Please use this identifier to cite or link to this item: http://hdl.handle.net/11455/6142
DC FieldValueLanguage
dc.contributor.advisor歐陽彥杰zh_TW
dc.contributor.advisorYen-Chieh Ouyangen_US
dc.contributor.author葉禮彬zh_TW
dc.contributor.authorYeh, Li-Binen_US
dc.date2000zh_TW
dc.date.accessioned2014-06-06T06:37:24Z-
dc.date.available2014-06-06T06:37:24Z-
dc.identifier.urihttp://hdl.handle.net/11455/6142-
dc.description.abstract近年來,在ATM網路中的可變位元速率之視訊(VBR video)已經證實具有自我相似性(self-similar)。若忽略這種特性,將會使我們在網路效能分析上產生過度樂觀的預估值,進而導致網路資源分配不足。換句話說,自我相似的特性對網路的效能有很大的影響。赫氏參數(Hurst parameter)是用來描述自我相似程度多寡的重要參數。為了要能判別網路的交通是否具有自我相似性,我們需要使用一些判別的方法來測試其赫氏參數值。本篇論文,我們使用小波轉換(wavelet transform)作為判別赫氏參數的方法。小波的重要特徵是它的觀念和使用方法容易,而且在於測量交通資料的斜率。小波的其它優點是我們可以任意的選擇vanishing moment來消除先前的取樣結果。此外,藉由小波係數的能量變化,我們可以正確地估計赫氏參數值。我們應用FIR多層類神經網路來分析和預測ATM網路中自我相似性交通流量的狀態,FIR多層類神經網路能準確地預測在下一個時槽即將進入網路的總資料數。利用此一資訊,我們可以估計在ATM網路中的可變位元速率之視訊交通的頻寬需求。實驗的結果顯示,我們可以得到最小的有效頻寬需求。另外我們也利用隨機事先偵測(random early detection)的方法來作交通壅塞控制,模擬的結果顯示,我們使用隨機事先偵測的方法可以得到更高的貫通率(throughput)。zh_TW
dc.description.abstractMore recently, the variable-bit-rate (VBR) video over ATM networks is found to exhibit self-similar characteristics. Neglecting the self-similarity would lead to overly optimistic performance predictions and inadequate network resources allocation. In other words, the presence of self-similarity has serious implications on the performance of networks. A key parameter characterizing self-similar processes is the Hurst parameter H, which is designed to capture the degree of self-similarity. In order to determine if a given time series exhibits self-similarity, a method is needed to estimate H for a given time series. In this thesis we present an estimation tool by using wavelet transform. An important feature of the wavelet-based tool is the conceptual and practical simplicity, consisting essentially in measuring the slope. The other advantage of wavelet-based tool is that we can arbitrary choose vanishing moment to remove pre-selected trends. Moreover, the Hurst parameter can be accurately estimated from the power-law behavior of wavelet coefficients. A finite-duration impulse response (FIR) multilayer network is used to predict the number of incoming cells at the next period. Based on the information provided by the FIR multilayer network, we apply the wavelet-based tool to measure the Hurst parameters of predicted data. Then we use the Norros formula to estimate the bandwidth requirement for each predicted data. The results show that the lowest effective bandwidth requirement for MPEG1 video traffic over ATM networks can be obtained. Then we use the random early detection (RED) algorithm for traffic congestion control based on the predictive bandwidth. The simulation results indicate that the actual data has higher throughput by using RED.en_US
dc.description.tableofcontents1 Introduction 1 1.1 Motivation 1 1.2 Overview of Wavelets 2 1.3 Contribution and Organization of Thesis 3 2 Characteristic of Self-similarity 5 2.1 Definition of Self-Similar Processes 5 2.1.1 Continuous Time Definition 6 2.1.2 Discrete Time Definition 7 2.2 Estimation of Self-Similar Traffic 8 2.2.1 R/S Plot 8 2.2.2 Variance Time Plot 11 2.3 Effective Bandwidth Requirement for Self-Similar Traffic 13 2.3.1 Norros Effective Bandwidth Formula 13 2.4 Summary 14 3 Discrete Wavelet Transform and Multiresolution Analysis 17 3.1 Definition of Discrete Wavelet Transform 18 3.2 Multiresolution Analysis (MRA) 20 3.3 Vanishing Wavelet Moments 23 3.4 Summary 24 4 Self-Similar Traffic Parameters Estimation Using Wavelet 25 4.1 The Long-Range Dependence Phenomenon 26 4.2 Discrete Wavelet Transform of Scaling Processes 27 4.2.1 Two Key Properties in the Scaling Processes 27 4.2.2 Wavelet Transform of Long-Range Dependence process 28 4.3 The estimstor alpha of Long-Range Dependence 29 4.4 Choice of Scales [J1,J2] 31 4.5 Select the Number of Vanishing Moments 34 4.6 Summary 35 5 Traffic Control Using RED and Simulation Results 40 5.1 Network Traffic Congestion Control 40 5.2 Simulation Results 42 5.2.1 Traffic Traces 42 5.2.2 Traffic Prediction Using FIR Multilayer Network 42 5.2.3 A Wavelet-based Estimation of Self-Similar Parameters 46 5.2.4 Effective Bandwidth for Self-Similar Traffic 54 5.2.5 Traffic Management by Random Early detection (RED) 56 6 Conclusions and Future Works 59 6.1 Conclusions 59 6.2 Future work 60 Bibliography 61en_US
dc.language.isoen_USzh_TW
dc.publisher電機工程學系zh_TW
dc.subjectHurst parameteren_US
dc.subject赫氏參數zh_TW
dc.subjectself-similarityen_US
dc.subjectcongestion controlen_US
dc.subjectrandom early detectionen_US
dc.subjectthroughputen_US
dc.subject自我相似性zh_TW
dc.subject壅塞控制zh_TW
dc.subject隨機事先偵測zh_TW
dc.subject貫通率zh_TW
dc.title可變位元速率視訊交通之控制—使用小波估計自我相似參數值zh_TW
dc.titleVariable Bit Rate (VBR) Video Traffic Control: A Wavelet Approach for Self-Similar Parameters Estimationen_US
dc.typeThesis and Dissertationzh_TW
item.openairecristypehttp://purl.org/coar/resource_type/c_18cf-
item.openairetypeThesis and Dissertation-
item.cerifentitytypePublications-
item.fulltextno fulltext-
item.languageiso639-1en_US-
item.grantfulltextnone-
Appears in Collections:電機工程學系所
Show simple item record
 
TAIR Related Article

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.