See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/336846687

Machine Learning in Antenna Design: An Overview on Machine Learning

Concept and Algorithms

Conference Paper · July 2019

DOI: 10.1109/HPCS48598.2019.9188224

CITATIONS READS
65 17,287

2 authors:

Hilal El Misilmani Tarek Naous

Beirut Arab University Georgia Institute of Technology
31 PUBLICATIONS 369 CITATIONS 22 PUBLICATIONS 256 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Hilal El Misilmani on 20 May 2020.

The user has requested enhancement of the downloaded file.

Machine Learning in Antenna Design: An Overview
on Machine Learning Concept and Algorithms
Hilal M. El Misilmani and Tarek Naous
Electrical & Computer Eng. Department
Beirut Arab University
Beirut, Lebanon
[email protected], [email protected]

Abstract—With the growth and wide variety of available data,

advanced processing, and affordable data storage, machine learn-
ing is witnessing great attention in finding optimized solutions in
various fields. Machine learning techniques are currently taking
a major part of the ongoing research, and expected to be the
key player in today’s technologies. This paper introduces and
investigates the applications of machine learning in antenna
design. It covers the major aspects of machine learning, including
its basic concept, differentiation with artificial intelligence and
deep learning, learning algorithms, its wide applications in
various technologies, with a main focus on its usage in antenna
design. The review also includes a comparison of the results
using machine learning in antenna design, compared to the
conventional design methods.

Index Terms—Machine learning, artificial intelligence, antenna

design, neural networks, electromagnetic simulations Fig. 1: Relationship between Artificial Intelligence, Machine
Learning, and Deep Learning
I. I NTRODUCTION
Artificial Intelligence (AI) is the art of enabling machines and investigated in this paper, machine learning is a large
to perform tasks that require human thinking abilities, such subset of artificial intelligence, as shown in Fig. 1. In fact,
as learning, decision making, and problem solving. In sim- machine learning can be considered as an approach to achieve
ple words, AI is implementing human thinking ability in AI applications.
machines. With the latest advances in big data availability, Machine Learning is briefly described as getting useful
software engineering capability and affordable high computing information out of data, achieved by developing reliable pre-
power, AI is becoming an essential part of today’s research. diction algorithms. These algorithms can be very powerful in
It is expected to affect most of our daily activities, with fun- optimization, but their success rely on the condition and size
damental changes in science and engineering, and enormous of the data collected. Therefore, machine learning is frequently
impacts on society, with the ability to create, transform, or associated with statistics and data analysis [2], [3].
optimize different aspects and applications of our daily lives. As for Neural Networks, they are defined as a type of
To be able to implement and achieve artificial intelligence, machine learning algorithms that tries to imitate how the
several capabilities need to be possessed by the computer or human brain works. They consist of layers of interconnected
machine. For instance, natural language processing is required nodes. Each node produces a nonlinear function of its input.
for successful communication in English and knowledge rep- Deep Neural Networks (DNNs) are neural networks that have
resentation is needed for storing information. To answer ques- more than one hidden layer [4], usually refereed to as Deep
tions and drawing conclusions using the information stored, Learning. Both of these algorithms are considered as types of
automated reasoning is required, whereas, to derive patterns machine learning algorithms.
in data and make predictions and adapt to new conditions, This paper presents an overview on machine learning, with a
machine learning is necessary. For object perception and major focus on investigating its usage in antenna design appli-
manipulation computer vision and robotics are needed. By cations. The concept of machine learning is studied along with
using intelligent algorithms and large amount of data sets, its different learning algorithms. Next, an extensive review
iterative processing allows AI software to learn automatically of several antenna designs and electromagnetic computational
from patterns or features [1]. methods is investigated. The methods used to design the
Artificial Intelligence and Machine Learning (ML) are often antennas using machine learning in each is presented, along
used interchangeably, however, as will be discussed hereafter with the outcome of each algorithm used.

