Physics-informed mode decomposition neural network for structured light in multimode fibers

Research output: Contribution to book/Conference proceedings/Anthology/ReportConference contributionContributedpeer-review

Abstract

We propose a novel approach for the referenceless mode decomposition of multimode fibers. A deep neural network with the interaction of a physical model achieves decomposition of 55 modes without pre-training. This paradigm shift is of great importance for space division multiplexing.

Details

Original languageEnglish
Title of host publication2023 IEEE Photonics Conference, IPC 2023 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers (IEEE)
Pages1-2
ISBN (electronic)979-8-3503-4722-7
Publication statusPublished - 2023
Peer-reviewedYes

Publication series

SeriesIEEE Photonics Conference (IPC)
ISSN2374-0140

Conference

Title2023 IEEE Photonics Conference
Abbreviated titleIPC 2023
Duration12 - 16 November 2023
Website
LocationHilton Orlando Buena Vista Palace
CityOrlando
CountryUnited States of America

Keywords

Keywords

  • deep learning, multimode fiber, physics-driven, scattering, space division multiplexing