Roh, Chiwoo (Korea Transport Institute(KOTI)), Cho, Sungeun (Ajou University), Park, Haneul (Ajou University), So, Jaehyun (Ajou University)

A Hybrid Traffic Crash Risk Analysis Modeling of Crash Data-Driven Analysis and Microscopic Simulations

Scheduled for presentation during the Regular Session "S32b-AI-Driven Traffic Monitoring, Safety, and Anomaly Detection" (FR-EA-T32), Friday, November 21, 2025, 14:50−14:50, Southport 2

2025 IEEE 28th International Conference on Intelligent Transportation Systems (ITSC), November 18-21, 2025, Gold Coast, Australia

This information is tentative and subject to change. Compiled on October 18, 2025

Abstract

This study proposes a hybrid framework integrating macroscopic crash frequency modeling with microscopic surrogate safety validation to identify high-risk traffic zones. Machine learning models, including Random Forest, XGBoost, and Deep Neural Networks (DNN), were trained separately for expressways and urban expressways using roadway and traffic attributes. High-risk segments were selected based on SHapley Additive exPlanations (SHAP) values, ensuring interpretable and data-driven selection. Microscopic simulations were performed using the Simulation of Urban Mobility (SUMO) platform, and surrogate safety measures—Time to Collision (TTC) and Deceleration Rate to Avoid Crash (DRAC)—validated the elevated risk levels. Results confirmed that SHAP-identified high-risk segments exhibited significantly higher critical event rates than normal segments. Driving behavior analysis further revealed distinct risk patterns: aggressive longitudinal control on expressways and frequent lane changes on urban expressways. The proposed framework offers a robust and interpretable method for proactive traffic safety management, and future studies will focus on integrating environmental and driver-related factors to enhance generalizability.