Decision-Tree Based Machine Learning Approach for the Design and Optimization of 5G n78 Sub-Band Antenna for WiMAX/WLAN Applications

In these days, it is inevitable to apply the machine learning techniques in all fields of science and engineering including electromagnetic applications. In this article a thoughtful attempt is made to introduce decision tree machine learning algorithm for the design of a microstrip patch antenna for wireless WiMAX/WLAN application at 5G-n78 sub-band frequency (3.3GHz).The proposed antenna is a simple rectangular patch of 0.304λ X 0.238λ dimensions designed on FR-4 substrate (ɛ
r
=4.4) by using line feeding technique. The antenna is radiating at 3.3GHz frequency with S
11
of -37.4dB, VSWR of 0.233 and gain of 4.06 dB. The data points were collected from the full wave electromagnetic solver High Frequency Structural Simulator (HFSS) and decision tree is built in R-language platform. The decision tree algorithm is applied for multiple length values of metal patch by targeting S
11.
The mean square error calculated is 3.459 & 3.75 for train and test data respectively, further decision tree is generated. The designed antenna has good potential to apply for wireless WiMAX/WLAN and 5G applications and the decision tree obtained is better suiting for the optimization of the designed antenna.