Spatially Disaggregated Modelling of Coherent Cross-Channel Interference in Dispersion-Managed Optical Links

Accumulation of non-linear noise in transmission of coherent signals over dispersion-managed network segments can be severe due to the small chromatic dispersion accumulated in each span due to optical inline dispersion compensation resulting in noise contributions originated in different spans to sum up coherently. Moreover, small chromatic dispersion slows down the signal to take on gaussian distribution. These effects are properly accounted within traditional modeling schemes such as the gaussian noise model family, especially when mixed fiber types are considered. In this work, we propose a machine learning-assisted, spatially disaggregated approach extending the classic GN model, able to provide instantaneous and accurate non-linear noise estimation of coherent channels propagating in sections of mixed fiber types, dispersion managed links.