Learning and Avoiding Disorder in Multimode Fibers

Multimode optical fibers (MMFs) have gained renewed interest in the past decade, emerging as a way to boost communication data rates context of an expected saturation current single-mode fiber-based networks. They are also attractive for endoscopic applications, offering possibility achieve similar information content multicore fibers, but with much smaller footprint, thus reducing invasiveness procedures. However, these advances hindered by unavoidable presence disorder that affects propagation light MMFs and limits their practical applications. We introduce here general framework study avoid effect wave-based systems demonstrate its application multimode fibers. experimentally find almost complete set channels resilient induced strong deformations. These deformation principal modes obtained only exploiting measurements weak perturbations harnessing generalized Wigner-Smith operator. explain this demonstrating that, even high level disorder, can be characterized just few key properties. results made possible thanks precise fast estimation modal transmission matrix fiber which relies on model-based optimization using deep learning frameworks.Received 25 November 2020Revised 29 April 2021Accepted 20 May 2021DOI:https://doi.org/10.1103/PhysRevX.11.021060Published American Physical Society under terms Creative Commons Attribution 4.0 International license. Further distribution work must maintain attribution author(s) published article’s title, journal citation, DOI.Published SocietyPhysics Subject Headings (PhySH)Research AreasClassical opticsGeometrical & wave opticsInformation theoryPhysical SystemsFibersOptical fibersTechniquesHolographyOptical techniquesGeneral PhysicsAtomic, Molecular Optical