Research on Fatigue State Recognition of Firefighters Based on Heart Rate Variability Features

To evaluate firefighters' fatigue states during rescue operations, this study utilizes a wearable wristband equipped with a heart rate sensor to collect PPI data, from which HRV features are extracted. A comprehensive approach is developed for recognizing firefighter fatigue states based on HRV features. Through difference analysis and correlation analysis, a subset of six HRV features is identified which is Mean HR, LF/HF, PNN50, SDSD, HF, and HFnu. Based on the selected features, multiple classification models are evaluated. CNN exhibite superior performance in all key metrics. The proposed HRV-based fatigue recognition framework shows great potential for real-time fatigue monitoring in high-risk occupations such as firefighting and provides a foundation for future work in physiological signal-based health assessment systems.