Effects and feasibility of shared mobility with shared autonomous vehicles: An investigation based on data-driven modeling approach

Abstract

Shared mobility is a promising travel mode in the era of autonomous driving. Travelers may no longer own a vehicle, but use shared autonomous vehicle (SAV) services. This study investigates the effects and feasibility of SAV-based shared mobility, which includes ride-sharing and car-sharing strategies, by using a data-driven modeling approach. Ride-sharing indicates that two trips with similar origin–destination information can be combined into a new one, whereas car-sharing indicates that trips can be fulfilled by a single vehicle consecutively. On the basis of license plate recognition data of Langfang, China, this study extracts the urban-scale vehicle travel demand information. Models for ride-sharing and car-sharing are formulated to generate SAV assignment strategies for fulfilling travel demands. This study reveals the prospects and potential problems of SAV-supported shared mobility at different development stages by setting a variety of scenarios with different participation levels of ride-sharing and car-sharing. The minimum fleet size to fulfil the vehicle travel demand in the road network and the total vehicle stock in the urban area are compared under different scenarios, and the effects of shared mobility on vehicle kilometers traveled (VKT) and parking demand are evaluated. This study also reveals the impacts of SAVs in a practical scenario, which is constructed based on an online survey. Results show that ride-sharing and car-sharing with high participation will lead to considerable benefits, i.e., reductions in fleet size, vehicle stock, and parking demand. Under the shared mobility scenario with 100% ride-sharing and car-sharing participation levels, one SAV can potentially replace 3.80 private conventional vehicles in the road network. However, ride-sharing and car-sharing exhibit opposite effects on VKT. Car-sharing alone increases VKT whereas car-sharing and ride-sharing together have the potential to decrease VKT. This study provides insights for understanding the development of shared mobility and facilitating the efficient utilization of SAVs.