Developing a Taxonomy of Human Errors and Violations That Lead to Crashes

Khattak, Asad; Ahmad, Numan; Wali, Behram; Dumbaugh, Eric · 2019 · ROSA P / Collaborative Sciences Center for Road Safety

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 predominance of human factors in traffic crashes, which account for over 90% of incidents, by developing a systematic Taxonomy of Driver Errors and Violations (TDEV). Motivated by the limitations of subjective police crash reports, the research aims to provide an objective, nuanced understanding of pre-crash driver behaviors and their interaction with roadway and built environments. The study seeks to quantify the contributions of human, vehicle, and environmental factors and explore how specific errors vary across different land-use contexts to inform safety countermeasures and automated vehicle development. The researchers utilized data from the SHRP2 Naturalistic Driving Study (NDS), which provides high-resolution, real-world sensor and video data from instrumented vehicles. The analysis focused on a subset of 9,593 trips, comprising 673 crashes, 1,331 near-crashes, and 7,589 baseline driving segments. The methodology involved classifying driver errors based on a perception-reaction framework into recognition errors, decision errors, performance errors, and intentional violations. A safety matrix was constructed to assess the simultaneous contributions of human, vehicle, and roadway factors. Additionally, path analysis was employed to uncover direct and indirect relationships between built-environment characteristics, specific driver errors, and crash propensity. The results indicate that human errors and violations contributed to 93% of observed crashes, while roadway and vehicle factors contributed to 17% and 1%, respectively. Recognition errors were the most prevalent, occurring in 39% of crashes, followed by decision errors in 34%. These two error types were particularly frequent (approximately 39% each) in areas with business or industrial structures. While recognition errors were the most common in both crashes and near-crashes, performance errors such as weak judgment showed a strong correlation with actual crash occurrence. Path analysis revealed that urban environments, due to their complexity, are associated with a 7.66% higher chance of crashes. These environments induce more recognition errors, which in turn increase crash likelihood by an additional 3.40%, resulting in a total effect of 11.06%. Similar mediating effects of recognition and decision errors were observed in school, playground, and construction zones. The significance of this work lies in its provision of a detailed, evidence-based taxonomy of driving errors derived from naturalistic data, overcoming the subjectivity of traditional crash reports. The findings highlight the critical role of recognition and decision errors, particularly in complex built environments, suggesting that locality-specific countermeasures are necessary. Furthermore, the study offers valuable insights for the development of connected and automated vehicles by identifying complex "fringe case" situations and error pathways that automated systems must be designed to handle. The research underscores the need for improved crash investigation methods that account for the interplay between driver behavior and environmental context.

Key finding

Human errors and violations contributed to 93% of observed crashes, with recognition and decision errors being the most common types.

Methodology

naturalistic

Sample size: 673

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).

StageOutcomeToolModelPromptAttemptsCompleted
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 partial 2 2026-06-10

Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified_with_issues.

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).