Fever is a primary clinical indicator of infections and inflammatory conditions, making its early detection essential for effective healthcare response. Conventional temperature measurement techniques, including contact thermometers and infrared scanners, require close proximity, manual operation, or direct line-of-sight, which restricts their usefulness in large-scale and continuous monitoring environments. This paper presents a novel, contactless fever detection framework based on WiFi Doppler signal distortion, enabling passive health monitoring without additional sensing hardware. The proposed system exploits variations in Channel State Information (CSI) and Doppler frequency shifts produced when WiFi signals interact with the human body. Changes in body temperature subtly influence RF signal amplitude, phase, and frequency characteristics due to thermal radiation and involuntary micro-movements. A structured processing pipeline comprising signal denoising, Doppler feature extraction, and statistical feature modeling is designed to capture these variations. Machine learning techniques are then applied to classify normal and elevated temperature conditions. The solution is non-invasive, cost-effective, privacy-preserving, and compatible with existing WiFi infrastructure 2,7. Experimental observations indicate a consistent relationship between Doppler-based RF features and body temperature variations, validating the feasibility of WiFi-assisted fever detection. This study contributes to wireless health sensing research and demonstrates the potential of RF signals for scalable and autonomous fever screening.
Chaudhari et al. (Thu,) studied this question.