本文關(guān)鍵詞： Storm GPS數(shù)據(jù) 算法 實(shí)時計(jì)算　出處：《西安電子科技大學(xué)》2015年碩士論文　論文類型：學(xué)位論文

【摘要】：伴隨著經(jīng)濟(jì)的不斷發(fā)展,物流行業(yè)如雨后春筍般蓬勃發(fā)展起來,物流車輛越來越多,物流車輛管理行業(yè)悄然興起,我們利用便攜式車載GPS實(shí)時采集車輛行駛過程中的時間、速度、經(jīng)度、緯度等信息,根據(jù)客戶的需求,設(shè)計(jì)算法,通過對這些數(shù)據(jù)進(jìn)行分析、處理,來實(shí)現(xiàn)對物流車輛的遠(yuǎn)程監(jiān)控、管理。面對GPS產(chǎn)生的海量數(shù)據(jù),傳統(tǒng)的處理方式已經(jīng)力不從心,大數(shù)據(jù)處理技術(shù)的出現(xiàn)為其提供了一個很好的解決途徑。Hadoop以及Storm從中脫穎而出,成為大數(shù)據(jù)處理的主要方法,Hadoop已經(jīng)發(fā)展相當(dāng)成熟,但其主要特點(diǎn)為對數(shù)據(jù)進(jìn)行并行批處理,實(shí)時性很低,無法滿足人們對于大數(shù)據(jù)處理實(shí)時性越來越高的要求,而Storm這個Apache基金會下的開源軟件正好可以滿足這種要求。Storm是一個分布式實(shí)時計(jì)算系統(tǒng),它對于實(shí)時計(jì)算的意義類似于Hadoop對于批處理的意義。Storm的實(shí)時性體現(xiàn)在它以流的方式處理大數(shù)據(jù),即Storm可以處理源源不斷發(fā)送過來的消息,這些消息以數(shù)據(jù)元組為基本單位,形成一條有向無界的數(shù)據(jù)流,處理完之后將結(jié)果寫入到某個存儲中去。而其分布式特性體現(xiàn)在它的處理組件是分布式的,而且處理延遲極低,所以可以作為一個通用的分布式RPC框架來使用。所以,為了更好地管理車輛并且為其提供更優(yōu)的服務(wù),我們選擇Storm這個平臺來對GPS車輛大數(shù)據(jù)進(jìn)行分析、處理。本文的核心內(nèi)容是將兩個GPS車輛數(shù)據(jù)處理算法在搭建好的Storm流式處理平臺上進(jìn)行實(shí)現(xiàn)。首先,對Storm的研究背景、意義、研究現(xiàn)狀等進(jìn)行了簡單介紹,詳細(xì)介紹了Storm的基本概念、特點(diǎn)、運(yùn)行機(jī)制、系統(tǒng)架構(gòu)、容錯性能以及集群安裝部署方法等內(nèi)容,至此,可以對Storm有一個基礎(chǔ)的了解。然后,詳細(xì)介紹了兩個GPS車輛數(shù)據(jù)處理算法——判斷點(diǎn)是否在中國區(qū)域內(nèi)的算法以及車輛軌跡匹配算法。最后,將這兩個算法放在搭建好的Storm流式處理平臺上進(jìn)行實(shí)現(xiàn),為其分別設(shè)計(jì)向集群提交的拓?fù)?從外部數(shù)據(jù)源讀取數(shù)據(jù)進(jìn)行處理,得到最終結(jié)果,以此來實(shí)現(xiàn)對車輛的監(jiān)控。
[Abstract]:With the continuous development of economy, the logistics industry is booming, more and more logistics vehicles, logistics vehicle management industry quietly rising, we use portable vehicle GPS real-time acquisition of vehicle driving time, Speed, longitude, latitude and other information, according to the needs of customers, design algorithms, through the analysis of these data, processing, to achieve the remote monitoring of logistics vehicles, management. Facing the massive data generated by GPS, The traditional processing methods have been unable to meet their expectations, and the emergence of big data's processing technology has provided it with a very good solution. Hadoop and Storm have emerged from the fore, and become the main method of big data's handling, Hadoop has developed quite maturely. But its main characteristic is to carry on the parallel batch processing to the data, the real time is very low, cannot satisfy the people to big data processing real-time request which is more and more high. And Storm, the open source software of the Apache Foundation, just meets this requirement. Storm is a distributed real-time computing system. Its meaning for real-time computing is similar to that of Hadoop for batch processing. A directed unbounded data stream is formed, and the results are written to a storage after processing. Its distributed nature is that its processing components are distributed and the processing latency is extremely low. So it can be used as a general distributed RPC framework. Therefore, in order to better manage vehicles and provide better service for them, we choose Storm as a platform to analyze GPS vehicle big data. Processing. The core of this paper is to implement two GPS vehicle data processing algorithms on the Storm flow processing platform. Firstly, the research background, significance and research status of Storm are briefly introduced. The basic concept, characteristics, running mechanism, system architecture, fault-tolerant performance and cluster installation and deployment methods of Storm are introduced in detail. At this point, we can have a basic understanding of Storm. This paper introduces in detail two GPS vehicle data processing algorithms, which are the algorithms for judging whether the points are in the Chinese region and the vehicle trajectory matching algorithm. Finally, the two algorithms are implemented on a Storm flow processing platform. The topology to be submitted to the cluster is designed respectively, and the data is read from the external data source for processing, and the final result is obtained, so as to realize the monitoring of the vehicle.
【學(xué)位授予單位】：西安電子科技大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2015
【分類號】：TP311.13

