高效穿越人群:一种受自然启发的交通规则
Moving efficiently through a crowd: a nature inspired traffic rule
AI总结:
本文提出一种受流场启发的偶极交通规则,使惰性智能体通过小幅移动为精英智能体让路,并通过二维周期域模拟刻画了自由流动、合作运动、冻结和集体漂移四种参数区域,验证了其在受限环境机器人群体中的应用潜力。
摘要:
在本文中,我们提出了一种受自然启发的交通规则,该规则有助于精英智能体高效地穿过由惰性智能体组成的人群。当一个物体在流体介质中游动,或一个侵入体被强行推过颗粒物质时,其周围会产生特征流场。我们表明,如果惰性智能体根据从这些特征流场导出的交通规则进行小幅移动,它们就能高效地重新组织并腾出足够空间,使精英智能体得以通过。本文所使用的交通规则是一个偶极场,它能够令人满意地捕捉运动侵入体周围流场的特征。我们通过在二维周期性域中进行数值模拟来研究这种偶极交通规则的有效性,在该模拟中,一个自推进的精英智能体试图穿过一群倾向于保持静止状态的惰性智能体。我们在广泛的交通规则强度和人群堆积密度范围内进行了模拟。我们对参数空间中的四个区域进行了刻画和分析,即自由流动区、合作运动区、冻结区和集体漂移区,并结合所观察到的集体行为讨论了该交通规则的后果。我们相信,所提出的方法可用于在受限环境中工作的机器人群体。
英文摘要:
In this article, we propose a traffic rule inspired from nature, that facilitates an elite agent to move efficiently through a crowd of inert agents. When an object swims in a fluid medium or an intruder is forced through granular matter, characteristic flow-fields are created around them. We show that if inert agents, made small movements based on a traffic rule derived from these characteristic flow-fields, they efficiently reorganize and transport enough space for the elite to pass through. The traffic rule used in the article is a dipole-field which satisfactorily captures the features of the flow-fields around a moving intruder. We study the effectiveness of this dipole traffic rule using numerical simulations in a 2D periodic domain, where one self-propelled elite agent tries to move through a crowd of inert agents that prefer to stay in a state of rest. Simulations are carried out for a wide range of strengths of the traffic rule and packing density of the crowd. We characterize and analyze four regions in the parameter space, free-flow, motion due to cooperation, frozen and collective drift regions, and discuss the consequence of the traffic rule in light of the collective behavior observed. We believe that the proposed method can be of use in a swarm of robots working in constrained environments.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。