Machine Learning and Instrument Autonomy Group

Principal Investigator

Peyman Tavallali

      POISE will develop autonomous technologies that enable coordinated, targeted, adaptive observations across multiple observing systems guided by the estimated improvement to our model-based understanding, predictive skills, and scientific understanding of highest priority Earth processes in the Earth Sciences. To achieve this vision, we will bring together adaptive model-driven observation selection via automated tasking and intelligent instruments.

      For example, tropical cyclones may manifest rapid intensification: a sudden increase of wind speeds with significant impact to planning and alert systems. Rapid intensification may be due to structures within the central eyewall (small scale) or global thermodynamic environmental states (large scale). POISE will be able to determine which potential observations would best discriminate between these hypotheses and optimally re-task smart assets considering cost and feasibility constraints. Intelligent instruments permit high-level observational requests such as “track and observe the eyewall convective structure” at a reconfigurable resolution. Alternatively, two smart instruments in a train could be coordinated, with the first detecting the eyewall and the second taking the higher resolution measurements. The resulting observations are assimilated into the model and fresh observation goals are generated.