Real-World Human Activity Recognition Using Radar Signals
Abstract
This research aims to design deep learning networks that parse sensory signals for solving several tasks such as noise removal, image generation, feature learning and representation for human activity detection and classification under fringe situations, such as the people are occluded or behind obstacles. Additionally, they aim to design an IoT hardware device, embedded with Radar and thermal sensors, that apply few-shot learning techniques to generalize the model to novel activities based on the collected dataset as the prior knowledge.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Feb 16, 2024
- Source ID
- FA23862314058
Entities
People
- Minh-triet Tran
Organizations
- Air Force Office of Scientific Research
- United States Air Force