Assessment of Combination of Automated Pupillometry and Heart Rate Variability to Detect Driving Fatigue

Shi, Lin; Zheng, Leilei; Jin, Danni; Lin, Zheng; Zhang, Qiaoling; Zhang, Mao · 2022 · Crossref

DOI: 10.3389/fpubh.2022.828428

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 addresses the challenge of detecting driving fatigue, a factor responsible for 20–30% of traffic accidents, by evaluating the efficacy of automated pupillometry and heart rate variability (HRV). While physiological monitoring systems exist, their practical application is often hindered by complexity. The authors investigated whether pupillary light reflex (PLR) parameters, potentially offering a less intrusive alternative to electroencephalogram or electromyogram systems, could effectively detect fatigue and how they compare to or complement HRV metrics. The researchers conducted a prospective observational study with 32 healthy volunteers who underwent a 90-minute monotonous simulated driving task designed to induce fatigue. Data were collected at baseline and at 30-minute intervals. Fatigue was assessed using the Karolinska Sleepiness Scale (KSS) and a custom Fatigue Grade scale, alongside behavioral performance metrics. PLR was measured using an automated pupillometer, capturing parameters such as maximum constriction velocity (MCV) and minimum pupil size. HRV was recorded using a standard monitor, analyzing indicators like standard deviation of NN intervals (SDNN) and frequency domain measures. Statistical analyses included repeated measures ANOVA, Spearman correlations, and receiver operating characteristic (ROC) curve analysis to determine diagnostic accuracy. The results confirmed that the simulated driving task successfully induced fatigue, evidenced by significantly increased KSS scores and deteriorated driving performance. Both PLR and HRV parameters showed significant variations correlated with increasing fatigue levels. Specifically, PLR metrics such as MCV and minimum pupil size changed significantly over time. Correlation analysis revealed that changes in KSS were moderately to highly correlated with changes in PLR parameters (e.g., MCV, minimum pupil size) and HRV indicators (e.g., SDNN) at later timepoints. ROC analysis demonstrated that PLR variations, particularly MCV, had excellent discriminatory power for detecting fatigue, achieving an area under the curve (AUC) of 0.835, with 85.00% sensitivity and 72.34% specificity. HRV parameter SDNN also performed well (AUC = 0.805). However, combining PLR and HRV metrics did not significantly improve detection performance compared to using PLR alone (AUC = 0.872 vs. 0.835, P > 0.05). The study concludes that PLR variations, specifically MCV, are a robust and practical indicator for detecting driving fatigue, performing comparably to HRV and combined models. The findings suggest that automated pupillometry could serve as a standalone, effective tool for developing commercialized driving fatigue detection systems, offering a viable alternative to more complex physiological monitoring methods.

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.

Summary generated by qwen3.6-27b-nvidia on 2026-08-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• drowsiness detection algorithms
• workload measurement
• drowsiness

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, tool software