Theoretical Analyses of Different Nonlinear Compensation Methods Based on Perturbation Theories in the Unrepeatered System With Raman Amplification

You Wang,Wei Li,Muyang Mei,Hao Zheng,Yaobing Chen,Zhongshuai Feng

doi:10.1109/jphot.2020.3002655

You Wang, Wei Li + Show 4 more

Open Access

https://doi.org/10.1109/jphot.2020.3002655

Copy DOI

Abstract

In this paper, the pre- and post-nonlinear compensation (NLC) methods based on regular perturbation (RP) theory are contrastively investigated with the original NLC as a bridge for analyses. Firstly, the numerical error functions of pre-NLC and original NLC are derived, revealing that the numerical error of pre-NLC is more severe due to the error accumulation. Secondly, we deduce the relevance of post-NLC and original NLC, which uncovers the essential difference is that the input condition of post-NLC possesses certain additional information, making the signal's intensity distribution after post-NLC more beneficial for the hard-decision. Meanwhile, the pre-, post- and original NLC methods based on enhanced regular perturbation (ERP) theory have also been discussed. Finally, the simulation is carried out with signal modulation of 16/32/64 QAM and baud-rate of 32 GBaud in an unrepeartered system with Raman amplification. The results agree with the analyses, the post-NLC is the best while the original NLC surpasses the pre-NLC. Additionally, we demonstrate an experiment with signal modulation of 16 QAM and baud-rate of 10 GBaud. The improvements of the Q2-factor of the RP-based and ERP-based post-NLC are about 0.6 dB and 1 dB compared with the electronic dispersion compensation (EDC).

Highlights

In coherent optical fiber communication systems, the nonlinear signal distortion caused by fiber nonlinearity is the main setback to expand the transmission distance and the communication capacity [1]–[5]
First we theoretically deduce the numerical error functions of the pre-nonlinear compensation (NLC) methods and original NLC methods, which demonstrate that the numerical error of the pre-NLC methods will accumulate along the practical transmission process and the integral infinitesimal is proportional to the distributed signal power along the fiber
In the unrepeartered system with Raman amplification, the signal power could reach as high as about 10 dBm at around 40 km away from the transmitter side, which causes the numerical error of the pre-NLC methods more severe than that of the original NLC methods

Summary

Introduction

In coherent optical fiber communication systems, the nonlinear signal distortion caused by fiber nonlinearity is the main setback to expand the transmission distance and the communication capacity [1]–[5]. Used mathematical method to solve the NLSE is the split step Fourier method (SSFM), based on which the BP technique is generally considered to implement, referred as SSFM-BP-NLC in the remaining parts This NLC method can perfectly compensate for the intra-channel nonlinear impairment [9], [10], its performance is constricted by the high computational complexity and the imperfection of the practical hardware. There are usually two ways to implement the RPBP-NLC, which are the pre-distortion nonlinear compensation (RPBP-pre-NLC) and the post nonlinear compensation (RPBP-post-NLC) These methods have been extensively studied in the multi-span long-haul fiber communication systems and both methods are believed to demonstrate effective compensation and lower computational complexity compared with the SSFM-BP-NLC method [22]–[24], [26].

NLC Based on Perturbation Theories

The Numerical Error Functions of RPBP-pre-NLC and ORP-NLC

The Relevance Between ORP-NLC and RPBP-post-NLC

Simulation

Experiment

Conclusion