Isolation Analysis of Miniaturized Metamaterial-Based MIMO Antenna for X-Band Radar Applications Using Machine Learning Model

A novel metamaterial-based circular patch multi-input multi-output (MIMO) antenna is designed with a ‘C’-shaped defected ground structure for high isolation. A 4 × 4 mm2 unit cell for a ring resonator has been designed and exhibited double negative material (DNG) properties from 1.0 to 2.92 GHz and 13.68 to 17.67 GHz and Mu negative material (MNG) from 4.70 to 13.67 GHz. The proposed antenna structure is designed by embedding the ring resonator-based meta-structure to a circular patch antenna and fabricated with dimensions 0.245λ0 × 0.409λ0 (15 × 25 mm2). The proposed antenna operating at 8.50 to 14.23 GHz for X and lower Ku bands is used in the Unmanned Arial Vehicle (UAV’s) applications. The spacing between elements is 0.088λ0 (5.4 mm) on an FR4 epoxy substrate, and the ‘C’-shaped structure on the back of the antenna improves the isolation of more than 24 dB in the operating band. Distance between the antenna elements plays a crucial role, and parameters affected by this are optimized by introducing machine learning. For future predictions, a linear regression model was created to optimize the parameters’ linear dependencies like isolation and return loss on the distance between the antenna elements. The radiation efficiency and gain of the antenna are enhanced by 92% and 6.02 dB at 13.22 GHz, respectively. The MIMO antenna’s simulated results of diversity and other parameters are in the acceptable range with the measured results used for X-band radar applications. The proposed decoupling technique is simple to understand and implement. © 2022, Electromagnetics Academy. All rights reserved.