EAGER-DynamicData: Dynamic Data Driven Distributed Simulation for Transportation System Applications on Emerging Computing Platforms

EAGER-DynamicData：新兴计算平台上运输系统应用的动态数据驱动分布式仿真

• 批准号: 1462503
• 负责人: Richard Fujimoto
• 金额: $ 22万
• 依托单位: Georgia Tech Research Corporation
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2015
• 资助国家: 美国
• 起止时间: 2015-09-15 至 2018-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1462503&HistoricalAwards=false
• 关键词: EAGER; DynamicData; Dynamic; Data; Driven

Major changes in urban transportation systems are now underway due to innovations occurring along several different dimensions. Smart cars, autonomous and semi-autonomous vehicles, electric vehicles, compact energy-efficient micro-vehicles, and increased deployment of vehicle-to-vehicle and vehicle-to-infrastructure communications are revolutionizing the vehicle fleet. New types of transportation services are emerging such as short-term vehicle rentals and commercial crowd-sourced taxi services that are increasing shared use and lessening reliance on private vehicle ownership. New sources of data and vast amounts of information concerning the transportation network itself are becoming available including crowd-sourced data provided by citizens through mobile devices and data collected from surveillance systems including unmanned aerial vehicles (UAVs). At the same time advances in computing and communications are enabling new, mobile high performance computing platforms that can be embedded within the transportation system. It is not well understood how these platforms can be best exploited in emerging transportation system applications to help address long-standing issues concerning safety, congestion, resource consumption, and pollution that significantly degrade the quality of life in the U.S. and throughout the world. New computational methods are needed to maximally exploit these new technologies to create the reliable, resilient, and efficient transportation systems demanded by modern cities.This project will explore synergies and challenges arising from the integration of three largely separate areas dynamic data analytics, embedded distributed simulations, and power-aware computing on emerging high performance mobile computing platforms for transportation system applications. The context for this research is power-tunable platforms termed micro-clusters that are small closely coupled multi-node computer systems assembled from mobile computing components. On a larger scale, distributed networked micro-clusters will form the computation core for many dynamic data driven application system (DDDAS) deployments in the future. This project seeks to develop deep understandings of key principles concerning the design and operation of micro-clusters executing integrated dynamic data analysis algorithms and embedded distributed simulations for emerging transportation system applications. A micro-cluster hardware testbed will be created as a key element of this project for experimental research in power aware dynamic data-driven distributed simulation and used for benchmarking. The results of this study will be used to develop a research agenda and follow-on research to develop approaches and techniques to minimize energy consumption while maintaining effectiveness of scalable DDDAS deployments. Beyond transportation system applications, platforms such as micro-clusters will become commonplace and widely used in the operation of real-world systems for a variety of applications besides transportation in areas such as manufacturing, logistics, and telecommunications. Project results will be incorporated into undergraduate and graduate courses in transportation, modeling and simulation and parallel computing. The project will engage underrepresented and K-12 groups in computing and engineering through participation in summer internship programs and summer camps.

由于沿着几个不同的维度发生的创新，城市交通系统正在发生重大变化。智能汽车、自动驾驶和半自动驾驶汽车、电动汽车、紧凑型节能微型汽车以及车对车和车对基础设施通信部署的增加正在彻底改变车队。新型交通服务正在出现，如短期车辆租赁和商业众包出租车服务，这些服务正在增加共享使用，减少对私人车辆所有权的依赖。关于交通网络本身的新的数据来源和大量信息正在变得可用，包括公民通过移动的设备提供的众包数据和从包括无人驾驶飞行器（UAV）在内的监视系统收集的数据。与此同时，计算和通信方面的进步正在使新的、移动的高性能计算平台成为可能，这些计算平台可以嵌入到运输系统中。目前还不清楚如何在新兴的交通系统应用中最好地利用这些平台，以帮助解决长期存在的安全、拥堵、资源消耗和污染问题，这些问题严重降低了美国和世界各地的生活质量。为了最大限度地利用这些新技术来创建现代城市所需的可靠、弹性和高效的交通系统，需要新的计算方法。该项目将探讨在新兴的高性能移动的计算平台上集成动态数据分析、嵌入式分布式仿真和电力感知计算这三个基本独立的领域所带来的协同效应和挑战。这项研究的背景是功率可调的平台称为微集群，是小型紧密耦合的多节点计算机系统组装从移动的计算组件。在更大的规模上，分布式网络微集群将成为未来许多动态数据驱动应用系统（DDDAS）部署的计算核心。该项目旨在深入了解有关微集群的设计和操作的关键原则，执行集成的动态数据分析算法和嵌入式分布式模拟新兴的交通系统应用。本项目将建立一个微集群硬件测试平台，用于功耗感知动态数据驱动分布式仿真的实验研究，并用于基准测试。这项研究的结果将用于制定研究议程和后续研究，以开发方法和技术，在保持可扩展DDDAS部署有效性的同时最大限度地减少能耗。除了运输系统应用之外，微集群等平台将变得司空见惯，并广泛用于现实世界系统的操作中，用于制造、物流和电信等领域的运输之外的各种应用。项目成果将被纳入本科和研究生课程的运输，建模和模拟和并行计算。该项目将通过参加暑期实习计划和夏令营，吸引代表性不足的K-12群体参与计算和工程。