This article has been accepted for publication in IEEE Transactions on Antennas and Propagation.

This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TAP.2025.3613531

Compact Series-Fed Monopulse Antenna Array

Based on Multi-feed Technology
Chuanming Zhu, Haikun Jia, Member, IEEE, Wei Deng, Senior Member, IEEE, Baoyong Chi, Senior
Member, IEEE


Abstract—A compact multi-feed monopulse antenna array is still high and has difficulty in achieving compatibility with
proposed in this paper. An equivalent circuit and the theory of front-end circuits. Subsequently, quasi-planar gap waveguide
multi-feed antenna array are developed to analyze the array factor (GWG)-based monopulse antennas [5]-[11] represent a
of the sum and difference beam. A close agreement between the promising approach in achieving compact size, low loss and
calculated and electromagnetic simulated radiation pattern is high gain. For example, planar Magic-Tee-based monopulse
observed. By controlling the magnitudes and phases of multiple
excited ports, the desired sum beam and difference beam can be comparator network is utilized in [5] to realize a 16  16
achieved. Different from the conventional four-quadrant antenna array, where the utilization of gap waveguide
symmetrical layout, only one subarray is employed to realize eliminates the galvanic contact between different layers. Butler
monopulse antenna array. In comparison with the conventional matrix composed of stereo lithography apparatus (SLA) is
single-feed monopulse antenna array, the beamwidth of the sum implemented to realize high-efficiency radial line slot
beam in the proposed multi-feed antenna array is almost twice as monopulse antenna [8]. However, it is accompanied by a bulky
wide, and the circuit size of the multi-feed antenna array is also structure and increased cost. Based on GWG structure, dual-
reduced by half. Without increasing the antenna array panel size,
polarized monopulse antenna array with SLL of lower than 19
the function of low-loss on-antenna power combining is achieved
in the proposed multi-feed antenna array. Meanwhile, the need for dB is demonstrated in [11]. It has been acknowledged that
a power combiner and passive impedance matching network is waveguide and GWG-based monopulse antenna arrays exhibit
eliminated. Moreover, the results show the effective isotropic the advantages of high gain and efficiency [2]-[11]. However,
radiated power (EIRP) of transmitter system with the proposed these designs face considerable challenges, including an
multi-feed antenna array exhibits an increase by 10  Log10(N) (N intricate and cumbersome comparator feeding network, high
 3) dB. The proposed multi-feed monopulse antenna array paves cost, rigorous fabrication accuracy, and difficulties in seamless
the way for the development of miniaturized mid-and long-range integration with other planar circuits.
monopulse radar system. As quasi-waveguide structure, SIW monopulse antenna
arrays [12]-[20] are widely studied as an efficient solution
Index Terms—Compact size, monopulse antenna, multi-feed owing to the characteristic of their superior low profile. The
antenna array, power combining.
sum and difference beams are achieved in SIW-based
monopulse antenna arrays by magic tee [12], [13], coupler and
phase shifter [14]-[17] or high-order modes [18]-[20] in SIW
I. INTRODUCTION cavity. Planar SIW magic-tee-based comparator network is
discussed in [13] to realize 4  4 slot monopulse antenna array
Mcommunication and satellites as the distance and angle
ONOPLUS radar [1] is widely used in modern
over a wide bandwidth. Nevertheless, complicated comparators
information of the targets can be obtained by using the sum and [12] increase the circuit size and design difficulty. SIW
difference beam, respectively. Recently, various methods have monopulse antenna with dual polarizations [17] is presented by
been applied to design monopulse antenna arrays [2]-[29]. using four series-fed slot subarrays oriented in a cruciform
Based on irregular interconnected waveguide, a W-band structure. The high-order cavity mode in SIW cavity is
monopulse antenna array with high gain and low sidelobe level employed as monopulse comparator to realize the sum and
(SLL) is realized in [2]. Wideband circularly polarized difference beam [20]. However, multiple subarrays [12]-[20]
monopulse antenna is reported in [3] using four waveguide occupy relatively large circuit sizes. Moreover, an impedance
couplers and one wideband phase shifter. Although waveguide- transformer is typically required in order to interface a chip with
based monopulse antennas [2], [3] features the advantages of SIW components [12]-[20], which could increase system size
low loss and high gain, their profile is higher than planar and design complexity. In contrast, microstrip lines are
counterpart and thus increase the difficulty of integration with typically capable of direct integration with to a chip, offering a
other front-end circuits. Reference [4] presents a high-gain 16 more compact and simplified design approach.
 16 circularly polarized monopulse antenna array implemented A single-layer corporate-series-fed monopulse 16  16
using 3D-printed waveguide technology. Likewise, its profile is antenna array is reported in [21], where multiple hybrid

