In-Cabin Non-Invasive Driver State Monitoring Using Vision-Based Eye Tracking and rPPG

As vehicles move toward conditional automation (SAE Level 2–3), holistic, real-time driver monitoring is essential for safety. We present a non-invasive, vision-based in-cabin solution integrating behavioral and physiological analysis. A multi-camera eye-tracking system detects distraction and drowsiness via real-time gaze, blink, and head pose metrics, using a robust facial landmark model. A simulator study achieved 1.6° gaze accuracy and 97% distraction detection. Complementing this, a remote photoplethysmography (rPPG) module estimates heart rate, breathing rate, IBI, and blood pressure from facial video using an edge-optimized video transformer. With synchronized NIR-illuminated cameras and compact hardware, this full-stack system delivers holistic, multi-modal driver state monitoring for enhanced safety and regulatory compliance.”