High dynamic range 100 Gbit/s PAM4 PON with SOA preamplifier using Gated Recurrent Neural Network equaliser

Murphy, Stephen; Jamali, Fariba; Townsend, Paul D.; Antony, Cleitus

High dynamic range 100 Gbit/s PAM4 PON with SOA preamplifier using Gated Recurrent Neural Network equaliser

Files

ecoc_final_draft_stephen_murphy_submission.pdf(629.96 KB)

Accepted Version

Date

2022

Authors

Murphy, Stephen

Jamali, Fariba

Townsend, Paul D.

Antony, Cleitus

Publisher

Institute of Electrical and Electronics Engineers (IEEE)

Abstract

We investigate parallel multi-symbol equalisation scheme for 100Gb/s/λ PAM4 using Gated Recurrent Neural Networks and exploit SOA preamplifier gain suppression to achieve 27 dB system dynamic range below hard-decision FEC BER limit of 3.8 × 10−3 using a receiver with two gain settings.

Keywords

Gated Recurrent Neural Networks

Citation

Murphy, S., Jamali, F., Townsend, P. D. and Antony, C. (2022) 'High dynamic range 100 Gbit/s PAM4 PON with SOA preamplifier using Gated Recurrent Neural Network equaliser', in Leuthold, J., Harder, C., Offrein, B. and Limberger, H. (eds.) European Conference on Optical Communication (ECOC) 2022, Technical Digest Series (Optica Publishing Group, 2022), paper Th1C.6.