AI has enormous potential to improve Air Force pilot training by providing actionable feedback to pilot trainees on the quality of their maneuvers and enabling instructor-less flying familiarization for early-stage trainees in low-cost simulators. Historically, AI challenges consisting of data, problem descriptions, and example code have been critical to creating AI revolutions. A joint MIT and USAF team (the USAF-MIT AI Accelerator) has developed such an AI challenge using real-world Air Force problems. The Maneuver ID challenge has assembled thousands of virtual reality simulator flight recordings collected by actual Air Force trainee pilots at Pilot Training Next (PTN). This dataset has been publicly released at Maneuver-ID.mit.edu and represents the first of its kind public release of USAF flight training data. Using this dataset we have applied a variety of AI methods to separate “good” vs “bad” simulator data as well as categorizing and characterizing maneuvers. The algorithms and software are described and are being released as baseline performance examples for others to build upon to enable the AI ecosystem for flight simulator training.
Keywords
AI,DEEP LEARNING,FLIGHT SIMULATION,FLIGHT TRAINING,MACHINE LEARNING,SIMULATIONS
Additional Keywords