基于稀疏因子图的联合编码调制优化设计

发布时间：2018-02-07 12:37

本文关键词： LDPC联合编码调制低密度格码广义低密度生成阵码稀疏因子图信号设计和优化　出处：《北京邮电大学》2014年博士论文　论文类型：学位论文

【摘要】：因子图是一种用于迭代接收机分析和设计的有效工具,联合编码调制是逼近高频谱效率下仙农信道容量的有效手段。本论文深入研究了三种基于稀疏因子图的联合编码调制方案。主要工作有： 1.提出了一种基于叠加映射规则的LDPC联合编码调制方案,包括：(1)根据信号的叠加映射规则设计优化了LDPC联合编码调制的信号星座,使得信道输入和输出信号之间互信息最大。(2)基于逐节点消息传递策略,给出了低复杂度的解映射算法,并且利用半解析的方法分析解映射器输出软消息的概率密度函数。(3)联合优化了叠加映射的LDPC联合编码调制的度分布。其主要创新是,联合编码和调制进行度分布优化。相对于传统的LDPC联合编码调制,所优化的LDPC联合编码调制具有更优的性能。 2.优化设计了AWGN信道和ISI信道下的低密度格码(LDLC),包括：(1)提出了用于AWGN信道的基于单个高斯分布近似(即将消息的概率密度函数近似为单个高斯分布)的低复杂度译码算法；根据单个高斯近似和全零格点的假设,优化了AWGN信道下LDLC码。(2)设计了用于码间干扰(ISI)信道的LDLC码。证明了在ISI信道下,通过将ISI引入LDLC码,会取得额外的标称编码增益。该编码增益可以通过成形操作获得。针对特定的ISI信道,优化设计了LDLC码,进行仿真验证。其主要创新在于,优化了AWGN信道和ISI信道下LDLC码的校验矩阵。因为在设计中自然地匹配了特定的信道条件,较之非二进制LDPC码,优化的AWGN信道下的LDLC码具有更好的阈值特性；而相对于ISI信道下的优化的LDPC码,优化的LDLC码具有更好的性能。 3.提出了广义低密度生成阵码(G-LDGM codes)的低复杂度的译码算法及其收敛性分析方法、信号优化设计和成形操作的设计方案。(1)推导出了基于稀疏生成矩阵的低复杂度的参数化高斯置信度传播译码算法,并且进行了收敛性分析,证明了该译码算法的收敛条件。(2)针对功率受限AWGN信道和功率不受限AWGN信道,优化设计了G-LDGM码。(3)设计了一种基于稀疏生成矩阵的成形算法,用于功率受限的通信信道。其主要创新是,优化了G-LDGM码的稀疏生成矩阵。相对于LDLC码,优化的G-LDGM码具有更好的阈值特性和更低的编码复杂度。
[Abstract]:Factor graphs are an effective tool for iterative receiver analysis and design. Joint coded modulation (JCM) is an effective method to approach Shannon channel capacity with high spectral efficiency. In this paper, three schemes of JCM based on sparse factor graph are studied. The main work is as follows:. 1. A LDPC joint coded modulation scheme based on the superposition mapping rule is proposed, including: 1) the signal constellation of LDPC joint coding modulation is optimized according to the superposition mapping rule of the signal. Based on the node-by-node message passing strategy, a low-complexity de-mapping algorithm is proposed, which makes the mutual information between the input and output signals of the channel maximum. And the probability density function of demapper output soft message is analyzed by semi-analytic method, and the degree distribution of LDPC joint coding modulation of superposition map is optimized. The main innovation is that, Compared with the traditional LDPC joint coding modulation, the optimized LDPC joint coding modulation has better performance. 2. The low density trellis codes in AWGN channel and ISI channel are optimized and designed. A low complexity decoding algorithm based on a single Gao Si distribution approximation (i.e. the probability density function of message is approximate to a single Gao Si distribution) is proposed for AWGN channel. Based on the assumption of a single Gao Si approximation and a full zero lattice point, the LDLC code for ISI) channel in AWGN channel is optimized. It is proved that in ISI channel, the ISI is introduced into the LDLC code. Additional nominal coding gain will be obtained. The coding gain can be obtained by shaping operation. LDLC codes are optimized for specific ISI channels and verified by simulation. The main innovation is that, The check matrix of LDLC codes in AWGN channel and ISI channel is optimized. Because the special channel conditions are naturally matched in the design, the LDLC code in the optimized AWGN channel has better threshold characteristics than that in the non-binary LDPC code. Compared with the optimized LDPC codes in ISI channel, the optimized LDLC codes have better performance. 3. A low complexity decoding algorithm for generalized low density generating matrix codes (G-LDGM codesand its convergence analysis method) is proposed. The design scheme of signal optimization design and shaping operation. (1) the low complexity parameterized Gao Si confidence propagation decoding algorithm based on sparse generation matrix is derived, and the convergence analysis is carried out. It is proved that the convergence condition of the decoding algorithm. (2) for the power constrained AWGN channel and the power unconstrained AWGN channel, the G-LDGM code. For power constrained communication channels, the main innovation is to optimize the sparse generation matrix of G-LDGM codes. Compared with LDLC codes, the optimized G-LDGM codes have better threshold characteristics and lower coding complexity.
【学位授予单位】：北京邮电大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TN911.22

【相似文献】