978-1-7281-4484-9/19/$31.00 ©2019 IEEE 600

 TABLE I: The Three Main Categories Of Learning – k Nearest Neighbors [9]: considered among the simplest
Learning Category Description of all machine learning algorithms. After memorizing the
A model is trained on a data set training set, the output of any new input is predicted by
Supervised Learning
Predictions are made on new inputs
A pattern is derived from the data
the algorithm based on the outputs of its closest neighbors
Unsupervised Learning in the training set.
after exploring it
Model takes decisions and learns In unsupervised learning, what the results should look like is
Reinforcement Learning
from its actions
not known, the algorithm derives a structure from the data after
identifying similarities in the inputs [7]. Unsupervised learning
algorithms include but not limited to: K-Means Clustering and
Dimentionality Reduction Algorithms.
– K-Means Clustering [3]: k-unique clusters are automati-
cally formed by this algorithm. It is a type of unsuper-
vised learning where variables in the data are grouped
together based on relationships among them.
– Dimentionality Reduction Algorithms [9]: such as the
Principle Component Analysis Algorithm (PCA) where
Fig. 2: Schematic of an artificial neural network the goal of this algorithm is minimizing the projection
error by reducing the reducing every feature’s distance to
a certain projection line.
II. M ACHINE L EARNING OVERVIEW In reinforcement learning, no labeled data set is received by
Machine learning is based on algorithms that can learn from the machine. Instead, information is collected after interacting
data without relying on rules-based programming [5]. It can be with the environment through different actions. The machine is
generally divided to three key categories: supervised learning, rewarded after each action, hence its objective is maximizing
unsupervised learning, and reinforcement learning [2], [6]–[8], this expected average reward where the action would become
listed in Table I. Additional learning scenarios that are less optimal. An example of a reinforcement learning model is the
widely used include: Semi-supervised learning, Transductive Markov Decision Process (MDP). [2], [8]
Inference, On-line learning and Active Learning [2]. Each Other machine learning algorithms exist with less widely
category has different learning algorithms that belong to it. usage, such as: Decision Trees [12], Boosting [13], Naive
In supervised learning, what a correct outputs looks like is Bayes [9], Bayesian Regularization [14], Kriging [15], and
already known. After training the learning algorithm on a given more [9].
When solving real world machine learning problems, the
data set, the algorithm generalizes to give accurate predictions
data set is split into three parts [9]: a training set used to train
to all possible inputs [7]. Supervised learning algorithms
the algorithm, a cross-validation set used for model selection,
include but not limited to: linear regression, logistic regression,
and a test set used for testing the performance of the prediction
artificial neural networks, and support vector machines.
algorithm, that is checking if it succeeds to generalize on
– Linear Regression [9]: consists of fitting a continuous inputs not previously seen.
linear function through the data from which the algorithm Machine learning techniques have been widely employed in
can make predictions on new inputs. communication technology, such as antenna selection in wire-
– Logistic Regression [9]: used in classification tasks where less communications [16], [17], smart grid networks where
it predicts the probability that a certain input corresponds machine learning can detect malicious events before occuring
to one of the already known classes. [18], wireless networks where ANNs can be used for predict-
– Artificial Neural Networks [9], [10]: in neural networks, ing the mobility patterns and content requests of wireless users
large interconnections of ”neurons”, which are simple [6], speech recognition where SVMs with kernel functions
computing cells, are employed to achieve good perfor- can be useful to improve its generalization performance [19],
mance. When complex functions with many features are context aware computing in IoT [20], and much more [21].
found, neural networks offer an alternate way to perform
ML. Neural networks are made of multiple layers, the III. M ACHINE L EARNING IN A NTENNA D ESIGN
input layer, the output layer, and hidden layers between Several papers have investigated the applications of ma-
the input and output layers, as shown in Figure 2. One of chine learning in antenna design. It is expected that machine
the many types of neural networks is the feed-forward learning can provide accelerated antenna design process while
neural network where a weighted sum of connected- maintaining high accuracy levels, with a minimization of
neurons output are received by each neuron as an input. error and time saving, along with a possible prediction of
– Support Vector Machines (SVMs) [10], [11]: another type the antenna behavior, a better computational efficiency and
of supervised learning algorithm. In particular, it is used reduced number of necessary simulations.
in classification and deals with the more difficult case of In general, in order to apply machine learning in antenna
non-linearly separable patterns by using kernel methods. design, the following steps can be done:

