Multimodal Fusion of Heart Rate and Gaze Data For Real-Time Driver Monitoring in Naturalistic Driving

Driver distraction and diminished alertness remain key contributors to road‑traffic accidents. This study introduces a real‑time multimodal driver‑monitoring framework that fuses heart rate (HR) signals with gaze‑based metrics. Five licensed drivers undertook 480 minutes of naturalistic driving in both urban and motorway settings while instrumented with Polar H10 chest‑strap ECG sensors and Pupil Labs Invisible eye‑tracking glasses. Inferential statistics (p<0.05) revealed significantly higher mean HR and more frequent gaze indicators of cognitive load during urban driving. Leveraging these findings, we designed a weighted‑fusion algorithm that combines HR, blink duration, fixation duration, and saccade kinematics into a single alertness score, which is then compared with a tunable binary threshold. A sensitivity analysis shows that adjusting feature weights and the decision threshold enables a controllable trade‑off between false alarms and missed detections. The results demonstrate the feasibility of integrating physiological and gaze information for practical, real‑time detection of reduced driver alertness, paving the way for adaptive in‑vehicle safety systems that proactively mitigate risk.