Research On Robustness-Enhanced Dsp Method In High-Speed Optical Communication System

In recent years,with the emergence of high-bandwidth services such as IPTV,mixed reality,and the Internet of Things,global data traffic has grown rapidly,and users’ demand for bandwidth in communication networks has continued to increase.As the cornerstone of network information transmission,optical fiber communication systems carry more than 90%of the world’s data traffic.Limited by the Shannon limit,fiber nonlinear effect,available spectrum resources and other factors,the capacity of optical communication systems is close to the theoretical upper limit.Therefore,there is an urgent need to carry out research on key technologies of high-speed optical communication systems and realize the expansion of optical fiber communication systems to cope with the problem of global traffic surge in the future.Probabilistic Shaping(PS)and channel impairment compensation based on machine learning are effective methods to improve the spectral efficiency and transmission rate of optical communication systems.However,when the original Digital Signal Processing(DSP)method for uniform distribution is used to compensate for phase noise,dispersion,nonlinear impairment,etc.in high-speed optical communication systems,there is a problem that it is incompatible with PS technology and machine learning algorithms.Therefore,how to enhance the robustness of digital signal processing methods in high-speed optical communication systems,and use the same method to accurately match and compensate signals with different channel conditions and different probability distributions,so as to improve the effectiveness and reliability of the system has become a hot spot in domestic and foreign research.Based on the theoretical basis of studying probabilistic shaping technology and DSP method in high-speed optical communication system,this dissertation focuses on the study of the robustness of DSP method for the purpose of enhancing the robustness of DSP method.The Optimized Principal Component-based Phase Estimation(OPCPE)algorithm robust to PS technology is studied,which solves the problem of accurate phase recovery of PS signals in the high signal-to-noise region.The QuasiMaxwell-Boltzmann(Quasi-MB)probabilistic shaping scheme friendly to the Carrier Phase Recovery(CPR)algorithm is studied,and the problem of optimizing the PS signal distribution characteristics to improve the recovery performance of the CPR algorithm is solved.The Phase Ambiguity(PA)chaotic encryption algorithm robust to blind equalizer and PS technology is studied,which solves the problem of high-sensitivity and high-reliability symbol-level encryption.The probabilistic weighted KMeans clustering algorithm robust to PS technology is studied,and the problem of accurate clustering of PS signals under the influence of strong nonlinearity is solved.The Polar K-Means optimization decision algorithm for turbulence-induced distortion is studied,which solves the problem of limited transmission distance of free space optical communication.The main work and innovation points of this dissertation are as follows.1.Highly robust phase recovery algorithm in coherent optical communication system.Aiming at the problem that the existing phase recovery algorithms in coherent optical communication system cannot accurately estimate the phase of the signal after strong probabilistically shaping,an OPCPE algorithm that is robust to PS technology is proposed,and a Quasi-MB probabilistic shaping scheme friendly to CPR algorithm is proposed.(1)An OPCPE algorithm that is robust to PS technology is proposed,by applying constellation mirroring and partial scaling optimization steps,the problem of incompatibility between the PCPE algorithm and the PS technique in the high signal-to-noise ratio range is solved,and the accuracy of the PCPE algorithm for phase recovery of probabilistically shaped signals with different shaping coefficients is significantly improved,and the performance advantages of the PCPE algorithm in terms of complexity and cycle-slip rate are effectively retained.The simulation results show that compared with the Blind Phase Search(BPS)algorithm,the OPCPE algorithm can achieve a signal-to-noise ratio gain of 0.74dB in a PS 64QAM system with a shaping factor of 0.03 when the line width of the transceiver laser is 500kHz.(2)A Quasi-MB probabilistic shaping scheme friendly to CPR algorithm is proposed to achieve better compatibility with CPR algorithms by assigning the same occurrence probability to constellation points located on the same square ring in the constellation diagram,so that the signal after probabilistic shaping exhibits characteristics similar to the time-division mixed modulation signal.Simulation results show that using the Quasi-MB distribution model instead of the traditional Maxwell Boltzmann distribution model for probabilistic shaping can effectively improve the Normalized Generalized Mutual Information(NGMI)performance after the recovery of the CPR algorithm,and achieve up to 51%and 21%NGMI improvement for the PCPE algorithm and BPS algorithm,respectively.2.Highly robust physical layer security algorithm in optical access network system.Aiming at the problem that the chaotic encryption algorithm based on Constellation Shifting(CS)scheme in optical access network system is incompatible with the Decision Directed-Least Mean Square(DD-LMS)algorithm and PS technology,a highly robust chaotic encryption algorithm based on phase ambiguity is proposed.By converting the chaotic sequence originally used to generate artificial noise masks into phase rotation keys and signal conjugate keys,the proposed algorithm effectively solves the problem of erroneous convergence when the chaotic encryption algorithm and blind equalization algorithm are used in combination,and realizes the effective retention of the shaping gain of probabilistic shaping technology,and enhances the robustness of the chaotic encryption algorithm to PS and DD-LMS.Experimental results show that compared with CS algorithm,the proposed algorithm can achieve a sensitivity gain of 0.7dB,and can effectively avoid the erroneous convergence of the blind equalization algorithm in the post-equalization process.3.Highly robust clustering algorithm in long-distance coherent optical communication system.Aiming at the problem that K-Means clustering algorithm and probabilistic shaping technology cannot be properly compatible in longdistance coherent optical communication systems,a weighted K-Means clustering algorithm that is robust to probabilistic shaping technology is proposed,which significantly improves the accuracy of the clustering algorithm in the process of centroid localization by introducing a weighting factor that follows the reciprocal of the MB distribution in the clustering process,and expands the dynamic range of the optimization decision method of probabilistic shaping signals under different channel conditions.Even under the influence of strong ASE noise or strong nonlinear effects,the proposed weighted K-Means clustering algorithm can still effectively realize the optimization decision of PS signals,and avoid the erroneous convergence problem of traditional K-Means algorithms.Experimental results show that in the case of Back-To-Back(BTB)transmission,the weighted K-Means algorithm can achieve an optical signal-to-noise ratio gain of about 0.6dB compared with the traditional K-Means algorithm.Under the condition of 375km transmission,the traditional K-Means algorithm can only operate normally when the dynamic range of launched optical power is-4dBm to 2dBm,while for the proposed weighted KMeans algorithm,it achieves correct convergence when the dynamic range is-5dBm to 5dBm.4.Highly robust optimization decision method in free space optical communication system.Aiming at the erroneous convergence problem of K-Means algorithm in clustering signals affected by turbulence-induced distortion in free space optical communication system,a Polar K-Means clustering algorithm is proposed,which can effectively improve the grouping accuracy of the clustering algorithm and increase the tolerance of the optimization decision method to turbulence distortion by converting the signal to the polar coordinate system for clustering and giving more weight to the phase component of the signal.Simulation results show that compared with the traditional Cartesian coordinate system-based decision scheme,the maximum achievable transmission distance corresponding to 16QAM signal can be increased by about 14.1%by using the proposed Polar KMeans clustering algorithm.