Design Optimization of a Double Stage Extended L-band EDFA Using Neural Networks Trained on Experimental Data

Konferenz: ECOC 2024 - 50th European Conference on Optical Communication
22.09.2024-26.09.2024 in Frankfurt, Germany

Tagungsband: ITG-Fb. 317: ECOC 2024

Seiten: 4Sprache: EnglischTyp: PDF

Autoren:
Rabbani, Hamed; Jalilpiran, Saber; LaRochelle, Sophie; Rusch, Leslie A.

Inhalt:
Designing a double-stage L-band amplifier with low pump powers, low noise figure, and flat gain is challenging. We tackle this with fast and accurate neural networks amplifier models trained on experimental data. Our approaches optimize for flat gain, while minimizing noise figure and pump power.