ASDL Student Selected as Finalist in FAA Data Challenge
ASDL Student Selected as Finalist in FAA Data Challenge
ASDL PhD Student Jeffrey Pattison has been selected as a finalist in the FAA Data Challenge based on his abstract “A Deep Learning Approach Using Social Media Data to Estimate Ground Risk of UAS in Urban Areas”. Jeffrey is one of ten finalist teams representing eight different U.S. universities. Dr. Dimitri Mavris and Dr. Michael Balchanos served as the faculty advisors, and Pattent LLC served as a partner.
The FAA Data Challenge is sponsored by the Federal Aviation Administration’s Chief Data Office and managed by the National Institute of Aerospace. This year’s challenge focused on the use of Artificial Intelligence/Machine Learning (AI/ML) and advanced analytics to explore aviation-related problems and opportunities. The challenge sought submissions that pushed the boundary and introduced novel approaches to aviation problems as the FAA moves further towards an info-centric National Airspace System.
Finalist teams will demonstrate their solutions at the Challenge Forum in June 2023. After reviewing and evaluating the Finalists’ Technical Papers and demonstrations, the lead university of the top three teams will be awarded a $10,000 cash prize. In addition, the team earning Best Overall Submission will receive an additional $15,000.
Jeffrey’s abstract addressed the need for risk evaluation for UAS operations to safely operate over people in urban areas. With UAS being introduced into the airspace relatively recently, there is insufficient flight data for UAS to make similar risk assessments as for manned aircraft. For this reason, modeling and simulation has become the best way to estimate UAS risk. Jeffrey applied a physics-based modeling and simulation method to approximate UAS risk, which will help generate training data so an ML algorithm can approximate the UAS risk in a more time efficient manner. Future work includes incorporating the ground risk assessment into this route planning framework and allowing an option to find a route below some target level of risk in place of the binary urgency option.