并行分布匹配算法仿真数据集

"由于标准M-QAM调制格式光信号中，各符号出现的概率相同，系统容量难以逼近香农极限，在加性高斯白噪声信道中，标准M-QAM信号的传输容量与香农容量间的差距约为1.53dB。围绕星座概率整形技术，对概率幅度整形系统和分布匹配器进行了深入地研究。对于并行分布匹配器的研究主要面向高阶QAM调制信号，基于恒定组成分布，研究概率幅度整形，实现对28Gbaud的16QAM、64QAM和256QAM信号的概率整形，通过VPI仿真软件搭建了仿真系统，仿真比较了不同信噪比下的广义信息量和LDPC译码前后的误码率，分析在净比特速率相同的条件下，概率整形信号相比于均匀分布信号的优势。
此外，基于上述的工作，对恒定组成分布匹配器的进行设计和仿真，针对CCDM的速率损失问题，基于此设计了并行结构的分布匹配器以改善CCDM在短块长度下的速率损失问题，采用VPI与matlab联合仿真分析了采用该成对优化并行分布匹配器的概率整形信号的广义互信息量和误码率性能，从系统角度说明了该方案相比于CCDM的优势。最后计算所提并行分布匹配器方案在不同输出块长度下的速率损失，并从理论和仿真两方面说明该方案的可行性及相比于CCDM的优势。以PS-64QAM信号为例，仿真分析了不同光信噪比条件下，n为30、50和100时，采用该成对优化并行分布匹配器的PS-64QAM信号的广义互信息量，以及n为100和无穷（n=50000）时采用CCDM的PS-64QAM信号的广义互信息量，并与均匀分布的64QAM信号作对比。相比于CCDM，采用该并行分布匹配器的PS-64QAM信号所需的光信噪比更小，可改善约0.12dB左右；在光信噪比一定的条件下，采用该成对优化并行分布匹配器的PS-64QAM信号，相比于采用CCDM的PS-64QAM信号，块长度可节约40%，说明了该成对优化并行分布匹配器可使用更少的块长度，实现和CCDM相同的性能。

For optical signals with standard M-QAM modulation format, since each symbol has equal probability of occurrence, the system capacity is difficult to approach the Shannon limit. In additive white Gaussian noise (AWGN) channels, the gap between the transmission capacity of standard M-QAM signals and the Shannon capacity is approximately 1.53 dB. This work conducted in-depth research on probabilistic amplitude shaping (PAS) systems and distribution matchers (DMs) focusing on constellation probability shaping (PS) technology.
For the research on parallel distribution matchers, it mainly targeted high-order QAM modulation signals. Based on constant composition distribution (CCD), probabilistic amplitude shaping was studied to realize probability shaping for 28Gbaud 16QAM, 64QAM and 256QAM signals. A simulation system was built using VPI simulation software, and the generalized mutual information (GMI) under different signal-to-noise ratios (SNRs) and the bit error rate (BER) before and after LDPC decoding were simulated and compared. The advantages of probability-shaped signals over uniformly distributed signals under the same net bit rate were analyzed.
In addition, based on the above work, the constant composition distribution matcher (CCDM) was designed and simulated. Aiming at the rate loss problem of CCDM, a parallel-structure distribution matcher was designed to alleviate the rate loss of CCDM under short block lengths. The joint simulation of VPI and Matlab was adopted to analyze the GMI and BER performance of probability-shaped signals using the proposed pairwise optimized parallel distribution matcher (PO-PDM), and the advantages of this scheme over CCDM were demonstrated from a system perspective. Finally, the rate loss of the proposed parallel distribution matcher scheme under different output block lengths was calculated, and the feasibility of this scheme and its advantages over CCDM were verified through both theoretical analysis and simulation.
Taking PS-64QAM signals as an example, the GMI of PS-64QAM signals using the PO-PDM was simulated and analyzed when n=30, 50 and 100 under different optical SNRs, as well as the GMI of PS-64QAM signals using CCDM when n=100 and infinity (n=50000), and the results were compared with those of uniformly distributed 64QAM signals. Compared with CCDM, the PS-64QAM signals using the proposed parallel distribution matcher require lower optical SNR, with an improvement of approximately 0.12 dB. Under a fixed optical SNR, the block length of PS-64QAM signals using the PO-PDM can be reduced by 40% compared with those using CCDM, which demonstrates that the PO-PDM can achieve the same performance as CCDM with a shorter block length.

搜集汇总

背景与挑战

背景概述

该数据集聚焦于光纤通信技术中的概率幅度整形和并行分布匹配算法研究，通过VPI和Matlab仿真生成了16QAM、64QAM和256QAM调制信号的性能数据，旨在解决恒定组成分布匹配器在短块长度下的速率损失问题。仿真结果显示，采用该并行分布匹配器可改善光信噪比约0.12dB，并在相同性能下节约40%的块长度，为高阶QAM调制系统的优化提供了关键实验数据。