5G compatible microstrip antenna desing and application with artifical intelligence based deep learning method

This study aims to explore high-performance antenna designs that meet the requirements of 5G technology. The research delves into the significance of 5G antennas, their applications in wireless communication, and the challenges encountered in their design. The necessity of incorporating artificial intelligence (AI) and deep learning (DL) based modeling techniques for antenna design in 5G applications is also emphasized. In the second section, key concepts related to antenna design are discussed, with a focus on the advantages of microstrip patch antennas. These antennas are highlighted for their ability to operate at high frequencies, small size, low cost, and ease of integration, making them ideal for a wide range of applications. The study specifically examines the designs of triangular-shaped microstrip patch antennas and stacked microstrip patch antennas, detailing their structural features and performance criteria. The third section explores how AI and deep learning techniques can be applied in antenna design. The historical development of AI, including contributions from Alan Turing and the concept of the Turing Test, is discussed. The progression from expert systems to modern deep learning methods is outlined, emphasizing the potential of machine learning and deep learning techniques to provide faster, more accurate, and optimized solutions in antenna design. The section also delves into the use of supervised and unsupervised learning algorithms in the design of microwave antennas. In the final section, the study focuses on the design and optimization of triangular-shaped and stacked microstrip patch antennas using deep learning-based surrogate models. For the triangular-shaped microstrip patch antenna, a deep learning-based regression model achieved an RME value of 4.8%. For the stacked microstrip patch antenna, a convolutional neural network (CNN) combined with Honey Bee Mating Optimization (HBMO) algorithms was employed to develop a highly accurate and generalizable surrogate model. The results demonstrate that the proposed deep learning-based models offer superior computational efficiency and performance for antenna design in sub-6 GHz 5G applications.