Crash risk associated with eyes-off-road duration by road control type and intersection type
DOI: 10.1016/j.jsr.2025.03.003
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 quantifies the crash risk associated with eyes-off-road (EOR) durations across different road control types and intersection types to inform alert timer settings for driver monitoring systems (DMSs). The research addresses two primary questions: whether DMS alert thresholds should differ between controlled and uncontrolled access roads, and whether they should vary between intersections and non-intersections. The motivation stems from the need to balance DMS sensitivity with specificity, ensuring alerts are triggered appropriately based on environmental risk factors without generating excessive false alarms. The analysis utilized data from the Second Strategic Highway Research Program Naturalistic Driving Study (SHRP 2 NDS), focusing on safety-critical events (SCEs), defined as crashes and near-crashes, compared against baseline driving periods. Eyeglance data was coded into 21 categories, aggregated into "eyes-on-road" and "EOR" metrics. The study evaluated three EOR metrics—total duration, average duration, and single longest duration—within 6-second and 15-second analysis windows. Road types were categorized as controlled access (e.g., interstates) or uncontrolled access (e.g., urban roads). On uncontrolled roads, maneuvers were further classified as turning at an intersection, going straight at an intersection, or driving on a straight segment. To account for discrepancies in crash composition and advanced driver assistance system (ADAS) effectiveness across scenarios, the authors applied a weight-adjustment method to calculate odds ratios (ORs). Key findings indicate that drivers on uncontrolled access roads exhibited longer EOR glances and higher relative crash risks (ORs) compared to those on controlled access roads. Consequently, lower DMS alert thresholds are recommended for uncontrolled roads. Regarding intersection types, the odds ratios for EOR were highest for turning maneuvers, despite a lower prevalence of long single glances (>2 seconds) in these samples. This suggests a need for stricter, lower-threshold alerts when turning at intersections. The study identified 2.7 seconds and 3.7 seconds as critical changing points for ORs, proposing these as potential DMS thresholds. At a 2.7-second threshold, the alert rate was one per 11 minutes, whereas a 3.7-second threshold resulted in one alert per 40 minutes. Additionally, the single longest EOR metric proved more predictive of crash risk than total or percentage EOR, and a 6-second analysis window was more informative than a 15-second window. The study concludes that DMS alert timer settings should be context-dependent, with lower thresholds for uncontrolled access roads and intersection-related maneuvers, particularly turning. These findings provide critical data for ADAS development and driver behavior education, highlighting the trade-off between alert sensitivity and false alarm rates. Limitations include small sample sizes for turning maneuvers and the need for further investigation into non-linear risk relationships and other road features, such as divided versus undivided roads.
Key finding
Crash risk associated with eyes-off-road duration is significantly higher on uncontrolled access roads and when turning at intersections, necessitating lower alert timer thresholds for driver monitoring systems in these specific contexts.
Methodology
dataset
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 | — | — | — | 1 | 2026-05-28 |
| archive | success | canonical_url | — | — | 4 | 2026-06-06 |
| extract | success | cached | — | — | 3 | 2026-06-10 |
| clean | success | clean | — | — | 1 | 2026-06-07 |
| chunk | success | chunk | — | — | 1 | 2026-06-07 |
| embed | success | embed | Qwen/Qwen3-Embedding-8B | — | 1 | 2026-06-07 |
| enrich | success | semantic_scholar | — | — | 4 | 2026-07-02 |
| promote | success | — | — | — | 1 | 2026-06-04 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 2 | 2026-06-10 |
| tag | success | vector_similarity | — | — | 15 | 2026-06-11 |
| verify | success | — | — | — | 2 | 2026-06-10 |
Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.
Topics
Ranked by relevance to this paper. Hover a topic for its definition.
- visual
- temporal
- looked but failed to see
- gaze based attention detection
- external distraction
- eye movements scanning
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: crash risk outcomes, behavioral performance data, observational prevalence