多归一化因子的LDPC译码算法

LDPC decoding algorithm with multiple normalization factors

为提高低密度奇偶校验码的译码性能,更好满足5G系统对于误码率的需求,提出多归一化因子最小和(multiple normalized dactors min-sum,MNF-MS)算法。此算法以归一化最小和(normalized min-sum,NMS)与密度演化最小和(density evolution min-sum,DE-MS)算法为基础,将蒙特卡罗仿真求归一化因子的方法进行简化,求得校验节点信息平均绝对值和最小绝对值的比值,这个比值加上一个大于零的权值得到对应校验节点的归一化因子,迭代次数平均分为5组,每组使用不同的权值。仿真结果表明,所提算法与现有算法相比,其译码性能更为优异。

In order to improve the decoding performance of low-density parity codes and better meet the requirements of 5G system for bit error rate,this paper proposed MNF-MS(multiple normalized factors min-sum)algorithm.This algorithm is based on NMS(normalized min-sum)and DE-MS(density evolution min-sum)algorithm and the method of calculating the normalization factor by Monte Carlo simulation is simplified.First,the ratio of the average absolute value of the information of the verification node to the minimum absolute value was obtained.Then,the ratio plus a weight greater than zero was used to get the normalization factor of the corresponding verification node.The number of iterations was divided into five groups,and each group used different weights.The simulation results show that the decoding performance of the algorithm proposed in this paper is about 0.15 dB better than that of the DE-MS algorithm when the error rate is 10^(-6).