基于变量节点LLR消息加权的改进最小和算法

Improved Min Sum Algorithm Based on Weighted Message LLR of Variable Nodes

为了提高低密度奇偶校验(LDPC)码的单最小值最小和(single-minimum Min-Sum,sm MS)算法的误码性能,提出了一种基于变量节点LLR(Log Likelihood Ratio)消息加权的改进最小和(Improved Min Sum algorithm based on weighted message LLR of variable nodes,IMS-WVN)算法。首先,将迭代次数所确定的次小值的估值参数与最小值相加后取代次小值,以增强sm MS算法校验节点的可靠度。然后,将变量节点输出LLR消息与迭代前LLR消息进行加权处理,降低变量节点的振荡幅度,降低平均译码迭代次数。仿真结果表明,在信噪比为3.2 d B时,IMS-WVN算法的误码性能比VWMS算法提升0.53 d B,当误码率为10-5时,IMS-WVN算法平均译码迭代次数较MS算法减少58%。

In order to improve the bit-error-rate performance of the single-minimum Min-Sum algorithm for decoding Low- density parity check （LDPC） codes, the IMS-WVN （Improved Min Sum algorithm based on weighted message LLR of variable nodes） was proposed in this paper, firstly, determined the estimation parameter of the sub-minimum value accorded to the number of decoding iterations, and added the minimum value to replace the sub-minimum so as to enhance the reliability of the check-node. Secondly, the currently message of variable-to-cheek node and the message of old variable-to- cheek node were weighted to decrease the oscillation of the variable node and decrease the average number of decoding iteration. The simulation results show that the IMS-WVN algorithm had improved 0. 53 dB than VWMS algorithm and the 3. 2 dB order of error rate, when the error rate was 10.5 , The average number of iterations of the IMS-WVN algorithm is 58% less than that of the MS algorithm.