Turbo 码新译码算法  

Novel Decoding Algorithm of Turbo Codes

作　　者：戴懿[1] 曾加[1] 郑志航 

机构地区：[1]上海交通大学电子信息学院

出　　处：《上海交通大学学报》1998年第6期32-37,60,共7页Journal of Shanghai Jiaotong University

摘　　要：Ｔｕｒｂｏ码是一种新的纠错编码，具有十分突出的纠错能力．Ｔｕｒｂｏ码编码端由两个或更多个卷积码并行级联构成，译码端则采用了一种基于软判决信息输入／输出的反馈迭代结构．介绍了Ｔｕｒｂｏ码原理以及现有主要的两种Ｔｕｒｂｏ码译码算法——ＭＡＰ和ＳＯＶＡ，推导并提出了两种改进的译码算法：ＡＬ－１和ＡＬ－２．计算机模拟和定性讨论表明，在白高斯噪声信道下，ＡＬ－１和ＡＬ－２算法既可大大减少计算复杂度，又保持了良好的译码性能，在整体上优于ＭＡＰ和ＳＯＶＡ算法．Turbo code achieves almost reliable data communication at signal to noise ratio very close to the Shannon limit. The turbo codes encoder consists of a parallel concatenation of two or more convolutional codes, and its decoder uses a soft input/soft outpur feedback decoding rule. After studying the turbo codes and its two main decoding algorithms——MAP and SOVA, we propose two modified decoding algorithms, named AL 1 and AL 2. Through the computer simulation and discussion, it is shown that, in AWGN channels, AL 1 and AL 2 not only reduce the computation comple xity, but also keep good error rate performance. In the whole, they are superior to MAP and SOVA algorithms respectively.

关 键 词：TURBO码 MAP算法 SOVA算法 AL-1算法 纠错编码 

分 类 号：TN911.22[电子电信—通信与信息系统]