Classification Analysis of Microstrip Antennas for 700, 2600, and 3600 MHz frequencies based on Adaptive Neuro-Fuzzy Inference System

Communication antennas find extensive application in wireless communication domains, where data exchange relies heavily on the transmission and reception of electromagnetic waves. Microstrip antennas (MSA) are commonly employed for high-frequency operations among the various antenna types. Despite the existing body of literature on this antenna, achieving an efficient output design remains challenging for engineers due to numerous factors that must be considered. Consequently, this study sought to contribute to the existing research by employing a novel methodology to attain an effective antenna design. This research used a rectangular MSA with a copper conductor and air substrate. The simulation uses operating frequencies typical of 5G systems: 700, 2600, and 3600 MHz. Matlab's antenna designer obtained various parameters such as reactance, resistance, and S11. These parameters were then processed using the adaptive neuro-fuzzy inference system (ANFIS) feature. The Tagaki-Sugeno ANFIS model, in particular, generated 27 fuzzy rules for each operating frequency, employing distinct test and training data sets. The collected data was analyzed to classify and compare the different S11 values. The analysis involved surface plots and scatter plots to examine the results, and the results demonstrated that the algorithms performed according to the anticipated outcomes, hence validating their performance.