1
Return

Zone-level urban crash prediction via multimodal heterogeneous graph transformer: Case studies in Chicago, New York, and San Francisco

delete2026-07-01
delete0
PRE
AI
J
Jiahui Zhao
X
Xuehao Zhai
K
Kequan Chen
Z
Zhibin Li
P
Pan Liu *
DOI：10.1016/j.trc.2026.105844delete
deleteOriginal
deleteOriginal request for help
deleteShare
deleteSave
Abstract

Abstract

En 中文
• A multimodal heterogeneous graph transformer is proposed for zone-level crash prediction. • Heterogeneous road structure, temporal mobility, and static urban context are jointly integrated. • A Spatial Collision-Avoidance Sampling Algorithm is developed to generate comparable spatial catchments. • The framework enables interpretable safety assessment through ablation and severity-cost analysis.
Keywords:
Multimodal heterogeneous graph transformer model
Urban crash analysis
Crash prediction
Road risk level

Journal

Transportation Research Part C-Emerging Technologies cover
Transportation Research Part C-Emerging Technologies
IF:
7.9
Papers:
4.7K
Citations:
3.2W

Organization

I
imperial college london
Scholars:
8.3K
Papers: 3.8K
Citations: 0
S
Southeast University
Scholars:
1.8W
Papers: 7.7K
Citations: 480
Cited Papers

Cited Papers

Citing Papers

Citing Papers