Joint Analysis of E-Bike Crash Types and Injury Severity at Different Locations Using a Copula-Based Approach

Abstract

E-bikes stand out as one of the most energy-efficient forms of motorized transport. However, a significant increase in e-bike traffic crashes has raised concerns about the factors influencing crash severity. Moreover, the specific impacts of different locations on e-bike crash injury severity remain elusive. This study investigated the factors influencing e-bike crash severity with a focus on the diverse effects of crash location and crash type. A copula-based joint framework was applied to capture the dependency structure between crash type and severity. The exogenous variables with location effects were captured by the interaction variables. This structure allowed exploration of how these variables and their location influenced crash severity in the pooled dataset for each crash type. We utilized data from the Meituan Shared E-bike Platform, comprising 1,150 crash records. An exhaustive set of independent variables, including environment conditions, locations, e-bike cyclist characteristics, and actions was considered. These copula parameters suggested the completely distinct nature of the relationship between crash types and their corresponding severities. Overall, the determinants of injury severities were found to vary between different crash types. Key common indicators such as heavy rain affected the severity of the rollover and motor vehicle crash in opposite ways owing to different stopping distances, mass, and kinetic energy. Younger riders tended to experience less severe outcomes in rollover crashes but were more likely to be involved in motor vehicle crashes. This study contributes to guiding transportation agencies in developing effective countermeasures to promote the safety of e-bikes.