Manuscript submitted at 22 May 2025. (Corresponding author: Haikun Jia)

The authors are with the School of Integrated Circuits, Tsinghua University,
Beijing 100084, China (e-mail: [email protected])

Authorized licensed use limited to: Onur Akarsu. Downloaded on October 06,2025 at 10:16:24 UTC from IEEE Xplore. Restrictions apply.
© 2025 IEEE. All rights reserved, including rights for text and data mining and training of artificial intelligence and similar technologies. Personal use is permitted,
but republication/redistribution requires IEEE permission. See https://www.ieee.org/publications/rights/index.html for more information.
 This article has been accepted for publication in IEEE Transactions on Antennas and Propagation. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TAP.2025.3613531

couplers are fabricated in the same layer of RF substrate that work. As shown in Fig. 1(b), only one subarray is utilized to
introduce spurious radiation and thus worse sidelobe level realize the desired sum and difference beam based on the multi-
(SLL). A series-fed 22 monopulse antenna array with multiple feed antenna technique. Compared with the conventional
polarizations is discussed in [22] by using two 180 ring hybrid single-feed monopulse antenna array, the beamwidth of the sum
couplers and four patch elements. Likewise, co-layer ring beam in the proposed multi-feed array is almost twice and the
hybrid couplers could produce parasitic radiation and thus circuit size is reduced by half. Moreover, EIRP of transmitter
affect the performance of patch array to a greater or lesser system with a multi-feed antenna array exhibits an increase by
extent. Subsequently, to mitigate aforementioned problems, the 10  Log10(N) (N  3) dB while maintaining the same
technique of via-coupled feeding is implemented in [23], where beamwidth as the system using a single-feed antenna array.
monopulse antenna consists of four ring couplers and four patch This paper is organized as follows. Section II introduces the
elements. Similarly, stripline hybrid quadrature comparator equivalent circuit and design theory of multi-feed antenna array.
networks are implemented in 4  80 series-fed monopulse Power control theory of dual-feed antenna array is developed to
antenna array [24] with Taylor weighting. Beam-diversity- analyze the sum and difference beam. Then the electromagnetic
controllable monopulse antenna arrays are realized in [25]-[28]. (EM) simulation is conducted to further validate the design
A 2  2 stacked-patch antenna array with single circular theory. In order to enhance the EIRP, a three-feed monopulse
polarization is discussed in [25], which has reconfigurable sum antenna array is designed based on the dual-feed antenna array.
and difference beam by using one tunable monopulse The simulated and measured results are discussed in Section III
comparator. Then a dual circularly polarized 2  2 stacked- and conclusions are provided in Section VI.
patch monopulse antenna array is implemented in [26] using
Subarray or element Subarray or element
four branch-line couplers and two rat-race couplers as a
comparator network. A multifunctional 8  2 phased array Monopulse comparator network
1 5
integrated with the function of monopulse is illustrated in [27],  beam 2
90TL
Coupler
90TL Coupler 6
where 2-bit pin diode-based phase shifter and a number of beam
couplers are employed as a monopulse comparator network. 3 7
90TL 90TL
Coupler Coupler 8
Likewise, one corporate-fed 2  2 stacked-patch monopulse
4
antenna array with flexible polarization [28] is realized using
two 180 couplers as a comparator network. Based on the Subarray or element Subarray or element
multilayer PCB, substrate integrated coaxial line (SICL) (a)

