IEEE Embedded Systems Letters, 2025 (SCI-Expanded)
Detecting individuals trapped after earthquakes is critical for guiding rescue operations, yet current search methods still face challenges, especially in large-scale emergencies. This work focuses on leveraging a low-cost, lightweight continuous wave radar (CWR) sensor for through-wall human detection, with the goal of enabling fast, scalable deployment in disaster zones. This solution facilitates deployment across different locations, supporting wide-area coverage, parallel search operations, and rapid rescue in time-critical situations. The proposed system uses machine learning models to accurately detect human presence and classify movement types behind walls, including static, low-energy, and high-energy movements. It achieves 97.1% accuracy in presence detection and 84.5% in static vs. movement classification, offering a practical and scalable solution for post-disaster search and rescue.