多机器人追捕中带双向筛选的动态任务分配算法研究  

Research on Dynamic Task Allocation Algorithm with Bilateral Selective in Multi-robot Capture

作　　者：刘政强 陈寿元[1] 邵增珍[1,2] 张永[1,2] 刘衍民[3] 

机构地区：[1]山东师范大学信息科学与工程学院,济南250014 [2]山东省物流优化与预测工程技术研究中心,济南250014 [3]遵义师范学院数学与计算科学学院,贵州遵义563002

出　　处：《小型微型计算机系统》2017年第7期1568-1572,共5页Journal of Chinese Computer Systems

基　　金：中国博士后科学基金项目(2016M592697)资助;山东省科技发展计划项目(2014GGH201022)资助;山东省经信委软科学计划项目(2015EI010)资助;国家自然科学基金项目(71461027)资助;贵州省自然科学基金(黔教合KY[2014]295)资助

摘　　要：多机器人执行并行追捕任务时,须通过任务分配形成若干子群,每个子群针对某一目标进行协作追捕.由于环境中目标数量及位置的变化,要求任务分配能够动态进行.提出基于合同网任务分配模型的带双向筛选机制的动态任务分配方法,以减少协商过程的通信开销.机器人自主动态构建任务数据集且在追捕过程中自动更新.招标机器人和投标机器人筛选数据集中的最优信息选择结盟,且根据任务数据集信息在不同合同网联盟间动态迁移.仿真结果表明,使用该方法在不同情况下均能高效地实现防入侵追捕任务,同时可有效降低协商过程通信开销.To capture for multiple targets in parallel, Multi-robots should be divided into some subgroups by task allocation, so that each subgroups can cooperatively work focusing on its target. Due to varying of target number and location in environment, dynamic task allocation is required. A novel dynamic task allocation approach bilateral selective mechanisms based on Contract Net Protocol is creatively proposed, which focuses on reducing the communication load during the task negotiation process. Robots get information and build robotic missions database. Tenderer Robots and Bidder Robots form an alliance by electing optimal date in database, and robots update database according to task situation so as to dynamically adjust robots to move in different alliance. Simulation results show that the proposed method can effectively realize anti-intrusion capture mission in different situation, and can effectively reduce the communication overhead of negotiation process.

关 键 词：多机器人 追捕 任务分配 合同网 双向筛选 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]