Real-Time Driver Monitoring using Multi-Modal Emotion Recognition using Machine Learning

Abstract

Road traffic safety requires continuous assessment of driver state, including fatigue and emotionaldistraction. This paper presents a robust real-time Driver Monitoring & Alert System, integrating multi-modalmachine learning for drowsiness and emotion analytics. The system logs behavioral metrics—facial landmarks,speech volume, and micro-expressions—using edge computation and stores entries in a scalable database. A modernanalytics dashboard visualizes trends, enables time-based filtering, and computes key summaries (e.g., averagedrowsiness, emotion distributions). Empirical results show high recognition accuracy, rapid dashboardresponsiveness, and practical potential for real-world automotive deployment. The project demonstrates advancesin real-time driver monitoring, combining intuitive visualization with rigorous machine learning methods.