A Multi‐Node Wearable Monitoring System for Firefighters With Edge AI and LoRa‐BLE Heterogeneous Communication

ABSTRACT Firefighters working in high‐risk environments, such as building fires and chemical accidents, are exposed to extreme thermal, toxic, and physical stressors, making real‐time monitoring of physiological and environmental states essential for safety assurance. This study proposes a multinode wearable monitoring system tailored for firefighting scenarios, featuring a distributed architecture composed of chest, wrist, and environmental nodes. A heterogeneous communication framework combining Bluetooth Low Energy (BLE) and LoRa is adopted to support collaborative sensing and long‐range data backhaul. A lightweight one‐dimensional convolutional neural network (1D‐CNN) is deployed on the ESP32‐S3 platform to enable real‐time edge‐side recognition of six representative firefighter postures. In parallel, an adaptive Physiological Strain Index (aPSI) model is implemented to quantify heat‐stress load through the fusion of chest temperature, wrist temperature, and heart rate. The environmental node monitors representative hazardous fireground gases and provides local audible and visual alarms for safety redundancy. Experimental results indicate stable communication performance and effective multimodal monitoring, demonstrating the system's engineering feasibility for firefighter safety support in complex operational environments.