多模态融合的多机器人任务分配算法研究  

A New Multi-Robot Task Allocation Algorithm Based on Multi-Modality Synthesis

作　　者：段俊花[1] 朱怡安[1] 黄姝娟[1,2] 陆伟[1,2] 

机构地区：[1]西北工业大学计算机学院,陕西西安710072 [2]西北工业大学软件与微电子学院,陕西西安710072

出　　处：《西北工业大学学报》2013年第6期974-978,共5页Journal of Northwestern Polytechnical University

基　　金：航天科技创新基金(2011XR160001);航空科学创新基金(20130753006);西北工业大学基础研究基金(JC20110283)资助

摘　　要：针对分布式任务分配中的资源竞争冲突问题以及解的质量和求解代价难以协调的问题,提出了一种基于多模态融合的多机器人分布式自组织任务分配算法。它将基于信息素的隐式交互机制、基于视觉的阈值模型任务分配机制和基于通信的显式协商协调机制等相结合,实现了未知环境下多机器人系统的多模态分布式自组织任务分配。仿真实验结果表明,多模态融合技术扩大了机器人的感知范围,提高了系统可靠性,而且在较少通信的情况下,有效解决了任务分配冲突问题,提高了任务执行效率。In distributed task allocation , it is hard to balance three problems：conflict, solution quality and solving price .Multimodal synthesis method is put forward for multi-robot distributed and self-organized task allocation . Three models of cooperation mechanism are combined; these three models are conflict problem , which is implicit cooperation mechanism based on pheromone , implicit task allocation realized by threshold method based on robot visual sense , and explicit task allocation based on improved contract net mechanism .The experiment results and their analysis show preliminarily that multimodal synthesis method can expand the scope of the robot perception , improve the system reliability , resolve the conflict problem by less communication , and improve the task execution efficiency .

关 键 词：分布式任务分配 多机器人系统 多模态 

分 类 号：TP302[自动化与计算机技术—计算机系统结构]