Investigating the Performance and Suitability of Neural Network Architectures for Nonlinearity Mitigation of Optical Signals

Dsilva, Vegenshanti ValerianVegenshanti ValerianDsilvaSackey, IsaacIsaacSackeyRonniger, GregorGregorRonnigerHünefeld, Guillermo vonGuillermo vonHünefeldChacko, BinoyBinoyChackoSchubert, ColjaColjaSchubertFreund, RonaldRonaldFreund2023-06-282023-06-282022https://publica.fraunhofer.de/handle/publica/4448872-s2.0-85146400072We compare three different neural network architectures for nonlinearity mitigation of 32 GBd OOK and QPSK signals after transmission over a dispersion-compensated link of 10-km SSMF and 10km DCF. OSNR gains up to 2.2 dB were achieved using reservoir networks, suitable for fast training.enInvestigating the Performance and Suitability of Neural Network Architectures for Nonlinearity Mitigation of Optical Signalsconference paper