WSN中基于壓縮網(wǎng)絡編碼的數(shù)據(jù)匯集技術研究
發(fā)布時間：2019-02-11 09:57
【摘要】：作為21世紀的前沿技術,無線傳感器網(wǎng)絡(Wireless Sensor Network,WSN)在工農(nóng)業(yè)控制、生物醫(yī)療、環(huán)境質(zhì)量監(jiān)測、搶險救災以及國防軍事等領域都扮演著重要角色。它融合了傳感技術、微電子工藝以及無線網(wǎng)絡技術,為人類認識世界和改變世界提供了一種很有效的工具。在WSN中,各分布式節(jié)點分別采集到局部數(shù)據(jù)后進行一定的處理,然后將處理過的數(shù)據(jù)通過某指定的路由路徑發(fā)送給匯聚節(jié)點,最后由匯聚節(jié)點對處理后的數(shù)據(jù)進行還原后得到區(qū)域數(shù)據(jù)。不同的路由路徑選擇方法以及中間節(jié)點對數(shù)據(jù)的處理方法,形成了各種不同的數(shù)據(jù)匯集方案。目前應用在WSN的數(shù)據(jù)匯集方案主要有兩種:基于特定路由的多跳存儲轉發(fā)、基于網(wǎng)絡編碼的數(shù)據(jù)傳輸與處理。其中前者設計簡單,可以漸進地感知數(shù)據(jù),但是會使得網(wǎng)絡中各節(jié)點處于能耗不均衡狀態(tài),嚴重影響網(wǎng)絡壽命;后者雖然可以提高的無線網(wǎng)絡的能耗均衡性,但解碼端卻存在“全有或全無”(All-or-Nothing,AON)問題,嚴重影響了網(wǎng)絡的可靠性。本文介紹了WSN的結構以及典型的數(shù)據(jù)匯集技術,并說明其存在的不足;分別對網(wǎng)絡編碼(Network Coding,NC)以及壓縮感知(Compressed Sensing,CS)的數(shù)學模型進行了說明,包括隨機線性網(wǎng)絡編碼(RLNC,Random Linear Network Coding)以及壓縮感知的非相干測量和重構算法;分析兩者之間的內(nèi)在聯(lián)系,利用WSN各節(jié)點感知數(shù)據(jù)的相關性以及無線傳輸?shù)膹V播特性,建立了一個漸進感知的高能效的WSN數(shù)據(jù)匯集方案——壓縮網(wǎng)絡編碼方案(Compressed Network Coding,CNC),簡單來說就是由于WSN中感知數(shù)據(jù)存在相關性,對于不滿秩的網(wǎng)絡編碼矩陣可以利用CS解碼算法重構出來一定的信息。在該方案的基礎上,分析了匯聚節(jié)點對感知數(shù)據(jù)的重建過程,結果表明,匯聚節(jié)點對于數(shù)據(jù)的重建是漸進的,CNC方案精確重構數(shù)據(jù)的成功率比一般的網(wǎng)絡編碼方案高了約15%以上,從而很好地解決了AON問題。然后通過定性和定量分析網(wǎng)絡各節(jié)點的能耗情況,說明該方案可以大幅改善WSN的能耗均衡問題以及減少網(wǎng)絡總能耗,從而提高了網(wǎng)絡壽命。在每個無線節(jié)點的能量都相同的情況下,采用了CNC方案的WSN的網(wǎng)絡壽命是采用了傳統(tǒng)的匯聚樹路由協(xié)議(Collection Tree Protocol,CTP)WSN的2.5倍以上。
[Abstract]:As a frontier technology in the 21st century, wireless sensor network (Wireless Sensor Network,WSN) plays an important role in the fields of industrial and agricultural control, biomedicine, environmental quality monitoring, emergency relief and defense and military. It combines sensing technology, microelectronics technology and wireless network technology, and provides a very effective tool for people to understand and change the world. In WSN, each distributed node collects local data and processes it to a certain extent, and then sends the processed data to the convergent node through a specified routing path. Finally, the region data is obtained after the data is restored by the convergent node. Different routing path selection methods and data processing methods of intermediate nodes form a variety of data aggregation schemes. At present, there are two kinds of data collection schemes used in WSN: multi-hop storage and forwarding based on specific route, data transmission and processing based on network coding. The former design is simple and can gradually perceive the data, but it will make the nodes in the network energy imbalance, which seriously affect the network life. Although the latter can improve the equalization of energy consumption in wireless networks, there exists the problem of "all or nothing" (All-or-Nothing,AON) in the decoder, which seriously affects the reliability of the network. This paper introduces the structure of WSN and the typical data collection technology, and explains its shortcomings. The mathematical models of network coding (Network Coding,NC) and compressed sensing (Compressed Sensing,CS) are described, including random linear network coding (RLNC,Random Linear Network Coding) and incoherent measurement and reconstruction algorithms of compressed sensing. Based on the analysis of the internal relationship between the two, a progressive perceptual WSN data aggregation scheme, compressed network coding scheme (Compressed Network Coding,CNC), is established by using the correlation of WSN nodes' perceptual data and the broadcast characteristics of wireless transmission. Simply speaking, because of the correlation of perceptual data in WSN, the network coding matrix with unsatisfactory rank can be reconstructed by CS decoding algorithm. On the basis of the proposed scheme, the process of data reconstruction is analyzed. The results show that the convergent node is progressive for data reconstruction. The success rate of accurately reconstructing data in CNC scheme is more than 15% higher than that in general network coding scheme, which solves the problem of AON well. Then through qualitative and quantitative analysis of the energy consumption of each node in the network, it is shown that the proposed scheme can greatly improve the energy balance problem of WSN and reduce the total energy consumption of the network, thus increasing the network lifetime. Under the condition that the energy of each wireless node is the same, the network lifetime of WSN using CNC scheme is more than 2.5 times that of WSN using traditional convergent tree routing protocol (Collection Tree Protocol,CTP).