【相似文獻(xiàn)】

相關(guān)期刊論文 前10條

1 徐宗健;線形工程設(shè)計(jì)中里程數(shù)據(jù)處理算法的研究[J];鐵路計(jì)算機(jī)應(yīng)用;1994年03期

2 周慧霞;李洪蓮;王選科;;關(guān)于導(dǎo)線網(wǎng)測量數(shù)據(jù)處理算法的研究[J];科技與企業(yè);2012年06期

3 晁寧;李言俊;;弱小目標(biāo)跟蹤中利用樣條曲線逼近的數(shù)據(jù)處理算法[J];紅外與激光工程;2010年03期

4 張良義;程晶晶;張嘉偉;薛志波;;NMR測井儀中數(shù)據(jù)處理算法的優(yōu)化實(shí)現(xiàn)[J];儀表技術(shù)與傳感器;2011年04期

5 譚杰鋒;周蘇峽;;基于SQL Server的綜合評價數(shù)據(jù)處理算法分析[J];北京教育學(xué)院學(xué)報(bào)(自然科學(xué)版);2007年04期

6 朱洪濤;李江明;王志勇;;一種基于藍(lán)牙串口模塊的數(shù)據(jù)處理算法[J];實(shí)驗(yàn)室研究與探索;2013年11期

7 譚杰鋒;周蘇峽;;基于SQL Server的綜合評價數(shù)據(jù)處理算法分析[J];甘肅聯(lián)合大學(xué)學(xué)報(bào)(自然科學(xué)版);2007年05期

8 高善清;魏新亮;;數(shù)字氣壓高度表數(shù)據(jù)處理算法分析[J];四川兵工學(xué)報(bào);2013年07期

9 楊帆;;雷達(dá)數(shù)據(jù)處理算法的仿真研究[J];電子元器件應(yīng)用;2012年09期

10 李梅娟;微機(jī)應(yīng)用系統(tǒng)中輸入信號的數(shù)據(jù)處理算法[J];鞍山師范學(xué)院學(xué)報(bào);1997年02期

相關(guān)會議論文 前2條

1 唐鋒;王向朝;張以謨;;白光掃描干涉系統(tǒng)數(shù)據(jù)處理算法的研究[A];第十一屆全國光學(xué)測試學(xué)術(shù)討論會論文（摘要集）[C];2006年

2 陳國維;;全自動生化分析儀中測量值的數(shù)據(jù)處理[A];海峽兩岸醫(yī)療儀器學(xué)術(shù)會議論文集[C];1995年

相關(guān)碩士學(xué)位論文 前2條

1 岳夢婕;面向智慧農(nóng)業(yè)種植領(lǐng)域數(shù)據(jù)處理算法的研究[D];浙江理工大學(xué);2016年

2 齊亞星;Storm平臺下GPS車輛數(shù)據(jù)處理算法的實(shí)現(xiàn)[D];西安電子科技大學(xué);2015年

，

本文編號：1507290