面向大规模数据服务的异构融合车联网架构与协议研究

Internet of Vehicles (IoV) is an integrated innovation direction between traditional transportation domain and emerging interdisciplinary domains including mobile computing, artificial intelligence and big data, etc. Based on the highly heterogeneous, highly dynamic and highly distributed features of IoV, this project refines the key scientific problems of heterogeneous network convergence and large-scale data services. The primary contents include the research on converged heterogeneous IoV architecture with key technologies including software-defined network (SDN), network function virtualization (NFV), network slicing and fog computing, etc.; the research on cross-layer cooperated large-scale data dissemination protocols with key technologies including adaptive power control, bandwidth multiplexing, cooperative routing and resource scheduling, etc.; the research on large-scale data collection protocols with key technologies including distributed architecture, dynamic clustering, parallel data aggregation, and distributed data transmission, etc. The project outcomes will facilitate the completeness of system architecture, communication protocol and theory system of IoV, and will be strong driving forces for IoV based applications with respect to technique research and development, standard establishment and industrial innovation. To sum up, this is an important research topic, which is in line with national strategic directions and the needs of people’s livelihood. This is a frontier research topic, which is dedicated to smart IoV applications. This is an interdisciplinary research topic, which integrates communication theory, networking theory, algorithm theory and optimization theory. This is a basic research topic, which is innovative in terms of idea, content and methodology. This project is meaningful and valuable.

车联网是传统交通领域与新兴移动计算、人工智能、大数据等交叉领域的融合创新方向。本项目基于高异构、高动态、高分布式车联网特征提炼出异构网络融合与大规模数据服务关键科学问题。主要研究包括基于软件定义网络、网络功能虚拟化、网络切片与雾计算等关键技术的异构融合车联网架构；基于自适应功率控制、带宽多路复用、协作路由及资源调度等关键技术的跨层协作大规模数据分发协议；基于分布式架构、动态分簇、数据并行聚合与分布式传输等关键技术的大规模数据采集协议。研究成果有助于完善车联网系统架构、通信协议与理论体系，有助于推动车联网相关应用的技术研发、标准制定与产业创新。综上，本项目是一项符合国家战略方针与实际民生需求的重要研究课题；是一项面向智慧车联网应用的前沿研究课题；是一项综合通信理论、网络理论、算法理论与优化理论的交叉研究课题，是一项在思想、内容和方法上具有创新性的基础研究课题，是有理论意义、有实用价值的。

为推动智能驾驶、无人驾驶、车路协同等基于车联网的新兴智能交通应用的发展，本项目深入研究了面向高异构、高动态、高分布式车联网环境中的网络融合与大规模数据服务的科学问题。主要研究成果包括: 1）提出了基于软件定义网络、网络功能虚拟化、网络切片与雾计算等关键技术的异构融合车联网新架构，该架构致力于实现复杂网络资源的抽象化与虚拟化、控制面与数据面的分离、以及逻辑集中调度与分布式调度的结合，进而增强系统的可靠性与可扩展性、提高服务部署的敏捷性与灵活性，为面向大规模数据服务打下坚实基础；2）设计了基于自适应功率控制、带宽多路复用、协作路由及资源调度等关键技术的跨层协作大规模数据分发协议，具体包括基于I2V/V2V通信的并行数据传输策略、基于SDN与雾计算的协作路由策略、以及基于异构融合网络的资源分配与数据调度策略，从而最大化系统面向大规模数据分发服务的性能；3）设计了基于分布式架构、动态分簇、数据并行聚合与分布式传输等关键技术的大规模数据采集协议，该协议综合考虑车联网具体应用需求，协同设计自适应分簇策略、数据并行聚合与上传策略，并基于MapReduce基本思想定制映射函数和归纳函数，最终实现面向大规模数据采集服务的高并行数据聚合与传输。本项目研究成果有助于完善车联网系统架构、通信协议与理论体系，有助于推动车联网相关应用的技术研发、标准制定与产业创新。在本项目支持下，共发表期刊论文24篇，国际会议论文18篇，培养毕业3名博士研究生，9名硕士研究生，获取授权专利5项，新增国家基金面上项目1项并荣获重庆市自然科学一等奖，组织参与国内外学术交流19余次，项目按照计划顺利完成各项指标任务。