Machine Learning-Based in-Band OSNR Estimation From Optical Spectra

Fabiano Locatelli,Salvatore Spadaro,Michela Svaluto Moreolo,Konstantinos Christodoulopoulos,Josep M Fabrega

doi:10.1109/lpt.2019.2950058

Abstract

Measuring the optical signal to noise ratio (OSNR) at certain network points is essential for failure handling, for single connection but also global network optimization. Estimating OSNR is inherently difficult in dense wavelength routed networks, where connections accumulate noise over different paths and tight filters do not allow the observation of the noise level at signal sides. We propose an in-band OSNR estimation process, which relies on a machine learning (ML) method, in particular on Gaussian process (GP) or support vector machine (SVM) regression. We acquired high-resolution optical spectra, through an experimental setup, using a Brillouin optical spectrum analyzer (BOSA), on which we applied our method and obtained excellent estimation accuracy. We also verified the accuracy of this approach for various resolution scenarios. To further validate it, we generated spectral data for different configurations and resolutions through simulations. This second validation confirmed the estimation quality of the proposed approach.

Highlights

THE optical signal to noise ratio (OSNR) is considered one of the most important signal quality parameters to measure
By interpolating the noise level at the sides of the considered channel, the optical spectrum analyzers (OSAs) allows the measurement of the amplified spontaneous emission (ASE) noise introduced by the optical amplifiers and other noisesensed impairments
In the first part of this section we report the results of the best performing machine learning (ML) model, which was Gaussian Model (GM)

Summary

Introduction

THE optical signal to noise ratio (OSNR) is considered one of the most important signal quality parameters to measure. By interpolating the noise level at the sides of the considered channel, the OSA allows the measurement of the amplified spontaneous emission (ASE) noise introduced by the optical amplifiers and other noisesensed impairments. Such measurements are typically taken offline to optimize a newly deployed connections or for troubleshooting failures. Issues arise in wavelength switched optical networks employing ultra-dense wavelength division multiplexing (ultraDWDM) or flex-grid filters [3]. In such networks, the channels exhibit different noise levels, according to their routes. Failure handling and dynamic network optimization in low margin and/or in disaggregated networks requires to measure the OSNR in-band and non-intrusively, as the network operates [6], [7]

Objectives

Methods

Results