Driver Fatigue Prediction Using HRV Measures and Machine Learning

A controlled driving simulation study collected physiological and subjective fatigue data from 40 participants. The dataset includes Heart Rate Variability (HRV) indicators combined with Karolinska Sleepiness Scale (KSS) and Rating of Fatigue (ROF) scores under normal and sleep-deprived conditions. The data was used to train machine learning models, with XGBoost achieving 77% accuracy for fatigue classification.

Use Cases

• Classifying driver fatigue states based on HRV parameters and subjective scores.
• Training machine learning models for real-time fatigue detection using physiological signals.
• Analyzing correlations between autonomic nervous system activity (via HRV) and self-reported fatigue.
• Investigating the impact of sleep deprivation versus prolonged driving on physiological and subjective fatigue measures.

Strengths

• Includes data from 40 participants across four driving sessions.
• Combines objective HRV measures with subjective KSS and ROF scores.
• Machine learning model performance is reported, with XGBoost achieving 77% accuracy.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Data collection was conducted in a controlled simulation environment, which may not fully reflect real-world driving conditions.

Provenance

Source

Maya Arlini Puspasari via figshare

Collection Method

Data collected from a controlled driving simulation experiment with 40 participants.

Freshness

Last updated 2026-06-03 04:26:46; freshness should be verified.