基于整数运算的LDPC码最小和译码算法  

Min-sum decoding algorithm for LDPC codes based on integer operation

作　　者：野晓东[1,2] 马林华[1] 王卫民[1] 李森[1] 

机构地区：[1]空军工程大学工程学院,陕西西安710038 [2]九四五三七部队,河南许昌461132

出　　处：《通信学报》2010年第6期106-111,共6页Journal on Communications

摘　　要：对低密度奇偶校验码(LDPC)在高斯信道下的译码算法进行了深入研究,提出了一种基于整数运算的最小和译码算法。在该算法中所有变量都用固定长度的整数表示,因此非常便于硬件实现,其性能在信噪比大于3dB时,与高精度浮点数的和积译码算法相当;同时对最小和算法提出了一个修正因子,加上修正因子,其译码性能有进一步的提高,特别是在低信噪比下改善较大;另外还发现在高信噪比下,该算法具有更低的错误平层。The LDPC codes decoding algorithm over additive white gaussian noise（AWGN） channels was studied,and a min-sum algorithm based on integer operation was proposed for decoding LDPC codes.All the variables were fixed length integers in this algorithm,which was very convenient for hardware implementation.At the signal-to-noise ratios（SNR） above 3dB,its performance was equal to that of sum-product algorithm whose variables were high precision floating point numbers.Also,a coefficient was designed to modify the min-sum algorithm,whose decoding performance was improved greatly especially at low SNR.And another achievement had been found that this algorithm had a lower error floor at high SNR.

关 键 词：纠错码 低密度奇偶校验码 最小和算法 整数运算 

分 类 号：TN911.22[电子电信—通信与信息系统]