Heuristic Optimization Based Learning Framework for the Design of Microstrip Patch Antenna

Abstract

In today s rapidly advancing wireless communication systems, the push for compact, high-performance antennas is driving the development of novel design and optimization methods. Microstrip Patch Antennas (MPAs) have emerged as a popular choice for many applications due to their slim profile, simple manufacturing process, and ability to integrate smoothly with circuit boards. Despite these advantages, MPAs face certain drawbacks, such as narrow bandwidth, reduced gain, and imperfect impedance matching, which can limit their effectiveness in more demanding communication scenarios. These limitations can hinder their application in advanced communication systems requiring wide bandwidth, high gain, and low return loss. This study investigates and evaluates three cutting-edge optimization techniques Northern Goshawk Optimization (NGO), a combined Particle Swarm Optimization and Simulated Annealing (PSO-SA) approach, and CryStAl algorithm coupled by an Extreme Learning Machine (ELM) to improve the design and functionality of different MPA configurations. newlineThe NGO algorithm is first utilized to fine-tune the geometric dimensions of slotted triangular MPAs, focusing on variables like slot length, slot width, and dielectric constant for improved performance. NGO, inspired by the hunting behaviors of the Northern Goshawk, enhances antenna design by performing a robust global search, which helps to avoid local optima in the optimization process. The MATLAB-implemented NGO algorithm iteratively fine-tunes antenna design parameters based on a fitness function evaluating gain, bandwidth, and return loss, achieving significant improvements in MPA performance, particularly in bandwidth expansion and return loss reduction. newlineThe second approach, PSO-SA hybrid optimization method combines the exploratory power of Particle Swarm Optimization (PSO) with Simulated Annealing (SA) for effective refinement, targeting the design of inset-fed newlinerectangular MPAs for Ku-band and C-band use.