Feasibility of Using In-Vehicle Video Data to Explore How to Modify Driver Behavior That Causes Nonrecurring Congestion
DOI: 10.17226/14509
archive: archived pipeline: cataloged verified
Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)
Summary
This report, part of the second Strategic Highway Research Program (SHRP 2), investigates the feasibility of using in-vehicle video data from naturalistic driving studies to identify driver behaviors that cause nonrecurring congestion. Nonrecurring congestion, resulting from random incidents like crashes and weather, accounts for approximately half of all traffic delays. The study aims to determine if existing video datasets can reveal the relationship between observable driver behavior and these disruptions, thereby informing countermeasures to improve travel time reliability. The research team evaluated domestic and international candidate studies that collected video, kinematic, and external data. They established evaluation criteria for feasibility, including legal restrictions, data comprehensiveness, video quality, and linkage to external data. Qualified datasets were subjected to manual video review and data reduction to identify contributing factors to crashes and near-crashes. The analysis focused on distinguishing between driver-related errors (decision, recognition, performance) and environmental factors. Additionally, the team developed a multimode statistical model for travel time reliability, designed to capture the probability of encountering congestion and the resulting travel time variations. Key findings indicate that most crashes and near-crashes are caused by driver inattention and errors. In the Road Departure Crash Warning System Field Operational Test, decision errors (e.g., following too closely) caused over 85% of freeway events, while recognition errors accounted for over 5%. In the 100-Car Study, recognition errors comprised 32% of crashes, and decision errors accounted for 28%. The study concluded that appropriate countermeasures could prevent a significant portion of these events: 91% of near-crashes in the Road Departure study and over 80% in the 100-Car Study were deemed correctable. However, the existing datasets lacked sufficient crash samples to directly model the impact of incidents on traffic conditions. Consequently, the proposed multimode travel time model, validated using 100-Car Study data, provided a superior fit to field data compared to traditional unimodal models, effectively linking model parameters to incident probabilities. The study concludes that while it is feasible to identify driver behaviors leading to crashes from video data, existing datasets have limitations, including insufficient camera views, image glare, and incomplete driver identification. The report recommends that future data collection efforts include comprehensive onboard equipment, such as five-view video cameras and GPS data, and implement strict protocols for driver identification and trip context recording. Integrating naturalistic driving data with external sources like weather and traffic volume is essential for accurately assessing the relationship between driver behavior modifications and the reduction of nonrecurring congestion.
Key finding
In-vehicle video data from naturalistic driving studies is feasible for identifying driver behaviors like inattention and decision errors that cause crashes and near crashes, which are key contributors to nonrecurring congestion.
Methodology
review
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 | failed | — | — | — | 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.
- naturalistic crash near crash
- traffic density
- exposure measurement
- work zones
- crash reconstruction hf
- distraction detection algorithms
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: observational prevalence
- Methodological Resource: dataset resource, tool software