A University of Alabama in Huntsville (UAH) researcher has received a $599,830 National Science Foundation grant to pioneer collaborative learning techniques for drones, enhancing applications from disaster response to infrastructure inspection.
Innovative Approach to UAV AI
Dr. Dinh Nguyen, assistant professor in UAH’s Department of Electrical and Computer Engineering, will lead the initiative, which is set to run through September 2028. Traditional machine-learning methods for drones often require sending raw data to centralized servers—a process limited by privacy concerns, bandwidth constraints, and delays.
This project leverages secure multi-modal federated learning (MFL), a distributed AI method allowing drones with diverse sensors to learn collaboratively without sharing raw data.
Applications Beyond Defense
While MFL can support defense and security operations, Dr. Nguyen emphasizes civilian uses, including:
Disaster response
Environmental monitoring
Infrastructure inspection
Cybersecurity
By training AI models locally on each UAV and sharing only model updates, drones can build a shared global model more efficiently and securely.
Privacy, Security, and Resilience
The system incorporates modality-aware differential privacy, secure aggregation, and attack detection mechanisms to safeguard sensitive data and prevent compromised drones from corrupting the shared model.
Dr. Nguyen explains, “These advances save bandwidth, reduce communication delays, and safeguard sensitive information. Overall, this system makes learning faster, more private, and more adaptable—ideal for missions with tight time, energy, and resource constraints.”
Huntsville at the Forefront
The research positions Huntsville and UAH at the cutting edge of drone AI technology, contributing to both regional innovation and broader national priorities in resilient, intelligent UAV networks.
Sign up for the Homely Huntsville newsletter to stay up to date on Huntsville.
