Research On The Phase Noise Compensation Methods In Coherent Optical OFDM System

Coherent optical orthogonal frequency division multiplexing(CO-OFDM)has been widely regarded as a promising candidate for future high-speed fiber transmission system.The technique CO-OFDM combines the advantages of coherent detection and OFDM modulation such as its high spectrum efficiency,high receiver sensitivity,and robustness against the fiber dispersion.However,different from the single carrier system,the CO-OFDM system has a longer symbol duration and is more sensitive to the phase noise mainly produced by the laser linewidth,which leads to rotation of all demodulated sub-carriers of OFDM symbol caused by a common angle namely common phase error(CPE)and an additive term called inter-carrier interference(ICI).On the one hand,in CO-OFDM systems,the phase noise generated by laser's linewidth will be converted into intensity noise after the transmission of the dispersion fiber,which limits the maximum transmission distance.On the other hand,the technique coherent detection requires precise tracking of the phase difference and frequency between the transmitted signal and the local oscillator output signal,which will be interfered by phase noise.In this way,the CPE and ICI phase noise would result in huge performance degradation,which makes the use of narrow linewidth laser be indispensable.Thus,the effective phase noise suppression methods are required for CO-OFDM to improve transmission performance.The research work and related results of this paper mainly focus on the principle,CPE compensation methods,ICI suppression methods and nonlinear compensation methods of the CO-OFDM communication system,which details are as follows:Firstly,the principles,construction and impairments of CO-OFDM systems are discussed detailly in this paper.And,the critical technologies used in the CO-OFDM systems are also analyzed in detail.To achieve an exhaustive research analysis,a CO-OFDM simulation system is conducted by the commercial Opti System and MATLAB.Secondly,for CPE compensation,the spectrum efficiency problem of the pilots aided method and the deviation problem of projection histogram(PH)assisted method are analyzed in this paper.In order to solve these problems,a two-dimensional projection histogram assisted and pilot-aided(PA-2PH)phase noise compensation scheme has been proposed in CO-OFDM systems for dealing with the CPE.Then,the feasibility and performance of PA-2PH are simulated in the simulation system.Thirdly,for nonlinear compensation,the limitations of traditional hard decision-based demodulation are analyzed in this paper.And then,the Gaussian mixture model(GMM)is applied to the nonlinear decision of CO-OFDM system according to the distribution characteristics of the receiver signal of the CO-OFDM system.This algorithm uses the soft decision based on probability to compensate the nonlinear effect by assimilating the machine learning thinking.The feasibility of using GMM for nonlinear compensation is verified by simulation analysis.Finally,for ICI suppression,the PCA and RBF-NN machine learning algorithms are combined in this paper.In which,the PCA algorithm is used to signal pre-processing in the receiver of CO-OFDM system,and the signals after processing with PCA algorithm are trained by RBF-NN to classify the constellation points.The simulation results show that the PCA and RBF-NN based method can effectively suppress ICI.