Motorcycle crashes often result in severe injuries on roads that affect people's lives physically, financially, and psychologically. These injuries could be notably harmful to drivers of all age groups. The main objective of this study is to investigate the risk factors contributing to the severity of crash injuries in different age groups. This Objective is achieved by developing accurate machine learning (ML) based prediction models. This research examines the relationship between potential risk factors of motorcycle-associated crashes using (ML) and Shapley Additive explanations (SHAP) technique. The SHAP technique further helped interpreting ML methods for traffic injury severity prediction. It indicates the significant non-linear interactions between dependent and independent variables. The data for this study was collected from the Provincial Emergency Response Service RESCUE 1122 for the Rawalpindi region (Pakistan) over three years (from 2017 to 2020). The Synthetic Minority Oversampling Technique (SMOTE) is employed to balance injury severity classes in the pre-processing phase. The results demonstrate that age, gender, posted speed limit, the number of lanes, and month of the year are positively associated with severe and fatal injuries. This research also assesses how the modeling framework varies between the ML and classical statistical methods. The predictive performance of proposed ML models was assessed using several evaluation metrics, and it is found that Catboost outperformed the XGBoost, Random Forest (RF) and Multinomial Logit (MNL) model. The findings of this study will assist road users, road safety authorities, stakeholders, policymakers, and decision-makers in obtaining substantial and essential guidance for reducing the severity of crash injuries in Pakistan and other countries with prevailing conditions.
Stakeholders expect emerging urban air mobility (UAM) services that use electric vertical takeoff and landing (eVTOL) aircraft to revolutionize transportation systems. However, to be effective, eVTOL facilities, known as vertiports, must seamlessly integrate with the existing multimodal transportation infrastructure. This research analyzes how to best integrate vertiports with the existing public transit network of a city, with San Francisco in the United States as a case study. This study developed a composite optimization model to identify the locations for an optimum number of vertiports within each of the eleven (11) supervisorial districts of San Francisco that would minimize the network distance to its bus stops. The results revealed that 38 locations for vertiports covered 3245 bus stops with an average walk time close to 10 min. Walking 10 min from each vertiport would cover 41.6% of the total road network length and 49.8% of all bus stop locations. Power-law regressions of the average distances as a function of the number of vertiports will help planners to make informed decisions about the optimal vertiport placements in each district by identifying diminishing returns in travel-time savings, and adjustments that consider income and population as additional demand factors.
Traditional hot spot and cluster analysis techniques based on the Euclidean distance may not be adequate for assessing high-risk locations related to crashes. This is because crashes occur on transportation networks where the spatial distance is network-based. Therefore, this research aims to conduct spatial analysis to identify clusters of high- and low-risk crash locations. Using vulnerable road users’ crash data of San Francisco, the first step in the workflow involves using Ripley’s K-and G-functions to detect the presence of clustering patterns and to identify their threshold distance. Next, the threshold distance is incorporated into the Getis-Ord Gi* method to identify local hot and cold spots. The analysis demonstrates that the network-constrained G-function can effectively define the appropriate threshold distances for spatial correlation analysis. This workflow can serve as an analytical template to aid planners in improving their threshold distance selection for hot spot analysis as it employs actual road-network distances to produce more accurate results, which is especially relevant when assessing discrete-data phenomena such as crashes.
26 November 2024 in Accident Analysis & Prevention