蘇炫榮Su, Hsuan-Jung臺灣大學:電信工程學研究所林松徵Lin, Song-JhengSong-JhengLin2010-07-012018-07-052010-07-012018-07-052008U0001-2607200823064500http://ntur.lib.ntu.edu.tw//handle/246246/188173本論文主要探討的問題在於，在二次高斯(Quadratic Gaussian)的假設下，如何為Wyner-Ziv編碼理論設計出一套實際的編碼方式。首先，本論文會介紹及討論Wyner-Ziv編碼理論。之後我們會分析在二次高斯的假設下Wyner-Ziv編碼的架構。這個假設之所以會引起廣泛的討論，是因為在此假設下的資料率失真(Rate-distortion)邊界，會等於在編碼器跟解碼器雙方都有附加資訊(Side information)的情況下的資料率失真邊界。在二次高斯Wyner-Ziv編碼的情況下，我們基於疊加編碼(Superposition coding)的理論，提出一種實際的編碼方式。我們稱這項技術為疊加二次高斯Wyner-Ziv編碼，並將它縮寫成SQG-WZC(Superposition quadratic Gaussian Wyner-Ziv coding)。再藉由隨機編碼(Random coding)，跟聯合典型性(Joint typicality)，來證明SQG-WZC確實可以達到Wyner-Ziv的資料率失真邊界。最後，我們將SQG-WZC延伸到向量的情況下，並且證明向量SQG-WZC可以達到Wyner-Ziv資料率失真邊界。We present practical codes designed for Wyner-Ziv coding in the quadratic Gaussian case. The structure of Wyner-Ziv coding is first introduced and discussed. Wyner-Ziv coding in the quadratic Gaussian case is then analyzed. This case is of interest since the rate distortion bound in this case is equal to the case that side information is known at both the encoder and decoder. For Wyner-Ziv coding in the quadratic Gaussian case, we propose a practical code design, which is based on superposition coding. This technique is named superposition quadratic Gaussian Wyner-Ziv coding and is abbreviated as SQG-WZC. SQG-WZC is able to achieve the Wyner-Ziv rate distortion bound by using random coding and joint typicality. Finally, we extend SQG-WZC to the vector case and show that the Wyner-Ziv rate distortion bound can also be achieved by vector SQG-WZC.1 Introduction 1.1 Motivation: Realizing the Quadratic Gaussian Wyner-Ziv Coding by Superposition Coding 3.2 Overview of Thesis 4.3 Notations 4 Theoretical Backgrounds 6.1 Rate Distortion Theory 6.2 Wyner-Ziv Coding 9.3 Introduction to Modulo-Additive Noise Channel and Its Capacity 12.3.1 Some Definitions and Lemmas of the Modulo Operation 13.3.2 Modulo-Additive Noise Channel Model and Its Capacity 14 Superposition Quadratic Gaussian Wyner-Ziv Coding 16.1 Intuition 17.2 Formulation of Superposition Quadratic Gaussian Wyner-Ziv Coding 22.2.1 Encoder and Decoder Structure of SQG-WZC 24.3 Random Coding Analysis for SQG-WZC 26.3.1 Parameter Selection 26.3.2 Random Coding Analysis for SQG-WZC 28 Vector Superposition Quadratic Gaussian Wyner-Ziv Coding 35.1 System Model 36.2 Formulation of Vector Superposition Quadratic Gaussian Wyner-Ziv Coding 39.2.1 Encoder and Decoder Structure for VSQG-WZC 39.3 Random Coding Analysis for VSQG-WZC 42.3.1 Parameter Selection 42.3.2 Random Coding Analysis for VSQG-WZC 43.4 Code Parameters Selection and the Rate Achievement 47 Conclusion 51ppendix 53 Proof of Theorem 3 53ibliography 56625419 bytesapplication/pdfen-US向量疊加二次高斯Wyner-Ziv編碼Wyner-Ziv編碼資料率失真邊界疊加編碼二次高斯假設VSQG-WZCWyner-Ziv codingrate-distortion boundsuperpositon codingquadratic Gaussian casethesishttp://ntur.lib.ntu.edu.tw/bitstream/246246/188173/1/ntu-97-R95942044-1.pdf