分层模型下的多Agent系统协作机制在RoboCup3D的应用  

Layered model based multi-agent system collaboration mechanism and its application in RoboCup 3D

作　　者：杭婷婷[1] 方木云[2] 

机构地区：[1]河海大学文天学院电气信息工程系,安徽马鞍山243031 [2]安徽工业大学计算机学院,安徽马鞍山243002

出　　处：《安庆师范大学学报（自然科学版）》2017年第2期72-75,共4页Journal of Anqing Normal University(Natural Science Edition)

基　　金：安徽省教学研究项目(2016jyxm0912);河海大学文天学院科研项目(WT16008)

摘　　要：为了提高多智能体协作的效率,让机器人足球队取得更好的成绩,以RoboCup3D机器人足球仿真系统为应用背景,提出了一种基于分层模型的多Agent系统协作机制。分层模型主要包括用来保证Agent通信的通信层,完成Agent角色分配的技术层和负责Agent阵型定位等决策的控制层。实验结果表明,团队协作能力往往是机器人足球比赛得分的关键,基于分层模型的协作机制对于提高机器人球队的整体对抗能力起到了明显作用。To improve the Multi-Agent collaboration efficiency, and let robot soccer team to achieve better results, we take the RoboCup3D robot soccer simulation system as the application background and propose a layered model based Multi- Agent system collaboration mechanism. Layered Model includes three layers： Connection Layer, Technique Layer and Controlling Layer. Connection Layer aims to ensure the communication of Agent; Technique Layer is to complete the role assignment of Agent; Controlling Layer is mainly responsible for the formation locate problem etc. The results show that team collaboration ability usually is the key factor of achieving scores in robot soccer competition. The collaboration mechanism based on layered model plays an important role in improving robot Soccer team＇ s whole Resistance ability.

关 键 词：人工智能 多智能体协作 分层模型 ROBOCUP3D 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]