【作者】 钟康平；

【导师】 李唐军；

【作者基本信息】 北京交通大学 ， 通信与信息系统， 2014， 博士

【摘要】 基于高阶调制码型、相干检测和数字信号处理技术的相干光通信,由于其能够实现高频谱效率、在电域实现传输损伤补偿等优势,成为下一代高速光通信系统发展方向。本论文结合国家863项目“160Gb/s一泵多纤光传输技术的研究”以及加拿大自然科学基金战略基金项目‘’Digital signal processing for fiber optic communication",围绕基于偏振复用16进制正交幅度调制(DP-16QAM)码型、相干检测和数字信号处理技术的相干光通信系统进行了深入的理论和实验研究。取得的主要研究成果如下：1.提出一种改进型四电平信号生成方案,实验生成了高质量14Gbaud/s、28Gbaud/s、32Gbaud/s、40Gbaud/s四电平信号。实验研究了基于改进型四电平信号生成方案的DP-16QAM发射机性能,实验结果显示112Gb/s和224Gb/sDP-16QAM发射机性能与理论值分别仅有2.7dB和3.5dB的光信噪比(OSNR)代价。搭建了112Gb/s DP-16QAM实验传输系统,实验研究了其传输性能。实验结果显示,在仅补偿光纤色散情况下单通道112Gb/s DP-16QAM传输系统最优入纤功率点为-3.5dBm；当前向纠错编码(FEC)阈值为3.8×10-时,实验实现了112Gb/sDP-16QAM信号1470km传输。2.开发了一套完整的适用于DP-16QAM的相干接收机数字信号处理算法,其中包括基于Gram-Schmidt正交法则的IQ非正交补偿算法；基于频域均衡器的色散补偿算法；数字时钟恢复；偏振复用信号的偏振解复用及动态均衡；载波频偏估计与补偿；载波相位噪声估计与补偿；判决及误码计算。仿真及实验验证了整套数字信号处理算法的性能。首次实验研究了基于Gram-Schmidt正交法则的IQ非正交补偿算法在DP-16QAM系统中的性能。3.研究了基于数字反向传输(DBP:Digital Back-Propagation)的光纤非线性补偿技术在DP-16QAM长距离相干光通信系统中的应用。分析了偏振复用传输系统中的光纤非线性效应。仿真与实验研究了该补偿技术中加权系数及补偿步长两个关键参数,结果显示：对于特定的传输系统,DBP的加权系数对补偿步长、入纤功率、传输距离等均不敏感。在综合考虑补偿性能及计算复杂度情况下,DBP的最优补偿步长为传输系统的中继距离。采用该补偿技术,112Gb/s DP-16QAM长传输系统的最优入纤功率被提升了2dB；当FEC阈值为3.8x10-时,实验实现了112Gb/s DP-16QAM信号2400km传输,最远传输距离比未采用该技术情况下提升近63%。4.提出三种基于改进型四相移键控分区算法(MP)的级联载波相位噪声估计(CPE)算法。研究了算法线宽容忍度、相位噪声估计性能、线性及非线性传输性能和硬件复杂度。仿真与实验结果表明：三种算法性能均与相位盲搜索(BPS)算法相近,硬件复杂度却比BPS算法分别降低了1.5-10倍。5.为了进一步降低硬件复杂度,提出一种基于准零乘法器算法的级联CPE算法。仿真与实验研究结果显示：级联CPE算法的线宽容忍度、相位噪声估计性能、线性及非线性传输性能均与BPS算法相当,但该算法硬件复杂度比BPS算法降低了17倍,是迄今为止报道的针对16QAM调制码型的硬件复杂度最低的前反馈级联CPE算法。6.提出将线性插值算法应用于载波相位噪声补偿技术中,提出基于线性相位插值(LPI)的增强型载波相位噪声补偿算法,理论与实验研究了该补偿算法性能。仿真与实验结果表明：LPI算法与滑动窗口(SW)补偿算法性能相当,好于数据块平均(BA)补偿算法；然而LPI算法的硬件复杂度仅与BA算法相近,远远低于SW补偿算法。更多还原

