Optimizing Last-Mile Public Transit Services by Leveraging Modular Autonomous Vehicles

Abstract

Improving the quality of first-mile and last-mile (FMLM) travel services can enhance users’ willingness to use public transportation. However, many current FMLM services often prioritize adapting to public transit over considering the passenger experience. New forms of shared mobility can provide added flexibility to FMLM systems when carefully designed; they enhance the feasibility of improving the welfare of users and operators. With the goal of reducing operational costs, modular autonomous vehicles (MAVs) can redefine vehicle routes and allow passengers to transfer between vehicles during the journey. This study evaluated the effects of introducing MAVs to build “seamless” mobility service in the context of FMLM, it involves coordinating between random boarding mechanisms and en route transfer mechanisms. An agent-based simulation was applied to assess the daily operation of the last-mile system using actual taxi trip data from New York City. Three different service designs were tested with different mechanisms: (1) single passenger service design; (2) carpooling service design; and (3) en route transfer service design. The results showed that, combining random boarding and en route transfer reduced vehicle demand (−29.34%) and lowered the total costs for users and operators (−14.09%). Meanwhile, passengers do not need to wait for the vehicle to arrive, and the average waiting time inside the vehicle is only 4 s. Additionally, the authors found a theoretical upper limit on the proportion of vehicles and passengers participating in en route transfer. This research can provide insights for the operation of MAVs and last-mile public transit services. This study provides actionable insights for transit planners and service operators looking to improve FMLM travel. By integrating modular autonomous vehicles with innovative service designs—such as random boarding combined with en route transfers—the research demonstrates how to create a more efficient and seamless mobility experience. In practical terms, the proposed system can reduce the number of vehicles required by nearly 30% and cut total operating costs by over 14% while keeping passenger waiting times almost negligible. These findings suggest that adopting such systems could significantly enhance service quality and lower expenses in urban transportation networks. Although the study uses advanced simulation techniques with real taxi data from New York City, its conclusions are directly applicable to cities aiming to modernize their transit services. Transit agencies can use these results to guide the deployment of MAVs and design last-mile services that improve overall efficiency, reduce congestion, and boost passenger satisfaction.