Skip Navigation

OU and Air Force Engineers Improve Natural Language Generation for Auto-Routing Aircraft

OU Oklahoma Aerospace and Defense Innovation Institute, The University of Oklahoma wordmark
  Four Navy P-8 Poseidon aircraft from Naval Air Station Jacksonville, Florida, line up with Air Force KC-135 Stratotankers on Tinker Air Force Base’s flightline. Air Force photo by April McDonald.

Four E-6 Mercury (TACAMO) aircraft line up with Air Force KC-135 Stratotankers on Tinker Air Force Base’s flightline. Air Force photo by April McDonald.

June 8, 2022

OU and Air Force Engineers Improve Natural Language Generation for Auto-Routing Aircraft

A research team, coordinated by the Oklahoma Aerospace and Defense Innovation Institute at the University of Oklahoma, is applying artificial intelligence and machine learning methods to improve autonomous routing of aircraft being operated by the U.S. Air Force.

“The Department of Defense analyzes massive amounts of information to orchestrate the defense and security actions of allied forces,” said Dean Hougen, Ph.D., an associate professor in the School of Computer Science in the Gallogly College of Engineering. “Routing platforms, like those used to set the flight plans for aircraft, depend on massive quantities of data that are ever evolving, and current routing methods are often heavily dependent on direct human involvement.”

Hougen is leading the project in partnership with the Oklahoma City Air Logistics Complex to develop a platform that will use natural language generation to improve autonomous routing capabilities. Co-investigators John Antonio and Lacey Schley, with OU engineering graduate students, are supporting the project. OC-ALC technical leads are Alexander Stringer and Brian Sun.

“This automated routing capability will be able to quickly ingest information and output a near optimal route based on the objectives of its human operators,” Hougen said. “Our prototype uses machine learning, demonstrated on publicly available Automatic Dependent Surveillance-Broadcast data and simulated data provided by the USAF, to develop routes that are autonomous, efficient, and understandable by human operators.”

The two-year project began Sept. 15, 2021 and is expected to be complete by Sept. 14, 2023.