Skip to main navigation Skip to search Skip to main content

A Nonlinear Regression-Based Approach for Estimating the Relaxation Parameters of GPR Traces

  • Saeed Aljabri
  • , Mohamed Deriche

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

Ground Penetrating Radar (GPR) techniques have been used extensively over the last decades in mapping underground objects by recording signal reflections due to the variations of objects' dielectric properties. Such an approach can be generalized to cover properties' analysis of the sub surface's materials. This is achieved by estimating the relaxation time (frequency) of the dielectric permittivity spectrum. By doing this, valuable information is obtained including, but not limited to, identifying moisture content of underground soil, water and clay content, rock classification, material ageing, etc. Several techniques have been proposed for estimating the relaxation time (frequency) under the ill-posedness condition. Existing techniques suffer from drawbacks including computation complexity and dependency on regularization functions that affect estimation accuracy. Non-linear regression technique have shown improvement in performance even the presence of noisy data. This paper proposes a new formulation using the Steepest Descent method. The simulation results showed that the proposed step size speeds the algorithm convergence in comparison to constant step size algorithm.

Original languageEnglish
Title of host publication2021 4th International Symposium on Advanced Electrical and Communication Technologies, ISAECT 2021
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781665437738
DOIs
StatePublished - 2021

Publication series

Name2021 4th International Symposium on Advanced Electrical and Communication Technologies, ISAECT 2021

Bibliographical note

Publisher Copyright:
© 2021 IEEE.

Keywords

  • Dielectric
  • GPR
  • Nonlinear regression
  • Permittivity

ASJC Scopus subject areas

  • Artificial Intelligence
  • Computer Networks and Communications
  • Energy Engineering and Power Technology
  • Renewable Energy, Sustainability and the Environment
  • Aerospace Engineering
  • Electrical and Electronic Engineering
  • Instrumentation

Fingerprint

Dive into the research topics of 'A Nonlinear Regression-Based Approach for Estimating the Relaxation Parameters of GPR Traces'. Together they form a unique fingerprint.

Cite this