Patterns of near-crash events in a naturalistic driving dataset: Applying rules mining
DOI: 10.1016/j.aap.2021.106346
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 associations between near-crash events and specific roadway geometry and trip features using naturalistic driving data. Motivated by the limitations of traditional crash data, which lacks detailed driving behavior information and requires long historical periods, the authors utilize surrogate safety measures to identify risky driving behaviors in real-time. The research specifically aims to differentiate patterns between two severity levels of near-crashes: "trivial" events (deceleration rates between -0.45g and -0.75g) and "non-trivial" events (deceleration rates ≤ -0.75g). By applying association rule mining, the study seeks to uncover co-occurrence patterns that can inform traffic safety countermeasures and connected vehicle applications. The methodology employs the Apriori algorithm to mine association rules from two integrated datasets: the Safety Pilot Model Deployment (SPMD) naturalistic driving data and the Highway Performance Monitoring System (HPMS) roadway inventory data. The SPMD data, collected in Michigan from 2012 to 2013, provided trajectory and maneuver data for 92 vehicles, while HPMS data supplied geometric features such as lane width, median type, and functional classification. After cleaning and spatially joining the datasets, 957 near-crash events were identified. The Apriori algorithm was configured with a minimum support of 0.1 and minimum confidence of 0.1 to generate meaningful rules, treating each variable category as an item and each event as a transaction. The results reveal distinct patterns for each severity level. Trivial near-crash events are strongly associated with roadways lacking medians and shoulders, lower functional classes (such as minor arterials), and non-peak driving hours. Specifically, principal arterial roads without protected medians during non-peak hours showed the highest lift value for trivial events. In contrast, non-trivial near-crash events are significantly linked to longer trips exceeding two hours and higher traffic volumes. A dominant rule for non-trivial events involves congestion on lower functional class roadways. Additionally, higher functional roadways with wide medians and shoulders were identified as an intriguing combination associated with non-trivial events, suggesting complex interactions between road design and severe braking maneuvers. The significance of this study lies in its detailed characterization of near-crash behaviors based on severity, offering insights beyond general crash correlations. By distinguishing between trivial and non-trivial events, the findings provide specific geometric and trip-based indicators for severe safety risks. These patterns can enhance the development of surrogate safety measures and real-time hazard detection systems for connected vehicles. The study demonstrates that association rule mining is an effective tool for extracting actionable safety insights from naturalistic driving data, particularly in identifying how road geometry and trip duration influence the severity of near-crash incidents.
Key finding
Non-trivial near-crash events are significantly more likely to occur during trips longer than two hours, whereas trivial near-crash events are predominantly associated with roadways lacking medians and shoulders.
Methodology
naturalistic
Sample size: 92
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. Discovered via author_sweep_intake on 2026-05-27.
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | author_sweep | — | — | 2 | 2026-05-27 |
| archive | success | unpaywall | — | — | 2 | 2026-06-04 |
| extract | success | cached | — | — | 3 | 2026-06-10 |
| clean | success | clean | — | — | 1 | 2026-06-04 |
| chunk | success | chunk | — | — | 1 | 2026-06-04 |
| embed | success | embed | Qwen/Qwen3-Embedding-8B | — | 1 | 2026-06-04 |
| enrich | success | semantic_scholar | — | — | 2 | 2026-06-04 |
| 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 | partial | — | — | — | 2 | 2026-06-10 |
Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified_with_issues.
Topics
Ranked by relevance to this paper. Hover a topic for its definition.
- naturalistic crash near crash
- incidence prevalence
- crash typology
- pre crash contributing factors
- induced exposure
- urban rural setting
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, observational prevalence
- Methodological Resource: dataset resource