針對短線性區塊碼的編解碼設計

Abstract

在現代通信應用的某些情境中，需要低延遲和高可靠度的傳送機制。因此， 對於短線性區塊碼的研究相當重要，本論文主要聚焦於探索 A* 解碼演算法對 短線性區塊碼應用軟式解碼的應用，為了實現最大似然性能 (maximum likelihood performance) 通常有大量的儲存空間需求、時間和計算複雜度。所以，我們的目 標是在保持相近解碼性能的同時減少解碼複雜度，本論文提出將里德-所羅門 碼與二進制碼或卷積碼或渦輪碼相結合的串接碼，以便使用 BCJR 解碼器進行 解碼，而其 soft output 可用於改善 A* 解碼演算法所需的 MRIP (most reliable and independent positions, 最可靠的獨立位置)。利用改善的 MRIP 可以大幅降低 A* 解 碼演算法的解碼複雜度而不犧牲解碼錯誤率。模擬結果證明我們所提方法的優越 性。

In certain scenarios within modern communication applications, there is a growing need for transmission mechanisms that prioritize low latency and high reliability. Consequently, it is important to study short linear block codes. This thesis primarily focuses on exploring the application of soft decoding techniques for short linear block codes, specifically utilizing the A* decoding algorithm. Achieving maximum likelihood (ML) performance usually requires significant storage space, time, and computational complexity. Therefore, our aim is to reduce decoding complexity while maintaining comparable de-coding performance. To achieve this goal, we propose concatenated codes for A* decoding, where we use a Reed-Solomon code as outer code and a binary convolutional code or a turbo code as inner code. With the soft output obtained by the BCJR algorithm for decoding the inner code, we are able to improve the most reliable and independent positions (MRIP) of A* decoding which can significantly reduce the decoding complexity while maintaining the near ML error performance. Simulation results demonstrate that superiority of the proposed method.