601
 Fig. 3: Reflectarray unit cell [26]
(a)

1) By multiple simulations, the electromagnetic character-

istics of an antenna are found out.
2) These characteristics are stored in a database and used
as a data set for training a certain machine learning
algorithm.
3) The antenna that gives the closest results is designed by
the algorithm after making predictions, depending on the
needs of the designer.
In this section, a detailed review of several antenna de-
sign papers using machine learning is provided. For a clear
investigation, the papers discussed here are divided to two
main categories: machine learning for parameter optimization,
and machine learning for enhancing evolutionary computation (b)
algorithms.
Fig. 4: Gain pattern comparison of SVM and MoM (a)
A. Machine Learning For Parameter Optimization elevation and (b) azimuth [26]
One approach of using machine learning in antenna design
is training a learning algorithm on data collected from previous
simulations to optimize the antenna parameters. The design of a rectangular microstrip antenna using Sup-
Based on Local-Periodicity (FW-LP), successful analysis port Vector Machines (SVMs) has been also presented in [28].
and optimization has already been achieved using a Full-Wave SVMs are used here for regression, and hence it may be called
analysis [22], [23], however, for space applications, faster Support Vector Regressors (SVR). The machine learning al-
computations are required. Although this has been addressed gorithm is trained on a dataset that includes measured values
previously using Artificial Neural Networks (ANNs) [24], of the input impedance, operation bandwidth and resonant
[25], the results were limited. In [26], the design of shaped- frequency of the antenna. Artificial Neural Networks are also
beam reflectarray unit cell using Support Vector Machine employed at the same time to compare the results with those
(SVM) has been presented. Two sets of four parallel dipoles obtained using SVR. It was shown that employing machine
are used in the design, shown in Fig. 3. To accelerate the reflec- learning algorithms in the design of microstrip antennas gives
tarray antennas design machine learning algorithm is used. The better and more accurate characterization than the theoretical
reflection coefficient matrix is first characterized using SVM. results. Comparing SVR and ANN, it was shown that SVR
The scattering parameters of the unit cell dimensions are then provides better computation efficiency since it has a faster
derived. The influence of the discretization of the angle of convergence rate, with less training time and test time needed
incidence is also taken into account in the simulations. The in comparison to ANN.
results in paper using SVM are then compared to those in The design of a planar inverted-F antenna (PIFA), shown
[27] based on the Method of Moments (MoM-LP) in terms in Fig. 5, with magneto dielectric Nano-composite (MDNC),
of speed and accuracy. The results showed that the sequential using the Bayesian Regularization algorithm as the neural
design using SVM can be accelerated by a factor of 880, and network learning process, has been presented in [14]. Starting
the parallel design can also by accelerated by a factor of 566. with the nano-magnetic material’s volume fraction and particle
Fig. 4 showing the gain pattern comparison of the SVM and radius, the different antenna parameters, such as radiation
the MoM-LP using the real angles of incidence. efficiency, gain, resonant frequency, bandwidth and others, can

602
 (a)

Fig. 7: A comparison of the accuracy and time saving using

Neural Networks [15]
(b)

