Comparisons Between Reliability-Based Iterative Min-Sum and Majority-Logic Decoding Algorithms for LDPC Codes
Comparisons Between Reliability-Based Iterative Min-Sum and Majority-Logic Decoding Algorithms for LDPC Codes
复制标题
基于可靠性的 LDPC 码迭代最小和与多数逻辑译码算法的比较
DOI：
10.1109/tcomm.2011.060911.100065
复制
发表时间：
2011-06
影响因子：
8.3
通讯作者：
Bai, Baoming
中科院分区：
文献类型：
--
作者：
Chen, Haiqiang;Zhang, Kai;Ma, Xiao;Bai, Baoming
基于Huang等人最近的工作，提出了一种改进的基于可靠性的迭代多数逻辑解码（MRBI-MLGD）算法，用于两类结构化LDPC码。与原算法相比，改进后的算法性能更好，但复杂度略有提高。然后提出了一种基于可靠性的迭代最小和译码算法。在RBI-MSD算法中，基于可靠性的整数消息在变量节点和检查节点之间进行处理和交换。主要的计算只包括二进制逻辑运算和整数加法。与传统的最小和算法不同，变量节点传递完整的消息而不是外部消息来检查节点。这可以减少内存负载和计算复杂性，但性能下降很小（或可以忽略不计）。仿真结果表明，与(M)RBI-MLGD算法相比，本文提出的RBI-MSD算法具有更好的误码性能、更快的译码收敛速度和更少的量化比特，且计算复杂度适度增加。此外，RBI-MSD算法也适用于随机LDPC码的解码，这与(M)RBI-MLGD算法有明显的区别。最后，我们指出MRBI-MLGD算法和RBI-MSD算法中使用的尺度因子可以通过离散密度进化进行优化。
A modified reliability-based iterative majority-logic decoding (MRBI-MLGD) algorithm for two classes of structured LDPC codes is presented based on a recent work by Huang et al. Compared with the original one, the modified algorithm has better performance with slightly increased complexity. Then a reliability-based iterative min-sum decoding (RBI-MSD) algorithm is presented. For the presented RBI-MSD algorithm, reliability-based integer messages are processed and exchanged between variable nodes and check nodes. The main computations include only binary logical operations and integer additions. Different from the conventional min-sum algorithm, the variable nodes pass full messages rather than extrinsic messages to check nodes. This can reduce the memory loads and the computational complexity but with a little (or negligible) performance degradation. Simulation results show that, compared with the (M)RBI-MLGD algorithms, the presented RBI-MSD algorithm achieves better error performance, faster decoding convergence rate and fewer quantization bits with moderate increased computational complexity. Furthermore, the RBI-MSD algorithm is also applicable to decoding random LDPC codes, a distinct difference from the (M)RBI-MLGD algorithms. Finally, we point out that the scaling factors employed in the MRBI-MLGD algorithm and the RBI-MSD algorithm can be optimized using discretized density evolution.
登录
查看更多内容
影响因子：
2.5
作者：
Kou, Y;Lin, S;Fossorier, MPC
通讯作者：
Fossorier, MPC
影响因子：
2.5
作者：
Richardson, TJ;Shokrollahi, MA;Urbanke, RL
通讯作者：
Urbanke, RL
DOI：
10.1007/978-3-540-69457-1_3
发表时间：
2009-01-01
期刊：
LDPC CODED MODULATIONS
影响因子：
--
作者：
通讯作者：
--
影响因子：
8.3
作者：
Qin Huang;Jingyu Kang;Li Zhang;Shu Lin;K. Abdel-Ghaffar
通讯作者：
Qin Huang;Jingyu Kang;Li Zhang;Shu Lin;K. Abdel-Ghaffar
DOI：
10.1109/isit.1995.550455
发表时间：
1995-09
期刊：
Proceedings of 1995 IEEE International Symposium on Information Theory
影响因子：
--
作者：
N. Wiberg;Hans-Andrea Loeliger;R. Koetter
通讯作者：
N. Wiberg;Hans-Andrea Loeliger;R. Koetter