Assessing safety in horizontal curves using surrogate safety measures and machine learning
DOI: 10.1038/s41598-025-97384-7
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 safety challenges associated with horizontal curves (HCs) on rural roads, which are critical sites for road crashes but difficult to analyze due to the rarity of crash events. Motivated by the high incidence of fatalities and injuries in rural areas, particularly in Iran, the research aims to comprehensively assess HC safety by integrating surrogate safety measures (SSMs) with machine learning. Unlike previous studies limited by data availability or methodological constraints, this approach leverages observational data to evaluate the simultaneous impact of driver characteristics, road specifications, and environmental conditions on safety outcomes. The methodology involved a field study conducted on multilane divided rural highways in Kerman Province, Iran. Data were collected using an instrumented vehicle equipped with high-precision GPS and video recording cameras, driven by 59 recruited participants during January 2021. The study focused on low-traffic conditions to minimize vehicle interaction effects. Two speed-based SSMs were utilized as proxies for crash risk: the maximum deceleration rate (MDR) and the difference between the speed in the tangent section preceding the curve and the speed within the curve (MSD). These metrics served as target variables, categorized into low and high risk based on established thresholds (1.5 m/s² for MDR and 15 km/hr for MSD). A decision tree classifier, specifically employing CHAID and CRT algorithms, was applied to a dataset of 359 records to determine the relative significance of various input variables, including geometric features, driver demographics, and environmental factors. The results identified curve radius and the speed in the tangent section before the horizontal curve as the most significant factors influencing safety. Critical thresholds were established for these variables, with a curve radius of 358 meters and a tangent speed of 23 m/s identified as key determinants of risk. Regarding driver characteristics, driving experience, crash history, and education level were found to be the most critical attributes for safely navigating HCs. Conversely, environmental conditions, such as temperature and daytime status, did not exhibit a significant impact on HC safety in this analysis. The decision tree models effectively ranked these features, providing a transparent interpretation of how specific geometric and behavioral factors contribute to safety risks. The significance of this research lies in its provision of actionable insights for improving road design and safety measures on rural highways. By identifying specific critical thresholds for curve radius and speed, the study offers practical guidelines for infrastructure planning and traffic management. Furthermore, the integration of SSMs with machine learning demonstrates a cost-effective and scalable method for assessing safety without relying on extensive crash databases or complex video processing. This holistic approach, which simultaneously considers driver, road, and environmental factors, addresses existing gaps in the literature and supports the development of more effective countermeasures to reduce crash risks in horizontal curves.
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 | OpenAlex-citations | — | — | 1 | 2026-06-24 |
| archive | success | unpaywall | — | — | 2 | 2026-06-26 |
| extract | success | cached | — | — | 2 | 2026-06-26 |
| clean | success | clean | — | — | 1 | 2026-06-25 |
| chunk | success | chunk | — | — | 1 | 2026-06-25 |
| embed | success | embed | Qwen/Qwen3-Embedding-8B | — | 1 | 2026-06-25 |
| promote | success | — | — | — | 1 | 2026-06-24 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 1 | 2026-06-26 |
| tag | success | vector_similarity | — | — | 6 | 2026-06-25 |
| verify | success | — | — | — | 1 | 2026-06-26 |
Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.
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: crash risk outcomes