Fig. 5: PIFA antenna structure: (a) top view, (b) side view [29] The design of reflectarrays using the Kriging algorithm
instead of the classical full-wave solvers has been presented in
[15]. Machine learning is used here to predict the response of
complex unit cells to design high-performance reflectarrays.
The scattering matrix of complex reflectarray elements is
predicted while reducing computational time. The model is
trained on a training set of N samples, where N ∈ [500,20000],
to learn the input output relationship. A comparison of the
accuracy and time saving using the proposed method is shown
in Fig. 7. It was concluded that using the Kriging model, a 99%
time saving can be reached, while maintaining a prediction
error below 5%.
The design of a W-Band slotted waveguide antenna array,
shown in Fig. 8, by employing artificial neural networks
has been also presented in [30]. The input layer of the
neural network consists of seven design parameters which
Fig. 6: Machine learning output vs target [29] are the lengths and orientation angles of the coupling slots
(lc , θ1 , θ2 , θ3 , θ4 , θ5 ), in addition to the length of the radiating
slots (lR ). A data set of 189 examples are obtained by simu-
be calculated with high accuracy using machine learning. For lations in HFSS and stored in a database which will be used
this design two databases are created. The first one contains for training, cross-validation and testing of the model. After
material properties (volume fraction and particle radius). The obtaining the optimized values of the design parameters, the
second one contains electrical properties of the material, antenna is fabricated using the SLA 3D printing techniques.
such as electric and magnetic loss tangent, permittivity and The simulated S11 characteristic as well as the gain of the
permeability. To create a relation between the performance antenna are compared to those measured and a good agreement
of the antenna and its properties, machine learning algorithm with some slight errors has been shown.
is trained using the two databases as input and output. New
models for permeability and permittivity of the composite B. Machine Learning For Enhancing Evolutionary Computa-
material are also proposed in [14]. Using these equations with tion Algorithms
only 42 samples created, a total error of only 7% is obtained Another approach of using machine learning in antenna de-
between machine learning data and the second database. Fig. sign is embedding it in Evolutionary Algorithms. Evolutionary
6 shows a comparison of the Machine Learning Output and Algorithms (EAs) are taking important interest in computing
the target (all of the data stored in second database) regression sciences. Using evolutionary mechanisms involved in natural
curve. It was concluded that machine learning techniques can selection, evolutionary algorithms allow to find approximate
be very useful in minimization of error and acceleration of the solutions to optimization problems. EAs have been used in
cycle time for new materials synthetization, and in expecting previous works [31]–[33] for the purpose of antenna design
the behavior of the antenna without performing extensive time and optimization. To improve their performance, machine
consuming simulations. Optimized results can be obtained by learning techniques have been used successfully in Evolution-
increasing the number of data samples used with the algorithm. ary algorithms. When using EAs, databases of adequate size

603
 (a)

Fig. 9: Regression plot [35]

(b)

Fig. 8: Slotted Waveguide Antenna (a) top view, (b) bottom

view. [30]

are produced which allows to use machine learning techniques

[34].
Particle Swarm Optimization Algorithm (PSOA), a type of
evolutionary algorithm, is used in [35] to design a multi-
band patch antenna using artificial neural networks. After the
geometrical parameters of the antenna are decided by the
PSOA, a mapping function is built by the ANN so that the
frequencies and associated bandwidths can be related to the
dimensional parameters of the antenna. The flow chart of the
optimizer used with four input parameters, is shown in Fig.
10. The results showed that the design process can be speed
up by eliminating the need for time-consuming simulations,
therefore reducing the computational burden significantly. The
design antenna has been also fabricated and tested, with good
analogy between the measured and simulated results in [35]. Fig. 10: Flow chart of proposed modified optimizer [35]
In [36], the design of an E-shaped antenna by combining
differential evolution (DE) with the Kriging algorithm, has
been presented. Six antenna parameters are optimized, the feed curves are also shown in Fig. 12. Combining machine learning
position Px, the slot position Ps, the patch width W, the slot with evolutionary algorithms has been proven to provide a
width Ws, the patch length L and the slot length Ls, shown in faster convergence rate with a similar solution quality in
Fig. 11. It is suggested that similar results of other optimization comparison to optimization methods, self-adaptive differential
techniques can be obtained while reducing the number of evolution [37] and wind driven optimization [38]. It was shown
necessary simulations by 80%. The magnitude of the S11 that the same optimization goals can be reached while reducing
parameter is desired to be minimized at frequencies 5.0GHz the number of simulations needed by self-adaptive differential
and 5.5GHz. After the proposed algorithm was run 5 times, evolution [37] and wind driven optimization [38] by 82.3%
optimum solutions were found, and good prediction accuracy and 77.9% respectively.
was exhibited by the model. The predicted and simulated S11 The design of an ultrawide band microstrip antenna by

