未知符元間干擾通道的維特比解調法則之研究

Abstract

最大概度序列估測在通訊系統有未知訊號間干擾時，常用來等化通道衰減及干擾，是常見的一種演算法機制。在接收端，我們應用最大概度序列估測在維特比解調器當作解調的機制。而實際通訊通道並非理想的白高斯雜訊通道,傳送通道在不同時間點並非個別獨立，而是往往之前傳送的狀態會影響到現在以及未來的狀態，我們稱這種模組為高斯馬可夫通道模組。我們要探討的是現在時間點的通道和前一時間點有關，為一階的高斯馬可夫通道參雜未知訊號間干擾,在論文中，高斯馬可夫在維特比解調時，提出多保留路徑的解調方式，其結果顯示可以得到較好的表現。然而，複雜度會因保留的路徑多寡而增加，使得實際硬體的應用需求相當大，因此我們想要提出預測量度且單一保留的構想。在模擬的時候，首先驗證參考論文的參數，並且說明各參數在通訊系統錯誤率方面的影響，衰減常數以及通道起始值在高斯馬可夫通道中所扮演的角色，在我們此篇論文，第一章為提出動機以及詳述參考論文採用的通道模組及解調機制，第二章用模擬來確認參考論文的參數。第三章則用數學分析來說明我們提出的預測量度演算法，第四章為我們對於推導預測量度的討論。第五章為我們的結論。

In digital communication equalization field, maximum likelihood sequence estimation(MLSE) is a popular method to combat with intersymbol interference(ISI). In receivers, we apply MLSE on Viterbi Demodulation, and transmission channels are not identical and independent between states, because previous states would influence temporary and future state. These channels are Gauss-Markov channels. Now, we discuss first order Gauss-Markov channel with unknown ISI, which temporary state would be effected by previous one state. In paper[1], multi-survivors on Viterbi Demodulation are applied, and it shows better performance. However, it grows complexity in proportion to the number of survivors so that increasing requirement of hardware. For the reason, we propose one prediction survivor on Viterbi Demodulation. On the processing of simulation, we must check parameters in paper[1] first, and describe what the roles they play in transmission system and explain how fading ratio constant and initial channel value effect BER. In the thesis, Chapter 1 is the decription of our motivation and paper[1] models. Chapter2 is the simulation for checking parameters in [1]. Chapter3 is the detail decription of our proposed prediction metric Viterbi Demodulation. Chapter is our simulation. Chapter5 is conclusion.