Neural net method for measuring electromagnetic parameters of artificial electromagnetic material

Abstract

The invention relates to a new method for measuring electromagnetic parameters, which is required to measure left-hand materials and artificial electromagnetic materials with periodic structures and have high measurement result precision, and the manufacture of a measurement sample is simple. A neural net arithmetic is a biological computation method which is suitable for a highly nonlinear problem, and is already widely applied to the other field of information processing computation and proved to be a better arithmetic. The method combining the neural net computation with a computing electromagnetics method computes transmission and reflection coefficients of measured materials by number value computation methods FEM and FDTD of computation electromagnetics and training neural nets by taking corresponding computing results as a training sequence. When the neural nets are sufficiently trained, the effective dielectric constant and the effective magnetic conductivity of the measured materials are computed through measurement values of the transmission and reflection coefficients.

Description

technical field [0001] The invention relates to the field of measuring equivalent electromagnetic parameters of artificial electromagnetic materials. More specifically, the invention is dedicated to a measurement method combining neural network algorithms with computational electromagnetics methods. Background technique [0002] Artificial electromagnetic materials have unique electromagnetic properties and potential application prospects, and their research has created a new field. With its development, many original technologies will get new developments and breakthroughs. It will be applied in many research fields: wireless communication technology, radar technology, nanotechnology and microelectronics technology, etc. Among them, the research on its design and manufacturing method, parameter testing method and application research has always been a hot issue in the field of artificial electromagnetic materials research at home and abroad. [0003] The research on the e...

AI Technical Summary

Problems solved by technology

[0004] First, there will be a phase error problem of the transmission coefficient. Although this problem has been solved to a certain extent, there will still be some errors at the half-wavelength frequency point;

[0005] Second, material samples are difficult to make, especially in the coaxial line measurement model, the test material requires high processing accuracy, and will cause damage to the material itself, therefore, some materials with complex periodic structures are often difficult to process into test sample;