Behavior Analysis of Traffic Accidents with High Fidelity Driving Simulator

Yang, Fan; Zhang, Kan; Sun, Xianghong · 2009 · OpenAlex-citations

DOI: 10.1061/41039(345)532

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 critical issue of traffic accidents, noting that approximately 74% of incidents in China are attributed to driver failure rather than road or vehicle conditions. The primary motivation was to develop an accurate system for testing driving skills to identify unskilled drivers and potentially reduce accident rates through targeted training. The researchers utilized a high-fidelity, immersive driving simulator (XuanAi QJ4B) to create a controlled environment that saves time and costs while avoiding the risks associated with real-world testing. The experimental design involved 19 male military drivers with varying levels of experience, ranging from less than one year to over five years. Participants completed 20 distinct driving routes derived from five real-world road types in China: highway, city, country, mountain, and mixed sections. Each road type was tested under two conditions: driving without traffic and driving with heavy traffic plus 2–5 emergency scenarios (e.g., pedestrians crossing, blocked roads). The simulator software recorded vehicle and driver data every 50 milliseconds, including task completion time, collisions, lane deviations, and instances of driving out of the road. After the tests, participants peer-evaluated each other’s driving skills, allowing the researchers to categorize drivers into good, middle, and bad skill groups. The results indicated that task completion time and driving experience were significant differentiators among skill levels, with statistically significant differences found in completion times ($F = 5.674, p = 0.014$) and experience years ($\chi^2 = 9.656, p = 0.008$). The frequency of driving out of the road also showed a trend toward significance ($p = 0.093$), while collision counts, lane deviations, and ignition events did not significantly distinguish skill levels. Analysis of specific road conditions revealed that city roads with traffic, mountain roads with traffic, and mixed mountain-city roads without traffic were particularly effective at differentiating driver performance based on completion time. In total, seven specific road configurations were identified as effective for skill assessment. The study concludes that a system utilizing these seven effective roads and three key variables (completion time, out-of-road frequency, and experience) can accurately evaluate driving skills. The findings suggest that complex environments, such as city roads with heavy traffic and rugged mountain roads, are superior for distinguishing between skilled and unskilled drivers. This simulator-based assessment tool has practical implications for driving schools and licensing organizations, offering a method to identify training needs and improve driver competence, thereby contributing to traffic safety. However, the authors note that future studies require larger participant samples to enhance the system's accuracy.

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.

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.

• simulator training transfer
• simulator validity fidelity
• human error taxonomy

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: validation psychometrics, tool software
• Theoretical Contribution: computational model