基于独立矢量分析的模分复用系统解复用技术研究

发布时间：2018-08-27 16:30

【摘要】：随着社会信息化进程的加快,移动互联网及大数据时代的到来,物联网、云计算、大型数据中心、高清视频及多媒体实时业务等带宽消耗型业务量迅速增长,使得全球网络数据流量以将近60%的速度迅速增长。所需的巨大网络吞吐量将对承载信息的骨干光网络传输系统的要求提高到全新的高度。然而,由于单模光纤传输介质中固有非线性效应及放大器的自发辐射噪声,使得单模光纤通信系统的容量越来越趋近香农极限。为了实现对骨干光网络的扩容,提出基于少模光纤的模分复用技术,其采用相互正交的模式作为信息传输载体,成倍的提高了光通信系统的传输容量。然而,由于在实际少模光纤传输系统中,存在模式随机耦合、差分模式群时延等损伤因素,使得传输系统的接收信号串扰严重,为了实现少模光纤模分复用系统的长距离大容量传输,需要在接收端采用有效的数字信号处理(DSP)算法实现对传输信号的解复用。本论文研究的主要内容为少模光纤模分复用系统传输信号的解复用,提出采用独立矢量分析(Independent vector analysis,IVA)算法对接收信号进行盲源分离,并分别将IVA算法应用于强度调制直接检测的模分复用系统及调制格式为QPSK相干检测的模分复用系统中,通过对接收信号解复用前后误码率的分析,衡量算法的解复用性能。本文的主要工作内容如下:首先,以Maxwell方程组为基本理论依据,对少模光纤传输介质中模式传输特性及截止条件进行推导。详细介绍了少模光纤传输链路的损伤因素:模式随机耦合、差分模式群时延,并分别在时域及频域内分析损伤因素对传输信号的影响。介绍了少模光纤传输系统的基本组成,并采用矩阵传输模型对传输链路进行建模。其次,对独立矢量分析算法的基本原理及流程进行详细介绍。IVA算法是将时域卷积混合信号变换到频域,在频域内对各个频点进行瞬时分离,实现混合信号的去串扰,其中在频域处理的过程,采用将同一源信号不同频点的数据作为向量进行处理,在各个频段瞬时分离的同时,解决频段间的次序不确定性问题。通过对分离准则的介绍,实现对IVA代价函数的简单推导,据此能够深刻理解IVA算法无需频段排序的理论依据,并得出分离矩阵的更新规则。此外,对IVA算法的适用条件进行约束,可以直观看出,模分复用系统中传输信号能够满足IVA算法使用的基本条件,同时介绍了IVA算法中的幅度不确定性及其解决方案。再次,依据传输链路建模理论,搭建基于强度调制直接检测(IM/DD)的少模光纤模分复用仿真系统,并采用IVA算法对IM/DD模分复用系统接收端混叠信号进行解复用。通过对接收信号解复用前后波形图眼图及误码率曲线的对比,衡量IVA算法的解复用效果。仿真结果表明IVA算法无需额外频点排序的情况下,具有良好的解复用性能。最后,将IVA算法应用于调制格式为QPSK的模分复用系统中,探究了IVA算法分离矩阵更新公式中非线性函数的选择对于混合QPSK信号的解复用性能的影响,通过解复用前后信号星座图的变化及系统误码率情况,分析IVA算法的解复用性能,并与频域独立成分分析算法的解复用性能进行对比。仿真结果显示,IVA算法与频域独立成分分析(FD-ICA)算法的解复用性能相当。并对IVA算法计算复杂度进行分析。
[Abstract]:With the acceleration of social informationization, the arrival of mobile Internet and big data era, bandwidth-consuming services such as Internet of Things, cloud computing, large data centers, high-definition video and multimedia real-time services are growing rapidly, which makes the global network data flow grow rapidly at a rate of nearly 60%. However, due to the inherent nonlinear effect in single-mode optical fiber transmission medium and the spontaneous emission noise of amplifier, the capacity of single-mode optical fiber communication system is getting closer and closer to Shannon's limit. In order to realize the expansion of backbone optical network, a few-mode optical fiber is proposed. Modular division multiplexing (DMA) technology, which uses orthogonal modes as information transmission carriers, doubles the transmission capacity of optical communication systems. However, due to the existence of mode random coupling, differential mode group delay and other damage factors, the received signal of transmission system crosstalk is serious, in order to achieve less. For the long-distance and large-capacity transmission of mode-division multiplexing (MODM) systems, effective digital signal processing (DSP) algorithms should be adopted at the receiver to realize the demultiplexing of the transmitted signals. VA) Blind source separation (BSS) algorithm is used to separate the received signals, and IVA algorithm is applied to the intensity modulation direct detection mode division multiplexing system and the modulation format is QPSK coherent detection mode division multiplexing system respectively. Firstly, based on the Maxwell equations, the mode propagation characteristics and cut-off conditions in the medium of small-mode fiber transmission are deduced. The damage factors of the transmission links of small-mode fiber, such as mode random coupling, differential mode group delay, are introduced in detail, and the influence of the damage factors on the transmission signal is analyzed in time domain and frequency domain respectively. Secondly, the basic principle and flow of the independent vector analysis algorithm are introduced in detail. IVA algorithm transforms the time domain convolution mixed signal into frequency domain, and separates each frequency point instantaneously in frequency domain to realize the mixed signal. In the process of crosstalk removal, the data of different frequency points of the same source signal are processed as vectors in frequency domain, and the order uncertainties between the frequency bands are solved while the data of different frequency points of the same source signal are separated instantaneously in each frequency band. In addition, constraints on the applicable conditions of the IVA algorithm show that the transmission signal in the modular division multiplexing system can satisfy the basic conditions of the IVA algorithm. At the same time, the amplitude uncertainty in the IVA algorithm and its solution are introduced. The simulation system of mode division multiplexing (MDM) based on intensity modulation direct detection (IM/DD) is built, and the overlapped signals of the receiving end of IM/DD MDM system are demultiplexed by IVA algorithm. The results show that the IVA algorithm has good demultiplexing performance without additional frequency sorting. Finally, the IVA algorithm is applied to the modular multiplexing system with QPSK modulation format, and the influence of the selection of nonlinear functions in the updating formula of the separation matrix of the IVA algorithm on the demultiplexing performance of the mixed QPSK signal is investigated. The performance of the IVA algorithm is analyzed and compared with that of the Frequency-Domain Independent Component Analysis (FD-ICA). The simulation results show that the performance of the IVA algorithm is equivalent to that of the Frequency-Domain Independent Component Analysis (FD-ICA).
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN929.11

【相似文献】