Coding Theorem for Systematic LDGM Codes Under List Decoding
Coding Theorem for Systematic LDGM Codes Under List Decoding
复制标题
列表译码下系统LDGM码的编码定理
DOI：
10.1109/itw.2018.8613510
复制
发表时间：
2018
期刊：
影响因子：
--
通讯作者：
Xiao Ma
中科院分区：
文献类型：
--
作者：
Wenchao Lin;Suihua Cai;Baodian Wei;Xiao Ma
本文关注三组系统低密度生成矩阵（LDGM）码，就误比特率（BER）而言，这三组码都已被证明可达到容量。然而，从本文构建的一个反例可以看出，这并不一定意味着它们在误帧率（FER）方面也能达到容量。然后我们表明，在二进制输入输出对称（BIOS）无记忆信道上进行列表译码时，第一组和第二组码可达到容量。我们指出，原则上，通过使用级联码，可以在速率损失可忽略的情况下消除由列表译码导致的含糊度。仿真结果表明，所考虑的卷积（空间耦合）LDGM码通过迭代置信传播译码算法可接近容量。
This paper is concerned with three ensembles of systematic low density generator matrix (LDGM) codes, all of which were provably capacity-achieving in terms of bit error rate (BER). This, however, does not necessarily imply that they achieve the capacity in terms of frame error rate (FER), as seen from a counterexample constructed in this paper. We then show that the first and second ensembles are capacity-achieving under list decoding over binary-input output symmetric (BIOS) memoryless channels. We point out that, in principle, the equivocation due to list decoding can be removed with negligible rate loss by the use of the concatenated codes. Simulation results show that the considered convolutional (spatially-coupled) LDGM code is capacity-approaching with an iterative belief propagation decoding algorithm.