Research On Demultiplexing Technology Of Mode Division Multiplexing System Based On Independent Vector Analysis

With the acceleration of the process of social information,the mobile networking and the arrival of large data age,Internet of things,cloud computing,large data centers,high-definition video and multimedia real-time business and other bandwidth consumption of business volume increased rapidly,making the global network data traffic to nearly 60% of the rapid growth rate.The huge network throughput will bring the requirements of backbone optical network transmission system carrying information to a new height.however,the capacity of the single-mode optical fiber communication system is getting closer to the Shannon limit,due to the inherent non-linear effect and the spontaneous emission noise of the amplifier in the single-mode optical fiber.In order to realize the expansion of the backbone optical network,the channel transmission capacity of the optical communication system is multiplied by the small amount of orthogonal mode as the information transmission carrier.In order to realize the expansion of the backbone optical network,mode division multiplexing based on few mode fiber is proposed,which doubled the channel transmission capacity of the optical communication system through the small amount of orthogonal mode.However,the existence of mode coupling,differential mode group delay and other damage factors makes the crosstalk of the received signal of the transmission system seriously.In order to realize the long-distance large-capacity transmission of few mode fiber mode multiplexing system,it is necessary to use the simple and effective digital signal processing(DSP)algorithm to complete the demultiplexing of the transmission signal.The main content of this paper is the demultiplexing of few mode fiber mode division multiplexing system.The independent vector analysis(IVA)algorithm is proposed to carry out blind source separation of received signals.IVA algorithm is applied to the Intensity modulation direct detection and QPSK coherent detection based on the mode multiplexing system.The demultiplexing performance of the algorithm is measured by analyzing the bit error rate of the received signal before and after demultiplexing.The main contents of this paper are as follows:Firstly,we derive the mode transmission characteristics and cutoff conditions in few mode fiber based on the Maxwell equations.The damage factors(mode random coupling and the difference mode group delay)of the optical fiber transmission link are introduced in detail and are analyzed the influence on the transmission signal in the time-frequency domain,respectively.The basic composition of the optical transmission system is introduced,and the transmission link is modeled by the matrix transmission model.Secondly,the basic principle and process of the independent vector analysis algorithm are introduced in detail.The IVA algorithm transforms the time domain convolutional mixed signal into frequency domain to separate each frequency point.The data of different frequency points of the same source signal are processed as a vector,which can solve the order of uncertainty between the frequency bands at the same time.Through the introduction of the separation criteria,the derivation of the IVA cost function is realized,and the theoretical basis of the IVA algorithm without permutation can be deeply understood,and the updating rule of the separation matrix are obtained.In addition,mode multiplexing system can meet the constraints of IVA algorithm,though the introduction of constraints.we introduce the problem of IVA algorithm in the amplitude of uncertainty and its solution.Thirdly,based on the theory of transmission link modeling,few mode fiber mode division multiplexing system based on intensity modulation direct detection(IM / DD)is constructed,and the IVA algorithm is used to complete the demultiplexing of the received signal of IM / DD mode multiplexing system.The performance of the IVA algorithm is measured by the waveforms and bit error rate of the signal before and after demultiplexing.The study shows that the IVA algorithm has good demultiplexing performance without the additional permutation.Finally,the IVA algorithm is applied to the mode multiplexing system with QPSK modulation format.The influence of the nonlinear function in the updating formula of the separation matrix on the demultiplexing performance to the QPSK signal is explored.The demultiplexing performance of the IVA algorithm is analyzed by the constellation and the bit error rate of the signal before and after demultiplexing,and compared with the the frequency domain independent component analysis algorithm.The computational complexity of IVA algorithm is analysed.