Studies on construction of error-correcting codes and soft-input soft-output decoding

纠错码的构造及软输入软输出译码研究

基本信息

批准号：
11650374
负责人：
YAMAGUCHI Kazuhiko
金额：
$ 1.15万
依托单位：
The University of Electro-Communications
依托单位国家：
日本
项目类别：
Grant-in-Aid for Scientific Research (C)
财政年份：
1999
资助国家：
日本
起止时间：
1999 至 2000
项目状态：
已结题

来源：
https://kaken.nii.ac.jp/grant/KAKENHI-PROJECT-11650374/
关键词：
error-correcting codes repetition decoding soft-input soft-output decoding Turbo code Turbo decoding concatenated code Viterbi decoding a posteriori probability decoding

项目摘要

The soft-input and soft-output decoding is important technology as decoding of parallel(turbo)and serial concatenated codes. The soft-in soft out decoding and the concatenated codes are key technology of reliability of the advanced communication technology. In this Study, it is aiming to propose the new decoding algorithm and new code construction. The alogorithm is based on the soft-in soft-out decoding and the code is suitable for new soft-in soft-out decoding. For the study, we construct the simulator of turbo decoding for the turbo code with memory channel, which is adaptable various data format of turbo codes and magnetic and mobile channel with/without turbo equalization.The soft-output information(i.e.reliability information)is the posterior probability of each information bit, and is calculated by maximum a posteriori(MAP)decoding, theoretically. The algorithms of MAP decoding(BCJR algorithm and so on)require complicated computation. That is not only number of operation but also memory space of working data. Therefore, sub-optimal, simplified soft-in soft-out decoding algorithms(i.e.SOVA and so on)have been studies. theretical studies of such a decoding have been studied from the viewpoints of learning theory, symbol dynamics, derivative geometry and so on.In this study, we find new viewpoint to achieve highly reliability of feedback decoding with soft-output. That is ordering of feed back scheme, and is related the variable threshold decoding of soft-input decoding of majority logic decodable codes.

软输入和软输出译码是并行(turbo)和串行级联码译码的重要技术。软入软出译码和级联编码是先进通信技术可靠性的关键技术。在本研究中，旨在提出新的解码算法和新的代码构造。该算法基于软输入软输出译码，代码适用于新型软输入软输出译码。为了研究，我们构建了带有存储通道的Turbo码的Turbo解码模拟器，该模拟器能够适应Turbo码的各种数据格式以及带/不带Turbo均衡的磁和移动信道。软输出信息（即可靠性信息）为每个信息比特的后验概率，理论上是通过最大后验（MAP）解码来计算的。 MAP解码算法(BCJR算法等)需要复杂的计算。这不仅是操作次数，也是工作数据的存储空间。因此，次优的、简化的软输入软输出译码算法（即SOVA等）得到了研究。人们从学习理论、符号动力学、导数几何等角度对这种译码进行了理论研究。在这项研究中，我们找到了实现高可靠性软输出反馈译码的新观点。即反馈方案的排序，并且与多数逻辑可译码码的软输入译码的可变阈值译码相关。