Experimental Evaluation of a Heart Rate–Based Warning System for Driver Sleepiness Detection

Abstract

Driver sleepiness is a major contributor to road accidents, making early detection and intervention essential for improving driving safety. This study presents an experimental evaluation of a heart rate (HR)–based warning system designed to identify physiological indicators of drowsiness and provide timely alerts to drivers. Fifteen licensed drivers (aged 21–26) completed four 30-minute sessions in a monotonous driving simulator while HR was continuously measured using a photoplethysmography (PPG) sensor and subjective sleepiness was assessed using the Karolinska Sleepiness Scale (KSS). Results showed a progressive decline in HR across baseline sessions (S1–S3: 83.70 → 76.05 → 70.17 bpm), corresponding with increasing sleepiness (KSS: 5.86 → 6.16 → 6.27). When the HRbased warning system was activated in Session 4, HR increased to 75.71 bpm while KSS decreased slightly to 6.06, indicating a partial reversal of drowsiness. Repeated-measures ANOVA confirmed significant effects of session on both HR and KSS (p < .05). These findings demonstrate that HR monitoring can effectively detect physiological changes associated with driver sleepiness and that a threshold-based auditory alert can elicit an acute arousal response. While results show promise, findings are based on a simulator study and validation with larger, real-world datasets is recommended.