Predicting driver fatigue using HRV measures and machine learning.
DOI: 10.3389/fspor.2026.1796180
archive: archived pipeline: cataloged verified
Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)
Summary
This study investigates the physiological and subjective markers of driver fatigue, specifically examining the distinct impacts of sleep deprivation and prolonged driving duration. Motivated by the high prevalence of fatigue-related traffic accidents in Indonesia, the research addresses the need for robust, objective fatigue detection methods that complement subjective assessments, which are prone to misreporting. The authors hypothesize that both insufficient sleep and extended driving time significantly increase fatigue, detectable through Heart Rate Variability (HRV) metrics and subjective scales. The experimental design involved 40 healthy participants (aged 18–25) in a controlled driving simulation. Using a within-subjects design, participants completed four 30-minute driving sessions under two conditions: normal sleep (≥7 hours) and sleep deprivation (≤4 hours). HRV data were recorded using Polar H10 chest sensors, with 5-minute segments analyzed at the end of each session to minimize artifacts. Subjective fatigue was assessed immediately post-session using the Karolinska Sleepiness Scale (KSS) and Rating of Fatigue (ROF). The study employed Repeated Measures ANOVA to analyze the effects of sleep and driving duration on HRV parameters (time-domain: Mean RR, SDNN, RMSSD; frequency-domain: LF, HF) and subjective scores. Additionally, machine learning models—Logistic Regression, Random Forest, and XGBoost—were trained to classify fatigue states (fatigued vs. non-fatigued) based on HRV features, using Group k-Fold cross-validation to prevent data leakage. Results indicated that driving duration had a stronger influence on physiological markers than sleep duration. Prolonged driving was associated with autonomic imbalance, evidenced by significant decreases in Mean RR, Mean HR, RMSSD, and HF. In contrast, sleep deprivation did not significantly alter most HRV indices but had robust effects on subjective measures, with ROF and KSS scores increasing significantly under sleep-deprived conditions. Significant correlations were found between HRV parameters and subjective fatigue ratings, aligning physiological changes with perceived drowsiness. In the machine learning analysis, the XGBoost ensemble model achieved the highest classification accuracy of 77%. Feature importance analysis identified Mean RR, Mean HR, and LF as the most influential predictors for fatigue detection. The findings demonstrate that integrating HRV metrics with ensemble learning improves the detection of fatigue patterns compared to conventional statistical methods. The study highlights that while driving duration drives physiological autonomic changes, sleep deprivation primarily impacts subjective perception. These results support the development of real-time, non-invasive fatigue monitoring systems that combine objective physiological data with machine learning to enhance driver safety, particularly in contexts involving sleep restriction and extended driving periods.
Provenance
The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | PubMed Central | — | — | 1 | 2026-06-19 |
| archive | success | unpaywall | — | — | 2 | 2026-06-25 |
| extract | success | cached | — | — | 2 | 2026-06-26 |
| clean | success | clean | — | — | 1 | 2026-06-20 |
| chunk | success | chunk | — | — | 1 | 2026-06-20 |
| embed | success | embed | Qwen/Qwen3-Embedding-8B | — | 1 | 2026-06-20 |
| enrich | success | openalex | — | — | 1 | 2026-06-20 |
| promote | success | — | — | — | 1 | 2026-06-19 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 1 | 2026-06-26 |
| tag | success | vector_similarity | — | — | 6 | 2026-06-20 |
| verify | partial | — | — | — | 1 | 2026-06-26 |
Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified_with_issues.
Topics
Ranked by relevance to this paper. Hover a topic for its definition.
Information type
What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).
- Empirical Findings: physiological data
- Methodological Resource: validation psychometrics
- Theoretical Contribution: computational model