Policies and Strategies for Evolving and Managing Automated Mobility

在政府和行业的推动下，自动车辆(AV)的部署正在迅速临近，政府正在放松法律，允许自动车辆在高速公路上运营，制造商和移动服务提供商都在大力投资于该技术及其应用的开发。AVS有望极大地提高现有交通系统的效率、安全性和便利性。然而，所有这些好处都取决于增值服务的市场渗透率是否足够高。在市场占有率较低的情况下，自动驾驶对提高运输系统效率的作用不大。更糟糕的是，及早部署AVs甚至可能会损害效率。过渡期预计会很长。如果我们能缩短它，AVS承诺的巨大好处就可以更快地实现。因此，这笔赠款旨在研究激励政策和创新的交通管理战略，以促进自动驾驶系统的开发和部署，以在自动驾驶系统部署的整个期间最大限度地发挥社会效益。具体地说，激励政策将培育自动驾驶汽车市场并加速其采用，而创新的交通管理计划旨在更好地利用自动驾驶汽车在交通流中的作用，并促进高占有率的移动服务，以在给定的市场份额下最大限度地发挥自动驾驶汽车的好处。激励政策和交通管理方案之间的协同作用可能会导致自动驾驶系统部署的螺旋式上升，特别是在自动驾驶系统对提高效率影响很小甚至是负面影响的初始部署时间。这笔赠款将为政府机构提供及时的支持，以更好地了解自动增值服务的影响和影响，并为其发展和部署提供指导。这笔助学金将涉及所有级别的学生和传统上代表人数较少的学生，并为新兴自动移动课程提供新的材料和案例研究。研究成果将通过各种媒体广泛传播。研究将分两个主要方向进行。首先是研究诸如税收抵免、补贴和对自动增值服务的优惠待遇等政策，这些政策可以激励自动增值服务的部署从较低的渗透率到较高的渗透率，以在整个规划范围内最大限度地发挥自动增值服务的好处。作为第一个推力的一部分，我们计划了一个连续时间的委托代理框架，在这个框架中，政府是提供激励政策的委托人，制造商是制定AVS零售价格的代理人。通过考虑这两个主体之间的相互作用，得到了最优激励机制。第二，我们将发展创新计划，以改善交通运输系统的社会福利。这些方案可能涉及基于间隔的拥堵定价，以惩罚原型AVs的过度间隔，或者涉及基于占用的定价，以促进高占有率机动性。同时，将开发一种分布式控制方案，利用交通流中的AVs作为控制执行器，在整个交通网络中分配交通需求，以减少拥堵。由于来自推力1的激励政策和来自推力2的交通管理战略的影响相互交织，在这两个推力中开发的模型的迭代应用可以规定一种明智的行动方案，以发展和管理自动化移动性。如果成功，这笔赠款将做出三个关键贡献：用于激励政策分析的连续时间委托代理方法，基于数据驱动的间隔或占有率的拥堵收费，以及用于管理交通流量的AVs的分布式控制。具体地说，我们将激励政策设计描述为不对称信息下的非零动态Stackelberg博弈，这是一类用传统方法很难解决的问题。我们提供了一个创新的框架来分离决策过程，使问题在数学上易于处理。我们的研究还提出了拥堵定价理论，为基于车辆轨迹和占有率的细粒度定价方案的设计提供了一个新的框架。它将模式从基于模型的定价转变为更多地由数据驱动。AVS用于管理网络流量的分布式控制丰富了交通控制文献，并为分布式、可扩展和有效的参与式交通控制提供了理论。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

The deployment of automated vehicles (AVs) is rapidly approaching with a push from governments who are relaxing laws to allow AVs to operate on highways, and industry, both manufacturers and mobility service providers, who are heavily investing in the development of the technology and its applications. AVs are expected to tremendously enhance the efficiency, safety and convenience of existing transportation systems. However, all these benefits hinge on the level of market penetration of AVs being sufficiently high. At low market shares, AVs exert little impact on enhancing transportation system efficiency. Worse yet, early deployment of AVs may even compromise the efficiency. The transition period is expected to be lengthy. If we can shorten it, the tremendous benefits promised by AVs can be realized sooner. This grant thus sets out to investigate incentivizing policies and innovative traffic management strategies to promote the development and deployment of AVs to maximize the social benefit over the entire duration of the AV deployment. Specifically, incentivizing policies will nurture the AV market and accelerate their adoption while innovative traffic management schemes aim to better utilize AVs in the traffic stream and promote high-occupancy mobility services to maximize the benefits of AVs at a given market share. The synergies between incentivizing policies and traffic management schemes may create an upward spiral for the AV deployment and particularly reduce the duration of initial deployment where AVs exert little or even negative impact on enhancing efficiency. This grant will provide timely support for government agencies to better understand the impacts and implications of AVs and provides guidance on their development and deployment. This grant will involve students at all levels and traditionally underrepresented students, and offer fresh materials and case studies for courses on emerging automated mobility. Research results will be broadly disseminated through a variety of media.The research will be conducted in two main thrusts. The first is the study of policies like tax credits, subsidies, and preferential treatments for AVs etc., which can incentivize the deployment of AVs from lower to higher penetration rates to maximize the benefits of AV deployment throughout a planning horizon. As part of the first thrust, we plan a continuous time principal-agent framework in which the government is the principal who offers an incentive policy, and the manufacturer is the agent who sets the retail prices of AVs. The optimal incentive mechanism is obtained by considering the interplay between these two entities. In the second thrust, we will develop innovative schemes to improve the social welfare of the transportation system. These schemes could involve headway-based congestion pricing for penalizing excessive headways of prototype AVs or occupancy-based pricing for promoting high occupancy mobility. In parallel, a distributed control scheme will be developed to use AVs in the traffic stream as control actuators to distribute traffic demand across the transportation network to reduce congestion. As the impacts of incentivizing policies from Thrust 1 and traffic management strategies from Thrust 2 are intertwined, an iterative application of the models developed in both thrusts can prescribe a wise course of actions to evolve and manage automated mobility. If successful, this grant makes three critical contributions: a continuous time principal-agent approach for incentive policy analysis, data-driven headway- or occupancy-based congestion pricing and distributed control of AVs for managing traffic flow. Specifically, we formulate the incentivizing policy design as a non-zero dynamic Stackelberg game under asymmetric information, a class of problems extremely difficult to solve using traditional techniques. We offer an innovative framework to decouple the decision-making processes to make the problem mathematically tractable. Our research also advances the theory of congestion pricing by providing a new framework of designing fine-grained pricing schemes based on vehicle trajectory and occupancy. It shifts the paradigm from model-based pricing to be more data-driven. The distributed control of AVs for managing network traffic flow enriches the traffic control literature and theorizes participatory traffic control that is distributed, scalable and effective.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.