EAGER: Towards a Paradigm Shift in Data Acquisition for Traffic Control

收集交通数据的过程一直被认为是城市的重要组成部分之一，因为交通数据不仅用于出行者信息系统，而且还作为支持城市街道交通控制的输入。在可预见的未来，我们设想，在一个智能互联的城市中，目前基于道路的交通数据收集基础设施，如环路、微波传感器或激光探测器，将被基于车辆的数据所取代。这种模式的转变将大大节省成本，并产生一种比现有数据更强大的新形式的交通数据。它将使我们的交通管理系统更有效率，使我们的城市更宜居。这个早期概念探索性研究资助（EAGER）项目的目的是加快导致这种范式转变的过程。该项目设想了一个系统，在该系统中，驾驶员通过向系统提供与时间相关的速度和定位信息作为“选票”来与交通控制进行交互，以换取可能的优先服务。通过“投票”让系统用户参与系统控制将引起对尖端系统工程技术的兴趣，特别是对K-12学生。该研究将为理解智能和互联社区中交通管理系统的机会提供重大进展，在智能和互联社区中，交通控制和个体驾驶员之间的互动是可行的。这种交互，目前只有特殊级别的车辆才能使用的功能，将从根本上改变算法开发的传统理论。它可能会影响用户平衡和系统最优理论，支配驾驶员的路径选择行为和最优系统控制在网络水平。该项目探索了一类新的算法，直接利用空间数据，而不是目前使用的现场数据，以支持交通控制算法。新算法与现有算法的不同之处在于，有关交通状态的“定性信息”将被用作交通控制的输入（例如，“很长的队列”而不是实际的队列长度）。这样的信息可以容易地从各个车辆的速度和定位数据推断。单个车辆和系统控制之间的交互将为系统用户提供机会，以告知系统其行程的价值。这种能力将导致需要重新思考用户平衡和其他旅行行为模型。

The process of collecting traffic data has long been considered as one of the important components of a city, because traffic data are used not only for traveler information systems but also as an input to support traffic control on urban streets. In the foreseeable future, we envision that in a smart and connected city the current roadway-based infrastructure for traffic data collection, such as loops, microwave sensors, or laser detectors, will be replaced by vehicle-based data. Such a paradigm shift would result in significant cost savings as well as the generation of a new form of traffic data that is more powerful than the existing one. It will make our traffic management systems more efficient and our cities more livable. The aim of this EArly-concept Grant for Exploratory Research (EAGER) project is to expedite the process that would lead to such a paradigm shift. The project envisions a system in which drivers interact with traffic control by providing their time-dependent speed and positioning information to the system as "votes" in exchange for possible priority service. Involving system users in system control through "voting" will engender interest in cutting-edge systems engineering techniques, especially for K-12 students. The research will provide significant advances in understanding opportunities for traffic management systems in smart and connected communities, where interaction between traffic control and individual drivers is practicable. Such interaction, a feature currently available only to special-class vehicles, would fundamentally change conventional theory for algorithm development. It could impact user-equilibrium and system optimum theories that govern route choice behavior of drivers and optimal system control at the network level. This project explores a new class of algorithms that directly utilize spatial data, as opposed to spot data used currently, to support traffic control algorithms. The new algorithms will differ from existing ones in that "qualitative information" about traffic states will be employed as an input to traffic control (e.g. "a very long queue" instead of the actual queue length). Such information can be readily inferred from speed and positioning data of individual vehicles. The interaction between individual vehicles and system control will provide system users with opportunities to inform the system of the value of their trips. Such capabilities will engender a need to rethink user equilibrium and other models of travel behavior.