Accurate Characterization of Graphene Reconfigurable Reflectarray Antenna Element by SVR

The analysis of the graphene reconfigurable reflectarray antenna electromagnetic (EM) response through the use of support vector regression (SVR) has been carried out in this paper. The EM response of the original graphene reflectarray antenna is transformed into a regression estimation problem by SVR. In this work, we consider up to 6 degrees of freedom: the width W and length L of the graphene patch, the graphene chemical potential mu(c), the working frequency f, and the elevation theta and azimuth angles phi of the incident field. Firstly, they are discretized separately, and then full-wave (FW) simulation software is used to calculate the reflectarray antenna unit cell under different parameters to obtain the training data set of SVR. Finally, SVR is used to obtain an alternative model of the graphene reconfigurable reflectarray antenna unit cell, which is used to quickly predict the scattering coefficient matrix in co-polarized and cross-polarized states. Through the comparison of SVR method with the radial basis function network (RBFN) and FW simulation results, the effectiveness of the proposed method in precision and efficiency is verified.