Intelligent Automation, Inc. — National Aeronautics and Space Administration SBIR Phase I: A3

Description

There are a number challenges to small UAS operations in the NAS. One of the most critical of these is access to microscale, hyperlocal weather estimates and predictions.Small UAS weigh less than 55 lbs have the risk of losing control in presence of light winds and gusts. The risk is amplified in urban areas due to presence of tall building and other man made objects. The phenomenon of urban canyon can cause high winds between building that are beyond the capabilities of current meso scale weather prediction models. High demand operations such as package delivery involves the UAS to take off from fulfilment center, travel to the destination and land where the customers are and takeoff for the return journey. Therefore, the ability to predict hyperlocal weather is a critical requirement for UAS in the NAS operations. The team propose to develop GUMP, a weather service that combines the coarse estimates of mesoscale prediction from NOAA weather products with high fidelity but localized ground station data using machine learning and computation fluid dynamics simulations. The innovative approach is scalable and transferrable to urban or rural areas. The computational workload is minimized by building on weather forecast publically available. GUMP fills a critical gap in weather prediction technology that will be beneficial for NASA researchers and industry UAS operators.