Premium
Unmagnetized plasma media parameters reconstruction using FDTD and gradient‐based minimization method
Author(s) -
Deng Haochuan,
Shi Lijuan,
Feng Xuejian,
Wei Xiao,
Yin Hongcheng
Publication year - 2019
Publication title -
international journal of numerical modelling: electronic networks, devices and fields
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.249
H-Index - 30
eISSN - 1099-1204
pISSN - 0894-3370
DOI - 10.1002/jnm.2695
Subject(s) - finite difference time domain method , conjugate gradient method , lagrange multiplier , nonlinear conjugate gradient method , gradient method , minification , inverse scattering problem , curl (programming language) , mathematics , time domain , mathematical analysis , inverse problem , physics , algorithm , mathematical optimization , computer science , optics , gradient descent , machine learning , artificial neural network , computer vision , programming language
A time‐domain inverse scattering algorithm for determining unknown plasma resonance frequency and electron collision frequency of unmagnetized plasma media object has been proposed. Finite‐difference time‐domain (FDTD) method in conjunction with Lagrange multipliers and gradient‐based minimization method are used in the proposed algorithm. The reconstruction of the parameters is based on the minimization of a cost function that describes the discrepancy between simulation and measured data of the electric field at the observation points. In order to obtain the gradient value of the unknown plasma parameters, an augmented cost function has been given by means of Lagrange multipliers with the Maxwell curl equations and constitutive relation of the unmagnetized plasma media. Then, the gradient value could be obtained after a series of transformations. At last, the Ploak‐Ribiere conjugate gradient algorithm has been applied to update the plasma parameters iteratively. The effectiveness of the proposed algorithm has been validated through the simulation test.