摘 要
傳感器節(jié)點的位置信息在無線傳感器網(wǎng)絡(luò)的監(jiān)測活動等應(yīng)用中起著至關(guān)重要的作用。而取得節(jié)點位置信息較簡便、快捷、精確的方法是通過手動設(shè)定或攜帶GPS定位設(shè)備等手段，但通過這種方式獲取的成本很高。因此，較好的方法是采用定位算法進行估計。本文將主要研究基于多維標度的無線傳感器網(wǎng)絡(luò)定位算法。
首先，本文在查閱大量相關(guān)文獻的基礎(chǔ)上，綜述了無線傳感器網(wǎng)絡(luò)的研究背景、研究意義及現(xiàn)狀，并介紹了無線傳感器網(wǎng)絡(luò)的結(jié)構(gòu)、特點以及典型的定位算法。
其次，介紹了多維標度技術(shù)及其在無線傳感器網(wǎng)絡(luò)定位算法中的應(yīng)用。在分析經(jīng)典MDS-MAP定位算法的基礎(chǔ)上，提出基于Hop-Euclidean的MDS-MAP(D)定位算法。該算法先采用分簇的算法，將大規(guī)模網(wǎng)絡(luò)分成多個具有簇首的局部網(wǎng)絡(luò)，在局部網(wǎng)絡(luò)中通過Hop-Euclidean算法計算鄰居節(jié)點間的歐氏距離來代替MDS-MAP算法中的所使用的最短路徑距離，這樣不僅提高了定位精度，而且有利于網(wǎng)絡(luò)的擴展。
再次，針對分布式加權(quán)MDS定位算法不能適應(yīng)網(wǎng)絡(luò)連通度變化、網(wǎng)絡(luò)拓撲不規(guī)則且收斂速度較慢的不足，提出一種改進算法。本文采用的加權(quán)機制與鄰居選擇機制綜合考慮1跳鄰居數(shù)目、節(jié)點自身定位精度與測距誤差，并且引入最速下降法優(yōu)化其目標代價函數(shù)。
最后，采用Matlab仿真平臺從定位誤差、拓撲結(jié)構(gòu)等方面對提出的兩種改進算法進行仿真分析并與原來算法做比較。仿真結(jié)果表明，提出的算法在定位精度提高的情況下對不規(guī)則、大規(guī)模網(wǎng)絡(luò)有很好的適應(yīng)性。

關(guān)鍵詞：無線傳感器網(wǎng)絡(luò)；多維標度；Hop-Euclidean算法；分布式加權(quán)；鄰居選擇機制




Abstract
Location information has played an increasingly important role in many applications of wireless sensor networks, such as monitoring activities and so on. The simple, quick and precise way to obtain location information is either to set up manually or to install GPS, which will waste a vast amount of time and human resources. A better way to obtain location information is to use the localization algorithm.In this paper, we mainly focuse on the research of wireless sensor network location algorithm based on multi-dimensional scaling.
Firstly, the research background, significance and status for WSN are summarized in this paper based on large amount of related literatures. And the framework, characteristics and typical localization algorithms of wireless sensor networks are introduced.
Secondly, in this paper, we give details of classical multidimensional scaling and its application in wireless sensor network location algorithm. Based on the analysis of the classic MDS-MAP of the location algorithm, this paper presents Hop-Euclidean based MDS-MAP(D) location algorithm, which uses the clustering method to divide large-scale network into several local networks with cluster head, and in the local positioning, using the Hop-Euclidean algorithm in place of the shortest path distance to calculate Euclidean distance between the nodes neighbor. Obviously, this algorithm not only increases the positioning accuracy, but also is beneficial to the expansion of the network.
Thirdly, in this paper, we propose an improved algorithm, which makes up for the shortage of the distributed weighted multi-dimensional scaling localization algorithm that can not adapt to the network connectivity change, the irregular network topology and the convergence speed is slow. We use the weighted mechanism and neighbor choice mechanism that comprehensively considerate the number of one hop neighbors, the node positioning precision and the location error. We also introduce the steepest descent method to ptimize the goal cost function.
Lastly, the two improved algorithms are tested by simulations via Matlab according to localization error, network topology and so on. Simulation results show that these algorithms proposed in this paper improve the localization accuracy and are more adaptable to the irregular topology, large-scale network.

Keywords: Wireless Sensor Network; Multi-dimensional scaling; Hop-Euclidean algorithm; Distribution weighted; Neighbors choice mechanism



目 錄
摘 要 I
Abstract III
第1章 緒 論 1
1.1 研究背景及研究意義 1
1.2 國內(nèi)外研究現(xiàn)狀 2
1.3 研究內(nèi)容及主要工作 4
1.4 論文結(jié)構(gòu)安排 5
第2章 無線傳感器網(wǎng)絡(luò)及其定位技術(shù) 7
2.1 無線傳感器網(wǎng)絡(luò)概述 7
2.1.1 無線傳感器網(wǎng)絡(luò)體系結(jié)構(gòu) 7
2.1.2 無線傳感器網(wǎng)絡(luò)的特點 8
2.2 定位技術(shù)的相關(guān)概念及術(shù)語 9
2.3 基于測距技術(shù)的定位 10
2.4 無需測距的定位算法 11
2.4.1 DV-Hop算法 11
2.4.2 Euclidean定位算法 12
2.4.3 Hop-Euclidean定位算法 13
2.4.4 MDS-MAP算法 13
2.5 本章小結(jié) 13
第3章 基于MDS-MAP(D)定位算法的改進 15
3.1 引言 15
3.2 多維標度技術(shù)和MDS-MAP定位算法 15
3.2.1 多維標度技術(shù)基本原理 15
3.2.2 MDS-MAP定位算法 16
3.3 基于Hop-Euclidean的MDS-MAP(D)定位算法 17
3.3.1 改進算法的相關(guān)研究工作 17
3.3.2 分簇算法 18
3.3.3 局部定位 20
3.3.4 求解全局坐標 23
3.3.5 基于Hop-Euclidean的MDS-MAP(D)算法定位過程 25
3.4 本章小結(jié) 25
第4章 分布式加權(quán)多維標度算法的改進 27
4.1 引言 27
4.2 分布式加權(quán)多維標度算法 27
4.2.1 分布式加權(quán)多維標度算法概述 27
4.2.2 dwMDS(G)算法定位過程 28
4.3 dwMDS(G)算法的改進 28
4.3.1 改進算法的相關(guān)研究工作 28
4.3.2 鄰居選擇機制 29
4.3.3 加權(quán)機制 30
..