Carrier Recovery Based On Kalman Filter In Coherent Optical Communication Systems

In recent years,the communication bandwidth has been improved significantly for meeting the demands of ever-growing high rate and high quality data services.To meet the needs,the optical fiber communication,one of the mainstream communications,is also driven towards long distance,high capacity and high spectral efficiency.Here,the coherent optical communication system combined with higher-order modulation format,polarization multiplexing,wavelength division multiplexing and digital signal processing is one of the most effective solutions of optical fiber communication.Firstly,this paper introduces the development of coherent optical communication system and digital signal processing algorithms.Then,we introduce in detail the structure of coherent optical communication system and the basic model of Kalman filter.Finally,the frequency offset compensation algorithm and carrier phase recovery algorithm are analyzed and studied.Subsequently,we discussed a joint-polarization parallel carrier phase recovery algorithm.Up to now,most of the frequency offset estimation algorithms need to remove the signal modulation information first,and then calculate frequency offset value.There are two methods to remove signal modulation information including the training sequence(TS)and power of M(M'th).In this paper,based on frequency domain FFT transform frequency offset estimation algorithm,we compared the simulation results of TS and M'th.In the case of large number of signals,both of the algorithms can achieve correctly frequency offset.The TS based algorithm has larger estimation range and bigger noise tolerance,but the application of TS reduces the spectrum efficiency and increases the complexity of the system.The M'th can achieve higher accuracy but only suit for specific modulation formats with small estimation range.In order to get high accuracy,the two methods need large number of signals which will increase computation complexity.For the past few years,a number of studies using Kalman filter for frequency offset estimation.Because it has high convergence speed,and can achieve high estimation precision with small number of signals.In the paper,the Kalman filter frequency offset estimation algorithm based on TS is studied.The estimation range and estimation precision are analyzed in the polarization-division multiplexing(PDM-16QAM)system.For carrier phase recovery algorithms,the Viterbi-Viterbi algorithm,blind phase search(BPS)carrier phase recovery algorithm and the Kalman filter(LKF)carrier phase recovery algorithm are mainly introduced.Furthermore,referring to the carrier phase recovery algorithm based on Kalman filter that published by Takashi Inoue,we discussed a joint-polarization paralleled carrier phase recovery algorithm.For the polarizationmultiplexing system,considering the signals in the two polarization have same phase noise.Therefore,we use the first and third ring symbols of both polarization to perform phase noise estimation for the PDM-16 QAM signals.The simulation results for 224 Gb/s PDM-16 QAM system demonstrate that the linewidth tolerance of joint-polarization LKF(2800 kHz)is about seven times of single-polarization LKF(400 kHz)and the joint-polarization LKF can process more symbols in one block than single-polarization LKF,which reduces algorithm computation complexity and enhance the real-time implemented.Finally,we experimentally demonstrate the results by using 224 Gb/s PDM-16 QAM transmission system.