NRI: FND: Extending Autonomy in Seemingly Sensory-Denied Environments Applied to Underwater Robots

美国的许多沿海社区人口稠密，高度城市化，为评估城市化和气候对沿海海洋的物理和生物状态的影响提供了关键位置。有效的研究需要利用自动驾驶车辆进行长期、持续的海洋监测，以更深入地了解这些过程。该项目旨在克服理论和技术挑战，使不同的自动驾驶车辆组能够在动态和感知拒绝的环境中执行智能采样，重点是海洋领域。 这些努力的成功将改善天气预报，水下运输动力学的理解，以及水生环境中各种物理现象的建模和预测。研究人员团队将通过利用佛罗里达国际大学（少数民族服务机构）的多元化学生群体，在STEM研究项目中吸引代表性不足的群体。教育活动旨在积极影响一代年轻的技术人员，并使他们具备强大的跨学科技能。该项目旨在通过创造新的测绘，定位，导航和机器人-人类通信技术，使自主水上车辆能够在传感器拒绝的动态环境中智能采样，从而推动机器人和海洋科学。研究活动分为四项任务。第一个任务是从物理和生物模型、水深测量和缺乏世界参考模型的原位样本中开发水生环境和特征的3D表示。第二个任务开发了一个新的框架，动态映射和本地化的水生机器人进行有针对性的采样动态演变的水生功能。任务3开发了对人类友好且机器人可消化的地图表示，以促进低数据交换的知识转移。任务4包括使用模拟、实验室实验和全尺寸现场试验对所开发的方法进行实验验证。计划中的地图表示，定位和导航的新框架可以在移动的机器人系统中找到应用，而不仅仅是水下车辆，如地下和太空机器人。该奖项反映了NSF的法定使命，并被认为值得通过使用基金会的智力价值和更广泛的影响审查标准进行评估来支持。

Many coastal communities in the U.S. are densely populated and highly urbanized, presenting critical locations for assessing the effects of urbanization and climate on the coastal ocean's physical and biological states. Effective study requires long-term, persistent ocean monitoring utilizing autonomous vehicles to enable a deeper understanding of these processes. This project aims to overcome the theoretical and technical challenges to enable heterogeneous groups of autonomous vehicles to perform intelligent sampling in dynamic and sensory-denied environments, with a focus on the maritime domain. The success of these endeavors will improve weather forecasting, underwater transport dynamics understanding, and modeling and prediction of various physical phenomena in aquatic environments. The team of researchers will engage underrepresented groups in STEM in the research program by leveraging a diverse student population at Florida International University (a minority-serving institution). The educational activities aim to positively impact a generation of young technologists and equip them with a strong interdisciplinary skillset.This project aims to advance robotics and ocean sciences by creating novel mapping, localization, navigation, and robot-human communication techniques that will enable autonomous aquatic vehicles to sample intelligently in sensor-denied, dynamic environments. The research activities are divided into four tasks. The first task develops 3D representations of aquatic environments and features from physical and biological models, bathymetry, and in situ samples that are devoid of world reference models. The second task develops a new framework for dynamic mapping and localization for aquatic robots to perform targeted sampling within dynamically evolving aquatic features. Task 3 develops map representations that are both human-friendly and robot-digestible to facilitate knowledge transfer with low data exchange. Task 4 includes the experimental validation of the developed approaches using simulation, laboratory experiments, and full-scale field trials. The planned novel frameworks to map representation, localization, and navigation can find applications in mobile robot systems beyond aquatic vehicles such as subterranean and space robotics.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.