實體層系統性猛禽碼之研究

Abstract

本論文討論使用於實體層之混合式自動重送請求(已下簡稱HARQ)之通道編碼，HARQ結合前饋式錯誤更正碼及自動重送請求，實體層猛禽碼為其一種無碼率之錯誤更正編碼。傳統上實體層猛禽碼能針對單一訊雜比做最佳化，文獻已證實於二位元輸入高斯雜訊通道中，可以適用所有訊雜比的通用猛禽碼並不存在，達到準通用的猛禽碼設計即為研究目標。而在非系統化的盧比轉換碼中，為了達到接近通道容量的效果，所需要的盧比轉換碼和低密度奇偶檢查碼的疊帶次數非常之高，並伴隨著長運算時間和高複雜度，此篇論文根據單一訊雜比設計相對應之系統性猛禽碼，並藉由互消息分析其疊帶次數和效能之間之轉換，在經由提前終止技術，設計一套能適用於廣泛雜訊比之逼近通道容量且又能在低復雜度完成之猛禽碼。

This thesis studies physical-layer channel coding with incremental redundancy over the additive white Gaussian noise channel together with a feedback channel. More specifically, we study the physical layer Raptor (PLR) Codes. Raptor Codes is a rateless code that is composed of a Luby transform (LT) codes and a Low-Density Parity-check (LDPC) code. For practical applications, the decoding complexity is an important issue. It has been shown that systematic PLR codes can be implemented with much lower complexity as compared to the nonsystematic counterparts. We propose a design to find a systematic PLR code with adaptive degree distributions which can obtain throughput at least as good as the nonsystematic counterpart. Moreover, we provide convergence analysis to predict the number of iterations needed for obtaining successful decoding. The analysis matches well with the throughput obtained from simulation.