comparator is employed to realize wideband monopulse Subarray

antenna array [29]. A conventional 2  2 monopulse antenna  beam beam
array [22], [23], [25], [26], [28] is typically composed of Semi-coax
Ch1 Ch2 Ch3
complicated monopulse comparators and four patch elements x Chip
with limited gain. Then, in order to increase antenna gain, patch 0°
y (b)
elements are replaced by subarrays [19], [20], [21], [24], [29].
As shown in Fig. 1(a), four-quadrant symmetrical layout is Fig. 1 (a) Conventional monopulse antenna array based on four-quadrant
typically implemented in the most of monopulse antenna arrays. symmetrical layout, (b) proposed compact multi-feed monopulse antenna array.
However, for the sum beam in azimuth ( = 0) or elevation
Semi-coaxial transition
plane, the beamwidth of the constituted array [2]-[29] is
reduced by half compared with single element or subarray. In Chip
1 2 3
general, wide-sector beam is more effective for detecting
multiple targets in consideration of random ones [30].
Meanwhile, the effective isotropic radiated power (EIRP) of
z
transmitter with a constituted array increases by only 10  x ° Antenna
Log10(2) dB in comparison with a single element or subarray. y
Furthermore, the realization of the desired sum and difference (a)
beam in traditional monopulse antenna arrays typically requires
at least two subarrays, which leads to an increase in circuit size. 0.508mm Rogers 5880
Much effort has been made to develop multi-feed antennas
0.1mm Rogers 4450f
[31]-[39] by adding multiple ports directly onto the antenna
with various functions of power-combining [31]-[37] or various 0.254mm Rogers 5880
polarizations [38] without any complicated power combiner. bump
Theory analysis and measurements are not found in [39].
However, the majority of existing research about multi-feed
Chip
antennas concentrates on individual antenna element itself (b)
rather than on the array as a whole. Fig. 2 (a) Proposed compact multi-feed monopulse antenna array. (a) exploded
A compact monopulse antenna array is proposed in this view, (b) side view.

Authorized licensed use limited to: Onur Akarsu. Downloaded on October 06,2025 at 10:16:24 UTC from IEEE Xplore. Restrictions apply.
© 2025 IEEE. All rights reserved, including rights for text and data mining and training of artificial intelligence and similar technologies. Personal use is permitted,
but republication/redistribution requires IEEE permission. See https://www.ieee.org/publications/rights/index.html for more information.
 This article has been accepted for publication in IEEE Transactions on Antennas and Propagation. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TAP.2025.3613531

II. MULTI-FEED MONOPULSE ANTENNA ARRAY A. Single-Feed Patch Antenna Array

Figs. 2(a) and (b) show the proposed compact multi-feed Due to the adopted symmetrical structure, it is assumed that
monopulse antenna array. It is composed of three-channel the patch array being characterized by an even number of patch
silicon chip, bump, 1  10 series-fed patch antenna array. elements is situated along the x-axis. Single-feed 1  10 antenna
Silicon chip is flip-chipped on 50  GCPW transmission line array is firstly designed. The lengths of feeding line and patch
(TL) of an RF board Rogers 5880 with a thickness of 0.254 mm. element are both about g/2 [40]-[42], where g is the guided
The diameter of bump is about 150 um. Subsequently, 1  10 wavelength at the center frequency f0 = 24 GHz. Both of lines
antenna array is fed by three semi-coaxial structures in Fig. 2(a). are lightly shorter than g/2 due to the open-end effect. One bent
Magnitudes and phases of three channels can be easily g/2 TL is inserted at the center-fed position to ensure all
controlled by an integrated chip. The sum beam is generated by elements in-phase. The normalized magnitude excitation
exciting all three ports with equal amplitude and in-phase coefficients of array are derived as 0.5987: 0.6282: 0.7640:
signals, whereas difference beam is realized by exciting port 1 0.9269: 1 after 20 dB Taylor distribution being applied.
and 3 with equal amplitude and out-of-phase signals. According to the above magnitude distribution that is nearly
proportional to the width of elements, the ones of patch are
0
Sythesized designed based on the extraction process of parameters [42].
Simulated Fig. 3 shows the synthesized and simulated normalized
-10 radiation pattern of center-feed antenna array. There is
difference between the theoretical and simulated response due
Normalized Gain (dB)

