的理论研究和模型测试提供了一个标准的实验平台。决策系统相当于现代真实足球比赛中的教练员，直接影响比赛的成败，能够体现出机器人足球系统的智能性，对策略系统的深入研究有重大的理论意义和现实应用价值。智能化是计算机发展的必然趋势，无论是计算机控制，还是商用民用软件，都要求含有越来越高的智能因素，因此人工智能领域的研究越来越受到重视。近几年人工智能成为研究热点之一。比赛中双方队员的位置瞬息万变，要求不仅能智能地躲避对方的拦截，带球突破，还要能同队配合，依照策略行事，这就对人工智能的控制提出了更高的要求。智能控制技术为自主机器人提供比赛必须的“决策”能力，即用来保证在复杂动态环境条件下，通过“知识”及时进行决策。实际上这个决策树就是足球教练的战术库。计算机仿真系统给自主机器人提供“学习”能力。它不但给自主机器人提供开发比赛策略的环境，而且通过它能开发各种比赛策略。首先在计算机上建立自主机器人球队，球场地的模型，然后在这个模型上开发足球比赛策略，经过反复比赛，实验之后，将成功的策略移植到真正的自主机器人上进行比赛。本论文以足球机器人系统为研究背景，主要针对人工智能的决策树的算法在决策系统中的运用，探索行之有效的方法来解决实际中的问题。首先，总文介绍了足球机器人系统的体系结构以及相关技术，提出了研究机器人足球的背景意义和研究的目的。然后介绍了本文在MLS的平台下面运行的机器人足球的仿真策略，详细的介绍了该平台的使用方法以及该平台与策略程序如何交互的方式。再次介绍了实现机器人足球的基本的数据接口，以及基本的函数的实现形式。提出了利用人工智能中的决策树来实现高级策略部分，并将这些决策树应用到具体的编程实现中，使程序具有智能性。最后总结了实验的结果，分析了该算法的优缺点，提出了以后需要改进的地方。更多还原

【Abstract】 Robot soccer is the project of robot research and race at recent years. It contains a lot of knowlegde at many regions, such as artifical intelligence、intelligence control、real time vision、wireless communication、machine and electronic integer. Robot soccer itself is a typical intelligent system. Also it offers a standard platform for intelligent system both in theory research and model testing. Strategy system likes a coach in real soccer. It will decide soccer’s win and lost and realize intelligence of robot soccer. The deep study in strategy system has important theory meaning and application.Integence is the necessary trend of computer development, no matter what is computer control、commercial/demotic software. All of them will demand more and more intelligence.Therefore the research of intelligence is attached important. Artifical intelligence has become one of most popular researches at recent years. Two part of Robots’ positions change immediately at soccer. The Robot not only must evade the hold up of opposites’, but also cooperate with teammates according to strategy system. It’s high demanding of intelligence control.Intelligence control offer strategetic ability for self-determination robot. It ensure that through knowledge come to a decision in time at complex dynamic environment.In actually the strategic tree is strategies of the soccer coach. Artifical system of computer offers self-determination robot ablities of study.It not only offers environment which robor soccer use, but also develops various strategies. First of all build self-determination robot soccer and venue model on computer, and then develop strategy of robot soccer on this model. After using strategy again and again, we can replant successful strategy to the real self-determination robot soccer.The papers which based on robot soccer system, find effective methods to resolve the problem of real life through strategtic tree of artifical intelligence apply in strategy system.At first, the papers introduce system structure and correlation technologies of robot soccer and bring forword background and research purpose of robot soccer.Also the papers introduce the data interface and basic functions which realize robot soccer. To realize advance strategy use strategtic tree of artifical intelligence, and apply them in concrete codes to make programmes intelligent.At last the papers summarize the outcome of experiments and analyse advantages and disadvantages of the arithmetic.更多还原