一种基于可传递置信模型的分布智能体决策融合方法  

New method to distributed agent decision fusion based on transferable belief model

作　　者：范波[1] 普杰信[1] 刘刚[1] 

机构地区：[1]河南科技大学电子信息工程学院,河南洛阳471003

出　　处：《计算机应用研究》2010年第2期443-445,449,共4页Application Research of Computers

基　　金：国家自然科学基金资助项目(60475021)

摘　　要：在分析与研究对抗性多机器人系统决策问题的基础上,提出了一种基于可传递置信模型的多智能体决策融合方法;构建了决策融合体系架构,分别设计了基于证据推理的观测智能体模型,基于TBM的决策智能体模型以及决策融合中心模型,给出了相应的算法。通过在机器人足球中的应用及仿真实验,体现了本方法在对抗性多机器人系统中决策制定的良好性能及效果。With the analysis and study of the decision in adversarial multi-robot system, this paper proposed a method to multi-agent decision fusion based on TBM. Built the architecture of decision fusion, in which designed observing agent model based on evidential reasoning, decision agent model based on TBM and decision fusion center model. Also, provided the associative algorithms. With the application and experiment in robot soccer, it is shown that this method has better performance and effectiveness for making decision in adversarial multi-robot system.

关 键 词：多智能体决策 信息融合 可传递置信模型 证据推理 机器人足球 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程] TP242.6[自动化与计算机技术—控制科学与工程]