Modeling Wrong-way Driving (WWD) crash severity on arterials in Florida

Identification of high-risk roadway segments for wrong-way driving crash using rare event modeling and data augmentation techniques

Wrong-Way Driving (WWD) crashes are relatively rare but more likely to produce fatalities and severe injuries than other crashes. WWD crash segment prediction task is challenging due to its rare nature, and very few roadway segments experience WWD events. WWD crashes involve complex interactions among roadway geometry, vehicle, environment, and drivers, and the effect of these complex interactions is not always observable and measurable. This study applied two advanced Machine Learning (ML) models to overcome the imbalanced dataset problem and identified local and global factors contributing to WWD crash segments. Five years (2015-2019) of WWD crash data from Florida state were used in this study for WWD model development. The first modeling approach applied four different hybrid data augmentation techniques to the training dataset before applying the XGBoost classification algorithm. In the second model, a rare event modeling approach using the Autoencoder-based anomaly detection method was applied to the original data to identify WWD roadway segments. A third model was applied based on the statistical method to compare the performance of ML models in predicting the WWD segments. The performance comparison of the adopted models showed that the XGBoost model with the Adaptive Synthetic Sampling (ADASYN) method performed best in terms of precision and recall values compared to the autoencoder-based anomaly detection method. The best-performing model was used for the feature analysis with an interpretable machine-learning technique. The SHapley Additive exPlanations (SHAP) values showed that high-intensity developed land use, length of roadway, log of Annual Average Daily traffic (AADT), and lane width were positively associated with WWD roadway segments. The results of this study can be used to deploy WWD countermeasures effectively.

Characterizing the effects of contributing factors on crash severity involving e-bicycles: a study based on police-reported data

Mitigating e-bicycle crash occurrence has become a great challenge across the world. It is of paramount importance for improving traffic safety to characterize the relationship between e-bicycle crash injury severities and contributing factors. This study positions itself at clarifying the roles of the factors in e-bicycle crashes from time, space, road, environment, rider and object characteristics. The partial proportional odds (PPOs) model as well as its elasticity analysis was employed to identify the influences based on 15,138 police-reported e-bicycle crashes in Shangyu District of Shaoxin City, China. The results evidenced that there were 12 factors having significant effects. Especially, the results emphasized the greater influences of rider gender, age, object hit and road type. Their maximum of the absolutes of elasticities was greater than 24%. Increased crash severity was associated with females, younger riders, and higher speed collisions. However, the remaining significant variables had minor effects (no more than 10%). The findings provide meaningful insights for advancing e-bicycle development, when making related policies and prioritizing safety countermeasures.

Investigating the Risk Factors Associated with Injury Severity in Pedestrian Crashes in Santiago, Chile

Pedestrians are vulnerable road users that are directly exposed to road traffic crashes with high odds of resulting in serious injuries and fatalities. Therefore, there is a critical need to identify the risk factors associated with injury severity in pedestrian crashes to promote safe and friendly walking environments for pedestrians. This study investigates the risk factors related to pedestrian, crash, and built environment characteristics that contribute to different injury severity levels in pedestrian crashes in Santiago, Chile from a spatial and statistical perspective. First, a GIS kernel density technique was used to identify spatial clusters with high concentrations of pedestrian crash fatalities and severe injuries. Subsequently, partial proportional odds models were developed using the crash dataset for the whole city and the identified spatial clusters to examine and compare the risk factors that significantly affect pedestrian crash injury severity. The model results reveal higher increases in the fatality probability within the spatial clusters for statistically significant contributing factors related to drunk driving, traffic signage disobedience, and imprudence of the pedestrian. The findings may be utilized in the development and implementation of effective public policies and preventive measures to help improve pedestrian safety in Santiago.