End-to-End Deep Learning of Long-Haul Coherent Optical Fiber Communications via Regular Perturbation Model

TL;DR

This paper introduces an end-to-end deep learning approach for coherent optical fiber communications, optimizing constellation shaping and nonlinear pre-emphasis to improve mutual information over long-distance links.

Contribution

It proposes a novel autoencoder-based learning method using a parallelizable perturbative channel model for optical communications.

Findings

Achieved a mutual information gain of 0.18 bits/sym./pol.

Simulated 64 GBd dual-polarization transmission over 30x80 km links.

Demonstrated the effectiveness of joint optimization in optical channels.

Abstract

We present a novel end-to-end autoencoder-based learning for coherent optical communications using a "parallelizable" perturbative channel model. We jointly optimized constellation shaping and nonlinear pre-emphasis achieving mutual information gain of 0.18 bits/sym./pol. simulating 64 GBd dual-polarization single-channel transmission over 30x80 km G.652 SMF link with EDFAs.