块衰落信道下基于陷阱集约束的Root-LDPC编码算法研究

发布时间：2019-01-30 17:38

【摘要】：随着无线移动通信技术的不断发展,人们对具有更好可靠性与有效性的通信传输技术的需求越来越大。现实生活中的无线信道,由于受到多径传播和衰落的影响,常被抽象成一种随时间缓慢变化的信道模型一一块衰落信道。近年来,构造基于块衰落信道具有良好码性能和全分集的根校验的低密度奇偶校验码(Root Check Low-Density Parity-Check, Root-LDPC)成为了研究热点。基于边扩展(Progressive Edge-Growth, PEG)算法的基础上,本文分析了几种能够影响Root-LDPC码性能的主要因素,详细介绍了陷阱集的图形结构及相关概念,并重点分析了陷阱集对Root-LDPC码错误平层及误帧率的影响。大量研究表明：对Root-LDPC码的纠错性能危害最大的陷阱集就是初等陷阱集(Elementary Trapping Sets, ETS)。在这些ETS中,那些尺寸较小的被称为显性初等陷阱集(Dominant Elementary Trapping Sets, dominant ETS)对Root-LDPC码的错误平层影响更大。而在显性初等陷阱集的Tanner图中基本不可能找到度数大于2的校验节点,即几乎所有Root-LDPC码的显性陷阱集都是初等陷阱集。为了解决这一问题,本文提出了一种块衰落信道下基于陷阱集约束的PEG算法来构造Root-LDPC码。该算法将一种有效的初等陷阱集扩展算法与传统的PEG算法相结合。首先应用PEG算法逐列构造校验矩阵,然后利用初等陷阱集扩展算法找到现有Tanner图中尺寸较小的初等陷阱集并对其进行扩展,以此避免了显性初等陷阱集对Root-LDPC码性能的影响。仿真结果表明：利用该改进算法构造的Root-LDPC码较原始及其他方法改进的PEG算法具有更低的错误平层和误码率。同时,随着码长减小,其误帧率也随之降低。然而,在运行仿真的过程中,我们发现该改进PEG算法是一种贪婪算法,随着码长增大,编译码复杂度也随之升高。准循环算法(Quasi-Cyclic, QC)是一种结构化的构造算法,具有较低的编译码复杂度。因此,在基于陷阱集约束的PEG算法基础上,本文提出了一个将QC算法与提出的初等陷阱集抑制PEG算法相结合的改进算法。实验表明：该改进算法与初等陷阱集抑制PEG算法具有相近的优良编码性能,相较现有PEG算法及其他改进算法具有更低的错误平层,并具有更短的运行仿真时间,达到了码的误帧率性能和编译码复杂度的平衡,具有较强的应用性。
[Abstract]:With the development of wireless mobile communication technology, there is a growing demand for more reliable and effective communication transmission technology. Due to the influence of multipath propagation and fading, wireless channels in real life are often abstracted into a fading channel model which changes slowly with time. In recent years, the construction of low-density parity-check codes (Root Check Low-Density Parity-Check, Root-LDPC) based on block fading channels with good code performance and full diversity has become a hot topic. Based on the edge expansion (Progressive Edge-Growth, PEG) algorithm, this paper analyzes several main factors that can affect the performance of Root-LDPC codes, and introduces the graphic structure and related concepts of the trap set in detail. The effect of trap set on Root-LDPC error level and frame error rate is analyzed. A large number of studies show that the primary trap set (Elementary Trapping Sets, ETS). Is the most harmful trap set to the error correction performance of Root-LDPC codes. In these ETS, the smaller size known as the dominant elementary trap set (Dominant Elementary Trapping Sets, dominant ETS) has more influence on the error level of Root-LDPC codes. However, it is almost impossible to find the check node with degree greater than 2 in the Tanner graph of the dominant elementary trap set, that is, almost all the dominant trap sets of Root-LDPC codes are elementary trap sets. In order to solve this problem, this paper proposes a PEG algorithm based on trap set constraints to construct Root-LDPC codes in block fading channels. This algorithm combines an effective elementary trap set expansion algorithm with the traditional PEG algorithm. First, the PEG algorithm is used to construct the checkout matrix, and then the elementary trap set extension algorithm is used to find and extend the small size elementary trap set in the existing Tanner graph, so as to avoid the influence of the dominant elementary trap set on the performance of Root-LDPC code. The simulation results show that the Root-LDPC code constructed by the improved algorithm has lower error level and bit error rate than the original and other improved PEG algorithm. At the same time, as the code length decreases, its frame error rate also decreases. However, in the course of running simulation, we find that the improved PEG algorithm is a greedy algorithm, and the complexity of encoding and decoding increases with the increase of code length. Quasi-cyclic algorithm (Quasi-Cyclic, QC) is a structured construction algorithm with low encoding and decoding complexity. Therefore, on the basis of PEG algorithm based on trap set constraint, this paper proposes an improved algorithm which combines the QC algorithm with the proposed elementary trap set suppression PEG algorithm. The experimental results show that the improved algorithm has similar excellent coding performance to the PEG algorithm, and has lower error level and shorter simulation time than the existing PEG algorithm and other improved algorithms. It achieves the balance of frame error rate performance and encoding and decoding complexity, and has a strong application.
【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】