面向多机器人动态任务分配的事件驱动免疫网络算法  

Event-driven immune network algorithm for multi-robot dynamic task allocation

作　　者：曹鹏飞[1,2] 郝矿荣[1,2] 丁永生[1,2] CAO Pengfei;HAO Kuangrong;DING Yongsheng(Engineering Research Center of Digitized Textile&Apparel Technology,Shanghai 201620,China;Ministry of EducationCollege of Information Science and Technology Donghua University)

机构地区：[1]东华大学信息科学与技术学院,上海201620 [2]数字化纺织服装技术教育部工程研究中心,上海201602

出　　处：《智能系统学报》2018年第6期952-958,共7页CAAI Transactions on Intelligent Systems

基　　金：上海市科学技术委员会国际合作项目(16510711100);国家自然科学基金项目(61503075;61473078);教育部长江学者奖励计划(2015-2019)

摘　　要：为实现多机器人系统的动态任务分配与协作,提出了一种面向多机器人动态任务分配的事件驱动免疫网络算法。将生物免疫网络的工作机理应用到多机器人动态任务分配算法中,借鉴Jerne的独特型免疫网络假说和Farmer提出的抗体激励动态方程,设计了多机器人任务分配与自主协作模型;基于事件驱动机制,设计了多机器人动态任务分配算法,并引入焦躁模型来解决任务死锁问题。仿真和实际多机器人系统实验结果表明,基于本文算法的多机器人系统在动态任务场景中具有较强的适应性和自主规划协调能力。To realize a dynamic task allocation and collaboration of multi-robot system,an event-driven immune network algorithm for multi-robot dynamic task allocation is presented.Inspired by the working mechanism of biological immune network,a multi-robot task allocation and autonomous cooperation model was established in this study,based on Jerne’s idiotypic network hypothesis and Farmer’s antibody dynamic equation.Then,a multi-robot dynamic task allocation algorithm was designed based on event-driven mechanism,and the anxiety model was used to solve the deadlock problem.Simulation and experimental results of multi-robot system show that a multi-robot system based on this algorithm has strong adaptability and independent planning and coordination in dynamic task environments.

关 键 词：免疫网络 多机器人系统 动态任务 自主协作 事件驱动 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]