Performance Prediction of Multiport Antenna Design for UAV-Assisted IoT Networks Using Machine Learning Algorithms

This paper provides the layout and examination of a two-port alumina ceramic-built filtering antenna for UAV-assisted IoT networks. Important features of the designed antenna are (a) aperture in combination with printed line provides the filtering response, that is, −20-dBi gain outside and 5.5 dBi within the working regime; (b) suspended partial reflecting surface helps to tilt the radiation pattern by ±45°; (c) metallic strips placement over dielectric block improves the isolation to above 25 dB; and (d) machine learning (ML) approach helps to predict the |S₁₁|/|S₁₂| parameters of the designed antenna. Experimental testing helps to verify the simulated/predicted outcomes. With a separation value greater than 25 dB, the antenna's operating frequency range is 5.2–5.95 GHz. Stable MIMO parameters and filtering features, along with large antenna coverage, make the designed antenna suitable for dedicated short-range communications (DSRCs) applications (5.5 GHz) as per UAV-IoT standards.