Autonomous mobility-on-demand in a rural area: Calibration, simulation and projection based on real-world data

In a pilot project, the operation of autonomous vehicle(s) (AV) in an Mobility-on-Demand (MoD) system without fixed routes is tested on public streets in Kelheim Germany, for the first time. The autonomous service is launched in September 2022 and is integrated into an existing service with human-driven vehicles. This work presents an agent-based model for the Kelheim county that is calibrated based on real demand data from the MoD operation. The model is used for demand prediction and impact assessment of several configurations of the AV service. The results suggest that the conventional service is not negatively affected but rather complemented by the AV service. Furthermore, the results show that with the expansion of the AV fleet, planned for 2023, the corresponding service area should also be expanded. This could lead to a higher efficiency in the AV operation compared to the conventional service. Similar to other findings in the literature, it is observed that the electric AV mainly replace trips with sustainable transport modes. In future, the results will be validated and the mode choice model will be further improved with the help of real AV demand data that will become available during the course of the project. • Open-source activity- and agent-based transport model for Kelheim County, Germany. • Real data-based calibration of a Mobility-on-Demand service. • Decision support for configuring a Mobility-on-Demand service with automated vehicles.