Published April 20, 2021 | Version v1
Conference paper Open

Experimental Verification of Complex-Valued Artificial Neural Network for Nonlinear Equalization in Coherent Optical Communication Systems

Creators

  • 1. Aston University
  • 2. Infinera
  • 3. Telecom Italia

Description

We propose a novel design of neural networks for mitigating the fiber nonlinearity, employing a structure based on physical modeling. The neural network achieved nearly 5 times BER reduction in field trial when transmitting WDM 200G DP-16QAM over a 612 km legacy link.

Files

_ECOC__ACCEPTED__Experimental_Verification_of_Complex_Valued_Artificial_Neural_Network_Nonlinear_Equalizer_in_Coherent_Optical_Communication_Systems.pdf

Files (302.5 kB)

Additional details

Funding

FONTE – Fibre optic nonlinear technologies 766115
European Commission
REAL-NET – REAL-time monitoring and mitigation of nonlinear effects in optical NETworks 813144
European Commission