Training strategies for learned Volterra MIMO equalizers in WDM systems

Nelson Castro, Sonia Boscolo, Andrew D. Ellis, Stylianos Sygletos

Research output: Unpublished contribution to conferencePosterpeer-review

Abstract

This research advances training strategies for a learned Volterra MIMO equaliser tailored for Wavelength Division Multiplexing (WDM) systems. Building upon prior success with minimal steps per span, our current focus is improving training efficiency. The equaliser integrates filters to address key signal impairments, and our work delves into refining training strategies for these filters. As part of the considered techniques, we explore the potential of applying transfer learning among the channels in the equaliser. This investigation yields valuable insights into the interplay of training strategies and filter optimisation, contributing to the enhanced performance of learned Volterra MIMO equalisers in WDM systems.
Original languageEnglish
Number of pages1
Publication statusPublished - 21 Feb 2024
EventTelecommunications, Optics & Photonics (TOP) Conference 2024 - London, United Kingdom
Duration: 21 Feb 202422 Feb 2024
https://topconference.com

Conference

ConferenceTelecommunications, Optics & Photonics (TOP) Conference 2024
Abbreviated titleTOP 2024
Country/TerritoryUnited Kingdom
CityLondon
Period21/02/2422/02/24
Internet address

Fingerprint

Dive into the research topics of 'Training strategies for learned Volterra MIMO equalizers in WDM systems'. Together they form a unique fingerprint.

Cite this