【Abstract】 Optical coherent communication systems, based on high-order modulation format, coherent detection and digital signal processing, has been considered as a promising technology for next generation high-speed optical communication systems due to its ability for achieving high spectral efficiency and compensation for impairments in electric field. In this thesis, we conduct theoretical and experimental investigation on the high-speed optical coherent communication systems based on dual polarization16ary quadrature amplitude modulation (DP-16QAM) modulation format, coherent detection and digital signal processing. This work is jointly supported by National High Technologies Development Program (863)"Key technologies in160Gb/s optical transmission systems using multi-fiber pump" and Canada Natural Sciences and Engineering Research Council strategic project grants "Digital signal processing for fiber-optic communications". The main contributions of this thesis are listed as follows:1. An improved4-level signal generation method is proposed. High quality4-level signal at baud-rates of14Gbaud/s,28Gbaud/s,32Gbaud/s and40Gbaud/s are experimentally demonstrated. With the improved4-level signal generation, Only2.7dB and3.5dB OSNR penalty for the112Gb/s and the224Gb/s DP-16QAM transmitter are investigated, respectively. The performance of a112Gb/s DP-16QAM transmission system is experimentally investigated. The optimal launch power is-3.5dBm. With FEC threshold of3.8×10-3, a maximum transmission distance of1470km is demonstrated for112Gb/s DP-16QAM signal only with linear impairments compensation.2. A complete set of digital signal processing algorithms for DP-16QAM modulation format is developed. Including IQ imbalanced compensation, resampling, fix frequency domain equalizer for dispersion compensation, digital square and filtering clock recovery, polarization recovery and adaptive equalization, carrier frequency offset estimation and compensation, carrier phase noise estimation and compensation, symbol decision and bit error rate calculation. The performance of the digital signal processing algorithms is verified both by simulation and experimental data. The performance of Gram-Schmidt orthogonalization procedure based IQ imbalance compensation algorithm is first experimentally demonstrated in DP-16QAM system.3. Digital back-propagation (DBP) based fiber nonlinearity compensation is numerically and experimentally studied in long-haul DP-16QAM transmission systems. The scaling factor and step-size of DBP are fully studied both numerically and experimentally. The results show that the scaling factor is insensitivity to step-size, launch power and transmission distance. Considering the balance between computation complexity and performance, the optimal step-size of DBP is1span/step. Using DBP with a step size of1step/span, a2dB improvement of optimal launch power and a maximum transmission distance of2400km is experimentally demonstrated for a112Gb/s DP-16QAM transmission system, which improves the maximum transmission distance by63%compared to the case only with CD compensation.4. A modified QPSK partitioning carrier phase estimation (CPE) algorithm is proposed. Combined with the QPSK partitioning algorithm, simplified QPSK partitioning algorithm or blind phase search (BPS) algorithm, three novel linewidth tolerant, low-complexity two-stage carrier phase estimation algorithms for the16-QAM modulation format are proposed. The linewidth tolerance, estimation performance, linear and nonlinear transmission performance is comparable to the BPS algorithm. Reductions in the hardware complexity by factors of about1.5-10are achieved in comparison to the BPS algorithm.5. In order to further reduce the complexity of CPE algorithms, a quasi-multiplier free CPE algorithm for16QAM modulation format is proposed. Combined with simplified QPSK partitioning algorithm, a linewidth-tolerant, low-complexity two-stage feed-forward CPE algorithm is introduced for DP16-QAM. Numeral and experimental results show that the proposed algorithm achieves comparable performance to the BPS algorithm but with reduction in the hardware complexity by a factor of17. To our best knowledge, this is the lowest hardware complexity feed-forward two-stage CPE algorithm for DP-16QAM reported so far.6. Linear interpolation is introduced to improving the performance of carrier phase compensation. Linear phase interpolation (LPI) based carrier phase compensation algorithm is proposed. Numeral and experimental results show that the performance of LPI algorithms are better than the block averaging based algorithms, and are comparable to the sliding window based algorithms. However, significant reduction in the hardware complexity is demonstrated for LPI based carrier phase compensation algorithms in comparison to sliding window based algorithms.更多还原