Journal of King Saud University – Engineering Sciences 34 (2022) 544–561

Contents lists available at ScienceDirect

Journal of King Saud University – Engineering Sciences

journal homepage: www.sciencedirect.com

Original article

Optimization of antenna array pattern for uniformly excited rectangular

array via a thinning method
Maryam H. Agha a,⇑, Maan A.S. Al-Adwany b, Oguz Bayat a, Hind Th. Hamdoon b
a
Department of Electrical and Computer Engineering, Altinbas University, Istanbul, Turkey
b
Computer and Information Engineering Department, Ninevah University, Mosul, Iraq

a r t i c l e i n f o a b s t r a c t

Article history: Antenna arrays have various applications, such as radar, military systems, and wireless communications.
Received 16 June 2020 The performance of the communication system depends heavily on the antenna array design process. The
Accepted 1 February 2021 major challenges in designing such arrays are the high sidelobe levels and the enormous number of ele-
Available online 23 February 2021
ments. Recently, optimization techniques that exploit the evolutionary algorithms have been used exten-
sively for thinning the antenna arrays to reduce power consumption and improve the radiation pattern
Keywords: by decreasing the sidelobe levels. However, due to the stochastic nature of the resolution algorithms,
Antenna radiation patterns
achieving a global optimal for such algorithms is not assured. Based on the combination theory, this
Planar arrays
Sidelobe level reduction
paper adopts an element-by-element removing strategy for thinning planar arrays. Without loss of gen-
Thinned arrays erality, a uniformly excited 25 elements rectangular array with isotropic elements is investigated as an
example. The proposed thinning method examines all the possible arrangements of the placements of
25 elements in the phased array aperture. The optimum position of elements is selected to achieve the
desired characteristics of the radiation pattern. The results show that with the 40% thinning percentage,
the same radiation pattern for the full array (without thinning) can be witnessed. Consequently, a con-
siderable reduction in resource utilization can be achieved in terms of power consumption, the number
of phase shifters, and the number of attenuators.
Ó 2021 The Authors. Production and hosting by Elsevier B.V. on behalf of King Saud University. This is an
open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction tion pattern requirements. Various adaptive beamforming tech-

niques have been developed to compute these excitation weights
Successful deployment of 5G wireless networks depends on the (Li and Stoica, 2006). It aims to reduce the error between the actual
efficient design of antenna arrays. The smart antenna arrays are and desired signal and increase the signal-to-interference ratio
composed of radiating individual elements (individual antennas) (SIR). In a strong interference environment, when the actual signal
configured in a geometrical way to work as one antenna to provide is weak, the SIR greatly affects the antenna array performance and
high directivity gain. They allow the main lobe of the radiation pat- must be considered (Banerjee and Dwivedi, 2018).
tern to be directed toward the desired signal. The nulls of the radi- Smart antenna arrays help to minimize interference while
ation pattern can be steered towards the directions of the maintaining an optimal coverage area and reducing the transmit
undesired signals or the sources of interference. The smart antenna power of both base station (BS) and mobile handset. In addition,
array’s elements should be weighted to adapt to the desired radia- more energy can be transmitted at higher frequencies by using
narrow beams of the same size as the physical aperture. Smart
antenna design enables different beams to use the same channel,
⇑ Corresponding author. therefore, minimizes co-channel interference (Feng and Zhang,
E-mail addresses: [email protected] (M.H. Agha), [email protected]. 1999; Mismar and Ismail, 1997).
tr (O. Bayat).
Antenna array thinning is used with the aim of removing some
Peer review under responsibility of King Saud University.
elements of the full array that provides several advantages such as
power consumption reduction, lower weight and cost, and mini-
mizing the complexity of the Beam Forming Network (BFN). The
thinned antenna arrays’ features are conducive to preventing
Production and hosting by Elsevier

https://doi.org/10.1016/j.jksues.2021.02.003
1018-3639/Ó 2021 The Authors. Production and hosting by Elsevier B.V. on behalf of King Saud University.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
M.H. Agha, Maan A.S. Al-Adwany, O. Bayat et al. Journal of King Saud University – Engineering Sciences 34 (2022) 544–561

