|
 |
First author Ryan Lowe shows deep neural network technology to research group |
|
A University of Maryland (UMD) graduate student working in the UMD A. James Clark School of Engineering MATRIX Lab is the first author on a research paper now published in the Journal of Aerospace Information Systems.
“Evaluating Single-Camera Range Estimation for Autonomous Air-to-Air Refueling” was written by Ryan Lowe (M.Eng. ’26, Robotics), Dr. Violet Mwaffo (United States Naval Academy), and Dr. Donald “Bucket” Costello (UMD Aero and MATRIX Lab). It outlines how AI can be used to help U.S. Navy aircraft autonomously refuel themselves mid-flight.
As the Navy expands its use of uncrewed aerial systems (UAS), those aircraft will need the same operational flexibility as manned platforms, including the ability to remain airborne for extended periods and operate far from traditional support infrastructure. Autonomous aerial refueling can help with this, however, the systems need to be able to operate even when GPS and radio frequency communications are unavailable or unreliable. One solution is using onboard sensors and AI rather than relying on those external navigation aids. To test this option, Lowe and his team are analyzing whether a single lens camera combined with a deep neural network could accurately estimate the distance between aircraft components during autonomous air-to-air refueling.
“I am proud of the contributions Dr. Costello, the MATRIX Lab team, and I made towards advancing autonomous aerial refueling through neural network-based vision and control,” Lowe said. “I am excited about the future of autonomous UAS technologies and look forward to teaming with these systems during my service in the Navy.”
 The team’s paper outlines their experiments, which were conducted in the MATRIX Lab’s Omni-domain Autonomous Systems Integration Space (OASIS). Researchers used the OASIS’ Vicon Vantage camera system to precisely measure the distance between aircraft components, then compared those measurements with distances estimated by a YOLO11m neural network AI system using camera imagery. The system demonstrated near-zero average error at close distances (5-17 feet) and errors under ten inches at further distances (up to 45 feet).
The results suggest that a lightweight, single-camera vision system could support autonomous aerial refueling, providing a practical path toward enabling future Navy uncrewed aircraft to refuel autonomously in flight.
“This work represents the first journal publication that was based on research conducted in the MATRIX Lab by our graduate students,” Dr. Costello said. “I am extremely proud of Ryan and the team.”
Lowe graduated with a master’s degree in robotics from the University of Maryland in May 2026. He finished his time with the MATRIX Lab in June 2026 and is now heading to flight school to train as a Naval Aviator.
Related Articles:
Sensor Advancement Breaks Barriers in Brain-Behavior Research
Generations of Graduates: A Full-Circle Celebration
University of Maryland Represented at International Forum
UMD Student Bridging Research Innovation and Education Advocacy
MATRIX-Affiliated Faculty Solving Challenges with Solutions from Nature
Clark School Seed Grant Awards Announced
MATRIX-Affiliated Faculty Solving Challenges From Sea to Space
MATRIX-Affiliated Faculty Solving Tomorrow's Challenges Today
Grand Challenges 2.0: Engineers Partner to Unlock Solutions
Capstone Design Expo 2026: Students Solve for Maryland & Beyond
June 25, 2026
|
