實體層猛禽碼之數種研究結果

Abstract

本論文針對討論使用於實體層之混合式自動重送請求（以下簡稱ＨＡＲＱ）之通道編碼，ＨＡＲＱ分為三種型態，本論文主要集中探討型態二，前次傳送之解碼失敗封包不會被丟棄，而重傳資料與前次傳輸不同，並合併兩者形成糾錯能力更強的錯誤更正碼。常見的型態二ＨＡＲＱ為穿刺型低密度奇偶檢查碼與猛禽碼，兩者各有優缺點，傳統猛禽碼使用高碼率之低密度奇偶檢查碼，會有較佳表現，我們試著以較低碼率之穿刺型低密度奇偶檢查碼取代之，並做出各種嘗試修正，些微提高在最佳化目標訊雜比點位的通道傳輸量並降低複雜度。 並在第二部份，我們考慮不同類型之調變系統，成功將非系統性猛禽碼用於高階脈衝編碼調變，因文獻已證實用於高斯雜訊通道中之通用猛禽碼並不存在，我們應用一個可適性非系統性猛禽碼之設計，使其在高階脈衝編碼調變依舊在一個廣泛的雜訊比範圍有很高的通道使用效率。

This thesis describes a study into coding and modulation techniques for incremental redundancy automatic repeat request (IR-HARQ) and high order modulation schemes designed to increase throughput over the AWGN channel. The research is based on two major classes in IR-HARQ: Raptor Codes and Punctured LDPC codes. In the first part of this thesis, an analysis of the conventional optimizing method is conducted for both schemes before presenting an efficient approach to combining systematic Raptor Codes and Punctured LDPC codes in order to construct a code that provides an improvement in SNR ranges lower than the specific optimized SNR compared with the same pre-code as conventional systematic Raptor Codes. In the second part, the linear programming in Raptor Codes is applied to the high order modulation. Adaptive degree distribution is used on Gray-mapped 4PAM and 8PAM modulation, which can be easily implemented on 16QAM and 64QAM modulation. The simulation shows the adaptive degree distribution is able to retain a high channel efficiency for wide SNR regions.