Experimental Demonstration of Machine Learning Transformer Based Equalization of Channel Nonlinearities
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:
Gautam, Naveenta; Agarwal, Prankush; Kaylor, C. Alex; Lall, Brejesh; Choudhary, Amol; Ralph, Stephen E.
Inhalt:
We experimentally demonstrate a Machine Learning (ML) Transformer-based nonlinear equalization for WDM 16QAM coherent optical communication systems. Q factor gains as high as 1.7 dB are demonstrated for multi-span links. Performance and computational efficiency are superior to digital backpropagation and other ML techniques.