Wasserstein Autoencoder Based End-to-End Learning Strategy of Geometric Shaping for an OAM Mode Division Multiplexing IM/DD Transmission

Zhaohui Cheng, Ran Gao*, Qi Xu, Fei Wang, Yi Cui, Xiangjun Xin

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

We propose a Wasserstein Autoencoder based end-to-end geometric shaping scheme for IM/DD OAM-MDM optical fiber communication system. Compared with traditional autoencoder, the BER decreased by up to 28% and 33% with two OAM modes.

Original languageEnglish
Title of host publication2023 Asia Communications and Photonics Conference/2023 International Photonics and Optoelectronics Meetings, ACP/POEM 2023
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9798350312614
DOIs
Publication statusPublished - 2023
Event2023 Asia Communications and Photonics Conference/2023 International Photonics and Optoelectronics Meetings, ACP/POEM 2023 - Wuhan, China
Duration: 4 Nov 20237 Nov 2023

Publication series

Name2023 Asia Communications and Photonics Conference/2023 International Photonics and Optoelectronics Meetings, ACP/POEM 2023

Conference

Conference2023 Asia Communications and Photonics Conference/2023 International Photonics and Optoelectronics Meetings, ACP/POEM 2023
Country/TerritoryChina
CityWuhan
Period4/11/237/11/23

Keywords

  • Wasserstein autoencoder
  • geometric shaping
  • machine learning
  • orbital angular momentum

Fingerprint

Dive into the research topics of 'Wasserstein Autoencoder Based End-to-End Learning Strategy of Geometric Shaping for an OAM Mode Division Multiplexing IM/DD Transmission'. Together they form a unique fingerprint.

Cite this