Prediction of the resonant frequency of square patch microstrip antenna with DGS using Machine Learning

Abstract

In this paper, two robust machine learning (ML) prediction methods, Gaussian Process Regression (GPR) and Artificial neural network (ANN), are used to predict the resonant frequency of square patch microstrip antenna (SPMA) with an equilateral triangular defect in the ground. The designed antenna can work in the band of 5.9206 GHz to 25.7625 GHz for the C, X, k u , K band applications. By varying the size of the square patch and triangular DGS, a total of 125 data samples were collected through the simulation process using CSTTM, and out of which 105 data samples were used to build two ML models. For validation of the authentication of these models, 20 data samples used. After preparing the model, testing is done for 20 data sets and result obtained from ANN. GPR model was compared with simulated resonant frequency and found that the GPR model gives a better result than the ANN model. The predicted outcomes show that ANN and GPR models can be used to predict the resonant frequency of SPMA in the range of 5.9206 GHz to 25.7625 GHz.