基于车辆监控系统的数据分析及其优化

发布时间：2018-06-03 03:39

本文选题：智能交通 + 车辆监控系统　；参考：《北京邮电大学》2015年硕士论文

【摘要】：如今,随着中国国民经济的快速增长,全国汽车保有量急剧增加,给城市交通造成了巨大压力,同时也给交通运营、交通调度、交通监管等部门的工作带来了巨大的挑战。一方面道路运行车辆的增多会加大车辆监控和管理的难度,使得工作效率降低；另一方面,道路运行车辆的增多势必需要相关部门投入更大的人力、物力、财力等进行监控和管理,大大提高了管理成本。对此,本论文进行了相关调研,调研结果表明,实现智能化的综合交通管理是解决上述问题的有效措施,而设计和开发可行高效的车辆监控系统是实现智能交通的关键性手段之一。因此,本论文拟设计和实现一种能被交通运营、交通调度、交通监管等部门或企业运用的车辆监控系统。让车辆的运行状态可见、可控,以提高工作效率,降低管理成本。然而,随着监控车辆数目的不断增加,本论文发现,初始设计的车辆监控系统不足以应付如此庞大的数据量,常常出现系统响应时间明显延长,性能明显下降等问题。因此,本论文拟在设计和开发车辆监控系统的基础上,结合Hadoop开源分布式平台,又设计并初步实现分布式车辆监控系统,缩短系统响应时间,优化系统使用性能并改善用户体验。具体而言,本论文主要包括以下三个方面： (1)对车辆监控系统的需求进行详细分析,针对车辆监控系统进行总体设计,主要包括架构设计、需求分析和技术选型。 (2)结合GPS、GIS、Web Service、Ajax等技术,利用J2EE平台、Java语言等,对车载终端采集到的车辆运行动态数据进行处理和分析,实现数据批量导入、实时监控、历史轨迹回放、行驶统计查询等相应功能,并完成了这些功能的测试。 (3)针对测试过程中发现的初步设计的车辆监控系统因数据量大而造成系统响应时间过长的问题,本论文结合Hadoop开源分布式平台,对车辆监控系统进行了改进,设计并初步实现了分布式车辆监控系统,以缩短系统的响应时间,优化系统的使用性能并改善用户体验。最后,本论文还通过实验将车辆监控系统部署在了搭建的Hadoop集群当中,对原有系统和分布式系统的系统响应时间进行了对比分析,证明了分布式车辆监控系统的可行性和有效性。
[Abstract]:Nowadays, with the rapid growth of China's national economy, the number of vehicles in the country has increased rapidly, which has caused great pressure to the urban traffic. At the same time, it also brings great challenges to the work of traffic operation, traffic dispatch, traffic supervision and other departments. On the one hand, the increase of road running vehicles will increase the difficulty of vehicle monitoring and management, and make the work of work. On the other hand, the increase of road running vehicles is bound to require the relevant departments to invest more manpower, material and financial resources to monitor and manage, greatly improving the management cost.
In this regard, this paper has carried out a related research, and the results show that the realization of intelligent integrated traffic management is an effective measure to solve the above problems, and the design and development of a feasible and efficient vehicle monitoring system is one of the key means to realize intelligent traffic.
Therefore, this paper intends to design and implement a vehicle monitoring system which can be used by traffic operation, traffic dispatch, traffic supervision and other departments or enterprises. It can make the running state of the vehicle visible and controllable, in order to improve the efficiency and reduce the management cost.
However, with the increasing number of monitoring vehicles, this paper finds that the initial design of vehicle monitoring system is not enough to cope with such a large amount of data. It often appears that the response time of the system is prolonged obviously and the performance is obviously reduced. Therefore, this paper is based on the Hadoop open source distribution on the basis of the design and development of the vehicle monitoring system. The platform also designs and initially implements distributed vehicle monitoring system, shortening system response time, optimizing system performance and improving user experience.
Specifically, this paper mainly includes the following three aspects:
(1) the needs of vehicle monitoring system are analyzed in detail, and the overall design of vehicle monitoring system is mainly designed, including architecture design, demand analysis and technology selection.
(2) combining GPS, GIS, Web Service, Ajax and other technologies, using J2EE platform, Java language and so on, the dynamic data of vehicle running collected by vehicle terminal are processed and analyzed, and the corresponding functions such as data batch import, real-time monitoring, historical track replay, driving statistics query and so on are implemented, and the tests of these functions are completed.
(3) in view of the problem that the system response time is too long due to the large amount of data in the vehicle monitoring system, which is found in the test process, this paper combines the Hadoop open source distributed platform, improves the vehicle monitoring system, designs and implements a distributed vehicle monitoring system, in order to shorten the response time of the system and optimize the system. In the end, the vehicle monitoring system is deployed in the Hadoop cluster, and the system response time of the original system and the distributed system is compared and analyzed, which proves the feasibility and effectiveness of the distributed vehicle monitoring system.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U495;TP277

【参考文献】