FAA research project to provide guidance for air traffic control human-AI/ML interaction
A new research project funded by the Federal Aviation Administration (FAA) is designed to improve the understanding and application of human factors in air traffic control (ATC) automation.
ISE Professor Phil Smith will serve as principal investigator for the team, which includes Virginia Tech, University of Michigan and Roth Cognitive Engineering, Inc.
“Nationally and internationally, there are numerous research efforts focused on introducing new capabilities enabled by applications of artificial intelligence (AI) and machine learning (ML) into aviation-focused software tools,” Dr. Smith says. “Given such proposals to make use of these technologies within software tools developed by the FAA, the FAA Human Factors Division identified a need to provide guidance on how to effectively design such tools to ensure effective human-automation interaction.
“The Human Factors Division also concluded that there was a need to better understand how to successfully integrate these tools into the highly distributed work system that supports air traffic management from a broader systems perspective.”
The two-year grant addresses the federal compliance code focusing on the human factors aspects of automation for air traffic controllers and other FAA operational staff.
“The goal of this work is to produce a highly accessible guidance document to raise awareness of the human factors issues critical to the design, implementation and integration of AI and ML technologies to support ATC operations,” according to Dr. Smith, “and provide practical guidance on system design, implementation, integration and evaluation that will enable the FAA air traffic organization to avoid pitfalls and adhere to best practices for the integration of AI/ML technologies into ATC automation.
“This guidance document is intended to be very practical and applied in nature, with a specific focus on ATC providing guidance for system designers and implementation teams in the project management office.”
In year one, the team will identify potential future applications of AI/ML technologies and their integration with other automation to support ATC operations. The second phase will involve the development of the guidance document.
Dr. Smith says the project will include the participation of graduate students “providing them an opportunity to develop expertise in AI/ML air traffic control and the design and functioning of the aviation system, more generally.”
“As an example,” he says, “the FAA has identified deficiencies characterized as untimely or missed alerts associated with strategic conflict probe and tactical conflict alert functions supporting air traffic controllers.
“In addition, only limited resolution assistance is available to controllers today. This need raises the question of whether and how to introduce AI and/or ML technologies into such conflict probes to effectively support the performance of air traffic controllers.”
Story by Nancy Richison