仿真机器人足球学习方法研究综述  

Overview of Robotsoccer Learning Method

作　　者：童亮[1] 陆际联[1] 

机构地区：[1]北京理工大学机器人研究中心,北京100081

出　　处：《计算机仿真》2004年第6期1-5,共5页Computer Simulation

摘　　要：仿真机器人足球赛是近几年在国际上迅速开展起来的高技术对抗活动 ,虽然历史不长 ,但由于它集高新技术、娱乐比赛于一体 ,所以引起了人们的广泛关注和极大兴趣。作为多智能体系统研究的重要手段 ,许多研究者从不同的侧面对该项技术进行了研究并取得了一定的成果。对仿真机器人足球系统的研究 ,目前包括系统组织结构设计、多智能体结构及协调机制研究、智能体技能学习和对手模型预测等内容。In recent years,simulating robotsoccer competition has been developped fastly as a kind of adversary high-tech action.Although its history is short,it has attracted many interests of people widely because of combination of high-tech and entertainment.As an important means of multiagent system research,many researchers have studied it from different perspectives and much progress has been achieved.Today,many researchers focus on the design of system construction,research of multiagent construction and method of cooperation,learning of technology of the agent and prediction of opponent's model.In this paper,we introduce and comment on the robotsoccer learning methods from the view of MAS learning and cooperation.

关 键 词：仿真机器人足球赛 多智能体系统 系统组织结构设计 协调机制 手模型预测 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]