jamming effects and improving the receiving of a signal (Pan et al., to thin the rectangular planar array with low SLL (Dib, 2015). This
2020). Such advantages are further reflected in the digital beam- was achieved by optimizing the amplitude excitations of the indi-
forming antenna array systems as a computational load reduction. vidual elements. In addition, A modified compact genetic algo-
The synthesis of decreasing the number of elements is a nonconvex rithm (M
cGA) is applied to synthesize the linear and planar-
and strongly nonlinear optimization problem. It is, therefore, very thinned arrays (Ha et al., 2015). The results show that the M
cGA,
difficult to provide optimal global solutions (i.e., good antenna with a low computational cost, can control the SLL of both planar
choices). In general, reducing the number of array elements and linear arrays. Moreover, to optimize planar thinned antenna
involves designing antenna arrays with nonuniform element spac- arrays with relatively less time, an inverse Fourier transform
ing. In this framework, different methods have been proposed in (IFT) has been effectively applied to thinned planar arrays with rel-
the literature to synthesize such antenna arrays, such as ant colony atively minimum SLL (Suo et al., 2016).
optimization (ACO) (Quevedo-Teruel and Rajo-Iglesias, 2006), sim- Among the various planar antenna arrays, concentric ring arrays
ulated annealing (SA) (Trucco et al., 1997; Trucco, 2002), genetic (CRAs) with a circular symmetrical beam played an important role
algorithm (GA) (Haupt, 1994), and particle swarm optimization in satellite communications, direction-of-arrival estimation, radio-
(PSO) (Hooker and Arora, 2010). Unfortunately, those methods astronomy etc. Taking advantage of the scan capability for all-
are time-consuming, particularly when dealing with planar or lin- azimuth and almost invariant azimuthal coverage, the algorithm
ear arrays synthesis composed of many antenna elements. Thus, of TLBO (Xiaowen et al., 2016) is used to synthesize sparse concen-
there is a need for more efficient methods to reduce processing tric ring arrays to generate a circularly symmetrical radiation pat-
time. A hybrid algorithm that combines the genetic algorithm with tern with lower SLL and fewer elements. Effective results were
a convex optimization (Fu et al., 2016) offers a successful solution achieved by applying TLBO for one- and two-dimensional thinning
for reducing both the sidelobe level (SLL) of the planar antenna arrays (Recioui, 2016).
array and the processing time. In addition, a hybrid method that Many techniques were used to thin the antenna arrays have
combines the ant colony optimization and the least mean square optimized the SLL at the antenna’s boresight. However, grating
algorithm (LMS) is used for the thinned fractal array synthesis lobes started to appear when the antenna arrays were steered at
(El-Khamy et al. 2017a; El-Khamy et al., 2017b). Their method
has succeeded in steering the nulls toward the interferences direc-
tion and the main lobe towards the desired angle and conserving Start
the identical radiation characteristics at different frequency bands,
reducing the number of elements.
In the antenna array thinning, the SLL can be optimized either Generate the pattern for full rectangular array
by optimizing the position of the elements and utilizing the uni-
formly excited elements or by optimizing the design parameters,
such as phase or/and amplitude excitations of the individual ele- Compute all possible antenna arrangements
for each thinning percentage
ments and fixing the position of the elements. A teaching
learning-based optimization (TLBO) has been effectively applied
Select the desired thinning percentage

Generate the radiation patterns for the all possible

antenna arrangements of the desired thinning percentage

For j=1 to size of possible antenna

arrangement for each thinning percentage

Compute SLL and HPBW for each pattern of

possible antenna arrangement

j=j+1 Set the desired minimum SLL, narrower

Fig. 1. Planar array configuration with M  N elements. HPBW, and narrower FNBW

NO Are the desired

SLL, HPBW,
FNBW
achieved?

YES

Output the optimized radiation pattern

End

Fig. 2. Configuration of the 25 elements rectangular array. Fig. 3. Flowchart of the element by element removing strategy.

545
M.H. Agha, Maan A.S. Al-Adwany, O. Bayat et al. Journal of King Saud University – Engineering Sciences 34 (2022) 544–561

angles from
40° to + 40°. The rest of the paper is organized as fol-
lows: Section 2 presents the mathematical modeling for planar
arrays and the combination theory. Section 3 describes the planar
array thinning based on the proposed element-by-element remov-
ing strategy. Section 4 presents the simulation results, whereas
Section 5 concludes the paper.

