CAREER: Resilient Low-Cost Robot Teams for Autonomous Aquatic Exploration

职业：用于自主水生探索的有弹性的低成本机器人团队

• 批准号: 2144624
• 负责人: Alberto Quattrini Li
• 金额: $ 55.37万
• 依托单位: Dartmouth College
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2022
• 资助国家: 美国
• 起止时间: 2022-03-01 至 2027-02-28
• 项目状态: 未结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2144624&HistoricalAwards=false
• 关键词: CAREER; Resilient; Low; Cost; Robot

This Faculty Early Career Development (CAREER) project’s main objective is to study and develop low-cost multirobot systems for aquatic environments exploration, towards democratizing aquatic autonomous robotics. With water covering 70 percent of the Earth, a large part of the economy, called “Blue Economy” and valued to be at least US$24 trillion, relies on a healthy aquatic world, requiring its study and monitoring. Robots could automate these tasks. However, to date, aquatic robots deployed in practice are expensive (in the order of US$100k-US$1M). Current multi-robot exploration algorithms are not robust underwater, as the intrinsic limitations posed by the aquatic domain and inexpensive robot configuration are not explicitly considered – such as absence of global localization and communication infrastructure (e.g., GPS and cellular network) and limited communication bandwidth (bits per seconds). This project will produce resilient cooperative exploration algorithms for teams of inexpensive Autonomous Surface and Underwater Vehicles, that explicitly consider those limitations. This project outcomes will advance the state of the art on low-cost mobile robot autonomy in extreme environments. More broadly, the project will contribute to lower the barrier to entry in aquatic robotics and catalyze a broader educational outreach and support tasks for a healthy aquatic world. This project integrates research activities with educational and outreach plans; example include targeting environmental monitoring tasks with real deployments, organizing a summer camp for K-12 students, and involving undergraduate/graduate students in research. The project will intellectually contribute on novel algorithmic developments to three fundamental robotics research themes and on a comprehensive system integration with real low-cost robots in real aquatic environments: (1) resilient cooperative multirobot 3D exploration by probabilistic modeling of localization and communication constraints, optimizing multiple conflicting objectives, and reasoning in a probabilistic representation to find a safe optimal path; (2) resilient communication by tightly coupling state estimation and planning to calculate information value and designing an optimization framework to minimize the use of the low-bandwidth communication channel, but ensure situational awareness; (3) graceful recovery by designing algorithms that calculate a non-myopic rendezvous trajectory based on explicit modeling of localization and communication and allow for adaptive coalition formation to avoid loss of the single robot and of the whole system; (4) integration of the algorithms with a team of inexpensive custom-made ASVs and AUVs, which will be rigorously tested in simulations, real pool and lake/ocean environments. The research expected outcomes are new algorithms for low-cost aquatic robots that will augment their exploration capabilities and allow simpler deployments.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.

该学院早期职业发展(CALEAR)项目的主要目标是研究和开发用于水环境探索的低成本多机器人系统，走向水上自主机器人的民主化。由于水覆盖了地球的70%，被称为“蓝色经济”的经济的很大一部分依赖于一个健康的水生世界，需要对其进行研究和监测。机器人可以自动完成这些任务。然而，到目前为止，部署在实践中的水上机器人价格昂贵(大约在10万美元到100万美元之间)。目前的多机器人探测算法在水下并不健壮，因为没有明确考虑水上领域和廉价机器人配置带来的内在限制--例如缺乏全球定位和通信基础设施(例如，GPS和蜂窝网络)和有限的通信带宽(比特/秒)。该项目将为廉价的自主水面和水下机器人团队产生弹性合作探索算法，明确考虑这些限制。该项目的成果将推动低成本移动机器人在极端环境中自主的技术水平。更广泛地说，该项目将有助于降低进入水上机器人的门槛，并促进更广泛的教育推广和支持健康的水上世界的任务。该项目将研究活动与教育和推广计划结合起来；例如，将环境监测任务与实际部署相结合，为K-12学生组织夏令营，并让本科生/研究生参与研究。该项目将在智力上对三个基本机器人研究主题的新算法开发以及与真实水上环境中真实低成本机器人的全面系统集成做出贡献：(1)弹性协作多机器人3D探索，通过对定位和通信约束的概率建模，优化多个冲突目标，并在概率表示中进行推理，以找到安全的最优路径；(2)弹性通信，通过紧密耦合状态估计和规划来计算信息价值，并设计优化框架，以最大限度地减少对低带宽通信信道的使用，但确保态势感知；(3)通过设计算法来优雅地恢复，该算法基于定位和通信的显式建模来计算非近视交会轨迹，并允许自适应联盟形成，以避免单个机器人和整个系统的损失；(4)将算法与一组廉价的定制ASV和AUV集成，这些算法将在模拟、真实池塘和湖泊/海洋环境中进行严格测试。这项研究的预期成果是低成本水上机器人的新算法，它将增强它们的探索能力，并允许更简单的部署。该项目由跨部门机器人基础研究计划支持，该计划由工程总监(ENG)和计算机和信息科学与工程(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。

项目成果

Sunflower: locating underwater robots from the air

• DOI: 10.1145/3498361.3539773
• 发表时间: 2022-06
• 期刊: Proceedings of the 20th Annual International Conference on Mobile Systems, Applications and Services
• 作者: Charles J. Carver;Qijia Shao;Samuel Lensgraf;A. Sniffen;Maxine Perroni-Scharf;Hunter Gallant;Alberto Quattrini Li;Xia Zhou

MARCOL: A Maritime Collision Avoidance Decision-Making Testbed

MARCOL：海上避碰决策测试平台

• DOI: 10.1609/aaai.v37i13.27076
• 发表时间: 2023
• 期刊: Proceedings of the AAAI Conference on Artificial Intelligence
• 作者: Jeong, Mingi;Quattrini Li, Alberto
• 通讯作者: Quattrini Li, Alberto

Underwater Exploration and Mapping

水下勘探与测绘

• DOI: 10.1109/auv53081.2022.9965805
• 发表时间: 2022
• 期刊: IEEE
• 作者: Joshi, Bharat;Xanthidis, Marios;Roznere, Monika;Burgdorfer, Nathaniel J.;Mordohai, Philippos;Quattrini Li, Alberto;Rekleitis, Ioannis
• 通讯作者: Rekleitis, Ioannis

Monocular Camera and Single-Beam Sonar-Based Underwater Collision-Free Navigation with Domain Randomization

基于单目相机和单波束声纳的水下无碰撞导航与域随机化

• 发表时间: 2022
• 期刊: International Symposium on Robotics Research (ISRR
• 作者: Yang, P.;Liu, H.;Roznere, M.;Quattrini Li, A.
• 通讯作者: Quattrini Li, A.

Deep Underwater Monocular Depth Estimation with Single-Beam Echosounder

使用单波束回声测深仪进行深水下单目深度估计

• DOI: 10.1109/icra48891.2023.10161439
• 发表时间: 2023
• 期刊: IEEE
• 作者: Liu, Haowen;Roznere, Monika;Quattrini Li, Alberto
• 通讯作者: Quattrini Li, Alberto