Trace-Back Structures of Soft-Output Decoding Algorithms for Turbo Codes

• Main Authors: Chia-Chu Cho, 卓佳鉅
• Other Authors: Chi-chao Chao
• Format: Others
• Language: en_US
• Published: 2000
• Online Access: http://ndltd.ncl.edu.tw/handle/31901807918346544430

碩士 === 國立清華大學 === 電機工程學系 === 88 === Turbo codes were recently proposed by Berrou, Glavieux, and Thitimajshima as a remarkable step forward in high-gain, low-complexity coding. Because of the excellent error correction performance, turbo codes are used on the Dedicated Channel in the 3rd Generation Partnership Project (3GPP) proposal. A parallel concatenated convolutional code with 8-state constituent encoders is used for data services requiring bit error rate between 10^{-3} and 10^{-6}. The key ingredients of turbo codes are the uses of the interleaver and iterative decoding. Various soft-input soft-output algorithms have been proposed and applied to iterative decoding. In this thesis, we first survey these soft-in soft-output algorithms. Based on an unified framework, we then propose a unified trace-back structure for these algorithms. Finally, comparisons of computational complexity and simulated performance for these algorithms are presented.