2. Mathematical modeling for planar arrays

Consider the geometry of the planar antenna array as illustrated

(a)
in Fig. 1. The antenna array consists of (M  N) elements that are
. positioned symmetrically about the origin in the x-y plane, where
the elements are assumed isotropic. Under those conditions, the
array factor equation can be written as follows:
X
N X
M
2m
1
AFðh; /Þ ¼ Am;n cos½  kdy sinhcos/
n¼1 m¼1
2
2n
1
 cos½  kdx sinhsin/ ð1Þ
2
where AF ðh; /Þ is an array factor, Am;n represents the excitation
amplitude of the element at the position ðn; mÞ, and k ¼ 2p=k. In
addition, dx and dy represent the inter-element spacing along the
x-axis and y-axis, respectively, which is equal to 0:5k. Moreover,
(b) h represents the elevation angle and / is the azimuth angle.
The normalized value of the maximum sidelobe level (FMSLL) of
Fig. 4. Simulation of the full rectangular antenna array: (a) layout of the 25 the planar array is formulated as follows:
elements full antenna array (b) full antenna array radiation pattern.
AFðAm;n ; h; /Þ
F MSLL ðAm;n ; h; /Þ ¼ max½20logj j ð2Þ
angles away from boresight. To overcome this issue when design- AF max
ing an antenna array with a reduced sidelobe level during scan- where AF max represents the maximum value of the array factor.
ning, one must also optimize the thinned antenna arrays
scanning them within the desired scan area. A modified binary 2.1. Description of the combination theory
coded genetic algorithm (MBC-GA) has been presented in (Jijenth
et al., 2017) to synthesise the planar array with reduced SLL at A combination is a subset of elements selected from a given set
antenna boresight and also when the antenna array is steered of distinct elements. In other words, the combination refers to how
away from boresight by angles 20° and 40°. many ways we can choose the items in any order.
The thinned antenna array also could be used for radar imaging. The number of the r -combinations that are selected from a set
Several conventional algorithms for optimization, such as the of n -distinct elements is given by (Chen and Koh, 1992):
genetic algorithm, have been suggested to reduce the SLL of the
thinned array radiation pattern for radar imaging (Chen et al.,
n n!
Cðn; rÞ ¼ ð Þ ¼ ð3Þ
2018). r r!ðn
rÞ!
This paper introduces an accurate method for thinning a plan- Where n is the number of elements in the set, and r is the num-
ner array. Based on the combination theory, the proposed thinning ber of elements in the subset.
method examines all the possible antenna array arrangements by In this paper, the goal of the element-by-element removing
following the element-by-element removing strategy. Thus, the strategy is to select the best arrangement of the 25 elements rect-
array arrangement that provides the best performance could be angular antenna array, as a case study, that gives the best radiation
attained. In addition, the proposed array thinning has shown good pattern. The combination theory is used here to compute all the
performance when applying main beam steering with scanning possible antenna array arrangements for each number of elements

Table 1
Number of possible antenna array arrangements for each number of the removed elements in 25 elements antenna array.

# of Removed elements Thinning # of possible antenna array arrangements # of Removed elements Thinning #of possible antenna array arrangement
Percentage Percentage
1 4% 25 13 52 % 5,200,300
2 8% 300 14 56 % 4,457,400
3 12 % 2300 15 60 % 3,268,760
4 16 % 12,650 16 64 % 2,042,975
5 20 % 53,130 17 68 % 1,081,575
6 24 % 177,100 18 72 % 480,700
7 28 % 480,700 19 76 % 177,100
8 32 % 1,081,575 20 80 % 53,130
9 36 % 2,042,975 21 84 % 12,650
10 40 % 3,268,760 22 88 % 2300
11 44 % 4,457,400 23 92 % 300
12 48 % 5,200,300 24 96 % 25
The total number of possible antennas arrangements = 33,554,432

546
M.H. Agha, Maan A.S. Al-Adwany, O. Bayat et al. Journal of King Saud University – Engineering Sciences 34 (2022) 544–561

to remove. Fig. 2 shows a 5  5 rectangular antenna array; the total

Table 2 number of antenna array elements is 25.
Comparison of SLL, HPBW and FNBW obtained by the full rectangular antenna array
Based on the proposed element-by-element removing strategy,
and thinning antenna array.
we first remove one element from the 25 elements, so the number
of combinations without repetitions can be computed by using Eq.
(3):
 
n 25!
¼ ¼ 25 ð4Þ
r 1!ð25
1Þ!
where n represents the number of antenna elements, whereas r
is the number of elements to remove.
When the number of the elements to remove is two out of 25
elements, the number of the combinations would be:
 
n 25!
¼ ¼ 300 ð5Þ
r 2!ð25
2Þ!
Also, the number of combinations resulting from removing
three elements will be:
n 25!
¼ ¼ 2300 ð6Þ
r 3!ð25
3Þ!
As illustrated above, removing one element gives 25 different
antenna arrangements, and removing two elements gives 300
different antenna arrangements, whereas removing three ele-
ments gives 2300 different antenna arrangements and so on.
By simple computation, one can prove that the 25 elements
antenna array has 33; 554; 432 different arrangements. Thus, all

Fig. 5. Simulation of the 40% thinned antenna array: (a) layout 1 of the thinned antenna array (b) the obtained antenna array radiation patterns.

547
M.H. Agha, Maan A.S. Al-Adwany, O. Bayat et al. Journal of King Saud University – Engineering Sciences 34 (2022) 544–561

Fig. 6. Simulation of the 44% thinned antenna array: (a) layout 1 of the thinned antenna array (b) the obtained antenna array radiation pattern.

the possible antenna array arrangements for each number of ele- 3.1. Optimization process
ments to remove can be computed depending on the combina-
tion theory. Array thinning means that several antenna array elements are
turned off to steer the the main beam towards the desired direc-
tion and reducing the SLL without losing the array performance.
All ON elements of the array are fed with the value ‘10 , and the
3. Planar array thinning based on the element-by-element
OFF elements are fed with the value ‘00 ., A flowchart of the opti-
removing strategy
mization process is presented in Fig. 3. to clarify the optimization
process simulation results.
The antenna arrays must be designed to have low sidelobe
This section discusses the ability of the proposed method to
levels, maximum gain, and narrow beamwidth to obtain the opti-
enhance the radiation pattern of the 25 elements rectangular planar
mum performance of antenna arrays. To accomplish this goal, the
array by achieving a low sidelobe level with fewer active radiating
radiation pattern of antenna arrays is optimized either by control-
elements. It is done by comparing the patterns of the thinned rectan-
ling the element positions or by controlling the design parameters,
gular arrays and the full array. In addition, the obtained results are
such as phase or/and amplitude excitations of the individual
compared with those presented in the literature that exploits the
elements.
evolutionary algorithms for thinning the antenna arrays. Further-
This work proposes the element-by-element removing strategy
more, the beam steering capability of the proposed method is inves-
to achieve a thinned array. This method aims to reduce the SLL by
tigated by steering the main beam with scanning angles of ± 40°.
controlling the spacing between the antenna elements of the array,
whereas the weights of individual elements are assumed to be con-
stant. As mentioned before, the thinning method examines all the 3.2. Generation of the full antenna array pattern
possible arrangements of placing 25 elements to select the antenna
arrangement that gives the optimum radiation pattern of the A radiation pattern of the 5x5 full antenna array has been gen-
antenna array. erated with 0:5 k inter-element spacing and with uniform ampli-
548
M.H. Agha, Maan A.S. Al-Adwany, O. Bayat et al. Journal of King Saud University – Engineering Sciences 34 (2022) 544–561

(a)

(b)
Fig. 7. Simulation of the 64% thinned antenna array: (a) layout of the thinned antenna array (b) the obtained antenna array radiation pattern.

tude and phase excitations. The generated radiation pattern has As illustrated in Table 2, for each thinning percentage, there are
achieved an SLL equal to
12 dB, and the values of the HPBW a number of possible antenna array arrangements. According to
and FNBW were found to be equal to 21° and 47°, respectively, each possible arrangement of the thinning antenna array, the radi-
as shown in Fig. 4. ation pattern that is characterized by specific SLL, HPBW, and
FNBW will be generated. Among the 33; 554; 432 possible arrange-
3.3. Generation of thinned antenna array patterns ments, the one that gives the optimum radiation pattern is chosen,
as will be shown later.
The thinning method used in this paper is based on element by The obtained results that are listed in Table 2 are arranged
element removing strategy according to the combination theory. based on the achievable SLL. For example, a
12 dB SLL can be
Thus, all the possible antenna arrangements for each number of accomplished by the thinning percentage 40% (layout1). It is
elements to remove could be computed. Table 1 shows the number observed that with this thinning percentage, one can achieve the
of elements to remove and the corresponding number of possible same radiation pattern as that for 0% (full array) as shown in
array arrangements. Table 2 (red colored), also shown in Fig. 5.
The radiation patterns of all the possible 33,554,432 antenna Fig. 6 shows that the thinning percentage of 44 % (layout1) can
array arrangements have been generated via a MATLAB code previ- reduce the SLL of the full antenna array to
13:5 dB, keeping the
ously programmed for this purpose. The program calculates the values of HPBW and FNBW close to that for the full array. Conse-
sidelobe level, half power beamwidth (HPBW), and the first null quently, the power consumption can be reduced by 44% compared
beamwidth (FNBW) for each radiation pattern. The obtained with that needed for the full array.
results are arranged in Table 2, which shows the SLL, HPBW, FNBW, In addition, the SLL of the full array could be reduced to
14:2
thinning percentage, and the number of elements to remove. dB when the thinning percentage is 64%. This thinning percentage
549
M.H. Agha, Maan A.S. Al-Adwany, O. Bayat et al. Journal of King Saud University – Engineering Sciences 34 (2022) 544–561

(a)

(b)
Fig. 8. Simulation of the 32% thinned antenna array: (a) layout 1 of the thinned antenna array (b) the obtained antenna array radiation pattern.

has an HPBW of 25° and FNBW of 60° as illustrated in Fig. 7. Com- Fig. 13 shows that increasing the thinning percentage to 68%
pared with the full array, the total power consumption can be reduces the SLL to
18:06 dB. It is clear that the values of the
reduced by 64%. HPBW and the FNBW have increased to 24° and 60°, respectively.
Also, the SLL of the full antenna array could be reduced to
14:6 In addition, by the 40% thinning percentage (layout 2), the SLL
dB by thinning percentage of 32 % (layout 1). This reduction in SLL could be reduced to
18:5 dB, as shown in Fig. 14. However, this is
has been obtained, keeping the HPBW and FNBW close to that of because of an increase in HPBW and FNBW by three and thirteen
the full antenna array, as shown in Fig. 8. degrees, respectively.
Referring to the results in Fig. 9, a
15:7 dB SLL was achieved by Moreover, an SLL of
22 dB could be achieved by 44% thinning
48% thinning percentage, but by increasing one degree in HPBW percentage (layout 2), as shown in Fig. 15. The HPBW and the
and seven degrees in FNBW. This means that removing about half FNBW were found equal to 24° and 60°, respectively.
the number of active radiating elements reduces power consump- Finally, a minimum SLL can be obtained by a 32% thinning per-
tion by 48%. centage (layout 2), where the achieved SLL is found to be
24 dB,
It is clear from Fig. 10 and Fig. 11 that the SLL reduces to
16 dB as shown in Fig. 16. However, this is at the expense of a significant
by both thinning percentages, 24% and 72 %, respectively. How- increase in the HPBW and FNBW to about 25° and 66°,
ever, that is on account of some increase in HPBW and FNBW; respectively.
where with the 24% thinning percentage, the HPBW is 23° and Some performance parameters are illustrated in Table 3 to com-
the FNBW is 55°; whereas with the 72% thinning percentage, the pare the performance of the proposed method with the previous
HPBW is 25° and the FNBW is 60°. related works (Jijenth et al., 2017; Suo et al., 2016; Ha et al.,
Further reduction in SLL could be achieved by the 36% thinning 2015; Dib, 2015). It can be observed that the SLL obtained by the
percentage, where the SLL is
17:06 dB, as shown in Fig. 12. It is proposed method with a 5x5 antenna array is less than that
observed that there is an increase in HPBW and the FNBW com- obtained by MBC-GA, IFT, M
cGA, and TLBO that applied to differ-
pared with that of the full array. ent planar array sizes. Moreover, fewer ON antenna elements are

550
M.H. Agha, Maan A.S. Al-Adwany, O. Bayat et al. Journal of King Saud University – Engineering Sciences 34 (2022) 544–561

(a)

(b)
Fig. 9. Simulation of the 48% thinned antenna array: (a) layout 1 of the thinned antenna array (b) the obtained antenna array radiation pattern.

used by the proposed method to achieve minimum SLL compared where h represents the maximum steerability away from bore-
with the other methods. Consequently, our proposed method out- sight, k represents the wavelength at the operation frequency, and
performs MBC-GA, IFT, M
cGA, and TLBO in terms of minimum SLL d represents the inter-element spacing. It can be noticed from
and the number of ON antenna elements. Our proposed method equation (7) that when the inter-element spacing equals to the
examines all the possible antenna arrangements to achieve an 0.5 k, the maximum scan range is 90°.
optimal solution. The main beam of the optimized layout of the proposed 5x5
antenna arrays can be scanned from
40° to the + 40° keeping
the performance of the thinned array (i.e., low sidelobe level and
3.4. Beam steering capability of the thinned array
power consumption). For example, layout 1 of the 40 % thinned
antenna array obtained in Fig. 5 is steered to specific angles in
A major problem in developing thinned arrays is limited scan
the azimuth plane, and the scanning performance is shown in
range, which is due to the formation of grating lobes. The maxi-
Fig. 17(a-d). It can be noticed that with the 40 % thinned array,
mum scan angle that a thinned planar array can be steered from
one can achieve the same radiation pattern as that for the full array
broadside is expressed as (Bray et al., 2002):
(with SLL equal to
12 dB) when the main beam is steered to
20°,
 
20°,30° and 40° (away from the boresight). In other words, there

1 k
h ¼ sin
1 ð7Þ is no degradation in SLL and no grating lobes formation when the
d
thinned array is steered away from boresight.

551
M.H. Agha, Maan A.S. Al-Adwany, O. Bayat et al. Journal of King Saud University – Engineering Sciences 34 (2022) 544–561

(a)

(b)
Fig. 10. Simulation of the 24% thinned antenna array: (a) layout of the thinned antenna array (b) the obtained antenna array radiation pattern.

Furthermore, the same antenna steering is applied to the 32 % to optimize the radiation pattern and reduce the power consump-
thinned antenna array (layout 2, Fig. 16), achieving the lowest tion of full antenna arrays.
SLL of
24 dB. The resulting radiation patterns are gathered in Without loss of generality, a 25 elements rectangular array is
Fig. 18(a-h). It can be observed that the minimum SLL is still studied. According to the combination theory, the possible antenna
achieved when the thinned array is steered to the specified angles. array arrangements of the 25 elements antenna array are
In addition, no grating lobes are configured in the entire visible 33; 554; 432. Using the MATLAB computer program, an algorithm
area of the antenna. Thus, the proposed approach serves the pur- is designed to select the array arrangement that produces the best
pose of array thinning, which is the ability to achieve minimum radiation pattern for achieving low SLL, narrow HPBW, and narrow
SLL with the lowest possible antenna elements at both the bore- FNBW. The obtained results demonstrated that the 40% thinning
sight and the ± 40° scan angles. It is worth mentioning that if the percentage resulted in a radiation pattern similar to that of the full
scanning angles are increased above ± 40°,the main beam width antenna array. In other words, the radiation pattern of the 25 ele-
will significantly increase, resulting in a reduction in the thinned ments can be achieved by only 15 active elements with a specific
array performance. However, this problem can be overcome by arrangement (as shown in Fig. 5).
increasing the array elements to add further degree of freedom Thus, by using such a geometric array, the supplied power can
in selecting elements position. This, in turn, will increase the steer- be reduced by 40% compared to that for the full antenna array.
ability of the thinned array to more than ± 40° scanning range with On the other hand, the number of phase shifters and attenuators
low SLL and narrow beamwidth. also can be reduced by 40% if such configuration is used for beam-
forming. In addition, the results demonstrated that different thin-
ning percentages reduce the SLL up to
24 dB with more
4. Conclusion efficient power consumption. However, it would be due to an
increase in the HPBW and the FNBW.
A new approach to antenna array thinning is proposed in this The achievement of the optimal solution for the proposed thin-
work based on the element by element removing strategy in order ning strategy is assured due to the covering of all possible antenna
552
M.H. Agha, Maan A.S. Al-Adwany, O. Bayat et al. Journal of King Saud University – Engineering Sciences 34 (2022) 544–561

(a)

(b)
Fig. 11. Simulation of the 72% thinned antenna array: (a) layout of the thinned antenna array (b) the obtained antenna array radiation pattern.

553
M.H. Agha, Maan A.S. Al-Adwany, O. Bayat et al. Journal of King Saud University – Engineering Sciences 34 (2022) 544–561

(a)

(b)

Fig. 12. Simulation of the 36% thinned antenna array: (a) layout of the thinned antenna array (b) the obtained antenna array radiation pattern.

554
M.H. Agha, Maan A.S. Al-Adwany, O. Bayat et al. Journal of King Saud University – Engineering Sciences 34 (2022) 544–561

(a)

(b)

Fig. 13. Simulation of the 68% thinned antenna array: (a) layout of the thinned antenna array (b) the obtained antenna array radiation pattern.

555
M.H. Agha, Maan A.S. Al-Adwany, O. Bayat et al. Journal of King Saud University – Engineering Sciences 34 (2022) 544–561

(a)

(b)
Fig. 14. Simulation of the 40% thinned antenna array: (a) layout 2 of the thinned antenna array (b) the obtained antenna array radiation pattern.

556
M.H. Agha, Maan A.S. Al-Adwany, O. Bayat et al. Journal of King Saud University – Engineering Sciences 34 (2022) 544–561

(a)

(b)
Fig. 15. Simulation of the 44% thinned antenna array: (a) layout 2 of the thinned antenna array (b) the obtained antenna array radiation pattern.

557
M.H. Agha, Maan A.S. Al-Adwany, O. Bayat et al. Journal of King Saud University – Engineering Sciences 34 (2022) 544–561

(a)

(b)

Fig. 16. Simulation of the 32% thinned antenna array: (a) layout 2 of the thinned antenna array (b) the obtained antenna array radiation pattern.

Table 3
Comparison of minimum SLL with the optimal number of ON elements obtained with different methods applied to different size planar array elements.

Technique/reference/year Array size SLL # of ON elements

(dB)
MBC-GA (Jijenth et al., 2017) 12x12 –23.07 76
IFT (Suo et al., 2016) 10x20
21.12 108
M
cGA (Ha et al., 2015) 12X12
19.4 69
24X24 –23.3 253
30X30
24.6 540
TLBO (Dib, 2015) 6x6
16.71 26
10x10
19.64 65
Element by elementremoving strategy 5X5
24.6 17
(proposed)

558
M.H. Agha, Maan A.S. Al-Adwany, O. Bayat et al. Journal of King Saud University – Engineering Sciences 34 (2022) 544–561

(a) (b)

(c) (d)

(e)

Fig. 17. Simulation of 40 % thinned array steering (layout 1) to the specific angles: (a)
40° in azimuth plane (b)
20° in azimuth plane (c)
10° in azimuth plane (d) 30° in
azimuth plane (e) 40° in the azimuth plane.

559
M.H. Agha, Maan A.S. Al-Adwany, O. Bayat et al. Journal of King Saud University – Engineering Sciences 34 (2022) 544–561

(a) (b)

(c) (d)

(e)
(f)

(g) (h)

Fig. 18. Simulation of 32% thinned array steering (layout 1) at scan angles: (a)
40° in azimuth plane (b)
30° in azimuth plane (c)
20° in azimuth plane (d)
10° in azimuth
plane (e) 10° in azimuth plane (f) 20° in azimuth plane (g) 30° in azimuth plane (h) 40° in azimuth plan.

560
M.H. Agha, Maan A.S. Al-Adwany, O. Bayat et al. Journal of King Saud University – Engineering Sciences 34 (2022) 544–561

array arrangements. Furthermore, the proposed array thinning has Feng, Z., Zhang, Z., 1999. Dynamic spatial channel assignment for smart antenna.
Wirel. Pers. Commun. 11, 79–87.
shown good performance when applying main beam steering with
Fu, Y., Guo, Z.-G., Wang, H., Gui, Y., Yang, G.-M., 2016. Optimization of planar
scanning angles from
40° to + 40°. It is worth mentioning that the thinned antenna array based on genetic and convex hybrid algorithm, in: 2016
disadvantage of the proposed method is the huge computations it Progress in Electromagnetic Research Symposium (PIERS). IEEE, pp. 1299–1303.
requires for the antenna array sizes larger than 25 elements. Ha, B.V., Mussetta, M., Pirinoli, P., Zich, R.E., 2015. Modified compact genetic
algorithm for thinned array synthesis. IEEE Antennas Wirel. Propag. Lett. 15,
1105–1108.
Declaration of Competing Interest Haupt, R.L., 1994. Thinned arrays using genetic algorithms. IEEE Trans. Antennas
Propag. 42, 993–999.
Hooker, J.W., Arora, R.K., 2010. Optimal thinning levels in linear arrays. IEEE
The authors declare that they have no known competing finan- Antennas Wirel. Propag. Lett. 9, 771–774.
cial interests or personal relationships that could have appeared Jijenth, M., Suman, K.K., Gangwar, V.S., Singh, A.K., Singh, S.P., 2017. A novel
to influence the work reported in this paper. technique based on modified genetic algorithm for the synthesis of thinned
planar antenna array with low peak side lobe level over desired scan volume,
in: 2017 IEEE MTT-S International Microwave and RF Conference (IMaRC). IEEE,
pp. 251–254.
References Li, J., Stoica, P., 2006. Robust adaptive beamforming. Wiley Online Library.
Mismar, M.J., Ismail, T.H., 1997. Linear Arrays Pattern Synthesis with Multiple Broad
Nulls Using Partial Amplitude Control. J. King Saud Univ. Sci. 9, 239–249.
Banerjee, S., Dwivedi, V.V., 2018. Effect of signal to interference ratio on adaptive Pan, J.-S., Tsai, P.-W., Huang, H.-C., 2020. Advances in intelligent information hiding
beamforming techniques. J. King Saud Univ. Sci. 30, 313–319. and multimedia signal processing. Springer.
Bray, M.G., Werner, D.H., Boeringer, D.W., Machuga, D.W., 2002. Optimization of Quevedo-Teruel, O., Rajo-Iglesias, E., 2006. Ant colony optimization in thinned array
thinned aperiodic linear phased arrays using genetic algorithms to reduce synthesis with minimum sidelobe level. IEEE Antennas Wirel. Propag. Lett. 5,
grating lobes during scanning. IEEE Trans. Antennas Propag. 50, 1732–1742. 349–352.
Chen, T., Jin, G., Dong, Z., 2018. Near-Field Antenna Pattern Optimization with Recioui, A., 2016. Design and Thinning of Linear and Planar Antenna Arrays Using a
Thinned Array Based on Genetic Algorithm, in: Proceedings of the 2018 Binary Teaching Learning Optimizer. Acta Phys. Pol. A 130, 7–8.
International Conference on Electronics and Electrical Engineering Technology. Suo, Y., Yin, S., Li, W., 2016. Iterative fft algorithm for thinning planar array, in: 2016
pp. 157–161. International Symposium on Antennas and Propagation (ISAP). IEEE, pp. 414–
Chen, C.C., Koh, K.-M., 1992. Principles and techniques in combinatorics. World 415.
Scientific. Trucco, A., 2002. Weighting and thinning wide-band arrays by simulated annealing.
Dib, N.I., 2015. Synthesis of thinned planar antenna arrays using teaching-learning- Ultrasonics 40, 485–489.
based optimization. Int. J. Microw. Wirel. Technol. 7, 557. Trucco, A., Omodei, E., Repetto, P., 1997. Synthesis of sparse planar arrays. Electron.
El-Khamy, S.E., Eltrass, A.S., El-Sayed, H.F., 2017a. Design of thinned fractal antenna Lett. 33, 1834–1835.
arrays for adaptive beam forming and sidelobe reduction. IET Microwaves, Xiaowen, Z., Qingshan, Y., Yunhua, Z., 2016. Application of TLBO to synthesis of
Antennas Propag. 12, 435–441. sparse concentric ring arrays, in: 2016 10th European Conference on Antennas
El-Khamy, S.E., Eltrass, A.S., El-Sayed, H.F., 2017b. Adaptive beamforming synthesis and Propagation (EuCAP). IEEE, pp. 1–5.
for thinned fractal antenna arrays, in: 2017 XXXIInd General Assembly and
Scientific Symposium of the International Union of Radio Science (URSI GASS).
IEEE, pp. 1–4.

561