604
 TABLE II: Comparison of the different machine learning techniques used in the investigated papers
Learning
Paper Antenna Type Compared To Results
Algorithm Used
Accelerated design process while
[26] Reflectarrays SVM MoM & ANN
maintaining high accuracy levels
Minimization of error and acceleration of
[14] Planar Inverted F-antenna (PIFA) Bayesian Regularization
cycle time for new materials synthetization
Time saving can reach 99.9% while
[15] Reflectarrays Kriging
maintaining a prediction error below 5%
Possible prediction of antenna behavior
[29] Planar Inverted F-Antenna (PIFA) ANN Conventional Simulations
without extensive electromagnetic simulations
Better computation efficiency with
[28] Rectangular Microstrip Antenna SVM ANN
a faster convergence rate
Computation of several antenna parameters
[30] Slotted Waveguide Antenna ANN Conventional Simulations with good agreement with the simulated and
fabricated results
Design process can be sped up by eliminating
[35] Antenna (SWA) ANN Conventional Simulations
the need for time-consuming simulations
Similar results of other optimization techniques
[36] Stacked Patch Antenna Kriging Conventional Simulations can be obtained while reducing the number of
necessary simulations by 80%
The optimum results were found without any
[39] E-Shaped Antenna Linear Regression Conventional Simulations
necessary simulations
The speed of the design and optimization
[40] Microstrip Antenna Gaussian Process ML Differential Evolution procedure by more than four times
compared with differential evolution

Fig. 12: The simulated and predicted S11 curves of the optimal
design [36]
Fig. 11: E-shaped patch antenna [36]

combining regression with an evolutionary algorithm has been

presented in [39]. Machine learning technique is employed
here along with evolutionary algorithm in the estimation of
fitness function behaviors: the bandwidth (BW), the return loss
(RL) and the central frequency division (CFD). The regressive
machine learning algorithm used allows to fit a curve through
a discrete set of known data points which are the antenna
parameters obtained previously by simulations. A prototype
algorithm has been used to find the optimized values for
Ws and Ls , shown in Fig. 13, while keeping other antenna Fig. 13: Microstrip antenna with investigated parameters [39]
parameters constant and considering design restrictions (BW
> 9.0GHz, RL < -20dB and CFD < 0.37Hz). It was shown
that, using 170 datasets, the optimum results were found larger number of datasets is used.
while meeting the required restrictions. The perception on the The design of inter-chip antenna, a four-element linear array
behavior of the objectives (BW, RL and CFD) increases as a antenna and a two-dimensional array have been presented in

