Abstract
We propose a deep-learning-based geometric constellation shaping algorithm for resisting the residual laser phase noise. In experiment, the proposed scheme outperforms 16QAM by 5.3 dB and non-deep-learning constellation shaping by 1.8 dB in OSNR.
© 2022 The Author(s)
PDF Article | Presentation VideoMore Like This
Andrej Rode, Benedikt Geiger, and Laurent Schmalen
Th2A.32 Optical Fiber Communication Conference (OFC) 2022
B. M. Oliveira, M. S. Neves, F. P. Guiomar, M. C. R. Medeiros, and P. P. Monteiro
SW4E.7 CLEO: Science and Innovations (CLEO:S&I) 2022
Vladislav Neskorniuk, Andrea Carnio, Domenico Marsella, Sergei K. Turitsyn, Jaroslaw E. Prilepsky, and Vahid Aref
SW4E.5 CLEO: Science and Innovations (CLEO:S&I) 2022