Experimental Study of Deep Neural Network Equalizers Performance in Optical Links

TL;DR

This paper presents an experimental evaluation of a novel convolutional-recurrent neural network equalizer that improves signal quality in optical fiber communications over long distances, outperforming previous neural network and digital backpropagation methods.

Contribution

The study introduces a new convolutional-recurrent neural network equalizer and demonstrates its superior performance in optical link transmission compared to existing neural and traditional methods.

Findings

Achieved 1dB Q-factor improvement in optical links.

Outperforms previous neural network equalizers and digital backpropagation.

Effective in both single-channel and WDM multi-channel transmissions.

Abstract

We propose a convolutional-recurrent channel equalizer and experimentally demonstrate 1dB Q-factor improvement both in single-channel and 96 x WDM, DP-16QAM transmission over 450km of TWC fiber. The new equalizer outperforms previous NN-based approaches and a 3-steps-per-span DBP.