AI-Enhanced Driver Fatigue Detection Through Multi-Parameter Analysis: Integration of Radar-Based Heart Rate Monitoring and Deep Learning Techniques

Abstract

This research presents an innovative transportation safety system integrating millimeter-wave radar sensing with artificial intelligence. Our approach employs the MR60BHA1 radar sensor mounted behind the driver's seat to non-intrusively monitor vital signs, combined with deep learning algorithms that analyze correlations between heart rate variability, respiratory patterns, body movements, and environmental factors. Trained on Thailand's first comprehensive driver fatigue database, our custom neural network architecture achieved 97.5% detection accuracy, significantly outperforming traditional single parameter monitoring approaches. The research contributes three key innovations: (1) a novel deep learning framework optimized for Thai driving conditions that correlates multiple radar-derived physiological parameters; (2) an adaptive feature extraction algorithm that identifies and weights the most significant fatigue indicators based on individual characteristics and local contexts; and (3) a comprehensive multiparameter database incorporating synchronized measurements calibrated for various Thai driving environments. This culturally adaptive monitoring technology establishes a new paradigm for intelligent transportation systems particularly suited for tropical regions.