An improvement to soft-decision iterative decoding algorithms for linear block codes

线性分组码软判决迭代译码算法的改进

基本信息

批准号：
12650391
负责人：
KASAMI Tadao
金额：
$ 1.66万
依托单位：
Hiroshima City University
依托单位国家：
日本
项目类别：
Grant-in-Aid for Scientific Research (C)
财政年份：
2000
资助国家：
日本
起止时间：
2000 至 2001
项目状态：
已结题

来源：
https://kaken.nii.ac.jp/grant/KAKENHI-PROJECT-12650391/
关键词：
linear block code soft-decision decoding ML decoding Recursive ML decoding iterative decoding algebraic decoding

项目摘要

1. We introduced "adaptive maximum likelihood decoding" (abbreviated as AMLD). In AMLD, the most likely local (MLL) sub-codeword in a section is constructed by concatenating the MLL sub-codewords in its subsection which are found adaptively. We implemented a prototype program for AMLD, and evaluated decoding complexity of AMLD for several RM codes and BCH codes. The average complexity is considerably smaller than those of conventional maximum likelihood decoding algorithms. It turns out that for codes with relatively large parameter, there exist singular received sequences which occur at extremely rare intervals, however, winch require very long steps to guaarantee the optimality. To reduce the complexity furthermore, we have devised an effective method which makes good use of adjacency relation between the MLL sub-codewords in adjacent subsections [1, 3, 5]. In addition, we have studied early termination conditions with no or almost no loss of error performance.2. We have devised an e … More ffective method of selecting successive search centers around which bounded distance decoding is carried out in a soft-decision iterative decoding algorithm. In conventional soft-decision iterative decoding algorithms by using bounded distance decoding, such as Chase II decoding algorithm, the duplication of search region and the decoding failure of bounded distance decoding degrade the error performance. For several BCH codes of lengths 128 and 256, simulation results [2, 4] show that this new selection method improves the error performance considerably by reducing the above negative effect.3. For concatenated coding scheme whose outer code is an interleaved (225, 223) RS code a good choice of its inner code is proposed. The low weight structures of the binary images of shortened (8, 5) RS codes over GF(28) are analyzed extensively. Consequently, we have found several more promising shortened RS codes as the inner code than the conventional codes [6].* [x] refers to the x-th paper in the references. Less

1。我们引入了“自适应最大似然解码”（缩写为AMLD）。在AMLD中，一节中最有可能的局部（MLL）子代码是通过在其小节中定位的MLL子代码来构建的，该子编号被适应。我们为AMLD实施了一个原型程序，并评估了针对多个RM代码和BCH代码的AMLD的解码复杂性。平均复杂性被认为小于常规最大似然解码算法的平均复杂性。事实证明，对于具有相对大参数的代码，存在奇异接收的序列，这些序列以极少数的间隔发生，但是，绞车需要很长的步骤来保证最佳性。为了降低复杂性，我们设计了一种有效的方法，该方法很好地利用了相邻小节中MLL子代码之间的邻接关系[1，3，5]。此外，我们已经研究了早期终止条件，没有或几乎没有错误性能损失。2。我们设计了一种E…选择成功的搜索中心的更故障方法，该中心围绕该中心，在软性迭代迭代解码算法中进行了有界距离解码。在常规的软核能迭代解码算法中，通过使用有界距离解码（例如Chase II解码算法），搜索区域的重复以及有限距离解码的解码故障降低了误差性能。对于长度128和256的几种BCH代码，模拟结果[2，4]表明，这种新选择方法通过降低上述负面效应可大大提高错误性能3。对于串联编码方案，其外部代码是交错的（225，223）RS代码，提出了其内部代码的一个很好的选择。广泛分析了GF（28）上缩短（8，5）RS代码的二进制图像的低重量结构。因此，我们发现比常规代码[6]。* [x]提到参考文献中的X纸。较少的