当前位置: X-MOL 学术Bioinspir. Biomim. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
A swarm design paradigm unifying swarm behaviors using minimalistic communication.
Bioinspiration & Biomimetics ( IF 3.4 ) Pub Date : 2020-03-05 , DOI: 10.1088/1748-3190/ab6ed9
Joshua Cherian Varughese 1 , Hannes Hornischer , Payam Zahadat , Ronald Thenius , Franz Wotawa , Thomas Schmickl
Affiliation  

Numerous nature inspired algorithms have been suggested to enable robotic swarms, mobile sensor networks and other multi-agent systems to exhibit various self-organized behaviors. Swarm intelligence and swarm robotics research have been underway for a few decades and have produced many such algorithms based on natural self-organizing systems. While a large body of research exists for variations and modifications in swarm intelligence algorithms, there have been few attempts to unify the underlying agent level design of these widely varying behaviors. In this work, a design paradigm for a swarm of agents is presented which can exhibit a wide range of collective behaviors at swarm level while using minimalistic single-bit communication at the agent level. The communication in the proposed paradigm is based on waves of "ping''-signals inspired by strategies for communication and self organization of slime mold (dictyostelium discoideum) and fireflies (lampyridae). The unification of common collective behaviors through this Wave Oriented Swarm Paradigm (WOSP) enables the control of swarms with minimalistic communication and yet allowing the emergence of diverse complex behaviors. It is demonstrated both in simulation and using real robotic experiments that even a single-bit communication channel between agents suffices for the design of a substantial set of behaviors. Ultimately, the reader will be enabled to combine different behaviours based on the paradigm to develop a control scheme for individual swarms.

中文翻译:

群体设计范式使用简约通信统一群体行为。

已经提出了许多受自然启发的算法,以使机器人群,移动传感器网络和其他多智能体系统表现出各种自组织行为。群体智能和群体机器人技术的研究已经进行了几十年,并且已经基于自然的自组织系统产生了许多这样的算法。尽管存在大量关于群体智能算法的变化和修改的研究,但很少有尝试统一这些广泛变化的行为的底层代理级别设计。在这项工作中,提出了一种用于群体代理的设计范例,该模型可以在群体层次上表现出广泛的集体行为,同时在代理层次上使用简约的单比特通信。拟议范式中的交流基于“ ping”波 -信号受粘液霉菌(dictyostelium discoideum)和萤火虫(扁豆科)的交流和自我组织策略的启发。通过这种基于波群的范式(WOSP)来统一常见的集体行为,可以通过最少的通信来控制群体,但也允许出现各种复杂的行为。无论是在仿真中还是在实际的机器人实验中都证明了,即使代理之间的单个位通信通道也足以设计大量的行为。最终,将使读者能够根据范例组合不同的行为,从而为单个群体制定控制方案。通过这种基于波群的范式(WOSP)来统一常见的集体行为,可以通过最少的通信来控制群体,但也允许出现各种复杂的行为。在仿真和使用实际的机器人实验中都证明了,即使代理之间的单个位通信通道也足以设计大量的行为。最终,将使读者能够根据范例组合不同的行为,从而为单个群体制定控制方案。通过这种基于波群的范式(WOSP)来统一常见的集体行为,可以通过最少的通信来控制群体,但也允许出现各种复杂的行为。在仿真和使用实际的机器人实验中都证明了,即使代理之间的单个位通信通道也足以设计大量的行为。最终,将使读者能够根据范例组合不同的行为,从而为单个群体制定控制方案。在仿真和使用实际的机器人实验中都证明了,即使代理之间的单个位通信通道也足以设计大量的行为。最终,将使读者能够根据范例组合不同的行为,从而为单个群体制定控制方案。在仿真和使用实际的机器人实验中都证明了,即使代理之间的单个位通信通道也足以设计大量的行为。最终,将使读者能够根据范例组合不同的行为,从而为单个群体制定控制方案。
更新日期:2020-03-07
down
wechat
bug