基于变量节点更新改进的最小和算法  

An Improved Min-Sum Algorithm Based on Variable Node Update

作　　者：范芳 冯雪林[2] 

机构地区：[1]重庆邮电大学通信与信息工程学院,重庆400065 [2]中国科学院计算技术研究所移动计算与新型终端北京市重点实验室,北京100190

出　　处：《无线电工程》2018年第2期121-125,共5页Radio Engineering

摘　　要：LDPC码是一种具有稀疏性且接近香农极限的线性分组码。目前被广泛应用的LDPC简化算法的译码性能损失较多,译码性能和复杂度折中的算法具有重要价值。对几种已知的译码算法进行了深入研究,提出一种基于变量节点更新改进的最小和算法,该算法与最小和算法、归一化最小和算法复杂度相当,但译码性能得到约0.5 d B和0.2 d B的提高,在中高信噪比区,更加接近LLR-BP算法的性能。The LDPC code is a linear block code which approaches Shannon’s limit. At present,the widely used LDPC simplificationalgorithm has much decoding performance loss,and the compromise algorithm of decoding performance and complexity is of great value.By studying several known decoding algorithms in depth,this paper proposed an improved min-sum algorithm based on variable node update. The complexity of proposed algorithm is equivalent to that of min-sum and normalization min-sum algorithms. Compared with the min-sum and normalization min-sum algorithm,the decoding performance of the proposed algorithm is increased by about 0. 5 dB and 0. 2 dB,and in the high signal-to-noise ratio region,its performance approaches that of LLR-BP algorithm.

关 键 词：低密度奇偶校验码 最小和 归一化 变量节点更新 

分 类 号：TN911.22[电子电信—通信与信息系统]