Driver Performance and Behavior in Adverse Weather Conditions: Microsimulation and Variable Speed Limit Implementation of the SHRP2 Naturalistic Driving Study Results - Phase 3
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 study addresses the limited understanding of driver behavior and performance during adverse weather conditions at a trajectory level. While previous research relied on aggregate traffic and weather data, this work utilizes disaggregate data from the Second Strategic Highway Research Program (SHRP2) Naturalistic Driving Study (NDS) to investigate safety and operational impacts. The research aims to fill existing gaps by analyzing human behavior in conjunction with weather, traffic, and roadway geometry, ultimately supporting the development of Weather-responsive Active Traffic Management (ATM) systems and Intelligent Transportation Systems (ITS) countermeasures. The methodology involved a multi-phase approach using SHRP2 NDS data. First, parametric and non-parametric models, including ordinal logit regression, association rules mining, and K-means clustering, were employed to analyze behavioral factors such as lane-keeping, lane changes, gap acceptance, and speed selection. Second, a unique radar-vision algorithm was developed to process trajectory-level data, continuously predicting driving states and estimating events. Third, machine learning techniques, including Random Forests, Support Vector Machines, and Artificial Neural Networks, were used to detect and predict lane change maneuvers through data fusion. Fourth, surrogate measures of safety were identified using binary logistic regression and non-parametric models. Fifth, cost-effective detection systems for real-time weather identification (snow, fog, and general adverse conditions) were developed using image processing and deep learning, resulting in the creation of "RoadweatherNet." Finally, these findings were integrated into weather-based microsimulation models using VISSIM to assess the performance of Variable Speed Limit (VSL) systems on Wyoming freeways. Key findings revealed specific behavioral adjustments in adverse weather, such as altered lane-keeping abilities, modified lane-changing durations based on driver aggressiveness, and distinct speed selection patterns. The radar-vision algorithm successfully processed trajectory data to define continuous driving segments and classify events. Machine learning models demonstrated high accuracy in detecting and predicting lane changes, with feature selection algorithms identifying critical kinematic and environmental inputs. The study also established reliable indicators for near-crashes and surrogate safety measures. The developed RoadweatherNet provided effective real-time detection of road-surface weather conditions using deep learning. Microsimulation results indicated that calibrated driving behavior parameters significantly affected speed-flow and speed-density relationships, validating the integration of weather-specific behaviors into simulation environments. The significance of this work lies in its contribution to the reliability and effectiveness of safety and operational countermeasures. By unlocking the potential of NDS data for adverse weather research, the study provides a foundation for next-generation traffic management, including Connected Vehicle Technologies and cooperative automated transportation. The integration of human behavior models into microsimulation tools allows for more accurate assessment of ITS strategies like VSL systems, offering unprecedented opportunities to improve roadway safety and operational efficiency under adverse weather conditions.
Key finding
The study successfully developed and validated machine learning-based models for weather detection, lane change prediction, and surrogate measure of safety identification, which were then used to calibrate microsimulation models demonstrating the potential of variable speed limit systems to improve safety in adverse weather.
Methodology
mixed_methods
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 (6 acquisition events logged).
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | rosap | — | — | 2 | 2026-05-23 |
| 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.
- speed choice
- weather rain fog snow
- traffic density
- naturalistic crash near crash
- work zones
- speed distance perception
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).
- Methodological Resource: dataset resource, validation psychometrics
- Theoretical Contribution: computational model