RoboCupRescue Agent仿真中的多智能体协作方法  

Multi-Agent Coordination in RoboCupRescue Agent Simulation

作　　者：黄小雨[1,2] 谈英姿[1,2] 许映秋[1,3] 孟庆法[1,3] 陈科[1,3] 管大琦[1,3] 

机构地区：[1]东南大学RoboCup机器人训练基地,南京210096 [2]东南大学自动化学院,南京210096 [3]东南大学机械工程学院,南京210096

出　　处：《系统仿真学报》2009年第11期3288-3294,共7页Journal of System Simulation

基　　金：国家大学生创新性实验计划项目(081028618)

摘　　要：对RoboCupRescue智能体仿真系统内的多智能体协作问题进行了研究和分析,提出一种由多种评价模型和协作方式组成的协作方法,并从同类和异类智能体两个方面给出了具体的实现。该实现主要采用了人工免疫控制、静态分配与动态调整结合以及基于共识的分布式控制等方式。在仿真测试中,各类智能体通过相互协作有效完成了灾难环境中的救援任务。The technology of multi-agent coordination in the RoboCupRescue Agent Simulation System was seriously analyzed. Then an approach consisting of several evaluation models and cooperation methods was put forward, of which the implement was described in detail fxom two aspects： coordination among the same kind agents and among the different kind agents. Some useful coordination methods were adopted synthetically for the implement, such as the Artificial Immune Control System, the Dynamic Adjustment System with static assignment and the Consensus based Distributed Control System. During the simulation, different agents complete the rescue tasks effectively by coordination in the disaster space.

关 键 词：RoboCupRescue 智能体 多智能体系统 协作 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]