Fast and accurate nonlinear interference in-band spectrum prediction for sparse channel allocation
Résumé
We propose and numerically evaluate a machine-learning-based nonlinear interference spectrum estimator for a coherent optical network. The solution shows a root-mean-squared error of about 0.13 dB compared with split-step Fourier simulation when estimating the nonlinear interference variance.
| Origine | Fichiers produits par l'(les) auteur(s) |
|---|