AI-Powered Antennas and Microwave Components

AI-Powered Antennas and Microwave Components

Hafsa Talpur, Badar Muneer, Mohsin Ali Memon, Nadeem Abbas, Abi Waqas
Copyright: © 2023 |Pages: 40
DOI: 10.4018/978-1-6684-7702-1.ch004
OnDemand:
(Individual Chapters)
Available
$33.75
List Price: $37.50
Current Special Offers
10% Discount:-$3.75
TOTAL SAVINGS: $3.75

Abstract

In wireless communication systems, high-performance antenna, microwave, and radio frequency design systems are essential to meet end-user requirements. As demand for these components increases, it's crucial to design optimized structures in a short amount of time with guaranteed best results. This has led to the need for a higher level of intelligence in the design process. Artificial intelligence (AI) techniques such as evolutionary algorithms (EAs), machine learning (ML), deep learning (DL), and knowledge representation have been widely used to find parameter values of antenna and microwave components, leading to optimized designs in minimum processing time and overcoming long processing times and poor results. This chapter focuses on the major AI methods in the area of antenna, microwave, and other radio frequency (RF) components, including phase shifters, intelligent reflective surfaces (RIS), waveguides, filters, stubs, etc. The chapter discusses different EAs and ML algorithms and their use in optimizing antenna and microwave designs.
Chapter Preview
Top

1. Introduction

In today's fast-paced world, AI is playing a significant role in the development field of microwaves, antennas, and radio frequency (RF) systems, becoming an integral part of the RF front-end in wireless communication systems. AI techniques in antenna design, including gain, radiation pattern, half power-beamwidth, and S-parameters estimation, provides optimized and accurate results compared to conventional rules based on design experience or computationally long simulations (Weiland et al., 2008).

Traditionally, optimizing sub-optimal antenna and microwave component designs involved a time-consuming process of parameter tuning based on hit and trial methods with no guaranteed results. To address these issues and reduce market time, automation of antenna and microwave component optimization is necessary to obtain nearly optimal designs in the shortest time possible. Antenna structure optimization can be done through local or global optimization methods, with the latter using EAs as the major group of AI for antenna design.

ML algorithms are widely used in many antenna applications. These algorithms have replaced computationally expensive electromagnetic simulations, saving time and effort. In addition, knowledge representation is also used in antenna design through the use of semantic web technologies such as the Ontology Web Language (OWL) illustrated in Figure 1. The concept of ontologies is the first step in automated machine-based antenna and microwave system designs (Goudos et al., 2022).

Figure 1.

Categories of AI

978-1-6684-7702-1.ch004.f01
Source: Goudos et al. (2022)
Top

2. Ai In Antennas And Microwave Component Designing

2.1. AI and Its Subsets

Artificial Intelligence at a Glance

AI refers to the methods and techniques used to enable computers or machines to think and respond intelligently in solving tasks. It is an interdisciplinary field that draws on applied mathematics and computer science. To determine whether a machine is intelligent or not, we can compare its responses to those of a human being, who has an innate understanding of what intelligence is. Initially, AI researchers focused on modeling natural neurons in the brain. In 1943, McCulloch and Pitts proposed the idea of artificial neurons that function like binary switches, turning on or off (Matsumura et al., 2019). In 1949, Donald Hebb developed a neural network-based learning algorithm, and in 1951, Dean Edmonds and Marvin Minsky built the first neural network (NN) machine, called the Stochastic Neural Analog Reinforcement Calculator (SNARC). Following these achievements, the term “artificial intelligence” is defined as the creation of machines that can think and respond to tasks intelligently, like humans.

AI, ML, and DL: A Comparison

Artificial Intelligence (AI), Machine Learning (ML), and Deep Learning (DL) are three interconnected elements such that ML is the subset of AI and DL is the subset of ML as shown in the figure. 2. ML is a field of research and study that enables humans to design and develop computers that are capable of learning from experience(data) and does not require time-consuming and costly procedures of computer programming. Its goal is to develop computer systems that may learn to solve tasks just like humans, without any direct instructions to solve the tasks. This learning mechanism is realized by using various methods from statistics and data analysis to analyze data and predict patterns to accomplish given tasks (H. M. El Misilmani et al., 2020).

Figure 2.

Subsets of AI

978-1-6684-7702-1.ch004.f02
Source: Valarmathi et al. (2021)

Artificial Neural Networks (ANN) is the famous ML model and part of NN. It works according to the processes of the human brain. Firstly in 1943, McCulloch, an American psychologist, and Pitts presented the M-P model based on an algorithm that uses neurons as a functioning device, hence this became the beginning of theoretical research for the NN model. Then Rosenblatt, in 1958, developed the perceptron model inspired by the M-P model (Cheng et al., 2016).

After all these achievements, the subcategory of ML i.e., DL was introduced. DL focused on configuring deep neural networks (DNN) and helps to train machines to understand complicated patterns inside large data. It has more than one hidden layer.

Complete Chapter List

Search this Book:
Reset