Federating Planning of Observations for Earth Science

We present an adaptive sensing method for federated planning observations of Earth science phenomena. Adaptive sensing is a technique that assists in studying these events by utilizing online analysis of the phenomena to determine the locations most advantageous for sensing. We use a federated planning framework to select observation requests for a set of heterogeneous assets and submit those requests to specific asset planners to produce detailed plans.We develop and demonstrate two constraint optimization algorithms to maximize the utility of the observation requests while following operational constraints of the assets. We show that adaptive sensing with federated planning is capable of collecting observations of significantly better quality than a baseline sampling approach.