基于随机交织的分布式信源编码的实际设计  

Practical design of distributed source coding based on random interleavers

作　　者：曾伟超[1] 杨胜天[1,2] 仇佩亮[1] 

机构地区：[1]浙江大学信息与电子工程学系,浙江杭州310027 [2]西南交通大学信息编码与传输四川省重点实验室,四川成都610031

出　　处：《浙江大学学报（工学版）》2009年第11期1994-1999,共6页Journal of Zhejiang University：Engineering Science

基　　金：国家自然科学基金资助项目(60772093;60802014;60872063);浙江省自然科学基金资助项目(Y106068);西南交通大学信息编码与传输四川省重点实验室开放研究基金资助课题(2008-01);高等学校博士学科点专项科研基金资助项目(200803351023)

摘　　要：提出一种基于低密度奇偶校验(LDPC)码和随机交织器的对称Slepian-Wolf编码的实际设计方案.与已有方案不同,该方案对于具有相同速率的信源节点可以采用相同的LDPC编码器,只须通过不同的随机交织器就可以进行区分,使得系统的实现复杂度明显降低,尤其是在节点数量较大的情况下.该方案运用渐进边增长(PEG)算法对LDPC码的Tanner图进行优化,而在译码端利用信源之间的相关性进行联合迭代译码.在2、3个相关信源以及单个信源的特殊情形下的仿真结果表明,对于相关的非均匀信源,该方案在性能上优于已有方案.A practical scheme of symmetric Slepian Wolf coding based on low-density parity-check （LDPC） codes and random interleavers was presented. The difference between the previous schemes and the proposed one lies in that the source nodes with the same encoding rate can use the same LDPC code but with different interleavers, so the design complexity of the system is reduced evidently, especially for a large amount of source nodes. In the scheme, progressive edge-growth （PEG） algorithm is used to optimize the Tanner graph of the LDPC code, and joint iterative decoding is performed at the decoder by exploiting the correlation among sources. The simulation for the cases of two and three correlated sources as well as the special case of one single source shows that the proposed scheme performs better than the existed schemes for the correlated non-uniform sources.

关 键 词：Slepian—Wolf编码 分布式信源编码 LDPC码 随机交织器 联合迭代译码 

分 类 号：TN911.21[电子电信—通信与信息系统]