Utilizing the sparsity of quasi-distributed sensing systems for sub-Nyquist signal reconstruction

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

Abstract

Quasi-distributed sensing, e.g. Quasi-Distributed Acoustic Sensing (Q-DAS), with optical fibers is commonly used for various applications. Its excellent performance is well known, however, it necessitates high sampling rates and expensive hardware for acquisition and processing. In this paper, we introduce a technique, based on Compressed Sensing (CS) theory, to locate discrete reflectors' along a sensing fiber with a smaller number of samples than required according to Nyquist criterion. The technique is based on the fact that the fiber profile consists of a limited number of discrete reflectors and significantly weaker reflections of Rayleigh back-scatterers, and thus can be approximated as a sparse signal. The task of reconstructing a sparse signal from a sub-Nyquist sampled signal using Orthogonal Matching Pursuit (OMP) is presented and tested experimentally.

Original languageEnglish
Title of host publicationSeventh European Workshop on Optical Fibre Sensors
EditorsKyriacos Kalli, Gilberto Brambilla, Sinead O'Keeffe
PublisherSPIE
ISBN (Electronic)9781510631236
DOIs
StatePublished - 2019
Event7th European Workshop on Optical Fibre Sensors, EWOFS 2019 - Limassol, Cyprus
Duration: 1 Oct 20194 Oct 2019

Publication series

NameProceedings of SPIE - The International Society for Optical Engineering
Volume11199
ISSN (Print)0277-786X
ISSN (Electronic)1996-756X

Conference

Conference7th European Workshop on Optical Fibre Sensors, EWOFS 2019
Country/TerritoryCyprus
CityLimassol
Period1/10/194/10/19

Keywords

  • Compressive Sensing
  • Fiber Bragg Gratings
  • Fiber Optic Sensors
  • Optical Frequency Domain Reflectometry

Fingerprint

Dive into the research topics of 'Utilizing the sparsity of quasi-distributed sensing systems for sub-Nyquist signal reconstruction'. Together they form a unique fingerprint.

Cite this