基于Storm的实时交通信息管理系统的设计与实现

发布时间：2018-08-26 15:36

【摘要】：智能交通产业发展至今,各种先进的信息采集技术(如微波采集技术、视频采集技术、环形线圈感应式采集技术等)被广泛地运用于交通数据采集,交通管理部门不仅具备了交通基础信息,还拥有了各类动态数据,如车辆实时营运信息、道路交通状况等,采集的数据类型包括属性数据、空间数据、影像数据等。区别于原有手段所采集的交通数据,通过这些新兴技术采集的流数据具有数据容量大、实时性高的特点。在这种交通数据实时的、持续增长的状态下,交通行业传统非实时、静态管理数据的信息处理方式已经难以适用。因此,如何实时处理传感器采集的流数据已经成为关注的焦点和亟需攻破的难关。在这样的背景下,本文通过对基于Storm、HBase等大数据实时技术的研究,提出了相应的解决方案,其主要研究工作和成果如下:(1)提出了基于Storm的实时交通信息管理系统的总体架构。将架构分为5个层面:数据采集层、数据接入层、数据处理与存储层、服务层和应用层,同时研究与设计了 Storm集群的高可用性方案。(2)在研究了 HBase原理之后,根据交通流数据具有数据量大、接入速度快以及数据源多等特点,通过对行键结构设计的优化,解决了流数据落地时出现的“热点”问题。(3)以自主开发的Storm交通流实时分析处理平台为基础,实现对交通流数据的实时处理,解决了海量交通流数据的道路车流平均速度计算以及计算结果离散化的问题。同时,本文还对隐马尔可夫模型进行深入研究,以分析计算所得的数据序列,完成对交通隐马尔可夫模型参数的构造。通过对数据序列的统计分析,使用EM算法完成模型参数的训练。最后使用Viterbi算法完成未来五分钟的交通状态预测。(4)最后,根据总体架构应用层需求,基于Storm实时计算平台,构建并实现了实时交通信息管理系统。本文对系统进行了详细的功能模块设计,并实现了流量统计、拥堵查询、车辆跟踪、交通拥堵预测等功能。
[Abstract]:With the development of intelligent transportation industry, a variety of advanced information acquisition technologies (such as microwave acquisition technology, video capture technology, loop coil induction acquisition technology, etc.) have been widely used in traffic data acquisition. Traffic management departments not only have basic traffic information, but also have all kinds of dynamic data, such as real-time operation information of vehicles, road traffic conditions, etc. The data types collected include attribute data, spatial data, image data and so on. Different from the traffic data collected by the original means, the stream data collected by these new technologies have the characteristics of large data capacity and high real-time performance. Under the condition of real-time and continuous growth of traffic data, the traditional non-real-time and static information processing methods of traffic industry are difficult to apply. Therefore, how to process the stream data collected by sensors in real time has become the focus of attention and urgent difficulties. Based on the research of big data real-time technology such as Storm,HBase, this paper puts forward the corresponding solutions. The main research work and results are as follows: (1) the overall framework of real-time traffic information management system based on Storm is proposed. The architecture is divided into five layers: data acquisition layer, data access layer, data processing and storage layer, service layer and application layer. At the same time, the high availability scheme of Storm cluster is studied and designed. (2) after studying the principle of HBase, According to the characteristics of traffic flow data, such as large amount of data, fast access speed and many data sources, the design of row key structure is optimized. It solves the "hot spot" problem when the flow data falls to the ground. (3) based on the Storm real-time traffic flow analysis and processing platform developed by ourselves, the real-time processing of traffic flow data is realized. The problem of calculating the average speed of traffic flow and the discretization of the result is solved. At the same time, the hidden Markov model is deeply studied in order to analyze the calculated data sequence and construct the parameters of the traffic hidden Markov model. Through the statistical analysis of the data sequence, the EM algorithm is used to complete the training of the model parameters. Finally, the Viterbi algorithm is used to predict the traffic state in the next five minutes. (4) finally, according to the requirements of the application layer of the overall architecture, a real-time traffic information management system is constructed and implemented based on the Storm real-time computing platform. In this paper, the detailed functional modules are designed, and the functions of traffic statistics, traffic congestion query, vehicle tracking, traffic congestion prediction and so on are realized.
【学位授予单位】：扬州大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP311.52

【相似文献】