原模图LDPC码的准循环扩展算法研究

发布时间：2018-01-06 14:24

本文关键词：原模图LDPC码的准循环扩展算法研究　出处：《东北大学》2014年硕士论文　论文类型：学位论文

更多相关文章： 原模LDPC码 准循环扩展 短环 连通度

【摘要】：低密度奇偶校验(Low Density Parity Check, LDPC)码性能逼近香农容限并具有较低的译码复杂度,为此受到越来越多的关注。在此基础上,美国喷气推进(Jet Propulsion Laboratory, JPT)实验室提出了原模图LDPC码,此种编码技术因具有编译码复杂度低和误码性能良好等特点,而成为一类重要的LDPC码,并已成为卫星数字电视传输标准(Digital Video Broadcasting-Satellite 2 DVB-S2) 以及 CCSDS (Consultative Committee for Space Data Systems, CCSDS)深空通信等通信标准中的信道编码方案。众所周知原模图准循环扩展算法(简称为PQCE算法)不仅影响构造原模图LDPC码的误码性能,而且还决定着原模图LDPC码的编译码器的硬件实现复杂度,但PQCE算法中仍有许多问题有待于研究,所以研究PQCE算法具有重要意义。本文的主要工作如下：(1)介绍了原模图LDPC码的研究背景、意义和研究现状,然后阐述了其基本理论、编译码方法、以及PEXIT (Protograph EXIT, PEXIT)图。(2)针对利用现有PQCE算法构造的校验矩阵中存在大量短环和较慢收敛速度问题,提出了PEG-PH-PQCE算法。该算法首先利用PEG (Progressive Edge Growth, PEG)去重边扩展算法获得基矩阵完成原模图的第一步扩展；然后利用PH准循环扩展算法完成第二步扩展,即通过PEG准循环扩展算法得到初始指数矩阵,之后利用登山(HillClimbing, HC)算法优化初始指数矩阵,最终获得性能较好的校验矩阵。仿真实验表明该算法构造的校验矩阵中短环数量少、算法收敛速度快。(3)针对利用现有PQCE算法构造的原模图LDPC码环之间的连通度较低问题,提出了PE-IPEG-PQCE算法。该算法首先通过PEG去重边扩展算法获得初始基矩阵和边交换操作优化初始基矩阵中的环分布,完成原模图的第一步扩展；然后引入ACE和短环数量作为标准扩展树图,给出了PEG准循环扩展算法,利用此算法完成第二步扩展,并能够获得到连通度较高的校验矩阵。仿真实验表明所提算法不仅能够有效地增大环之问的连通度而且能够减少短环数,从而提高了原模图LDPC码的误码性能。
[Abstract]:The performance of low Density Parity check (LDPC) codes approximates Shannon's tolerance and has low decoding complexity. More and more attention has been paid to this. On this basis, Jet Propulsion Laboratory of the United States is promoted. The prototype LDPC code is proposed by JPTLab. This coding technique is a kind of important LDPC codes because of its low complexity and good error-code performance. Digital Video Broadcasting-Satellite 2 DVB-S2) has become the standard of satellite digital TV transmission. And Committee for Space Data Systems. As we all know, the quasi-cyclic expansion algorithm (PQCE algorithm) not only affects the error performance of constructing the original mode map LDPC code, but also the channel coding scheme in the communication standards such as deep space communication. And it also determines the complexity of the hardware implementation of the LDPC code, but there are still many problems in the PQCE algorithm to be studied. So it is of great significance to study the PQCE algorithm. The main work of this paper is as follows: 1) introduce the background, significance and research status of the primitive LDPC code, and then explain its basic theory. Encoding and decoding methods, and PEXIT / Protograph EXIT. In view of the check matrix constructed by using the existing PQCE algorithm, there are a large number of short rings and slow convergence rate problems. A PEG-PH-PQCE algorithm is proposed, which uses PEG Progressive Edge Growth at first. The peg) algorithm obtains the first step extension of the base matrix to complete the original mode graph. Then the PH quasi-cyclic expansion algorithm is used to complete the second step expansion, that is, the initial exponential matrix is obtained by the PEG quasi-cyclic expansion algorithm, and then the climbing algorithm is used. HCI) algorithm optimizes the initial exponential matrix and finally obtains a good performance check matrix. The simulation results show that the number of short loops in the algorithm is less. The convergence rate of the algorithm is fast. (3) aiming at the problem of low connectivity between LDPC code rings of primitive mode graph constructed by existing PQCE algorithm. In this paper, PE-IPEG-PQCE algorithm is proposed. Firstly, the initial basis matrix and the edge exchange operation are obtained by using the PEG resetting extension algorithm to optimize the ring distribution in the initial base matrix. Complete the first step extension of the original mode diagram; Then the ACE and the number of short rings are introduced as the standard extension tree graph, and the PEG quasi-cyclic expansion algorithm is given, and the second step is completed by using this algorithm. The simulation results show that the proposed algorithm can not only increase the connectivity of rings but also reduce the number of short rings. Thus, the error performance of the original mode diagram LDPC code is improved.
【学位授予单位】：东北大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN911.22

【相似文献】