Design and Evaluation of Radiation Pattern Reconfigurable Antenna using Bayesian Machine Learning Optimization

The fundamental benefit of a beam reconfigurable antenna is the ability to adjust the beam direction and beam steering. The present paper presents a radiation pattern reconfigurable antenna using four PIN diodes to regulate the radiation. This paper proposes a radiation pattern reconfigurable antenna with four PIN diodes to control the radiation characteristics. The arm lengths L1, L2, L3 and L4 of the antenna are optimized using a Bayesian machine learning model for the S-band applications to improve the radiation power and minimize the return loss. The designed antenna exhibits commercially acceptable gain and reflection coefficient parameters. The antenna radiates at [Formula: see text], [Formula: see text], [Formula: see text] and [Formula: see text] main beam angles under four different switching conditions with great radiation purity and isolation. The antenna is designed to operate at a resonant frequency (fc) of 2.4 GHz in the S-band with an operating bandwidth from 2.2 to 2.55 GHz. This design offers a miniature structure with good performance and well-controlled radiation patterns.