Research On Receiver Signal Processing Technology In The Probabilistically Shaped Optical Communication System

With the development of new technologies and applications such as 5G,big data,Internet of Things,and virtual reality,optical communication system capacity and dynamic service bandwidth have higher requirements.In the last century,Shannon proposed that a continuous Gaussian distribution source can achieve the maximum communication capacity of an additive white Gaussian noise channel,and probabilistic shaping technology was born from this.Although the current probabilistic shaping coding scheme has matured day by day,the research on signal processing around probabilistically shaped optical systems is still in its infancy.The existing signal processing schemes for standard quadrature amplitude modulation have performance degradation and failure problems under the probabilistically shaped system.To this end,this thesis focuses on the research of signal processing technology in three directions,including parameter estimation,carrier recovery and polarization demultiplexing of probabilistically shaped optical signals.Focusing on the problem of probabilistically shaped optical signal parameter estimation,a parameter estimation scheme based on kernel density estimation is proposed,which realizes three signal parameters estimation of probabilistically shaped optical systems including signal modulation format identification,shaping parameter and signalto-noise ratio estimation.Among them,probabilistical shaping parameter estimation will be used to improve the subsequent carrier recovery algorithm.Numerical simulation shows that the proposed scheme can realize parameter estimation within a wide range of signal-to-noise ratio and probabilistical shaping parameters.Among the estimation results of the three modulation formats,the recognition accuracy of the modulation format can reach more than 90%,and the SNR estimation error is less than 2dB,the estimation error of information entropy is less than 0.2 bits/symbol.Focusing on the problem of carrier recovery,frequency offset estimation and carrier phase noise recovery are studied respectively.By improving the existing radius-oriented fourth-power fast Fourier frequency offset estimation algorithm,the frequency offset estimation independent of the shaping parameter is realized.Numerical simulation shows that the improved frequency offset estimation algorithm can work in a larger signal-tonoise ratio range,and the frequency offset estimation error of Megahertz can be achieved.At the same time,an modified normalized blind phase search algorithm is proposed.Compared with the traditional blind phase search algorithm,the block length required by this algorithm is reduced by about 25%.and Under the conditions of 500kHz laser linewidth,mutual information also has at least 0.2 bits/symbol improvement.Focusing on the polarization demultiplexing of probabilistic shaping optical communication systems,an independent component analysis algorithm based on maximum likelihood estimation is proposed.Numerical simulations show that the proposed polarization demultiplexing algorithm will not be affected by the probability shaping parameters.Compared with the traditional constant modulus algorithm and radius decision dequalization,the pre-decoding error rate of the probabilistic shaped signal is reduced by at least two orders of magnitude,and there is no performance degradation or failure due to the change of the shaping parameter.This paper focuses on the signal processing technology of receiver in the probabilistically shaped optical communication system,and carries out the research of signal parameter estimation,carrier recovery and polarization demultiplexing.Parameter estimation based on kernel density estimation,the shaping parameter-independent carrier recovery algorithm and polarization demultiplexing algorithm based on independent component analysis were proposed in this thesis.Compared with the existing schemes,the scheme proposed in this paper is not only suitable for probabilistically shaped signals,but also for different probabilistic shaping optical communication systems without predicting the parameter information of the signals.