-20 to the element not being the ideal point, and mutual coupling
between elements [43]. Subsequently, impedance matching can
be realized by tuning the parameter r of the impedance
-30
transformer. Figs. 4(a) and (b) show the simulated return loss
and radiation pattern against the parameter r of impedance
-40 r transformer. As the parameter r increases, the operated
transformer
bandwidth also expands. However, sidelobe level (SLL) in Fig.
-50 4(b) deteriorates due to the parasitic radiation introduced by an
-90 -60 -30 0 30 60 90 increased-size impedance transformer [21]. Therefore, a trade-
Theta () off between the operated bandwidth and the SLL of antenna
Fig. 3. Synthesized and simulated normalized radiation of single-feed 110 array needs to be made, with r set to be 0.5mm.
antenna array.
0 r transformer
wf
w1 w2 w3 w4 w5 l5 l4 l3 l2 l1
-5 1 2 3 Port 1 4 5 6 7 Port 2 8 9 10
(a)
-10 1:M g/2 g
g g M:1 g/2 g
S11 (dB)

r = 0.2 mm Gpa1 Gpa2 Gpa3 Port1 Gpa4 Gpa5 Gpa6 Gpa7 Port2 Gpa8 Gpa9 Gpa10
-15
r = 0.3 mm
r = 0.4 mm (b)
-20 r = 0.5 mm Fig. 5. (a) Dual-feed antenna array structure and (b) its equivalent circuit at f0.
r = 0.6 mm
r = 0.7 mm
-25
23.0 23.5 24.0 24.5 25.0 B. Theory of Dual-Feed Patch Antenna Array
Frequency (GHz) Compared with single-feed antenna in Fig. 3, two coaxial
(a) ports are fed onto the center of feeding line between the
20
r = 0.2 mm radiating elements i and i+1 (i = 3, 7) in Fig. 5(a). With the
r = 0.3 mm elements being resonant, their imaginary parts cancel [40]
r = 0.4 mm which is not shown in the equivalent circuit in Fig. 5(b). The
10
r = 0.5 mm
r = 0.6 mm input admittance of port 1 Gin1 or port 2 Gin2 is derived as
Gain (dB)

r = 0.7 mm
0 N
Gin1  Gin 2   G paj  GL M (1)
j 1
-10
where Gpa is the radiating conductance of patch element and GL
-20 is the conductance of load [40]. M is the turn ratio of the
-90 -60 -30 0 30 60 90 transformer, which is realized by a circle patch. Due to the
Theta () symmetry of array, the conductance Gpaj equals to GpaN+1-j (j =
(b)
Fig. 4. Simulated (a) return loss, (b) radiation pattern against the parameter r.
1, 2…N/2). According to [40], [43], the parameters in Fig. 5(b)

