Data-driven Modelling of EDFAs by Neural Networks

/in /by

Publication Date: 3/13/2023

Event: SubOptic 2023

Reference: pp. 1-6, 2023

Authors: Fatih Yaman, NEC Laboratories America, Inc.; Hussam G. Batshon, NEC Laboratories America, Inc.; Daisuke Katsukura, NEC Corporation; Shinsuke Fujisawa, NEC Corporation; Eduardo Mateo, NEC Corporation; Takanori Inoue, NEC Corporation; Yoshihisa Inada, NEC Corporation

Abstract: Dependence of EDFA gain shape on input power and input spectrum shape is modelled using a simple neural network-based architecture for amplifiers with different gains and output powers. The model can predict the gain within ±0.1 dB. Even though the model has good success predicting the performance of the particular EDFA it is trained with, it is not as successful when used to predict a different EDFA, or even the same EDFA with a different pump power. However, retraining the model with a small amount of supplementary data from a second EDFA makes the model able to predict the performance of the second EDFA with little loss in performance.