【學位授予單位】：哈爾濱工業(yè)大學
【學位級別】：碩士
【學位授予年份】：2014
【分類號】：TP212.9;TN929.5
本文編號：2419624
[Abstract]:As a frontier technology in the 21st century, wireless sensor network (Wireless Sensor Network,WSN) plays an important role in the fields of industrial and agricultural control, biomedicine, environmental quality monitoring, emergency relief and defense and military. It combines sensing technology, microelectronics technology and wireless network technology, and provides a very effective tool for people to understand and change the world. In WSN, each distributed node collects local data and processes it to a certain extent, and then sends the processed data to the convergent node through a specified routing path. Finally, the region data is obtained after the data is restored by the convergent node. Different routing path selection methods and data processing methods of intermediate nodes form a variety of data aggregation schemes. At present, there are two kinds of data collection schemes used in WSN: multi-hop storage and forwarding based on specific route, data transmission and processing based on network coding. The former design is simple and can gradually perceive the data, but it will make the nodes in the network energy imbalance, which seriously affect the network life. Although the latter can improve the equalization of energy consumption in wireless networks, there exists the problem of "all or nothing" (All-or-Nothing,AON) in the decoder, which seriously affects the reliability of the network. This paper introduces the structure of WSN and the typical data collection technology, and explains its shortcomings. The mathematical models of network coding (Network Coding,NC) and compressed sensing (Compressed Sensing,CS) are described, including random linear network coding (RLNC,Random Linear Network Coding) and incoherent measurement and reconstruction algorithms of compressed sensing. Based on the analysis of the internal relationship between the two, a progressive perceptual WSN data aggregation scheme, compressed network coding scheme (Compressed Network Coding,CNC), is established by using the correlation of WSN nodes' perceptual data and the broadcast characteristics of wireless transmission. Simply speaking, because of the correlation of perceptual data in WSN, the network coding matrix with unsatisfactory rank can be reconstructed by CS decoding algorithm. On the basis of the proposed scheme, the process of data reconstruction is analyzed. The results show that the convergent node is progressive for data reconstruction. The success rate of accurately reconstructing data in CNC scheme is more than 15% higher than that in general network coding scheme, which solves the problem of AON well. Then through qualitative and quantitative analysis of the energy consumption of each node in the network, it is shown that the proposed scheme can greatly improve the energy balance problem of WSN and reduce the total energy consumption of the network, thus increasing the network lifetime. Under the condition that the energy of each wireless node is the same, the network lifetime of WSN using CNC scheme is more than 2.5 times that of WSN using traditional convergent tree routing protocol (Collection Tree Protocol,CTP).
【學位授予單位】：哈爾濱工業(yè)大學
【學位級別】：碩士
【學位授予年份】：2014
【分類號】：TP212.9;TN929.5
【參考文獻】
相關期刊論文 前1條
1 熊志強;黃佳慶;劉威;楊宗凱;;無線網(wǎng)絡編碼綜述[J];計算機科學;2007年03期
，本文編號：2419624
本文鏈接：http://sikaile.net/kejilunwen/wltx/2419624.html
最近更新
教材專著
