Thermal Source Localization Using WiFi Sensing

Thermal source localization is crucial in detecting fires and other anomalies, allowing emergency responders to quickly locate the source of the hazard to minimize damage. With recent advancements in WiFi sensing technology for thermal detection and the emergence of Integrated Sensing and Communication (ISAC) on WiFi systems, we are interested in exploring the potential of using WiFi sensing technology for thermal source localization using commercially available Internet of Things (IoT) WiFi hardware. Our study investigates the potential of using WiFi Channel State Information (CSI) to locate thermal sources. Using the cost-effective and power-efficient ESP32 microcontroller, we propose a WiFi sensing-based system to predict the location of a heat source. We demonstrate that the WiFi CSI signatures can be used to identify localized temperature changes in a timely manner. Our key contribution is effectively filtering out undesired components from the CSI and using smart subcarrier selection to form the feature set for the machine learning algorithm. We present a Support Vector Machine (SVM)–based classifier that achieves up to 99% accuracy in predicting the relevant heat source location. We also identify the optimal WiFi sensing device configuration, demonstrate our technology’s fast response time, and shed insights on the impact of heat source distance on localization performance. Our findings offer promising solutions for low-cost, environmentally friendly monitoring systems.