Comparison of Two Eye-Gaze Based Real-Time Driver Distraction Detection Algorithms in a Small-Scale Field Operational Test
DOI: 10.17077/drivingassessment.1297
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 driver distraction in real-time using eye-gaze tracking, a critical issue given the significant contribution of distraction to road casualties. The research compares two distinct algorithms for distraction detection: the "Percent Road Centre" (PRC) metric, which measures the percentage of gaze fixations within a central road area, and the "AttenD" algorithm, which utilizes a 3D vehicle model and a time-based attention buffer to monitor glances toward driving-relevant zones. The primary objective was to evaluate the performance of these algorithms in a naturalistic setting and determine if real-time distraction warnings could effectively modify driver attention. The methodology involved a small-scale field operational test with seven high-mileage drivers who drove an instrumented Saab 9-3 equipped with the SmartEye Pro 4.0 eye-tracking system. The study comprised a baseline phase where the warning system operated silently, followed a treatment phase where haptic warnings (seat vibrations) were issued when the AttenD algorithm classified the driver as distracted. Data analysis focused on PRC values and AttenD buffer levels within a 20-second window surrounding distraction warnings, across various speed intervals. The PRC was calculated using a four-second sliding window, while the AttenD algorithm tracked gaze duration relative to specific vehicle zones. The results indicated that both algorithms functioned similarly in identifying distraction events. Drivers classified as fully attentive by AttenD exhibited higher PRC values (approximately 70%) compared to those classified as distracted. PRC values dropped significantly, reaching their lowest point when the AttenD attention buffer reached zero, confirming a strong correlation between the two metrics. However, the study found no evidence that the haptic warnings influenced driver attention; neither PRC values nor the AttenD buffer increased more rapidly after a warning than during baseline conditions. Additionally, driving speed did not significantly alter the behavior of either algorithm, although PRC values were slightly lower at the highest speed intervals. The significance of this research lies in its validation of gaze-based metrics for distraction detection while highlighting the limitations of current warning systems. The findings suggest that while PRC and AttenD are effective at identifying when drivers look away from the road, the specific haptic warning used failed to prompt immediate corrective gaze behavior. The authors conclude that stand-alone warning systems may be insufficient and advocate for integrated sensor fusion approaches. Furthermore, the study underscores the difficulty of establishing a "ground truth" for distraction, suggesting that future research should focus on extreme glance behaviors and the integration of distraction warnings with broader driver assistance systems to improve traffic safety.
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 | Crossref | — | — | 1 | 2026-06-05 |
| archive | success | canonical_url | — | — | 1 | 2026-06-09 |
| extract | success | pdftotext | — | — | 2 | 2026-06-09 |
| clean | success | clean | — | — | 1 | 2026-06-09 |
| chunk | success | chunk | — | — | 1 | 2026-06-09 |
| embed | success | embed | Qwen/Qwen3-Embedding-8B | — | 1 | 2026-06-09 |
| enrich | success | openalex | — | — | 3 | 2026-07-02 |
| promote | success | — | — | — | 1 | 2026-06-05 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 1 | 2026-06-09 |
| tag | success | vector_similarity | — | — | 8 | 2026-06-11 |
| verify | success | — | — | — | 1 | 2026-06-09 |
Summary generated by qwen3.6-27b-prismaquant on 2026-06-09; verification: verified.
Topics
Ranked by relevance to this paper. Hover a topic for its definition.
- gaze based attention detection
- visual
- distraction detection algorithms
- attention allocation
- temporal
- peripheral attention
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: behavioral performance data
- Methodological Resource: measurement protocol, tool software