Abstract
Foreign Object Debris (FOD) can create significant safety implications for aircraft and personnel and is a continuous concern in the airport environment. The FAA Airport Technology Research and Development Branch conducted a research effort to explore the feasibility and maturity of using commercially available small unmanned aircraft systems (sUAS) and artificial intelligence/machine learning (AI/ML) algorithms to detect FOD on airport surfaces. The objectives of this research effort were to develop a novel, proof-of-concept sUAS-based FOD detection workflow using AI/ML algorithms and to assess the workflow to determine whether it is capable of meeting all, some, or none of the requirements in FAA Advisory Circular (AC) 150/5220-24, Airport Foreign Object Debris Detection Equipment.
The research team conducted initial testing of this workflow at Cape May County Airport and validation testing at Atlantic City International Airport. The sUAS-based FOD detection workflow, which used the FastFlow ML deep learning algorithm, was capable of meeting some of the AC 150/5220-24 requirements, including achieving a 96% detection rate for FOD items specified in the AC. However, further research and development will be needed for this technology to meet the full set of AC 150/5220-24 requirements, including reducing the false positive rate, reducing the data processing time, implementing a software interface for displaying and recording FOD detection alerts, and detecting FOD in low-light and inclement weather conditions.