面向车载自组网Anycast的机会性网络资源多尺度优化配备机理研究

Internet of cars has attracted much attention in recent years. Dynamic vehicle information gathering is one of the most representative applications for Internet of cars. However, its development has been restricted much due to the shortage of bandwidth resources. The vehicular ad hoc network has relative low transmission cost and is especially suitable for uploading mass data onto internet. In such type applications, anycast is usually used to routing data from many sources to relative few destinations. In VANETs, vehicles move quite randomly, and the network performance is very hard to guarantee. Considering there is comparative little work investigating anycast routing mechanism, we set up this project. Targeting at improving data uploading throughput, and with the principle of refining communication opportunities utilization, we investigate anycast routing mechanism for large scale dynamic vehicle condition collection. We study on micro-scale how to optimize the usage of communication opportunities, and also study on macro-scale how to allocate communication opportunities among different packets. Besides then, we also propose a light-weight implementation mechanism for anycast routing. The research results of this project will offer a full range of solutions for mass data collection through VANETs, as well as provide an important idea to address the serious bandwidth shortage problem.Internet of cars has attracted much attention in recent years. Dynamic vehicle information gathering is one of the most representative applications for Internet of cars. But its scale has been restricted much do to the shortage of bandwidth resources. The vehicular ad hoc network has relative low transmission cost and is especially suitable for uploading mass data onto internet. In such type applications, anycast is usually used to routing data from many sources to relative few destinations. But in VANETs, vehicles move quite randomly, and the network performance is hard to guarantee. Considering there is comparative little work investigating anycast routing mechanism, we set up this project. Targeting at improving data uploading throughput, with the principle of refining communication opportunities utilization, we investigate anycast routing mechanism for large scale dynamic vehicle condition collection. We study on micro-scale how to optimize the usage of communication opportunities, and also study on macro-scale how to allocate communication opportunities among different packets. Besides then, we also propose a light-weight implementation mechanism for anycast routing. The research results of this project will offer a full range of solutions for mass data collection through VANETs, as well as provide an important idea to address the serious bandwidth shortage problem.

近年来，车联网应用吸引了社会各界广泛关注。车辆动态信息采集是车联网的重要应用，但由于车辆数目众多，动态信息数据量庞大，通信资源紧缺成为制约其深入发展的重要因素。车载自组织网络因为数据传输成本较低，尤其适合于大量数据的上传。Anycast是一种较常见的自组网消息上传方式，但由于车辆移动随机性强，使得路由性能很难保障。目前国内外围绕车载自组网Anycast的理论研究成果较少，因此我们设立本项目对Anycast机制开展研究。本项目以提高数据上传总体吞吐量为目标，以优化车辆间通信机会配置为原则，设计用于大规模车况持续采集的Anycast消息转发机制，重点从微观尺度上研究通信机会的优化使用，从宏观尺度上研究通信机会的均衡分配，并设计轻量级的Anycast转发实现方法。本项目的研究成果将为基于车载自组网的大数据量信息采集提供一套完整的解决方案，为解决大规模车辆动态信息采集带宽资源不足问题提供重要思路。

项目面向车载自组织网络和移动物联网的数据采集应用需求研究相关的数据采集、数据上传机制。.在项目进展过程中，根据车载自组织网络和物联网行业的研究现状对初始研究内容进行了微调，目前完成的主要研究内容包括车联网数据传输算法、无线网络传输调度算法、物联网资源标识与共享体系、物联网数据异常检测与分析四个方面。.主要输出研究成果包括：.1）一种基于SDN的车联网架构，将软件定义网络（SDN）的实现思想应用到车联网中，将网络架构划分为应用层、控制层、数据转发层三个层次，并在此基础上提出了一种基于车辆未来轨迹和经验移动模型相结合的预测性数据采集机制；.2）一种面向城市数据采集质量优化的数据采集方法和上传机制，针对车载传感网的数据采集，以数据重构质量为优化目标，使车辆感知节点能够自适应的优化感知参数以及数据上传策略，降低数据传输开销；.3）基于梯度的背压式路由与传输调度策略。将物联网节点剩余能量作为背压式路由策略的一维，设计背压调度效用值和背压调度策略，以实现传感器网络中节点数据传输的负载均衡；.4）一种基于本体的物联网资源描述模型，从属性、状态、控制、历史信息和隐私5个基本属性对物联网设备资源进行描述，同时能组合以上属性用来描述更复杂的资源。基于该框架，物联网资源能通过REST接口对外发布。.5）基于分布式共识的物联网数据可靠性保障机制。基于区块链可靠分布式账本特点，通过设计一种分层的PoW共识算法实现物联网数据的可靠存储，并为数据来源可信提供保障。.6）一种基于可燃气体嗅探的火患检测方法，通过低成本的物联网节点搭载典型可燃气体传感器，对小型区域内的可燃气体进行广谱的扩散行为检测，以达到火险隐患预警的目的。