Atmosphere and ocean data improve our understanding of the climate, but nature-based sampling can be expensive. Tools like weather balloons and ocean buoys are low-cost solutions, but they are fundamentally underactuated—meaning, they rely on underlying flows in the water or air to navigate. This project will develop planning and learning algorithms for measuring uncertain flows with underactuated agents, enabling active data gathering for critical regions. The project will also serve to deepen a partnership between researchers at Cornell and WindBorne Systems, an advanced weather balloon start-up, and to initiate a relationship with the Naval Research Laboratory

Investigators: Sarah Dean, Cornell Bowers Computing and Information Science; Alexander Terenin, Center for Data Science for Enterprise and Society