Integrating Shared Autonomous Vehicles into Existing Transportation Services: Evidence from a Paratransit Service in Arlington, Texas

This study investigates the potential benefits of integrating shared autonomous vehicles and an existing transportation system by exploring a recently initiated project that integrates autonomous vehicles (AVs) with an existing on-demand ridesharing service, Via, in Arlington, Texas. We first identified the spatial patterns of the ridership on a localized scale, using geographically weighted regression (GWR) for the existing paratransit service, Handitran. Assuming that the existing ridership will be combined in the future with shared autonomous vehicles, we looked at integration options, based on the spatial patterns of supply and demand and payment options for the riders. The results suggest that the paratransit service, Handitran, is currently used by a small proportion of the eligible population, whose travel patterns differ based on their age. For instance, younger users usually ride Handitran for traveling to work, recreational activities, and routine chores, while senior riders often use the service for medical and recreational trips. The results of the GWR model indicate that the major determinants of Handitran usage are the population’s percentage of older adults, racial distribution, and household vehicle ownership; the coefficients of these factors vary across the city. Hot-spot analyses’ results reveal that integrating the services will improve the efficiency of the existing transportation system by responding to the excess rider demand, particularly in the downtown area. Finally, we describe the implications of implementing policies for AV integration for cities, service providers, and other stakeholders and suggest future research topics.