基于扩展卡尔曼滤波的高频谱效率光信号处理技术

发布时间：2018-09-12 12:56

【摘要】：近年来,基于卡尔曼滤波器的光信号协同估计技术针对光信号分模块数字信号处理中的环节代价、处理延时和调优参数耦合等问题提出了完美的解决方案,因此迅速成为了本领域的研究热点。光信号协同估计技术可以将多个数字损伤补偿结构简并成单个DSP结构同时进行跟踪和补偿。本文重点围绕偏振解复用、载波相位恢复、频偏估计领域协同估计技术进行深入研究,评估且提出多种协同估计方法,并结合数值仿真和实验平台等手段,对基于卡尔曼滤波的光信号协同估计技术的性能表现、应用范围进行验证。在偏振态和载波相位协同估计技术研究中,本文首先提出一种用于TDHQ信号(QPSK/16QAM混合信号)的盲和偏振态、相位协同估计方案。该方案基于扩展卡尔曼滤波(EKF),并且可以同时跟踪偏振态和载波相位。此外,不需要知道混合信号的帧结构组成。因此,该方案可以在收敛和跟踪阶段以全盲的方式进行。文章对系统性能与混合格式参数、环境变量的依赖关系进行了深入研究,具体参数和变量包括功率比(PR),帧长比(FR),相位噪声,残留频偏和偏振态旋转速率;其次本文将基于EKF的偏振态和相位协同估计技术扩展运用到高阶QAM系统中。通过考虑各种系统/信道条件(包括激光线宽,频率偏移和光纤信道偏振态旋转)下滤波器调谐参数的影响,详细研究了16QAM和64QAM情况下的EKF性能。通过仿真结果提出了不同系统/信道条件下的参数配置策略,研究显示EKF性能对调制阶数的依赖性。在载波相位和频偏协同估计技术中,本文修改和建模了印度理工Ankita Jain发表的基于EKF的非线性效应、载波相位噪声和频偏协同估计技术。在不考虑非线性效应的条件下,将该方案修改并应用于在112Gb/s PD-16QAM和PD-QPSK系统。通过仿真对该协同估计方案在两种调制格式信号下可以跟踪的最大频偏和相位范围进行研究,并探讨和评估了该方案的估计准确度。在前面工作积累的成果上,本文提出一种基于EKF的偏振态、载波相位和频偏三种协同估计方案。它可以在DP-QPSK和DP-16QAM两种调制格式信号下同时实现跟踪偏振状态,频率偏移和相位噪声。该算法具有自身快速偏振态跟踪功能,对激光器线宽,频率偏移的容忍性高等优势,已在典型系统环境下通过数值仿真和112Gb/s PDM-QPSK相干光通信实验系统确认其跟踪性能。因此,它还有望运用到双偏振高阶调制格式,对其快速时变激光频率漂移和偏振状态进行实时性能追踪。
[Abstract]:In recent years, the cooperative estimation technology of optical signal based on Kalman filter has put forward a perfect solution to the problems of link cost, processing delay and tuning parameter coupling in the digital signal processing of optical signal submodule. Therefore, it has become the research hotspot in this field. The optical signal cooperative estimation technique can degenerate multiple digital damage compensation structures into a single DSP structure to track and compensate at the same time. In this paper, we focus on the research of polarization demultiplexing, carrier phase recovery, frequency offset estimation, and evaluate and put forward a variety of cooperative estimation methods, combined with numerical simulation and experimental platform, etc. The performance of optical signal cooperative estimation based on Kalman filter is verified. In the research of polarization state and carrier phase cooperative estimation, this paper first proposes a blind and polarization state and phase cooperative estimation scheme for TDHQ signal (QPSK/16QAM mixed signal). The scheme is based on extended Kalman filter (EKF),) and can track both polarization and carrier phase simultaneously. In addition, there is no need to know the frame structure of the mixed signal. Therefore, the scheme can be carried out in a completely blind manner in the convergence and tracking stages. The dependence of system performance on mixed format parameters and environment variables is studied in this paper. The specific parameters and variables include power ratio (PR), frame length ratio (FR), phase noise residual frequency offset and polarization rotation rate. Secondly, the polarization state and phase cooperative estimation based on EKF is extended to high order QAM systems. By considering the influence of filter tuning parameters under various system / channel conditions (including laser linewidth, frequency offset and polarization rotation of fiber channel), the EKF performance of 16QAM and 64QAM is studied in detail. Based on the simulation results, a parameter configuration strategy under different system / channel conditions is proposed, and the dependence of EKF performance on modulation order is studied. In the technology of carrier phase and frequency offset cooperative estimation, this paper modifies and models the nonlinear effect based on EKF, carrier phase noise and frequency offset co-estimation technology published by Ankita Jain. The scheme is modified and applied to 112Gb/s PD-16QAM and PD-QPSK systems without considering the nonlinear effect. The maximum frequency offset and phase range of the cooperative estimation scheme can be tracked under two modulation formats by simulation, and the estimation accuracy of the scheme is discussed and evaluated. Based on the previous achievements, this paper proposes three cooperative estimation schemes based on EKF: polarization state, carrier phase and frequency offset. It can track polarization state, frequency offset and phase noise simultaneously in DP-QPSK and DP-16QAM modulation formats. The algorithm has the advantages of fast polarization state tracking, high tolerance to laser linewidth and frequency offset, and has been confirmed by numerical simulation and 112Gb/s PDM-QPSK coherent optical communication experimental system in typical system environment. Therefore, it is also expected to be used in dual-polarization high-order modulation scheme to track the fast time-varying laser frequency drift and polarization state in real time.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN929.1

【参考文献】