Autonomous vehicles are expected to significantly alter the way that cities approach transportation. Because they do not require an operator, they would likely be significantly cheaper to run, but more expensive to purchase. They are also anticipated to provide significant benefits in efficiency and emissions reduction, both of which may be magnified by also converting from conventional to electric engines. Such Automated Electric Vehicles (AEVs) would even be able to route themselves to charging stations, allowing for still more efficient operations.



While such technology is still a significant distance from being actualized on a meaningful scale, it is possible to consider and proactively address some key business model concerns. This paper, funded by the US Department of Energy Vehicle Technologies Office, seeks to define a fleet-sizing algorithm and to understand optimal charging-system planning.



By optimizing both of these factors, an AEV-based system can minimize the down-time of waiting for vehicles to adequately charge and return to their route, as well as avoiding unnecessary fleet expenditures for redundant vehicles.



The researchers created a network model and performed a mathematical analysis to understand how theoretical AEVs would behave given certain environmental parameters. This knowledge was then applied to a case study with a 25-node network and approximately 15,000 trips per day. Three benchmarks were used for comparison to evaluate the efficacy of the researchers' proposed strategy: MinTime, where AEVs always chose the most direct routes to their destinations, MinOperations, where AEVs chose the route that minimized the operation cost, and NoRelocation, where vehicles were not relocated by an operator throughout the day.



The analysis found that a total of 2,500 AEVs and 300 chargers were needed to meet the demand of 15,000 trips per day. This means that each AEV can complete 6 rides and each charger can fuel 8 AEVs on average.