Authorized licensed use limited to: Onur Akarsu. Downloaded on October 06,2025 at 10:16:24 UTC from IEEE Xplore. Restrictions apply.
© 2025 IEEE. All rights reserved, including rights for text and data mining and training of artificial intelligence and similar technologies. Personal use is permitted,
but republication/redistribution requires IEEE permission. See https://www.ieee.org/publications/rights/index.html for more information.
 This article has been accepted for publication in IEEE Transactions on Antennas and Propagation. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TAP.2025.3613531

can be obtained as Gpa1 = 0.002, Gpa2 = 0.0282, Gpa3 = 0.00235, when reflection coefficients S11 is small, the radiated power Pr
Gpa4 = 0.003, Gpa5 = 0.0011, M = 0.6. Two ports are can be reduced as
simultaneously excited for dual-feed antenna array. Thus,
active reflectional coefficient S iia for two ports is obtained as Pr  1  S21
2
(4)
2
Siia   Sij (2) where S21 actually denotes the fractional power flowing toward
j 1
to the subsequent section. It is assumed that the ratio of power
where Sij (j = 1, 2) denotes the passive transmission coefficient between the left and right sides of port 1 is approximated as the
from port j to port i. Based on the single-feed antenna, active ratio of their corresponding input admittances Ginl and Ginr,
reflectional coefficients S iia of dual-feed counterpart can be well respectively. Thus, the radiated power Pri of element i is
matched by turning the parameter r of the transformer in Fig. estimated as
5(a). The final physical parameters of dual-feed antenna array
are obtained as: w1 = 3.0mm, w2 = 3.5 mm, w3 = 4.35mm, w4 = Pri 
Ginl
 Pin  i (5)
4.8mm, w5 = 5 mm, wf = 0.48 mm, l1 = 4.05mm, l2 = l3 = l4 = Ginl  Ginr
3.97 mm, l5 = 4.03mm. Fig. 6 compares the impedance
matching of single-feed EM-simulated antenna with that of where Pin is the input power, and i is the radiation efficiency
dual-feed circuit and its corresponding EM-simulated of element i [47]. Radiation efficiency is nearly proportional to
counterpart. The results indicate that good matching is realized
the width of patch element [48]. For the element j (j  i) in the
in both single-and dual-feed antenna arrays, thereby validating
left side of port 1, the radiated power can be expressed as
the circuit model in Fig. 5(b).
 j 1 j 1
0 Prj  1   1 Pri  m   1 
i j i  j 1
m 
 m  i 1 m  i 1

j 1

 1      1
i j i  j 1
Reflection coefficients (dB)

-10 m  j 1  j (6)
m i  2 
Similarly, for the element j (j > i) in the right of port 1, the
-20 radiated power can be derived as

Single-feed EM  j 1 j 1
Prj  1   1 Pr (i 1)   m   1   m 
j  i 1 j i
-30 Dual-feed circuit  m i  2 m i  2

Dual-feed EM
j 2

 1      1
j  i 1 j i
m  j 1   j (7)
-40 m i 3 
23.0 23.5 24.0 24.5 25.0
Frequency (GHz) where Pr(i+1) = Pini+1Ginr / (Ginl + Ginr) is the radiated power of
Fig. 6. Comparisons of reflection coefficients of single-feed EM, dual-feed the element i+1. Based on the equations (3)-(7), the normalized
circuit and EM-simulated antenna.
magnitudes asi (i = 1, 2…10) of elements 1~10 of dual-feed
antenna with only port 1 excited is derived as 0.5503: 0.7211:
Pr1 Pr2 Pri Ginl Ginr Pri+1 PrN-1 PrN 0.9320: 1: 0.8281: 0.6633: 0.5237: 0.4000: 0.3024: 0.2390.
Similarly, due to the symmetry of two feeding ports, the
Pin Port1 magnitudes asi of the elements with only port 2 excited can be
written as follows: 0.2390: 0.3024: 0.4000: 0.5237: 0.6633:
Fig. 7. Model of power control of patch antenna array. 0.8281: 1: 0.9320: 0.7211: 0.5503. Similarly, the magnitudes asi
of patch elements under different feeding positions can also be
Subsequently, the radiation behavior is analyzed. Radiation obtained.
power control of single-feed antenna is shown in Fig. 7 since C. Dual-Feed Patch Antenna Array for Sum Beam
the one of dual-feed antenna array can be approximated as the
The sum beam is generated by simultaneously exciting two
linear superposition of two asymmetrical single-feed
ports with equal magnitude and in-phase signals. The array
counterparts. At resonance frequency f0, the power incident on
each patch element is partly radiated and partly transmitted to factor fs(, ) for single-feed antenna array is written as follows:
the subsequent sections [44], [45]. It is assumed that the mutual
N
coupling of elements, losses of metal and substrate is negligible. f s ( , )=  asi  e  j  d (sin  cos  ) (8)
The normalized radiated power Pr by each patch element is i 1
derived as [46]
2 2 where asi is the magnitude of the ith element, and asi is equal to
1  S11  S 21
Pr  2
(3) aSN+1-i due to the symmetry of center-feed antenna. However,
1  S11 for dual-feed antenna, the magnitude ai of element is actually

