Automated Runway Situational Awareness and Alerting

Thrust AI, in partnership with MIT Lincoln Labs, proposes to research and develop a proof-of-concept system architecture leveraging machine learning and modern extra-visual spectra sensing to accomplish real-time air traffic control (ATC) detection, recognition, and observation of aircraft in the terminal environment. The STTR Phase I effort will determine the technical feasibility, to the maximum extent possible, by which an automated optical system may perform ATC observational tasks in place of traditional (“out-the-window”) gear confirmation, safety observer/chase ship spotting, and pattern monitoring. The solution proposed herein would provide unprecedented situational awareness—fusing optical, radio, and ADSB sensor inputs to provide the decision-maker “human-in-the-loop” with an automated, multispectral, and slewable feed of tower-controlled aircraft beyond the capability of the human eye. Sensor inputs to the computer vision model will encompass the visual, infrared, and thermal spectra to provide hyper-observation capability in low-visibility environments (night, weather, etc). Advantages over the current “out-the-window” model include automation of ATC observational tasks and sensing beyond the human eye (extra-visual spectra, resolution at far distances, and proximity to aircraft). Delivery of real-time advanced analytics furthers the AF S&T Strategic Capability of Rapid, Effective Decision-Making. Sensors clusters will be hosted on the airfield to ascertain a more complete situational picture of the runway environment—further aiding safety and accident investigation boards to prevent aviation mishaps.