基于Robocuo的智能协作防守  

Intelligent Cooperative-defense Strategy Based on Robocup Platform

作　　者：章小兵[1] 陈黎[1] 刘艳春[1] 

机构地区：[1]安徽工业大学电气信息学院,安徽马鞍山243002

出　　处：《安徽工业大学学报（自然科学版）》2011年第3期272-276,共5页Journal of Anhui University of Technology（Natural Science）

摘　　要：在机器人足球比赛中防守是一个重要的部分,由于防守状态时处于被动,而且难以对对手建立精确的模型,协作防守就成为防守策略中一个难解决的问题。采用模糊学习策略,对带球对手的防守动作进行学习,从而提高个人防守成功率。根据所在防守区域环境的分析和统计,求出对手无球队员的危险系数和带球对手防守的失败率,从而选择效能较高的协作防守策略。实验结果显示智能体的整体防守成功率得到一定提高,证明此策略的可行性。Defense is an important part in robocup.It is passive when in defend state and difficult to build a accurate model of the rival,so cooperative defense is a difficult problem in the defense strategy.The fuzzy-learning strategy is adopted to learn the defense action of the rival with ball to improve success rates of the defense.The danger coefficient of the rivals without ball and defense failure rate of the rival with ball can be worked out according to the analysis and calculation of the defend area environment,then the better efficiency cooperative defense strategy can be selected.The final experimental result shows the whole defend success rates of the agent have improved,and testify the the feasibility of this strategy.

关 键 词：机器人足球 模糊学习 危险系数 协作防守 

分 类 号：TP24[自动化与计算机技术—检测技术与自动化装置]