Turbulence correction with artificial neural networks

TL;DR

This paper presents a neural network-based method to correct turbulence-induced distortions in optical modes, improving free-space optical communication robustness by using intensity profile measurements and simulation validation.

Contribution

It introduces a novel neural network approach that corrects optical mode distortions caused by turbulence using only intensity measurements, enhancing optical communication reliability.

Findings

Corrected optical modes closely match target profiles with near-zero mean square error.

The neural network approach is simple, robust, and effective in simulations.

Potential to significantly improve free-space optical link robustness.

Abstract

We design an optical feedback network making use of machine learning techniques and demonstrate via simulations its ability to correct for the effects of turbulent propagation on optical modes. This artificial neural network scheme only relies on measuring the intensity profile of the distorted modes, making the approach simple and robust. The network results in the generation of various mode profiles at the transmitter that, after propagation through turbulence, closely resemble the desired target mode. The corrected optical mode profiles at the receiver are found to be nearly identical to the desired profiles, with near-zero mean square error indices. We are hopeful that the present results combining the fields of machine learning and optical communications will greatly enhance the robustness of free-space optical links.