Search

Road traffic safety is confronted with unique challenges worldwide due to the varied road network, high population density, and the rapid increase in vehicle numbers. This necessitates an effective evaluation of traffic safety at the road segment level to improve traffic safety management. Existing safety indicators, such as crash frequency (which lacks real-time availability) and Surrogate Safety Measures (focused on micro-level interactions), fall short in quantifying the overall safety status of road segments. To overcome these limitations, we introduce the concept of entropy, typically used to measure disorder in complex systems, into road traffic safety assessment. Our research presents a novel macroscopic safety indicator named 'velocity entropy,' along with its calculation method, which involves discretization and the Parzen window technique. The efficacy of velocity entropy is validated using the UTE dataset, a high-fidelity trajectory dataset from China, showing a more significant correlation with traffic conflicts than other macroscopic traffic properties. Notably, velocity entropy can track the temporal and spatial evolution of traffic safety, providing insights into the dynamic development of traffic risks. Overall, velocity entropy offers an innovative, practical, and theoretically meaningful tool for assessing and predicting road safety. [ABSTRACT FROM AUTHOR]

Addressing the global challenge of traffic crashes necessitates transcending traditional statistical models, which often fail to fully capture the interactions between factors causing crashes. This oversight restricts the predictive accuracy and adaptability of current methodologies. Additionally, there is a notable gap in research that examines the links between behavior-cause relationships and crash injury severity. Our study deploys Natural Language Processing (NLP) and Frequent Pattern (FP) growth algorithm to mine crash narratives for behavior-cause connections, combines with the predictive strength of eXtreme Gradient Boosting (XGBoost) and the interpretative clarity offered by SHapley Additive exPlanations (SHAP), our approach not only predicts crash injury severity with satisfactory precision but also explains the influence of specific behavior-cause and environment conditions on crash outcomes. The integration of NLP and XGBoost, complemented by SHAP insights, has shown promising results with an accuracy of 0.79, outperforming traditional discrete choice models and competes closely with other machine learning approaches, including Support Vector Machines, Random Forest, Categorical Boosting (CatBoost), and Light Gradient Boosting Machine (LightGBM). Through detailed textual analysis and the establishment of a behavior-cause matrix, identifying five broad crash causes linked to 141 specific crash cause with behaviors, we uncover critical patterns such as the prominence of distracted driving in severe crashes. This comprehensive approach not only fills a critical research gap by linking behavior-cause relationships with injury severity but also sets the stage for developing targeted interventions to enhance road safety. • Merges NLP and XGBoost model for crash injury severity prediction. • Identifies crash causes and related driver behaviors using NLP and FP-growth. • Aids in developing traffic safety interventions based on SHAP interpretability. [ABSTRACT FROM AUTHOR]