605
[40]. A new method for designing antennas called surrogate and computational techniques have been also studied. It was
model assisted differential evolution for antenna synthesis seen that machine learning can provide accelerated antenna
(SADEA) has been proposed. This method combines Gaussian design process while maintaining high accuracy levels, with a
Process Machine Learning and a Differential Evolution (DE) minimization of error and time saving, along with a possible
Algorithm. Using this method in the design of the three differ- prediction of the antenna behavior, a better computational
ent antennas, the results showed that, using machine learning, efficiency and reduced number of necessary simulations.
SADEA enhanced the speed of the design and optimization
procedure by more than four times compared with DE. In R EFERENCES
[41] Less time consumption has been also achieved in the [1] S. J. Russell and P. Norvig, Artificial Intelligence: A Modern Approach,
design of an isotropic wire antenna by combining a machine 3rd Edition. Pearson Education, 2016.
learning technique with an evolutionary algorithm, the fitness [2] M. Mohri, A. Rostamizadeh and A. Talwalkar, Foundations of Machine
Learning. The MIT Press, 2012.
proportionate selection. [3] P, Harrington, Machine Learning in Action. Manning Publications, 2012.
By employing Support Vector Machines (SVMs), the design [4] R. Jiang, X. Wang, S. Cao, J. Zhao and X. Li, “Deep Neural Networks
of a regular microstrip antenna has been presented in [42]. for Channel Estimation in Underwater Acoustic OFDM Systems,” IEEE
Access, vol. 7, pp. 23579–23594, 2019.
Better characterization with higher accuracy is achieved by [5] D. Pyle and C. S. Jose, “An executive’s guide to machine learning,”
using SVMs. A comparison of the obtained results with the tech. rep., McKinsey & Company, June 2015.
results of a different approach that uses Artificial Neural [6] M. Chen, U. Challita, W. Saad , C. Yin and M. Debbah, “Machine
learning for wireless networks with artificial intelligence: A tutorial on
Networks (ANNs) is also presented. neural networks,” CoRR, vol. abs/1710.02913, 2017.
A summary of the different papers using machine learning [7] S. Marsland, Machine Learning: An Algorithmic Perspective, Second
in antenna design are summarized in Table II. Edition. Chapman and Hall/CRC, 2014.
[8] A. Burkov, The Hundred-Page Machine Learning Book. Andriy Burkov,
2019.
IV. C HALLENGES IN M ACHINE L EARNING
[9] S. Shalev-Shwartz and S. Ben-David, Understanding Machine Learning:
Although Machine Learning is very useful, it comes with From Theory to Algorithms. Cambridge University Press, 2014.
[10] S. O. Haykin, Neural Networks and Learning Machines. Prentice Hall,
a lot of challenges. The most common challenges include but 2008.
are not limited to: [11] K. Crammer and Y. Singer, “On the algorithmic implementation of
1) The choice of learning algorithm: It is not easy to multiclass kernel-based vector machines,” Journal of Machine Learning
Research, vol. 2, p. 265–292, 2001.
decide what algorithm to choose as there are a great [12] L. Hyafil and R. L. Rivest, “Constructing optimal binary decision trees
number of them. This depends directly on what is being is NP-complete,” Information Processing Letters, vol. 5, no. 1, p. 15–17,
predicted and also on the type of data acquired. A good 1976.
[13] Y. Freund and R. Schapire, “A decision-theoretic generalization of on-
