Electromagnetic Response Prediction of Reflectarray Antenna Elements Based on Support Vector Regression

Abstract

In this letter, support vector regression (SVR) is used to predict the electromagnetic (EM) response of a complex shaped reflectarray (RA) unit cell. The calculation of the scattering coefficients of passive RA elements with periodic intervals is firstly transformed into a regression estimation problem, and then an analysis model is established by SVR to quickly predict the EM response of the unit cells. To this end, the full-wave (FW) simulation software is used to obtain a set of random samples of the scattering coefficient matrix of the RA antenna unit cell, which is used for SVR training. Under the same conditions, the radial basis function network (RBFN) is also used to predict the EM response of the elements, and the comparison results show the effectiveness and accuracy of the proposed method.