中山大学学报自然科学版 ›› 2010, Vol. 49 ›› Issue (3): 47-51.

• 研究论文 • 上一篇    下一篇

基于校验节点的LDPC码的消息加权均值串行译码算法

韩国军1,2, 刘星成1

  

  1. (1中山大学电子与通信工程系, 广东 广州510275) (2广东工业大学信息工程学院,广东 广州510006)
  • 收稿日期:2009-05-24 修回日期:1900-01-01 出版日期:2010-05-25 发布日期:2010-05-25
  • 通讯作者: 刘星成

Serial Update Scheduling Based on Weighted Average Message of Check Nodes for LDPC Codes

HAN Guojun1,2, LIU Xingcheng1

  

  1. (1Department of Electronic and Communications Engineering, Sun Yatsen University, Guangzhou 510275, China)(2Faculty Information Engineering, Guangdong University of Technology, Guangzhou 510006, China)
  • Received:2009-05-24 Revised:1900-01-01 Online:2010-05-25 Published:2010-05-25

摘要:

中短长度LDPC码在采用BP算法译码时,由于短环的存在会导致某些比特LLR值出现振荡现象,从而影响了译码性能。针对这一现象,提出了一种采用消息加权均值来修正校验节点到变量节点消息传递的改进译码算法。该译码算法将上次校验节点到变量节点的消息和本次计算所得的校验节点到变量节点的消息进行加权求和后作为当前最新的消息。通过对本次计算所得的校验节点到变量节点的消息赋予比上次校验节点到变量节点消息更高的权值,从而减弱消息加权均值对正常BP消息传递的影响,有效抑制比特LLR值的振荡。仿真结果表明,与传统的串行迭代译码算法相比,改进的译码算法在计算复杂度增加极少的情况下,有效提高了中高信噪比区的译码性能。



关键词: 低密度校验(LDPC)码, 置信传播, 迭代译码, 因子图

Abstract:

For short and middle length LDPC codes, the oscillation of LLR owing to cycles degrades the performance of BP algorithm. In order to minimize the oscillation of LLR during BP decoding process, an improved BP algorithm with serial update scheduling using weighted average message to update the message of checktovariable node is proposed. Currently computed message of checktovariable node and the message of old checktovariable node are weighted and then added up to update the message of checktovariable node. By higher weighted currently computed message of checktovariable node than the message of old checktovariable node, the improved algorithm effectively minimizes oscillation of LLR, at the same time, keeps normal BP message propagation and low compute complexity. Simulation results show that, for short and middle length LDPC codes, the improved algorithm achieves more noticeable error performance than CSBP algorithm at middle to high signal to noise ratio.

Key words: LDPC codes, beliefpropagation, lterative decoding, factor graph

中图分类号: