Synthesis of new antenna arrays with arbitrary geometries based on the superformula

International Journal of Electrical and Computer Engineering (IJECE)
Vol. 12, No. 6, December 2022, pp. 6228~6238
ISSN: 2088-8708, DOI: 10.11591/ijece.v12i6.pp6228-6238  6228
Journal homepage: http://ijece.iaescore.com
Synthesis of new antenna arrays with arbitrary geometries
based on the superformula
Anas A. Amaireh1
, Nihad I. Dib2
, Asem S. Al-Zoubi3
1
Electrical and Computer Engineering Department, The University of Oklahoma, Norman, United State
2
Electrical Engineering Department, Jordan University of Science and Technology, Irbid, Jordan
3
Telecommunications Engineering Department, Yarmouk University, Irbid, Jordan
Article Info ABSTRACT
Article history:
Received Sep 20, 2021
Revised Jun 7, 2022
Accepted Jul 5, 2022
The synthesis of antenna arrays with low sidelobe levels is needed to
enhance the communication systems’ efficiency. In this paper, new arbitrary
geometries that improve the ability of the antenna arrays to minimize the
sidelobe level, are proposed. We employ the well-known superformula
equation in the antenna arrays field by implementing the equation in the
general array factor equation. Three metaheuristic optimization algorithms
are used to synthesize the antenna arrays and their geometries; antlion
optimization (ALO) algorithm, grasshopper optimization algorithm (GOA),
and a new hybrid algorithm based on ALO and GOA. All the proposed
algorithms are high-performance computational methods, which proved their
efficiency for solving different real-world optimization problems. 15 design
examples are presented and compared to prove validity with the most
general standard geometry: elliptical antenna array (EAA). It is observed
that the proposed geometries outperform EAA geometries by 4.5 dB and
10.9 dB in the worst and best scenarios, respectively, which proves the
advantage and superiority of our approach.
Keywords:
Antenna arrays
Antlion optimization algorithm
Grasshopper optimization
algorithm
Metaheuristics algorithms
Superformula equation
This is an open access article under the CC BY-SA license.
Corresponding Author:
Anas A. Amaireh
Electrical and Computer Engineering Department, The University of Oklahoma
Norman 73019, United State
Email: anas@ou.edu
1. INTRODUCTION
Several standard geometries like; line, circle, and ellipse, have been widely used in the synthesis of
different antenna array designs in the literature, due to the availability and the ease of implementation of their
parametric equations. These shapes have been implied in hundreds of articles to design antenna arrays, such
as; linear antenna arrays (LAA) [1]–[6], circular antenna arrays (CAA) [7]–[12], and elliptical antenna arrays
(EAA) [13]–[18], and to achieve different objectives like minimizing the maximum sidelobe level or
increasing the directivity of the array. Results’ variation of diverse geometries proves the significant role of
the array’s shape in determining the desired objective in the antenna arrays. Such observation is the main
inspiration to think of creating new optimal geometries that will improve antenna array characteristics.
In 2003, Gielis [19] proposed a new geometrical equation, called the superformula. This geometrical
approach has been used to model and understand numerous abstract, natural, and man-made shapes [19].
Superformula equation can describe different shapes in nature, which can be achieved by modifying the
parameters of the equation that generate various natural polygons [19]. In the literature, the superformula
equation has been implemented, before, in antenna designs [20]–[22].
In this paper, our goal is to minimize the sidelobe level in a new arbitrary antenna array radiation
pattern using three optimization algorithms: antlion optimization (ALO) [23], grasshopper optimization

Int J Elec & Comp Eng ISSN: 2088-8708 
Synthesis of new antenna arrays with arbitrary geometries based on the superformula (Anas A. Amaireh)
6229
algorithm (GOA) [24], and a new hybrid optimization algorithm based on ALO and GOA [11]. The rest of
the paper is organized. A brief overview of our new proposed hybrid algorithm is presented in section 2. The
description of the superformula equation and its implementation in the array factor equation are discussed in
section 3. The fitness function and numerical results are detailed in sections 4 and 5, respectively. Finally,
section 6 presents the conclusions of the paper.
2. THE NEW PROPOSED HYBRID ALGORITHM
In study [11], authors proposed a new hybrid algorithm based on two evolutionary algorithms; ALO
algorithm [23] and GOA [24]. The main inspiration behind proposing such hybridization is combining the
characteristics and overcoming the drawbacks of ALO and GOA. The main characteristic of ALO is the
robustness in exploiting the global optima, which has been verified in many articles in the literature
[2], [13], [25], [26]. On the other hand, the social forces in the grasshopper’s swarm show a strong ability to
effectively explore the whole search space in GOA [27]–[29]. Thus, these features give the idea to combine
ALO and GOA in a new hybrid algorithm, which makes a big improvement in their performance.
Moreover, the benefits of hybridization can be shown in overcoming the drawbacks of ALO and
GOA. The usage of the roulette wheel selection method is the main drawback in ALO, since it may cause:
early convergence, loss of diversity, and not enough pressure to select the fittest search agents among the
same fitness search agents. While the disadvantage of GOA exists in the c parameter equation, which
weakens the exploitation process in the algorithm.
Our proposed hybrid algorithm has the ability to efficiently explore and exploit the search space to
reach the global optimum, due to merging the characteristics of both ALO and GOA. The hybrid algorithm
has some factors that enhance the capability of exploration and exploitation processes like the population
nature that reduces local optima stagnation, the repulsion and attraction forces in GOA algorithm, the random
walks and roulette wheel selection method in ALO algorithm, choosing diverse samples of average and less
fitness search agents from next position’s matrix, and the modifications of c parameter in GOA algorithm.
These factors and modifications lead to better diversity of the search agents all over the search space and a
high probability for local optima stagnation avoidance. Moreover, the intensity of search agents in the
proposed algorithm has been decreased rapidly compared with ALO and GOA, due to the modification of the
c parameter and its combination with other shrinking factors. Therefore, the hybrid algorithm guarantees fast
and mature convergence compared with ALO and GOA.
The equation (1) has been used in our proposed algorithms:
𝑋𝑖
𝑑
= 𝑐 (∑ 𝑐
𝑢𝑏𝑑−𝑙𝑏𝑑
2
𝑠(|𝑋𝑗
𝑑
− 𝑋𝑖
𝑑
|)
𝑁
𝑗=1
𝑗≠𝑖
𝑋𝑗 − 𝑋𝑖
𝑑𝑖𝑗
) (1)
where 𝑋𝑖
𝑑
defines the next position for 𝑖𝑡ℎ
grasshopper and 𝑑𝑡ℎ
dimension, 𝑢𝑏𝑑 and 𝑙𝑏𝑑 are the upper and
lower bounds in the dth
dimension, respectively, and 𝑠(𝑟) = 𝑓𝑒
−𝑟
𝑙 − 𝑒−𝑟
where f represents the intensity of
attraction force and l indicates the attractive length scale.
𝑖𝑓 𝑙 ≤ (𝐿)/2
𝑐 = 𝑐𝑚𝑎𝑥 − 𝑙
𝑐𝑚𝑎𝑥−𝑐𝑚𝑖𝑛
(
𝐿
2
)
(2)
𝑖𝑓 𝑙 >
𝐿
2
𝑐 = (𝐿 − 𝑙 + 1) × (
𝑐𝑚𝑖𝑛
𝐿×10𝑙) (3)
Where 𝑐𝑚𝑎𝑥 and 𝑐𝑚𝑖𝑛 are the maximum and minimum values, respectively, L indicates the maximum
number of iterations, and l is the current iteration [24].
𝑐𝑡
=
𝐶𝑡
𝐼
(4)
𝑑𝑡
=
𝑑𝑡
𝐼
(5)

 ISSN: 2088-8708
Int J Elec & Comp Eng, Vol. 12, No. 6, December 2022: 6228-6238
6230
Where 𝑐𝑡
and 𝑑𝑡
indicate the minimum and maximum of all variables at tth
iteration, respectively, and 𝐼
represents a ratio, such that 𝐼 = 10𝑤 𝑡
𝑇
where t represents the current iteration, T indicates the maximum
number of iterations, and w is a constant that depends on the current iteration [23].
3. SUPERFORMULA AND ARRAY FACTOR EQUATIONS
The two-dimensional superformula representation in polar coordinates is given [19]:
𝜌(𝜑) =
1
{[((|
1
𝑎
cos(𝜑
𝑚1
4
)|)
𝑛2
+(|
1
𝑏
sin(𝜑
𝑚2
4
)|)
𝑛3
)]
1 𝑛1
⁄
}
(6)
where 𝑛1, 𝑛2, 𝑛3, 𝑚1 and 𝑚2 are real numbers. a and b are real numbers, excluding zero, that represent the
major and minor axes, respectively. Variables 𝑚1 and 𝑚2 are responsible for increasing the rotational
symmetry of the shape, while 𝑛2 and 𝑛3 determine whether the shape is inscribed or circumscribed within the
unit circle [19].
The effect of adjusting superformula parameters is shown in Figure 1, which mentions 6 different
shapes generated using the superformula equation with their specific parameters’ values. Figure 1(a)
represents generating an ellipse by tuning the values of 𝑚1, 𝑚2, 𝑛1, 𝑛2, 𝑛3, a, and b into 4, 4, 2, 2, 2, 0.9, and
0.5, respectively, while the equality between a and b, with keeping all other parameters the same, would give
a circle as shown in Figure 1(b). An equisetum stem and square shapes are shown in Figures 1(c) and 1(d)
when the superformula parameters are changed into (7, 6, 6, 10, 0.9, 0.9), and (4, 100, 100, 100, 0.9, 0.9),
respectively. Figure 1(e) represents a starfish with the following superformula parameters values
(5, 1.7, 1.7, 0.1, 0.9, 0.9), and to achieve a rotational shape as Figure 1(f), the parameters must be tuned as
(13/6, 0.3, 0.3, 0.3, 0.5, 0.5).
(a) (b) (c)
(c) (d) (f)
Figure 1. Shapes generated by superformula equation. The numbers inside the brackets refer to (𝑚 (𝑚1=𝑚2),
𝑛1, 𝑛2, 𝑛3, a, b). (a) ellipse (4, 2, 2, 2, 0.9, 0.5), (b) circle (4, 2, 2, 2, 0.6, 0.6), (c) equisetum stem (7, 6, 6, 10,
0.9, 0.9), (d) square (4, 100, 100, 100, 0.9, 0.9), (e) starfish (5, 1.7, 1.7, 0.1, 0.9, 0.9), and (f) rotational shape
(13/6, 0.3, 0.3, 0.3, 0.5, 0.5)
Based on [30], the general array factor for any antenna array can be described as (7):
𝐴𝐹(𝜃, 𝜑) = ∑ 𝐼𝑛𝑒𝑗(𝛼𝑛+𝐾𝜌𝑛.𝑎
̂𝑟)
𝑁
𝑛=1 (7)
Unit

6231
where N represents the number of elements, which are assumed to lie on the XY-plane, 𝐼𝑛 and 𝛼𝑛 are the
excitation current and phase for 𝑛𝑡ℎ
element, respectively. The wavenumber 𝐾 =
2𝜋
𝜆
, where 𝜆 is the
wavelength, 𝜌𝑛
represents the position vector of the 𝑛𝑡ℎ
element:
𝜌𝑛
= 𝑥𝑛𝑎
̂𝑥 + 𝑦𝑛𝑎
̂𝑦 (8)
and 𝑎
̂𝑟 is the unit vector for the observation point, which can be written as (9).
𝑎
̂𝑟 = sin 𝜃 cos 𝜑 𝑎
̂𝑥 + sin 𝜃 sin 𝜑 𝑎
̂𝑦 + cos 𝜃 𝑎
̂𝑧 (9)
The x and y coordinates of (6) are presented:
𝑥𝑛 = {[((|
1
𝑎
𝑐𝑜𝑠 (𝜑𝑛
𝑚1
4
)|)
𝑛2
+ (|
1
𝑏
𝑠𝑖𝑛 (𝜑𝑛
𝑚2
4
)|)
𝑛3
)]
−1 𝑛1
⁄
} 𝑐𝑜𝑠 𝜑𝑛
𝑦𝑛 = {[((|
1
𝑎
𝑚1
4
)|)
𝑛2
+ (|
1
𝑏
𝑚2
4
)|)
𝑛3
)]
−1 𝑛1
⁄
} 𝑠𝑖𝑛 𝜑𝑛 (10)
where 𝜑𝑛 is the angular position of the antenna elements. For uniform array distribution, 𝜑𝑛 is given:
𝜑𝑛 = 2𝜋
(𝑛−1)
𝑁
(11)
using (8)-(10), the array factor expression can be written:
𝐴𝐹(𝜃, 𝜑) = ∑ 𝐼𝑛𝑒𝑥𝑝 {𝑗𝛼𝑛 + 𝑗𝐾 ({[((|
1
𝑎
𝑚1
4
)|)
𝑛2
+
𝑁
𝑛=1
(|
1
𝑏
𝑚2
4
)|)
𝑛3
)]
−1 𝑛1
⁄
} cos 𝜑𝑛 sin 𝜃 cos 𝜑 + {[((|
1
𝑎
𝑚1
4
)|)
𝑛2
+
(|
1
𝑏
𝑚2
4
)|)
𝑛3
)]
−1 𝑛1
⁄
} sin 𝜑𝑛 sin 𝜃 sin 𝜑)} (12)
where 𝜃 represents the elevation angle, which is measured from the positive z-axis. 𝜃 is assumed to equal
90°
, since the array factor in the x-y plane will be of interest. It has been assumed, in this paper, that the main
lobe is directed along the positive x-axis, such that; 𝜃𝑜=90°
and 𝜑𝑜 = 0°
. To achieve this, the excitation
phase is assumed to be as (13).
𝛼𝑛 = − 𝐾(𝑥𝑛 sin 𝜃𝑜 cos 𝜑𝑜 + 𝑦𝑛 sin 𝜃𝑜 sin 𝜑𝑜) (13)
4. FITNESS FUNCTION
The main objective of this paper is to minimize the maximum sidelobe level in the array factor of
the proposed antenna arrays. In order to accomplish this, the following fitness function is used [11]:
𝐹𝑖𝑡𝑛𝑒𝑠𝑠 = (𝑊1𝐹1 + 𝑊2𝐹2)/|𝐴𝐹𝑚𝑎𝑥|2
(14)
𝐹1 = |𝐴𝐹(𝜑𝑛𝑢1)|2
+ |𝐴𝐹(𝜑𝑛𝑢2)|2
(15)
𝐹2 = max { |𝐴𝐹(𝜑𝑚𝑠1)|2
, |𝐴𝐹(𝜑𝑚𝑠2)|2} (16)
𝜑𝑛𝑢1 and 𝜑𝑛𝑢2 represent the angles that define the first null beamwidth, FNBW=𝜑𝑛𝑢2 − 𝜑𝑛𝑢1 = 2𝜑𝑛𝑢2.
Furthermore, 𝜑𝑚𝑠1 and 𝜑𝑚𝑠2 are the angles from -180o
to 𝜑𝑛𝑢1, and from 𝜑𝑛𝑢2 to 180o
, respectively, in
which the maximum side lobe level (SLL) is accomplished during the optimization process. The function 𝐹1
minimizes the array factor at 𝜑𝑛𝑢1 and 𝜑𝑛𝑢2 to define the major lobe in the radiation pattern. 𝐹2 minimizes
the radiation pattern in the sidelobe region around the major lobe. 𝐴𝐹𝑚𝑎𝑥 is the maximum value of the array
factor at (𝜃𝑜, 𝜑𝑜). 𝑊1 and 𝑊2 are weighting factors. The purpose of the optimization in this paper is to obtain
new creative geometries for antenna arrays, toward achieving the lowest maximum SLL. Therefore, we are
going to optimize the superformula parameters (𝑛1, 𝑛2, 𝑛3, 𝑚1, 𝑚2, a, b) to get the most suitable geometry
for antenna elements distribution.

 ISSN: 2088-8708
6232
5. NUMERICAL RESULTS
In this section, three different subsections are discussed; 8-element, 12-element, and 20-element.
Three different cases have been optimized using ALO, GOA, and the hybrid algorithm. The first case
discusses creating new geometries by optimizing superformula parameters only (𝑚1, 𝑚2, 𝑛1, 𝑛2, 𝑛3, a, b),
assuming that all elements have unity excitation currents and uniform angular position distribution depending
on (11). So, all the obtained results, in this case, will be compared with uniform EAA results [13]. The
second case discusses the optimization of array elements’ excitation currents along with optimizing the
superformula parameters. In this case, uniform distribution for the antenna elements’ positions has been
assumed depending on (11). The results for this case will be compared with the results of optimizing the
excitation currents in EAA [13]. The third case presents the optimization of the elements’ angular positions
and superformula parameters, assuming unity amplitude currents for all elements. This case’s results are
compared with the corresponding angular positions optimization examples in EAA [13]. It is worth
mentioning, that the EAA has been chosen for the comparison over the CAA due to the generality of the EAA.
All the results in this paper have been optimized through 20 independent runs using 1,000 iterations
and 50 search agents. Three parameters, four parameters, five parameters, and seven parameters of the
superformula equation are optimized for most of the examples. In three parameters optimization, the
parameters: 𝑚, 𝑛2 and 𝑛3 are optimized, assuming that 𝑚1=𝑚2 and 𝑛1=2 (since the ellipse has these
assumptions). While in four parameters optimization; the parameters: 𝑚1, 𝑛1, 𝑛2 and 𝑛3 are optimized, with
𝑚1=𝑚2. In five parameters optimization, all superformula parameters are optimized except a and b. While in
seven parameters optimization, all superformula parameters are optimized. Note that, the major and minor
axes (a and b) have been assumed to equal 0.5 and 0.433 for 8-element antenna array examples, 1.15 and
0.9959 for 12-element examples, and 1.6 and 1.3856 for 20-element examples. Moreover, the range of
optimized parameters is assumed as [1, 50] for 𝑚1, 𝑚2 and 𝑛1, and [-50, 50] for 𝑛2 and 𝑛3. It is worth
mentioning that all distances are in terms of 𝜆.
5.1. Optimizing 8 elements
5.1.1. Optimizing superformula parameters
In this example, an 8-element antenna array with unity amplitudes and uniform angular positions is
optimized to suppress the maximum SLL of the array factor. This can be achieved by optimizing different
parameters in the superformula equation. So, the aim here is to optimize the geometry of the 8-element array, to
get a better maximum SLL compared with standard geometries like EAA. Table 1 shows the optimum values of
superformula parameters and their corresponding maximum SLL, compared with uniform EAA. According to
the obtained results, optimizing 4-parameters, the algorithms ALO, GOA, and the hybrid outperform the
uniform EAA shape by almost 11 dB. In other words, without any change in the excitation currents or the
angular positions, the maximum SLL is improved significantly. Figure 2(a) shows the radiation patterns for
optimizing three and four superformula (S.F.) parameters based on ALO, the hybrid and GOA algorithms
compared with uniform EAA. Figure 2(b) presents the convergence curves for the proposed methods, which
shows that they reach the global optima with a small number of iterations. Box-and-whisker plot is represented
in Figure 2(c), which proves the stability of the hybrid method and ALO over 20 independent runs. Figure 3
represents elements distributions along three-parameter optimized superformula shape based on ALO,
three-parameter optimized based on the hybrid method, four-parameter optimized based on ALO,
four-parameter optimized based on GOA, and four-parameter optimized based on the hybrid method.
5.1.2. Optimizing elements amplitudes and superformula parameters
The optimization of excitation currents and superformula parameters is discussed in this example.
The optimal values of 8 elements currents and three, four, and seven superformula parameters, based on the
hybrid method, compared with 8-element EAA results are tabulated in Table 2. Figure 4(a) illustrates the
array factors depending on the results in Table 2. It can be noticed that generating new creative shapes gives
maximum SLL less than the best results of the optimized EAA by almost 6 dB, which confirms the
significance of generating new geometries of antenna arrays. Figure 4(b) shows the distribution of the
8 elements along the optimized superformula shapes.
Table 1. Optimum values of superformula parameters for 8-element optimization
Method and optimized parameters 𝜑𝑛𝑢2 = 51° [𝑚1, 𝑛2, 𝑚2, 𝑛3,𝑛1, a, b] Max. SLL (dB)
Three-parameter (ALO) [24.0052, 2.82534, 24.0052, 1.59748, 2, 0.5, 0.4330] -15.25
Three-parameter (Hybrid) [24.0099, 1.17679, 24.0099, 3.85461, 2, 0.5, 0.4330] -16.66
Four-parameter (ALO) [36, 44.071, 36, 41.702, 30.3222, 0.5, 0.4330] -17.95
Four-parameter (Hybrid) [12, 18.4426, 12, 17.555, 12.8075, 0.5, 0.4330] -17.94
Four-parameter (GOA) [44.0000, 39.1136, 44.0000, 36.4979, 26.7506, 2, 0.5, 0.4330] -17.66
EAA (Uniform) [4, 2, 4, 2, 2, 0.5, 0.4330] -7.76

6233
(a) (b)
(c)
Figure 2. Results for optimizing superformula parameters for 8-element antenna array (a) radiation patterns
for 8-element optimization, (b) convergence curves for 8-element optimization, and (c) box-and-whisker
plots for 8-element optimization in 20 runs
Figure 3. Distribution of 8-element array over optimized shapes based on ALO, GOA, and the hybrid method

 ISSN: 2088-8708
6234
Table 2. Optimum values of excitation currents and superformula parameters for 8-element optimization
N=8
𝜑𝑛𝑢2 = 51°
[𝑚1, 𝑛2, 𝑚2, 𝑛3, 𝑛1, a, b]
[ I1, I2, I3, …, I8]
Max. SLL (dB)
Three-parameter (Hybrid) [7.8031, 2.6905, 7.8031, 1.558, 2, 0.5, 0.4330]
[0.6223, 0.9365, 0.2873, 1.000, 0.6537, 0.9806, 0.2981, 0.9183]
-19.10
Four-parameter (Hybrid) [15.9703, 24.8769, 15.9703, -1.98988, 21.2323, 0.5, 0.4330]
[1.000, 0.7896, 0.7519, 0.8200, 0.0060, 0.8380, 0.7196, 0.8077]
-20.71
Seven-parameter (Hybrid) [15.9746, 15.8241, 20.0043, -5.17621, 20.5931, 0.35375, 0.564119]
[1.000, 0.7004, 0.6741, 0.8938, 0.0903, 0.8358, 0.7965, 0.6413]
-20.93
EAA (ALO) [4, 2, 4, 2, 2, 0.5, 0.4330]
[0.5379, 0.9011, 0.0667, 0.9246, 0.5856, 1.0000, 0.0249, 0.9764]
-14.64
EAA (Hybrid) [4, 2, 4, 2, 2, 0.5, 0.4330]
[0.5666, 1.0000, 0.0516, 0.9700, 0.5609, 0.9586, 0.0690, 0.9887]
-14.60
(a)
(b)
Figure 4. Results for optimizing excitation amplitudes and superformula parameters for 8-element antenna
array (a) radiation patterns for 8-element array and (b) distribution of 8-element array over S.F. optimized
shapes based on the hybrid method
5.2.1. Optimizing angular positions and superformula parameters
S.F. parameters and the angular positions for 12-element antenna array are going to be optimized in
this example. Five and seven parameters of the SF equation are optimized with 12 elements angular positions
based on the hybrid algorithm as mentioned in Table 3. The maximum SLL value obtained by optimizing the
seven-parameter is -16.12 dB, and -15.44 dB for optimizing the five-parameter. It is obvious that using the
SF enhances the maximum SLL by almost 6 dB compared with EAA results. Figure 5(a) shows the radiation
patterns for optimizing five-parameter, seven-parameter, and EAAs based on ALO, GOA, and the hybrid
method. Figure 5(b) represents 12 elements distributions for optimizing five-parameter and seven-parameter SF.

6235
Table 3. Optimum values of angular positions and S.F. parameters for 12-element optimization
N=12
𝜑𝑛𝑢 =22°
[𝑚1, 𝑛2, 𝑚2, 𝑛3,𝑛1, a, b]
[∅1, ∅2, ∅3, …, ∅12] in degrees
Max. SLL (dB)
Five-parameter
(Hybrid)
[4.11638, 23.5199, 4.63125, 24.5776, 24.7026, 1.15, 0.9959]
[1.8360, 29.2719, 73.0149, 132.5429, 152.8216, 175.9628, 195.9229, 254.4374, 281.2805,
316.2005, 325.4463, 345.7353]
-15.44
Seven-
parameter
(Hybrid)
[9.25799, -4.91652, 10.0473, -9.52872, 12.3331, 0.356268, 0.66011]
[7.0744, 86.9813, 100.8709, 115.5467, 136.2431, 169.4018, 202.4140, 209.2568,
227.7020, 234.5620, 248.9834, 254.8533]
-16.12
EAA (ALO) [4, 2, 4, 2, 2, 1.15, 0.9959]
[13.5053, 39.4896, 85.4927, 95.4538, 141.8509, 167.4591, 188.0407, 210.0000, 240.0111,
299.8681, 330.0000, 351.8085]
-9.52
EAA (GOA) [4, 2, 4, 2, 2, 1.15, 0.9959]
[9.2464, 43.3007, 60.0002, 93.9722, 139.9140, 164.3480, 189.0621, 210.0000, 269.7800,
299.1580, 330.0000, 356.2058]
-9.27
EAA (Hybrid) [4, 2, 4, 2, 2, 1.15, 0.9959]
[7.2016, 42.3489, 60.1202, 92.8993, 140.9393, 166.8977, 191.8173, 210.0000, 268.6000,
300.0000, 330.0000, 355.1231]
-9.42
(a) (b)
Figure 5. Results of optimizing angular positions and S.F. parameters for 12-element antenna array
(a) radiation patterns for 12-element arrays and (b) distribution of 12-element
array over S.F. optimized shapes
5.3.1. Optimizing superformula parameters
In this example, the number of array elements is increased to 20. So, the SF parameters are
optimized to generate the optimal geometry of a 20-element array that has minimum suppression for SLL.
Table 4 shows the optimum values for four and seven parameters using our proposed hybrid method, ALO,
and GOA. The best maximum SLL is obtained by optimizing the seven-parameter using the hybrid
algorithm, with max SLL=-16.01 dB, which is 10 dB less than the results of uniform EAA.
Figure 6(a) shows the radiation patterns of the 20-element array for different SF parameters
optimization. Figures 6(b) and 6(c) show convergence curves and box and whisker plots for ALO, GOA, and
the hybrid method, respectively. The uniform distribution of the obtained shapes for optimizing:
four-parameter based on ALO, four-parameter based on the hybrid method, seven-parameter based on ALO,
seven-parameter based on GOA, and seven-parameter based on the hybrid algorithm, are illustrated in
Figure 7. These results, again, prove that optimizing array shapes will significantly improve the max SLL
without changing the excitation currents or angular positions.
Table 4. Optimum values of S.F. parameters for 20-element optimization
𝜑𝑛𝑢2 = 16° [𝑚1, 𝑛2, 𝑚2, 𝑛3,𝑛1, a, b] Max. SLL (dB)
Four-parameter (ALO) [44, 5.63766, 44, 46.3647, 14.9658, 1.6, 1.3856] -14.49
Four-parameter (Hybrid) [36, 5.8958, 36, 49.1193, 15.6256, 1.6, 1.3856] -14.58
Seven-parameter (ALO) [24.0644, 13.1663, 18.3572, 21.9164, 21.0494, 0.931817, 1.76903] -15.78
Seven-parameter (GOA) [21.3997, 8.5576, 2.6021, 11.1608, 24.2988, 1.5287, 1.2400] -12.78
Seven-parameter (Hybrid) [20.0259, 10.76, 1.99925, -8.7371, 14.3647, 1.29043, 0.233034] -16.01
EAA (Uniform) [4, 2, 4, 2, 2, 1.6, 1.3856] -6.88

 ISSN: 2088-8708
6236
(a) (b)
(c)
Figure 6. Results of optimizing S.F. parameters for 20-element antenna array (a) radiation patterns of
optimizing 20-element array, (b) convergence curves for 20-element optimization, and (c) box and whisker
plot for 20-element optimization
Figure 7. Distribution of 20-element array over S.F. optimized shapes based on ALO, GOA,
and the hybrid method
Fitness
function
value

6237
6. CONCLUSION
In this paper, new arbitrary geometries for antenna arrays were proposed. Three evolutionary
algorithms, ALO, GOA, and the hybrid algorithm based on ALO and GOA, were used in the synthesis of the
arbitrary antenna arrays. The objective function was to reduce the maximum sidelobe level with the
constraint of a fixed major lobe beamwidth. Three cases were investigated in this paper; 8-element,
12-element, and 20-element antenna array geometries. Additionally, three different examples were
illustrated; optimizing only the superformula parameters, optimizing superformula parameters and excitation
currents of the antenna array, and optimizing superformula parameters with the angular position of the
antenna array. The results of all cases and examples prove the capability of our geometries to outperform the
standard geometries with a huge difference.
ACKNOWLEDGEMENTS
This research did not receive any specific grant from funding agencies in the public, commercial, or
not-for-profit sectors. The authors declare no conflict of interest in this research work.
REFERENCES
[1] B. Basu and G. K. Mahanti, “Fire fly and artificial bees colony algorithm for synthesis of scanned and broadside linear array
antenna,” Progress In Electromagnetics Research B, vol. 32, pp. 169–190, 2011, doi: 10.2528/PIERB11053108.
[2] A. A. Amaireh, A. Alzoubi, and N. I. Dib, “Design of linear antenna arrays using antlion and grasshopper optimization
algorithms,” in 2017 IEEE Jordan Conference on Applied Electrical Engineering and Computing Technologies (AEECT), Oct.
2017, pp. 1–6, doi: 10.1109/AEECT.2017.8257746.
[3] M. H. Bataineh and J. I. Ababneh, “Synthesis of aperiodic linear phased antenna arrays using particle swarm optimization,”
Electromagnetics, vol. 26, no. 7, pp. 531–541, Oct. 2006, doi: 10.1080/02726340600872948.
[4] S. U. Khan, M. K. A. Rahim, M. Aminu-Baba, and N. A. Murad, “Correction of failure in linear antenna arrays with greedy
sparseness constrained optimization technique,” Plos One, vol. 12, no. 12, Dec. 2017, doi: 10.1371/journal.pone.0189240.
[5] A. A. Amaireh, A. S. Al-Zoubi, and N. I. Dib, “The optimal synthesis of scanned linear antenna arrays,” International Journal of
Electrical and Computer Engineering (IJECE), vol. 10, no. 2, pp. 1477–1484, Apr. 2020, doi: 10.11591/ijece.v10i2.pp1477-1484.
[6] A. S. Al-Zoubi, A. A. Amaireh, and N. I. Dib, “Comparative and comprehensive study of linear antenna arrays’ synthesis,”
International Journal of Electrical and Computer Engineering (IJECE), vol. 12, no. 3, pp. 2645–2654, Jun. 2022, doi:
10.11591/ijece.v12i3.pp2645-2654.
[7] K. Guney and S. Basbug, “A parallel implementation of seeker optimization algorithm for designing circular and concentric
circular antenna arrays,” Applied Soft Computing, vol. 22, pp. 287–296, Sep. 2014, doi: 10.1016/j.asoc.2014.05.026.
[8] H. Guo, L. Cui, X. Zhang, and Y. Jiao, “Synthesis of non-uniform circular antenna arrays with multiple constraints,” International
Journal of Microwave and Wireless Technologies, vol. 10, no. 3, pp. 368–375, Apr. 2018, doi: 10.1017/S1759078717001295.
[9] M. A. Panduro, C. A. Brizuela, L. I. Balderas, and D. A. Acosta, “A comparison of genetic algorithms, particle swarm
optimization and the differential evolution method for the design of scannable circular antenna arrays,” Progress In
Electromagnetics Research B, vol. 13, pp. 171–186, 2009, doi: 10.2528/PIERB09011308.
[10] K. R. Mahmoud, M. I. Eladawy, R. Bansal, S. H. Zainud-Deen, and S. M. M. Ibrahem, “Analysis of uniform circular arrays for
adaptive beamforming applications using particle swarm optimization algorithm,” International Journal of RF and Microwave
Computer-Aided Engineering, vol. 18, no. 1, pp. 42–52, Jan. 2008, doi: 10.1002/mmce.20265.
[11] A. A. Amaireh, A. S. Al-Zoubi, and N. I. Dib, “Sidelobe-level suppression for circular antenna array via new hybrid optimization
algorithm based on antlion and grasshopper optimization algorithms,” Progress In Electromagnetics Research C, vol. 93,
pp. 49–63, 2019, doi: 10.2528%2Fpierc19040909.
[12] P. Civicioglu, “Circular antenna array design by using evolutionary search algorithms,” Progress In Electromagnetics Research
B, vol. 54, pp. 265–284, 2013, doi: 10.2528/PIERB13050112.
[13] N. Dib, A. Amaireh, and A. Al-Zoubi, “On the optimal synthesis of elliptical antenna arrays,” International Journal of
Electronics, vol. 106, no. 1, pp. 121–133, Jan. 2019, doi: 10.1080/00207217.2018.1512658.
[14] R. A. Sadeghzadeh, A. A. L. Neyestanak, M. Naser‐Moghadasi, and M. Ghiamy, “A comparison of various hybrid elliptical
antenna arrays,” Iranian journal of electrical and computer engineering, vol. 7, pp. 98–106, 2008.
[15] K. Guney and A. Durmus, “Elliptical antenna array synthesis using backtracking search optimisation algorithm,” Defence Science
Journal, vol. 66, no. 3, Apr. 2016, doi: 10.14429/dsj.66.9583.
[16] A. A. Amaireh, N. I. Dib, and A. S. Al-Zoubi, “The optimal synthesis of concentric elliptical antenna arrays,” International
Journal of Electronics, vol. 107, no. 3, pp. 461–479, Mar. 2020, doi: 10.1080/00207217.2019.1661028.
[17] R. Bera and J. S. Roy, “Thinning of elliptical and concentric elliptical antenna arrays using particle swarm optimization,”
Microwave Review, pp. 1–7, 2013.
[18] G. G. Lema, G. T. Tesfamariam, and M. I. Mohammed, “A novel elliptical-cylindrical antenna array for radar applications,” IEEE
Transactions on Antennas and Propagation, vol. 64, no. 5, pp. 1681–1688, May 2016, doi: 10.1109/TAP.2016.2539370.
[19] J. Gielis, “A generic geometric transformation that unifies a wide range of natural and abstract shapes,” American Journal of
Botany, vol. 90, no. 3, pp. 333–338, Mar. 2003, doi: 10.3732/ajb.90.3.333.
[20] I. A. Shittu, M. Hussein, and O. Al Aidaros, “A miniaturized surpershape UWB microstrip patch antenna design,” in 2021 IEEE
International Symposium on Antennas and Propagation and USNC-URSI Radio Science Meeting (APS/URSI), Dec. 2021,
pp. 739–740, doi: 10.1109/APS/URSI47566.2021.9703951.
[21] D. F. Mamedes and J. Bornemann, “Gain enhancement of bio-inspired antenna using FSS for 28 GHz 5G application,” in 2021
SBMO/IEEE MTT-S International Microwave and Optoelectronics Conference (IMOC), Oct. 2021, pp. 1–3, doi:
10.1109/IMOC53012.2021.9624742.
[22] G. Martins, P. Pinho, and C. Loss, “Star-shaped supershaped patch antenna for 5G,” in 2021 IEEE International Symposium on
Antennas and Propagation and USNC-URSI Radio Science Meeting (APS/URSI), Dec. 2021, pp. 1913–1914, doi:
10.1109/APS/URSI47566.2021.9704129.

 ISSN: 2088-8708
6238
[23] S. Mirjalili, “The ant lion optimizer,” Advances in Engineering Software, vol. 83, pp. 80–98, May 2015, doi:
10.1016/j.advengsoft.2015.01.010.
[24] S. Saremi, S. Mirjalili, and A. Lewis, “Grasshopper optimisation algorithm: theory and application,” Advances in Engineering
Software, vol. 105, pp. 30–47, Mar. 2017, doi: 10.1016/j.advengsoft.2017.01.004.
[25] E. S. Ali, S. M. Abd Elazim, and A. Y. Abdelaziz, “Ant lion optimization algorithm for renewable distributed generations,”
Energy, vol. 116, pp. 445–458, Dec. 2016, doi: 10.1016/j.energy.2016.09.104.
[26] Z. Zhang, R. Yang, and Y. Fang, “LSTM network based on on antlion optimization and its application in flight trajectory
prediction,” in 2018 2nd IEEE Advanced Information Management,Communicates,Electronic and Automation Control
Conference (IMCEC), May 2018, pp. 1658–1662, doi: 10.1109/IMCEC.2018.8469476.
[27] M. Jamil and S. Mittal, “Hourly load shifting approach for demand side management in smart grid using grasshopper optimisation
algorithm,” IET Generation, Transmission & Distribution, vol. 14, no. 5, pp. 808–815, Mar. 2020, doi: 10.1049/iet-
gtd.2019.0566.
[28] Z. Wu and D. Shen, “Parameter identification of photovoltaic cell model based on improved grasshopper optimization algorithm,”
Optik, vol. 247, p. 167979, Dec. 2021, doi: 10.1016/j.ijleo.2021.167979.
[29] D. Potnuru and A. S. L. V. Tummala, “Implementation of grasshopper optimization algorithm for controlling a BLDC motor
drive,” Soft computing in data analytics, pp. 369–376, 2019, doi: 10.1007/978-981-13-0514-6_37.
[30] W. L. Stutzman and G. A. Thiele, Antenna theory and design, 3rd Editio. Wiley, 2012.
BIOGRAPHIES OF AUTHORS
Anas A. Amaireh is a doctoral student in Electrical and Computer Engineering at
the University of Oklahoma, USA. He obtained his B.Sc. and M.Sc. degrees in Engineering
Technology (major in Wireless Communication Engineering) from Yarmouk University,
Jordan in 2015 and 2018, respectively. His research interests are antenna arrays, numerical
techniques in electromagnetics, metaheuristic algorithms, machine learning, and neural
networks. He can be contacted at email: anas@ou.edu.
Nihad I. Dib obtained his B. Sc. and M.Sc. in Electrical Engineering from Kuwait
University in 1985 and 1987, respectively. He obtained his Ph.D. in EE (major in
Electromagnetics) in 1992 from University of Michigan, Ann Arbor. Then, he worked as an
assistant research scientist in the radiation laboratory at the same school. In Sep. 1995, he
joined the EE department at Jordan University of Science and Technology (JUST) as an
assistant professor and became a full professor in Aug. 2006. His research interests are in
computational electromagnetics, antennas and modeling of planar microwave circuits. He can
be contacted at email: nihad@just.edu.jo.
Asem S. Al-Zoubi received his B.Sc. degree in Electrical Engineering from
Eastern Mediterranean University, Cyprus in 1993, the M.Sc. degree in Electrical Engineering
from Jordan University of Science and Technology, Jordan in 1998, and the Ph.D. degree in
Electrical Engineering/Electromagnetics from the University of Mississippi, USA, in 2008.
Currently, he is an Associate Professor with the Department of Telecommunications
Engineering in Yarmouk University, Jordan. His current research interests include dielectric
resonator antennas, microstrip antennas, and numerical techniques in electromagnetics. Mr.
Al-Zoubi is a member of the IEEE. He can be contacted at email: asem@yu.edu.jo.

Synthesis of new antenna arrays with arbitrary geometries based on the superformula

More Related Content

Similar to Synthesis of new antenna arrays with arbitrary geometries based on the superformula (20)

More from IJECEIAES (20)

Recently uploaded (20)

Synthesis of new antenna arrays with arbitrary geometries based on the superformula