一种低错误平层LDPC码构造方法  

A construction method of LDPC codes with the low error floor

作　　者：袁建国[1] 汪哲[1] 高文春[1] 吴英冬 郭乔[1] 胡潇月 YUAN Jianguo WANG Zhe GAO Wenchun WU Yingdong GUO Qiao HU Xiaoyue(Key Laboratory of Optical Communication and Networks, Chongqing University of Posts and Tel Chongqing 400065, P. R. Chin)

机构地区：[1]重庆邮电大学光通信与网络重点实验室,重庆400065

出　　处：《重庆邮电大学学报（自然科学版）》2017年第1期15-18,共4页Journal of Chongqing University of Posts and Telecommunications(Natural Science Edition)

基　　金：国家自然科学基金项目(61472464);重庆市基础与前沿研究计划项目(cstc2015jcyj A0554);2016年重庆邮电大学大学生科研训练计划项目(A2016-61)~~

摘　　要：针对低密度奇偶校验(low-density parity-check,LDPC)码在高信噪比区域可能存在错误平层的缺点,提出一种具有低错误平层LDPC码的新颖构造方法。在该方法中,基本矩阵由渐进边增长(progressive edge growth,PEG)算法搜索构造,通过在基本矩阵相应的Tanner图中增加校验节点,并将其与拥有最小额外信息度(extrinsic message degree,EMD)短环的变量节点相连来增大短环的连通性。另外,提出了一种基于伽罗华域的循环移位系数矩阵设计方案,无需计算机搜索即可完全避免4环的出现,降低算法复杂度。为了对该方法的可行性进行验证,分别对变量节点的度分布是规则和非规则的基本矩阵进行改进,在高斯白噪声(additive white gaussian noise,AWGN)信道下,采用置信传播(belief propagation,BP)迭代译码算法对改进后的码型进行仿真分析,仿真结果表明,利用该法所构造的码型可有效改善在高信噪比区域的错误平层。Considering that low density parity check（ LDPC） codes may exist the fault of the error floor in the high signal noise ratio（ SNR） region,a construction method of LDPC codes with the low error floor was proposed. In the proposed method,a basic matrix is constructed and searched by the progressive edge growth（ PEG） algorithm,and the connectivity of the short cycle can be effectively increased by way of adding the new check nodes of the corresponding Tanner diagram in the basic matrix and connecting them with the variable nodes of the short cycle which has the minimum extrinsic message degree（ EMD） value. In addition,a novel cyclic shift coefficient matrix scheme based on the Galois field method was presented,which can avoid the phenomenon of the girth-4 without the computer search,thus the algorithm complexity is reduced. In order to verify the feasibility of this construction method,the basic matrixes with the regular and irregular degree distributions of the variable nodes are improved respectively. The improved codes are simulated and analyzed under the addictive white gaussian noise（ AWGN） channel and the belief propagation（ BP） decoding algorithm. The simulation result shows that the LDPC codes constructed by the proposed method can effectively reduce the error floor in the high signal noise ratio region.

关 键 词：渐进边增长(PEG)算法 额外信息度(EMD) 低密度奇偶校验(LDPC)码 错误平层 

分 类 号：TN911.22[电子电信—通信与信息系统]