摘要
针对LDPC码识别过程中的稀疏校验矩阵重建问题,研究并提出了3种算法。在分析和比较LDPC码与一般分组码识别模型的基础上,将LDPC码的识别问题定义为寻找码字对偶空间下某组稀疏基的数学问题。通过以校验向量行重作为优化对象,先后设计和实现了了2-阶行间线性变换、p-阶行间线性变换、线性关系有限穷举的3种矩阵稀疏化算法,力求实现无误码条件下对适度码长长度LDPC码校验矩阵的有效重建。测试结果表明,该算法适用于包括802.16e、802.11n、DVB-S2、GJB7296、GB20600在内的多种LDPC码标准。
To solve the problem of restructuring sparse parity-check matrix in low-density parity-check（LDPC） recognition processing, three algorithms are proposed. Through analyzing and comparing the LDPC recognition model with the tradition coding recognition models, the former one is defined as a problem of finding a group sparse-base which spans the dual-space of the coding. Then, by making the weight of check-vector as the optimized object, the 2-order linear transformation algorithm, p-order linear transformation algorithm, and linear relationship exhaustive searching algorithm are proposed to restructure sparse parity-check matrix of a LDPC code with suitable code length in an error free environment. The result of simulations show that these algorithms fit most of LDPC standards, including 802.16 e, 802.11 n, DVB-S2, GJB7296, GB20600 and so on.
出处
《电子科技大学学报》
EI
CAS
CSCD
北大核心
2016年第2期191-196,共6页
Journal of University of Electronic Science and Technology of China
基金
国家自然科学基金(61172140)
关键词
信道编码识别
LDPC识别
校验矩阵
稀疏化
channel coding recognition
LDPC recognition
parity-check matrix
sparse