Authorized licensed use limited to: Onur Akarsu. Downloaded on October 06,2025 at 10:16:24 UTC from IEEE Xplore. Restrictions apply.
© 2025 IEEE. All rights reserved, including rights for text and data mining and training of artificial intelligence and similar technologies. Personal use is permitted,
but republication/redistribution requires IEEE permission. See https://www.ieee.org/publications/rights/index.html for more information.
 This article has been accepted for publication in IEEE Transactions on Antennas and Propagation. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TAP.2025.3613531

not equal to aN+1-i due to the radiation losses of patch elements. (d) exhibit smaller magnitudes and out-of-phase compared with
The array factor fsd(, ) for dual-feed antenna array is derived other patch elements. When two ports are simultaneously
as follows: excited, the current of dual-feed antenna array in Fig. 9(b) is
almost the superposition of the ones of asymmetrical feed
f sd ( , )=   a p1_ i  a p 2 _ i   e  j  d (sin  cos  )
N
(9) antenna array in Figs. 9(c) and (d). In Fig. 10, compared with
i 1 center-feed array, the enhanced magnitudes in the center region
of dual-feed array result in a reduction in SLL near broadside
where ap1_i and ap2_i are the magnitude of patch element under and a slight increase in the half-power beamwidth (HPBW).
the excited port 1 and port 2, respectively. Due to two ports The realized gains of both configurations are nearly identical,
being symmetrical about the center of array, fsd(, ) can be indicating that the function of power combining is realized in
further simplified as follows: the proposed dual-feed array. Meanwhile, different from the
a p1_ i  a p 2 _ N 1i , i  1, 2... N 2 (10) halved HPBW of the sum beam in monopulse arrays [20]-[29],
the HPBW of the proposed dual-feed antenna array is nearly
N /2 identical to that of single-feed counterpart.
f sd ( , )= 2(a p1_ i  a p1_ N 1i )  e j  d (sin  cos  ) (11)
i 1 2 3 4 5 6 7 8 9 10
1

Based on equations (8)-(11), Fig. 8 compares the calculated

radiation pattern of asymmetrical single-feed array with the one Port 1
(a)
of dual-feed counterpart. It can be seen that the radiation pattern
of dual-feed array is the linear superposition of the patterns
obtained when only port 1 and port 2 are individually excited.
Compared with the single-feed array, sidelobe level (SLL) of Port 1 Port 2
dual-feed counterpart decreases due to the enlarged magnitude (b)

in the middle elements.

