標題: 具有嵌入式循環冗餘檢驗特性之錯誤更正技術於混合式自動重傳請求機制下之研究
CRC-embedded Error Correction for Hybrid ARQ
作者: 孫玉婷
Sun, Yu-Ting
王忠炫
Wang, Chung-Hsuan
電信工程研究所
關鍵字: 迴旋碼;循環冗餘檢驗碼;混合式自動重傳請求;錯誤更正技術;convolutional code;CRC;hybrid ARQ;error correction
公開日期: 2008
摘要: 在使用串接碼的通訊系統下, 循環冗餘檢驗碼是一種通常被當作內部碼用於錯誤偵測的截短碼而迴旋碼則被視為外部碼。然而,如果吾人不僅想使用循環冗餘檢驗碼的錯誤偵測能力而且也想使用其錯誤更正能力,則最直接增加效能的想法就是利用迴旋碼和循環冗餘檢驗碼的聯合解碼來取代傳統的迴旋解碼。其中一種方式就是結合循環冗餘檢驗碼和迴旋碼的解碼網格。然而,雖然使用循環冗餘檢驗碼和迴旋碼的結合網格作解碼的聯合解碼法可以降低區塊錯誤率,但在實際運用上是不可行的。因此,在利用結合網格做解碼的情況下,吾人提出一種針對循環冗餘檢驗碼修正後的編碼方式以及在從結合網格造出的縮小網格上的一種新解碼系統。模擬結果顯示,不只修正後的循環冗餘檢驗解碼程序且提出的解碼系統都會比傳統的重傳系統得到較低的錯誤率。和表單維特比演算法的比較以及提出的解碼方式也會在此論文中呈現。
As we know, a cyclic redundancy code (CRC) is a shortened cyclic code which is com- monly used as an inner code for error detection and an convolutional code is considered as an outer code in the communication system employing a serially concatenated code. How- ever, if we utilize not only the error detecting capability but also the error correcting ability of CRC, it is straightforward to improve the performance by employing the joint decoding of convolutional codes and CRC instead of the conventional convolutional decoding. One of methods is to combine the decoding trellis of CRC and convolutional codes. Nevertheless, the joint decoding which applies a combined trellis composed of CRC and convolutional codes was infeasible in practical implementation although the block error rate can be de- ceased. Accordingly, under the circumstance of utilizing the combined trellis for decoding, we propose a modi?ed encoding method for CRC codewords and a new decoding scheme on the reduced trellis constructed from the combined trellis. Simulation results exhibit that not only the modi?ed CRC encoding process but also the proposed decoding scheme achieves lower error rate than the original retransmission schemes. The comparison between list Viterbi algorithm (LVA) and the proposed decoding scheme will be illustrated in this thesis in addition.
URI: http://140.113.39.130/cdrfb3/record/nctu/#GT079613517
http://hdl.handle.net/11536/41955
Appears in Collections:Thesis


Files in This Item:

  1. 351702.pdf