Research On Carrier Phase Recovery Technology Of High Speed Coherent Optical Communication Based On Kalman Filter

The development of communication technology has brought tremendous changes to the society,the Internet of Things and 5G have gradually entered people's vision,with the improvement of information infrastructure and the development of information technology,the increase in transmission traffic has put forward higher requirements for optical transmission networks.Coherent optical communication has become the focus of attention and research in the field of optical communication due to its advantages of high speed and large bandwidth.The digital signal processing technology at the receiver can compensate for various damages in the signal transmission,thereby greatly increasing the transmission distance.Kalman filter algorithm has become one of the hotspots of digital signal processing in coherent optical communication due to its good real-time performance and low complexity.This article focuses on the topic of using kalman filter to perform carrier recovery in coherent optical communication systems,analyzes the causes of carrier noise in coherent optical communication,introduces classic signal processing algorithms,and the influence of carrier phase noise on transmission performance is studied.Carrier phase recovery is performed using extended kalman filter and linear kalman filter algorithms,and then the system-level functional simulation of the algorithm is performed in conjunction with DSP Builder.The main work is as follows:(1)The QPSK and QAM coherent optical communication transmission simulation system is built.Using this simulation platform,the differential coding method of 16 QAM signal is studied,the effects of carrier frequency offset and phase noise on the signal are analyzed,a variety of carrier phase recovery algorithms are compared,and the feasibility of the algorithm is verified by MATLAB.(2)A dual-polarization carrier phase recovery algorithm(XY-EKF)based on extended kalman filtering is proposed and the back-to-back and 320 km optical fiber transmission simulated in a 224Gbit/s DP-16 QAM transmission system.The results show that compared with the classical extended kalman filter algorithm,XY-EKF algorithm reduces the computational complexity by 1/3 and a certain linewidth tolerance is improved.(3)Carrier phase recovery is carried out using linear kalman filter combined with maximum likelihood estimation(KF/ML).At the same time,the method of inserting training sequence and initial value update solved the problem of slow convergence during block filtering.The performance of the two algorithms in the case of double polarization is compared,the results show that the KF/ML algorithm has higher linewidth tolerance,while the block kalman filter algorithm has higher recovery accuracy and lower complexity.(4)Simulink and DSP Builder design software in MATLAB are used to perform system-level simulation of KF/ML algorithm.Through the DSP Builder device library on the Simulink platform,each functional module of the KF/ML algorithm is designed using the components in it,and the results of the software are compared and analyzed.