practice is to always visualize the data before choosing line learning and an application to boosting,” European Conference on
the algorithm. Computational Learning Theory (EuroCOLT), p. 23–37, 1995.
2) Problem Formulation: Beginning with wrong assump- [14] C. Gianfagna, M. Swaminathan, P. M. Raj, R. Tummala and G. An-
tonini, “Enabling antenna design with nano-magnetic materials using
tions usually lead to worthless results that can cost a machine learning,” 2015 IEEE Nanotechnology Materials and Devices
lot of time. It is necessary to know what area of the Conference (NMDC), pp. 1–5, Anchorage, AK, 2015.
problem is best to spend time working on. [15] L. Tenuti, G. Oliveri, D. Bresciani and A. Massa, “Advanced learning-
based approaches for reflectarrays design,” 2017 11th European Confer-
3) Getting enough data: Some data can be hard to find ence on Antennas and Propagation (EUCAP), pp. 84–87, Paris, 2017.
or obtain. In antenna design, multiple simulations are [16] X. Wang, P. Wang and X. Wang, “Adaptive Sparse Array Reconfigura-
needed in order to obtain a training set. tion based on Machine Learning Algorithms,” 2018 IEEE International
Conference on Acoustics, Speech and Signal Processing (ICASSP),
4) Pre-processing of data: To ensure that the learning pp. 1159–1163, Calgary, AB, 2018.
algorithm performs adequately, multiple steps need to [17] J. Joung, “Machine Learning-Based Antenna Selection in Wireless Com-
be performed on the data, such as data cleaning, nor- munications,” IEEE Communications Letters, vol. 20, no. 11, pp. 2241–
2244, Nov. 2016.
malization and feature selection, which would cost time [18] Z. M. Fadlullah and M. M. Fouda and N. Kato and X. Shen and Y.
in case of very large datasets. Nozaki, “An early warning system against malicious activities for smart
5) Debugging the algorithm: Knowing what to do next can grid communications,” IEEE Network, vol. 25, no. 5, pp. 50–55, 2011.
[19] B. A. Sonkamble and D. D. Doye, “Use of Support Vector Machines
also be challenging. When problems, such as high bias through Linear-Polynomial (LP) Kernel for Speech Recognition,” Inter-
and high variance occur, it is crucial to know what national Conference on Advances in Mobile Network, Communication
steps to take. This requires following some diagnosis and Its Applications, pp. 46–49, 2012.
[20] C. Perera and A. Zaslavsky and P. Christen and D. Georgakopoulos,
techniques such as plotting the learning curves. “Context Aware Computing for The Internet of Things: A Survey,” IEEE
Communications Surveys Tutorials, vol. 16, no. 1, pp. 414–454, 2014.
V. C ONCLUSION [21] S. S. W. Samek and T. Wiegand, “The convergence of machine learning
This paper presented an overview on machine learning with and communications,” CoRR, vol. abs/1708.08299, 2017.
[22] J. Huang and J. Antonio Encinar, Reflectarray Antennas. Wiley-IEEE
an investigation of its concepts, differentiation with artificial Press, 2007.
intelligence and deep learning, its different learning algo- [23] D. R. Prado, M. Arrebola, M. R. Pino, R. Florencio, R. R. Boix, J. A.
rithms and techniques. An extensive investigation is presented Encinar and F. Las-Heras, “Efficient Crosspolar Optimization of Shaped-
Beam Dual-Polarized Reflectarrays Using Full- Wave Analysis for the
on the usage of machine learning in antenna design, for Antenna Element Characterization,” IEEE Transactions on Antennas and
which its advantages with respect to the traditional design Propagation, vol. 65, no. 2, pp. 623–635, 2017.

