Analysis and synthesis of L- and T-shaped flexible compact microstrip antennas using regression-based machine learning approaches

Abstract

Abstract The purpose of this study was to analyze and synthesize L-shaped compact microstrip antennas (LCMA) and T-shaped compact microstrip antennas using regression-based machine learning algorithms (TCMA). This was accomplished by simulating 3808 LCMAs and 900 TCMAs operating at UHF and SHF frequencies with different physical and electrical characteristics. The acquired data was utilized to create a data set containing the antennas’ physical and electrical characteristics, as well as their resonant frequencies in the TM010 mode. Four baseline regression models and seven machine learning models were developed to determine the resonance frequency of antennas and the values of the physical parameters required for a particular frequency. To examine the efficacy of machine learning models, three-dimensional LCMAs and TCMAs were created using polylactic acid (PLA) and felt-based flexible substrates, as well as copper tape. The results illustrate the feasibility of using machine learning models for LCMA and TCMA analysis and synthesis.