Driver performance and behavior in adverse weather conditions : an investigation using the SHRP2 naturalistic driving study data—phase 1.
archive: archived pipeline: cataloged verified
Get this paper ↗ (full text — opens at the source; we link to it, we don't host it)
Summary
This report presents the findings of Phase 1 of a study investigating driver performance and behavior during adverse weather conditions, specifically focusing on heavy rain. The research was motivated by the significant safety risks associated with inclement weather, which contributes to over 24% of crashes and severely impacts visibility and pavement conditions. While previous literature established that adverse weather elevates crash risk, it lacked detailed data on how drivers actively adjust their behavior—such as speed choice, lane maintenance, and headway—in response to these conditions. This gap hindered the development of effective safety countermeasures, particularly Variable Speed Limit (VSL) systems, which currently rely on traffic simulation and historical data rather than real-time driver behavior. The study aimed to determine if the Second Strategic Highway Research Program (SHRP2) Naturalistic Driving Study (NDS) and Roadway Information Database (RID) could effectively identify weather-related trips and characterize driver responses to reduced visibility. The researchers utilized a subset of SHRP2 data from Florida and Washington, two sites with significant rainfall. The methodology involved linking NDS data, which captures vehicle dynamics, driver actions, and video feeds, with RID data containing roadway and weather information. To isolate trips occurring in rainy conditions, the team used windshield wiper usage duration as a proxy, defining "rainy" trips as those where wipers were used for more than 10 minutes. This approach yielded 5,013 trips from 143 participants across both states. The analysis focused on freeway segments and examined driver behaviors including speed selection, acceleration, lane wandering, and headway adaptation. The study also employed machine vision for visibility estimation and used ordered probit models to analyze speed behavior relative to weather conditions. The results demonstrated that the NDS and RID datasets could effectively identify inclement weather trips and characterize driver responses. The analysis revealed distinct differences in driver behavior between clear weather and heavy rain conditions. Specifically, the study found that drivers adjusted their speeds and headways in response to reduced visibility, though the extent of this adaptation varied. The ordered probit models provided statistical evidence of how weather conditions influence speed selection, confirming that drivers do modify their behavior, but often insufficiently to mitigate the increased risk. The study also identified surrogate measures for weather-related crashes, such as specific patterns of lane wandering and sudden braking, which were more prevalent during heavy rain. These findings provided objective insights into the "dynamic visual acuity" and reaction processes of drivers under stress, filling a critical data gap in safety modeling. The significance of this research lies in its potential to improve the design and implementation of Variable Speed Limit (VSL) systems and Connected Vehicle (CV) technologies. By providing empirical data on how drivers actually behave during adverse weather, the study supports the development of VSL algorithms that account for driver behavior rather than relying solely on environmental conditions. This approach can lead to more realistic and effective speed limits that better match driver capabilities and road conditions, thereby reducing traffic injuries and fatalities. The successful proof-of-concept in Phase 1 validates the use of SHRP2 data for this purpose, paving the way for further research and the national adoption of improved safety countermeasures.
Key finding
Drivers significantly reduced their speeds during heavy rain compared to clear weather conditions on freeways.
Methodology
naturalistic
Sample size: 143
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 bulk_ingest_rosap on 2026-05-23 (7 acquisition events logged).
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | author_sweep | — | — | 3 | 2026-05-28 |
| archive | success | — | — | — | 1 | 2026-05-23 |
| extract | success | cached | — | — | 2 | 2026-06-10 |
| clean | success | — | — | — | 1 | 2026-06-01 |
| chunk | success | — | — | — | 1 | 2026-06-01 |
| embed | success | — | — | — | 1 | 2026-06-02 |
| enrich | success | — | — | — | 1 | 2026-05-23 |
| promote | success | — | — | — | 1 | 2026-05-23 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 3 | 2026-06-10 |
| tag | success | vector_similarity | — | — | 19 | 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.
- weather rain fog snow
- naturalistic crash near crash
- speed choice
- incidence prevalence
- urban rural setting
- traffic density
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