Rice信道下LDPC码译码粘滞点研究  

Research of Fixed Point of LDPC Decoding over Rice Channel

作　　者：徐华[1] 徐澄圻[1] 

机构地区：[1]南京邮电大学通信与信息工程学院,江苏南京210003

出　　处：《南京邮电大学学报（自然科学版）》2007年第1期6-10,19,共6页Journal of Nanjing University of Posts and Telecommunications：Natural Science Edition

摘　　要：LDPC码译码误比特率与迭代次数关系曲线及其粘滞点的研究,对于进一步深入分析译码机理和加快译码收敛的速度有着重要的意义。文章以R ice信道为研究背景,提出了一种采用基于offset BP_based译码算法的DDE分析结合差分进化技术的非正则LDPC码的优化方法,据此得出了Rice信道下最大比特节点次数为10、码率为1/2和1/3的优化次数分布对以及相应的码门限值;通过仿真得到了在R ice信道下LDPC码译码粘滞点分析的相关曲线,并从译码粘滞点角度研究了LDPC码的门限现象。通过对各阶节点的译码粘滞点分析结果表明,高阶节点相对于低阶节点而言,对于译码有较大贡献,据此提出了改进译码算法的两种思路。Research of the fixed point which appears in decoding process of LDPC codes is important to further analyze the decoding mechanism and speed up decoding convergence. Firstly, a new optimization scheme which combines DDE （Discrete Density Evolution ）for offset BP-based decoding algorithm with differential evolution is presented for optimizing irregular LDPC codes. Then the optimized degree distribution pair and threshold of LDPC codes over Rice channel arc given, with the maximum bit node degree 10 and code rate R = 1/2,1/3. Furthermore, we analyze a series of simulated curves related to LDPC decoding fixed point and threshold phenomenon. Simulation results show that the node with higher degree contributes more to the decoding for LDPC codes. Therefore, we tentatively put forward two kinds of thinking to improve LDPC decoding algorithm.

关 键 词：LDPC码 Rice信道模型 离散密瘦进化 粘滞点 

分 类 号：TN911.22[电子电信—通信与信息系统] TN911.7[电子电信—信息与通信工程]