606
[24] P. Robustillo, J. Zapata, J. A. Encinar and J. Rubio, “ANN Charac-
terization of Multi-Layer Reflectarray Elements for Contoured-Beam
Space Antennas in the Ku-Band,” IEEE Transactions on Antennas and
Propagation, vol. 60, no. 7, pp. 3205–3214, 2012.
[25] V. Richard R., Loison, R. Gillard, H. Legay and M. Romier, “Loss
analysis of a reflectarray cell using ANNs with accurate magnitude pre-
diction,” 2017 11th European Conference on Antennas and Propagation
(EUCAP), pp. 2396–2399, 2017.
[26] D. R. Prado, J. A. López-Fernández, M. Arrebola and G. Goussetis,
“Efficient Shaped-Beam Reflectarray Design Using Machine Learning
Techniques,” 2018 15th European Radar Conference (EuRAD), pp. 525–
528, Madrid, 2018.
[27] R. Florencio, R. R. Boix and J. A. Encinar, “Enhanced MoM Analysis
of the Scattering by Periodic Strip Gratings in Multilayered Substrates,”
IEEE Transactions on Antennas and Propagation, vol. 61, no. 10,
pp. 5088–5099, 2013.
[28] N. T. Tokan and F. Gune, “Support Vector Characterization of the
Microstrip Antennas Based on Measurements,” Progress In Electromag-
netics Research B, vol. 5, pp. 49–61, 2008.
[29] C. Gianfagna, H. Yu, M. Swaminathan, R. Pulugurtha, R. Tum-
mala, and G. Antonini, “Machine-Learning Approach for Design
of Nanomagnetic-Based Antennas,” Journal of Electronic Materials,
vol. 46, no. 8, pp. 4963–4975, 2017.
[30] J. Tak, A. Kantemur, Y. Sharma and H. Xin, “A 3-D-PrintedW-Band
Slotted Waveguide Array Antenna Optimized Using Machine Learning,”
IEEE Antennas and Wireless Propagation Letters, vol. 17, no. 11,
pp. 2008–2012, Nov. 2018.
[31] A. Hoorfar, “Evolutionary Programming in Electromagnetic Optimiza-
tion: A Review,” IEEE Transactions on Antennas and Propagation,
vol. 55, no. 3, pp. 523–537, March 2007.
[32] J. M. Johnson and Y. Rahmat-Samii, “Genetic algorithms and method
of moments (GA/MOM) for the design of integrated antennas,” IEEE
Transactions on Antennas and Propagation, vol. 47, no. 10, pp. 1606–
1614, Oct. 1999.
[33] M. D. Gregory, Z. Bayraktar and D. H. Werner, “Fast Optimization
of Electromagnetic Design Problems Using the Covariance Matrix
Adaptation Evolutionary Strategy,” IEEE Transactions on Antennas and
Propagation, vol. 59, no. 4, pp. 1275–1285, April 2011.
[34] J. Zhang, Z. Zhan, Y. Lin, N. Chen, Y. Gong, J. Zhong, H. S. H. Chung,
Y. Li and Y. Shi, “Evolutionary Computation Meets Machine Learning:
A Survey,” IEEE Computational Intelligence Magazine, vol. 6, no. 4,
pp. 68–75, Nov. 2011.
[35] S. K. Jain, “Bandwidth Enhancement Of Patch Antennas Using Neu-
ral Network Dependent Modified Optimizer,” International Journal of
Microwave and Wireless Technologies, vol. 8, no. 7, pp. 1111–1119,
2016.
[36] X. H. Chen, X. X. Guo, J. M. Pei and W. Y. Man, “A hybrid algorithm of
differential evolution and machine learning for electromagnetic structure
optimization,” 2017 32nd Youth Academic Annual Conference of Chinese
Association of Automation (YAC), pp. 755–759, Hefei, 2017.
[37] M. D. Gregory and Z. Bayraktar and D. H. Werner, “Fast Optimization
of Electromagnetic Design Problems Using the Covariance Matrix
Adaptation Evolutionary Strategy,” IEEE Transactions on Antennas and
Propagation, vol. 59, no. 4, pp. 1275–1285, 2011.
[38] Z. Bayraktar and M. Komurcu and J. A. Bossard and D. H. Werner,
“The Wind Driven Optimization Technique and its Application in
Electromagnetics,” IEEE Transactions on Antennas and Propagation,
vol. 61, no. 5, pp. 2745–2757, 2013.
[39] C. R. Silva and S. R. Martins, “An Adaptive Evolutionary Algorithm
for UWB Microstrip Antennas Optimization Using a Machine Learning
Technique.,” Microw. Opt. Technol. Lett., vol. 55, no. 8, pp. 1864–1868,
2013.
[40] B. Liu, H. Aliakbarian, Z. Ma, G. A. E. Vandenbosch, G. Gielen and P.
Excell, “An Efficient Method for Antenna Design Optimization Based
on Evolutionary Computation and Machine Learning Techniques,” IEEE
Transactions on Antennas and Propagation, vol. 62, no. 1, pp. 7–18, Jan.
2014.
[41] Saifullah and B. Ahmed, “Machine learning for isotropic antenna
design,” 2018 22nd International Microwave and Radar Conference
(MIKON), pp. 683–687, Poznan, 2018.
[42] N. T. Tokan and F. Gunes, “Support vector design of the microstrip
antenna,” 2008 IEEE 16th Signal Processing, Communication and
Applications Conference, pp. 1–4, Aydin, 2008.

607

View publication stats