Advanced Parasitic Microstrip Antenna Design for IoT and 5G Networks Using Attention-Enhanced Graph Convolutional Model

Abstract

The growing demand for high-speed, reliable wireless communication has accelerated advancements in antenna technology for 5G and IoT applications. However, existing antenna designs often face challenges such as limited bandwidth, inadequate gain, and poor impedance matching, which hinder their ability to meet the stringent performance requirements of modern networks. To address these limitations, this research presents a high-gain, wideband parasitic microstrip antenna designed for 5G and IoT applications, operating at 26 GHz within the 5G new radio Frequency Range 2 (FR2) band n258. The proposed antenna incorporates a miniaturized parasitic patch design featuring eight microstrip patches arranged around a centrally probe-fed active patch in a squared configuration. These parasitic patches are electromagnetically coupled via the magnetic and electric fields generated by the active patch, achieving a compact array with a total dimension of 24 × 24 mm². To enhance antenna performance, a Multi-Layer Attention Graph Convolutional Network (MLAGCN) is utilized to effectively extract key features from the input data, whereas the Gooseneck Barnacle Optimization (GBO) algorithm iteratively fine-tunes the design parameters. The antenna achieves a maximum gain of 12 dB and an efficiency exceeding 95% within the frequency range of 23–28 GHz. This integrated design and optimization approach facilitates cutting-edge performance in terms of bandwidth, gain, and reliability, meeting the rigorous demands of 5G/6G, IoT, and other next-generation services, as well as extending network coverage. This research proposes a compact parasitic microstrip antenna optimized for 5G/IoT using an MLAGCN and GBO. The MLAGCN captures interdependent antenna parameters, whereas GBO fine-tunes design variables for enhanced impedance matching and bandwidth. Results show significant improvement in return loss and wideband performance compared with conventional designs.