CAREER: Environmentally-Mediated Coordination in Natural and Robot Swarms

从仓库到娱乐灯光秀，数以千计的机器人和无人机同时操作，自主机器人群正变得无处不在。这场技术革命使得将目光投向并行工作的机器人之外，着眼于真正的群体智能是至关重要的，因为在群体智能中，整体超过了各个部分的总和。作为概念的证明，自然群通过在空间和时间上将信息集成和传播到它们的共享环境中，表现出可扩展、容错和自适应的特性。在机器人群中采用这些概念可以补充现有的控制体系结构，并可能导致系统的效率低于使用集中式控制体系结构的系统，但部署起来要快得多，成本也低得多，对单个故障具有弹性，更容易从初始部署调整到全面部署，并可以适应不断变化的任务或环境。虽然越来越受欢迎，但这种分布式协调有很多方面，人们仍然知之甚少--特别是作为一种旨在实现生物水平的复原力和适应性的工程蜂群的设计工具。为了解决这些不足，这个学院早期职业发展(CALEAR)研究项目将扩展环境中介协调的概念，从工作在静态环境中的完美机器人，到包括动态环境和处理错误和硬件可靠性的现实界限的方法。这项工作将产生一个动态环境中的蜂群模型，该模型可以整合、扩散、衰减和过滤来自生物模型系统的特征的信息，以及实际的机器人实验。实际验证包括集体机器人建造，蜜蜂跟踪气味羽流，以及可编程物质中的应变调节行为。这个项目还包括通过新的方法确保和增加如此多样化的劳动力，包括包容性的、共享的在线机器人课程；面向公众的跨代扩展计划；跨学科的学生项目；以及为机器人学、生物学和建筑学领域的研究人员举办的研讨会。该项目由跨部门机器人基础研究计划支持，该计划由工程指导委员会(ENG)和计算机和信息科学与工程指导委员会(CEISE)共同管理和资助。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。

Autonomous robot swarms are becoming ubiquitous with thousands of robots and drones operating simultaneously in places from warehouses to entertainment light shows. This technological revolution makes it critical to look beyond robots working in parallel, and towards actual swarm intelligence where the whole is more than the sum of its parts. As a proof of concept, natural swarms exhibit scalable, error tolerant, and adaptive properties by integrating and propagation information into their shared environment over space and time. Adoption of these concepts in robot swarms can complement existing control architectures, and may result in systems that are less efficient than those with centralized control architectures, but are much faster and inexpensive to deploy, are resilient to individual failures, resize easier from initial to full-scale deployment, and can adapt to changing tasks or environments. Although gaining in popularity, this type of distributed coordination has many facets and is still poorly understood - especially as a design tool for engineered swarms that aim to achieve biological levels of resilience and adaptability. To address these shortcomings, this Faculty Early Career Development (CAREER) research project will extend upon the concept of environmentally-mediated coordination from working on perfect robots operating in static environments, to include dynamic environments and ways to deal with realistic bounds on error and hardware reliability. The work will result in a model of swarms in dynamic environments that act to integrate, diffuse, decay, and filter information derived from characterization of a biological model system, as well as practical robot experiments. Practical verification involves collective robotic construction, odor plume tracking by honey bees, and strain-mediated behaviors in programmable matter. This project further involves aims to secure and increase such a diverse workforce through novel methods for inclusive, shared, online robotics curricula; cross-generational outreach programs for the public; interdisciplinary student projects; and workshops for researchers across the fields of robotics, biology, and architecture.This project is supported by the cross-directorate Foundational Research in Robotics program, jointly managed and funded by the Directorates for Engineering (ENG) and Computer and Information Science and Engineering (CISE).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.