Machine learning based compact MIMO antenna array for 38 GHz millimeter wave application with robust isolation and high efficiency performance

• Compact multiple-input multiple-output (MIMO) antenna array with a wide bandwidth of 35.181-39.689 (4.508) GHz for 5G applications. • High isolation achievement up to 27 dB, as well as an impressive maximum gain of 8.09 dB and an efficiency level of 98.2%. • RLC equivalent circuit model using ADS software • Machine learning verification The performance of wireless 5G communication networks can be enhanced by combining multiple-input multiple-output (MIMO) antennas with machine learning (ML). The suggested antenna, which is constructed on a Rogers 5880 substrate, is well-suited for usage in the high bands of 5G applications due to its 27 dB isolation and bandwidth of 35.181-39.689 (4.508) GHz within a -10 dB range. Besides being compact (measuring 37 mm × 24 mm), it boasts an impressive maximum gain of 8.09 dB and an efficiency level of 98.2%. The methods explored in this research are the RLC equivalent circuit model and simulations with validated measurements. An advanced design system (ADS) is utilized to compare the return loss because of CST to create a model like the suggested MPA. The next step is extensive data sampling using CST MWS simulation, followed by applying supervised regression ML techniques. Lasso regression yields excellent results in terms of accuracy and achieves the lowest degree of error when testing the bandwidth prediction. Considering everything, the antenna stands out as a top choice for the 5G communication system's high band. Designing a small MIMO antenna for 38 GHz mm-wave 5G applications presents challenges because it requires balancing high performance while minimizing mutual coupling between closely spaced elements and dealing with high-frequency complexities.