Inverse Problem for Electromagnetic Propagation in a Dielectric Medium using Markov Chain Monte Carlo Method (Preprint)

Abstract

This paper is concerned with a stochastic inverse methodology arising in electromagnetic imaging. Nondestructive testing using guided microwaves covers a wide range of industrial applications including early detection of anomalies in conducting materials. The focus of this paper is the identification of electromagnetic material parameters and emphasis is on one dimensional scattering of a dielectric slab. The direct problem can be solved numerically using the finite-difference time-domain method (FDTD). The Markov Chain Monte Carlo method (MCMC) is applied to the inversion problem. Some successful results of computational experiments are demonstrated in order to show the feasibility and applicability of the proposed method.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Aug 01, 2012
Accession Number
ADA565876

Entities

People

  • Fumio Kojima
  • Jeremy S. Knopp

Organizations

  • Air Force Research Laboratory

Tags

Communities of Interest

  • Advanced Electronics
  • Biomedical

DTIC Thesaurus Topics

  • Air Force
  • Air Force Research Laboratories
  • Algorithms
  • Computational Science
  • Detection
  • Dielectric Permittivity
  • Equations
  • Finite Difference Time Domain
  • Inverse Problems
  • Markov Chains
  • Materials
  • Monte Carlo Method
  • Nondestructive Testing
  • Probability
  • Scattering
  • Structural Health Monitoring
  • Time Domain

Readers

  • Finite Element Method (FEM) for solving Partial Differential Equations (PDEs)
  • Neural Network Machine Learning.
  • Plasma Physics / Magnetohydrodynamics