Research On Digital Signal Processing Algorithm In High Speed Coherent Optical Communication System

In recent years,coherent optical communication has become a mainstream technology in optical communication due to its advantages of large channel capacity,high sensitivity,long transmission distance and the possibility of adopting any modulation format.However,in coherent optical communication,the received signal at the receiving end suffers from various impairment s due to the unsatisfactory physical devices.In order to compensate for the received signal,a variety of digital signal processing(DSP)algorithms are used at the receiving end to recover the signal.This paper focuses on the compensation algorithm of polarization mode dispersion(PMD)in single-carrier coherent optical communication and the channel equalization algorithm in coherent optical orthogonal frequency division multiplexing(CO-OFDM)communication system.Firstly,we proposed two modified Radius-Directed Constant Modulus Algorithms(RD-CMA)based on Independent Component Analysis(ICA)for compensating the PMD in single-carrier coherent optical communication systems.Compared with the RD-CMA algorithm,the algorithms we proposed have a similar performance of PMD compensation,but have a faster convergence speed and overcome the singularity problem.Compared with the traditional ICA algorithm,the algorithms we proposed can continuously track the polarization variation in the PMD,and do not need to consider the permutation indeterminacy and complex uncertain scaling problems.Then,we proposed two blind channel equalization algorithms based on single-source ICA algorithm for the CO-OFDM communication system.Compared to the traditional channel equalization algorithm based on training symbols,the algorithms we proposed have a similar performance on channel equalization.Since the training symbols are not used,the spectral efficiency is improved effectively.Finally,the algorithms we proposed are simplified from the perspective of the number of symbols used by the ICA algorithm and the number of iterations.The computational complexity of algorithms is reduced.