CAREER: Transferring biological networks emergent principles to drone swarm collaborative algorithms

考虑搜救行动或大规模环境监测等场景，无人机必须自主导航，适应动态障碍，并协作优化其行动。在网络物理系统领域，在争取分散决策时，挑战加剧。这些算法的成功开发可以为对社会福利至关重要的应用程序产生潜在的变革性范例。值得注意的是，自然界中的生物，如黏菌和真菌，能够发展分散的、协调的网络，比工程师更好地优化交通，解决迷宫，探测远处的质量，甚至记住周期性事件。这种方法在应对动态环境条件和确保随着种群规模的扩大而可伸缩性方面至关重要。这项研究旨在通过利用先进的计算框架来模拟生物网络的功能，为无人机群开发新的协作组织算法。我们的战略包括将集体行为洞察从形成网络的有机体转移到为个别无人机制定规则。该提案解决了群体协作算法中的关键知识空白，侧重于理解指导从微观尺度到宏观尺度的群体行为转变的原则这一科学挑战。在工程方面，它的目标是开发健壮的机器学习程序，以准确地将观察到的行为传输到合成系统，并增强超级计算能力，以提高可扩展性。其新奇之处在于采用了自下而上的生物学观点，将展示涌现的模拟数据映射到无人机群的计算和通信限制。此外，这个项目填补了公众对工程中群体协调和紧急行为理解方面的关键教育空白。该倡议的开源工具旨在加快群体力学的基础研究，并加强各级STEM教育。该项目的重点是将复杂的概念翻译成日常语言，通过专门的课程和虚拟体验(如《蜂群任务》)影响下一代STEM工程师，并以普通公众为目标，针对最年轻的观众举办展览，如《你比黏菌更聪明吗？》。该计划包括指导宾夕法尼亚州的一名高中教师、宾夕法尼亚州立大学的研究生和本科生，以及创建关于网络物理系统中涌现概念的免费教学材料。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

Consider scenarios like search and rescue operations or large-scale environmental monitoring where drones must autonomously navigate, adapt to dynamic obstacles, and collaboratively optimize their actions. In the field of cyberphysical systems, the challenge intensifies when striving for decentralized decision-making. The successful development of these algorithms can produce potentially transformative paradigms for applications critical to societal welfare. Remarkably, organisms in nature, such as slime molds and fungi, are able to develop decentralized, coordinated networks that optimize transport better than engineers, solve mazes, detect masses at a distance, or even memorize periodic events. This approach is pivotal in addressing dynamic environmental conditions and ensuring scalability as the swarm size expands. This research aims to develop novel collaborative organization algorithms for drone swarms by leveraging advanced computational frameworks that mimic the functionalities of biological networks. Our strategy involves transferring collective behavior insights from network-forming organisms to formulate rules for individual drones. The proposal addresses critical knowledge gaps in swarm collaborative algorithms, focusing on the scientific challenge of understanding the principles guiding the transition from microscale to macroscale swarm behavior. On the engineering front, it aims to develop robust machine learning procedures for accurately transferring observed behaviors to synthetic systems and enhance supercomputing capabilities for improved scalability. The novelty lies in adopting a bottom-up biological perspective, mapping simulation data showcasing emergence to the computational and communication constraints of a drone swarm. Additionally, this project fills critical educational gaps in the public understanding of swarm coordination and emergent behavior in engineering. The initiative's open-source tools aim to accelerate basic research in swarm mechanics and enhance STEM education at various levels. With a focus on translating complex concepts into everyday language, the project impacts the next generation of STEM engineers through specialized courses, virtual experiences like "Swarm Quest", and also targets the general public with exhibitions like "Are you smarter than a slime mold?", aimed at the youngest audience. The initiative involves mentoring a high-school teacher in Pennsylvania, graduate and undergraduate students at Penn State, and the creation of free instructional material on the concept of emergence in cyberphysical systems.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.