Dynamic traffic demand has been a longstanding challenge for the conventional transit system design and operation. The recent development of autonomous vehicles (AVs) makes it increasingly realistic to develop the next generation of transportation systems with the potential to improve operational performance and flexibility. In this study, we propose an innovative transit system with autonomous modular buses (AMBs) that is adaptive to dynamic traffic demands and not restricted to fixed routes and timetables. A unique transfer operation, termed as “in-motion transfer”, is introduced in this paper to transfer passengers between coupled modular buses in motion. A two-stage model is developed to facilitate in-motion transfer operations in optimally designing passenger transfer plans and AMB trajectories at intersections. In the proposed AMB system, all passengers can travel in the shortest path smoothly without having to actually alight and transfer between different bus lines. Numerical experiments demonstrate that the proposed transit system results in shorter travel time and a significantly reduced average number of transfers. While enjoying the above-mentioned benefits, the modular, adaptive, and autonomous transit system (MAATS) does not impose substantially higher energy consumption in comparison to the conventional bus system.

动态交通需求一直是传统交通系统设计和运营面临的长期挑战。自动驾驶汽车 (AV) 的最新发展使得开发具有提高运营性能和灵活性潜力的下一代交通系统变得越来越现实。在这项研究中，我们提出了一种具有自主模块化巴士 (AMB) 的创新交通系统，该系统适应动态交通需求，不受固定路线和时间表的限制。本文介绍了一种独特的换乘操作，称为“动态换乘”，用于在运动的耦合模块化巴士之间换乘乘客。开发了一个两阶段模型，以促进动态换乘操作，以优化设计交叉路口的乘客换乘计划和 AMB 轨迹。在提议的 AMB 系统中，所有乘客都可以在最短的路径上顺利行驶，而无需在不同的公交线路之间实际下车和换乘。数值实验表明，所提出的交通系统导致更短的旅行时间和显着减少的平均换乘次数。在享受上述优势的同时，模块化、自适应和自主交通系统 (MAATS) 与传统公交系统相比并没有显着增加能耗。