Efficient Optimization of Dual Port Circularly Polarized 1*2 Array based Microstrip Antenna using Machine Learning Algorithms

This work studies the design and optimisation of a two-port high gain printed antenna using machine learning algorithms. The following are a few crucial characteristics of the suggested radiating structure: (i) In order to shorten the simulation time, machine learning (ML) techniques based on ANN and random forests are used to predict the reflection coefficient (RC) (|S11|) features of the designed radiator; (ii) Circularly polarized waves in the 10.38-10.45 GHz range are produced by using a stair slot; (iii) A 1*2 array structure yields high gain, or 7.5 dBi, within the operating frequency range; and (iv) polarization diversity concept is provided by opposite orientation of aperture, which raises the isolation level to more than 30 dB. It is established that the intended radiator functions between 10.37-10.55 GHz by comparing the findings of the HFSS and CST with the predictions made by the ML algorithm. The suitability of the suggested radiator for X-band applications is approved by good values for the diversity and far-field characteristics.