Development of systemwide pedestrian safety performance function using stratified random sampling and a proxy measure of pedestrian exposure

Predicting pedestrian-involved crash severity using inception-v3 deep learning model

This research leverages a novel deep learning model, Inception-v3, to predict pedestrian crash severity using data collected over five years (2016-2021) from Louisiana. The final dataset incorporates forty different variables related to pedestrian attributes, environmental conditions, and vehicular specifics. Crash severity was classified into three categories: fatal, injury, and no injury. The Boruta algorithm was applied to determine the importance of variables and investigate contributing factors to pedestrian crash severity, revealing several associated aspects, including pedestrian gender, pedestrian and driver impairment, posted speed limits, alcohol involvement, pedestrian age, visibility obstruction, roadway lighting conditions, and both pedestrian and driver conditions, including distraction and inattentiveness. To address data imbalance, the study employed Random Under Sampling (RUS) and the Synthetic Minority Oversampling Technique (SMOTE). The DeepInsight technique transformed numeric data into images. Subsequently, five crash severity prediction models were developed with Inception-v3, considering various scenarios, including original, under-sampled, over-sampled, a combination of under and over-sampled data, and the top twenty-five important variables. Results indicated that the model applying both over and under sampling outperforms models based on other data balancing techniques in terms of several performance metrics, including accuracy, sensitivity, precision, specificity, false negative ratio (FNR), false positive ratio (FPR), and F1-score. This model achieved prediction accuracies of 93.5%, 77.5%, and 85.9% for fatal, injury, and no injury categories, respectively. Additionally, comparative analysis based on several performance metrics and McNemar's tests demonstrated that the predictive performance of the Inception-v3 deep learning model is statistically superior compared to traditional machine learning and statistical models. The insights from this research can be effectively harnessed by safety professionals, emergency service providers, traffic management centers, and vehicle manufacturers to enhance their safety measures and applications.

Built-environment risk assessment for pedestrians near bus-stops: a case study in Delhi

Pedestrian safety is a serious concern in the developing nations of the world. It is evident from the past studies that built-environment characteristics near bus-stops, play a crucial role on the frequency and overall share of pedestrian deaths and injuries in road traffic crashes. The present study aims to identify critical built-environment features around vulnerable bus-stops in an Indian city and evaluate the odds of risk that prevails on the safety of pedestrians near bus stops. Hotspot analysis was conducted to finalise 177 bus stop sites within high-crash clusters in the study area. Built-environment attributes considered were based on sidewalk, crosswalk and bus stop conditions near such vulnerable locations. This study includes a video graphic and manual field survey conducted during the day and night-time. Logistic regression was applied to estimate the impact of built environment features on pedestrian crashes. Width and disability friendliness of sidewalks, presence of bus bays and on-street parking have significant impacts on pedestrian fatalities at locations with a higher share of pedestrian fatalities during the day. On the other hand, presence of zebra crossings at junctions, proper bus stop lighting and high sidewalks reduce the odds of pedestrian crashes at night near bus stops.