群机器人多目标搜索中带闭环调节的动态任务分工

Dynamic Task Allocation with Closed-Loop Adjusting in Swarm Robotic Search for Multiple Targets

群机器人在并行化地同时搜索多个目标时,须通过任务分工形成若干个子群联盟,每个子群分别针对一个确定性目标协同搜索.非结构化环境中搜索对象和搜索主体的变化,要求任务分工动态进行.故在基于响应阈值分工模型基础上,提出一种带闭环调节的动态分工方法.根据机器人或者通过检测目标信号、或者通过邻域通信获得目标认知的特点,将元任务分类并构造个性化任务集,基于概率原则评估针对目标激励的响应并自主选取意向目标.具有共同意向的机器人自组织地缔结子群联盟.然后用平均距离法度量子群联盟内的机器人资源配置水平,作为负反馈引入任务分工模型,调节机器人在不同子群联盟间动态迁移.仿真结果表明,使用本文方法对群机器人动态分工后进行协同搜索,搜索效率较现有方法有显著提高.

To search for multiple targets in parallel and simultaneously with swarm robots, those robots should be divided into some sub-swarm-coalitions by job allocation, so that each sub-swarm can cooperatively work focusing on its desired target. Due to varying of search subjects and search objects in unstructured environment, dynamic task allocation is required. Thus a dynamic allocation method with closed-loop adjusting based on the existing response threshold allocation model is proposed. First, a personalized task set is constructed including some meta tasks for each robot, according to cognition receiving by means of either detected signals （I-type target, direct） or informed information of target through neighborhood communications （//-type target, indirect）. Then responses to target stimuli are evaluated based on probability principle, and a desired target is selected independently from task set. Those robots having the same desired target confederate independently with each other to work cooperatively. Then, the level of robot resource configuration is measured with an average distance approach and the result is linked into allocation model as a negative feedback. This mechanism can makes dynamic migration occurred among different sub-swarm-coalitions on demand. Results from simulations reveal that overall search efficiency is improved by using our proposed method, comparing with the existing methods.