English

Analysis and Optimisation of Distribution Matching for the Nonlinear Fibre Channel

Signal Processing 2019-07-08 v1 Information Theory math.IT

Abstract

Enhanced Gaussian noise models are used to demonstrate that the per-block SNR after fibre transmission varies significantly due to the variable-composition nature of modern probabilistic shaping schemes. We propose a nonlinearity-optimised distribution matcher that improves the average and worst-case SNR by 0.14 and 0.22 dB, respectively.

Keywords

Cite

@article{arxiv.1907.02846,
  title  = {Analysis and Optimisation of Distribution Matching for the Nonlinear Fibre Channel},
  author = {Tobias Fehenberger and Alex Alvarado},
  journal= {arXiv preprint arXiv:1907.02846},
  year   = {2019}
}