0
Port 1 Port 1
(c)
Port 2
Dual port
-10
Normalized Gain (dB)

Port 2
(d)
-20

Port 1
-30 (e)
Fig. 9. Simulated current vector distributions of (a) center, and dual-feed
antenna array under different port excitations (b) both, (c) only port 1, (d) only
port 2, (e) asymmetric single-feed antenna array without right-side bent TL.
-40
-90 -60 -30 0 30 60 90
20
Theta () center-feed
Fig. 8. Calculated radiation patterns of asymmetrical single-and dual-feed dual-feed
antenna array.
10 only port 1
Figs. 9(a)-(d) show the EM-simulated current distributions of only port 2
center-and dual-feed antennas when different ports are excited. 0
Gain (dBi)

It can be found that when only port 1 or port 2 is excited, the

current distributions in Figs. 9(c) and (d) is not symmetrical due
to the losses of radiation and mismatch. As illustrated in Figs. -10
9(c) and 9(e), the magnitudes of patch elements 8~10 near
unexcited port are much smaller than other elements due to the -20
mismatching loss. This means that the number of active
elements under asymmetrical feeding is less than the one in
symmetrical feeding. Therefore, minor magnitudes of patch -30
-90 -60 -30 0 30 60 90
elements located outside of unexcited port, resulting from
impedance mismatching, are designated as negligible. It is also Theta ()
the reason why the operated bandwidth of dual-feed antenna is Fig. 10. Comparison of simulated radiation patterns of center-feed antenna with
wider than the one of single-feed counterpart in Fig. 6. dual-feed counterpart under different port excitations.
As shown in Fig. 10, when only port 1 or port 2 is excited,
the reduced maximum gain is not at broadside because three Then the impact of feeding positions on the sum beam of
patch elements farthest from the excited port in Figs. 9(c) and dual-feed antenna array is discussed. In Fig. 11, the calculated

Authorized licensed use limited to: Onur Akarsu. Downloaded on October 06,2025 at 10:16:24 UTC from IEEE Xplore. Restrictions apply.
© 2025 IEEE. All rights reserved, including rights for text and data mining and training of artificial intelligence and similar technologies. Personal use is permitted,
but republication/redistribution requires IEEE permission. See https://www.ieee.org/publications/rights/index.html for more information.
 This article has been accepted for publication in IEEE Transactions on Antennas and Propagation. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TAP.2025.3613531

radiation pattern of a dual-feed antenna for the sum beam is 1 2 3 4 5 6 7 8 9 10

investigated under different feeding positions of two excited

ports. It can be seen that the general trend, except for two Port 1 Port 2
excited ports located between elements 4 and 5, is that the closer (a)
two excited ports are, the smaller the SLL of dual-feed antenna
array is and the wider the beamwidth is. Figs. 12 and 13 show
EM-simulated radiation patterns and current distributions. The
simulated radiation pattern in Fig. 11 is overall consistent with Port 1 Port 2
the calculated one depicted in Fig. 12. The reason underlying (b)
this phenomenon is that the decrease of the distance between
two simultaneously excited ports results in the increase of the
difference between the magnitudes of center and side patch Port 1 Port 2
elements. As illustrated in Figs. 13(a)-(c), it can be further (c)
verified by a comparative analysis of the current distributions
for different feeding positions. However, when two excited
ports are located between elements 4 and 5, SLL of dual-feed
antenna deteriorates. This phenomenon is caused by the Port 1 Port 2
diminished discrepancy between the magnitudes of center and (d)
peripheral patch elements. Fig. 13. Current vector distributions of dual-feed antenna for sum beam with
two simultaneously excited ports of equal magnitude and in-phase, located
between the elements (a) 1 and 2 (9 and 10), (b) 2 and 3 (8 and 9), (c) 3 and 4
0 (7 and 8), (d) 4 and 5 (6 and 7).
element 1/2 element 3/4
element 2/3 element 4/5
element 1/2 element 3/4
-10 element 2/3 element 4/5
0
Normalized Gain (dB)

-20
-10
Normalized Gain (dB)

-30
-20

-40 -30

-50 -40
-90 -60 -30 0 30 60 90
Theta ()
-50
Fig. 11. Calculated radiation patterns of dual-feed antenna for sum beam with -90 -60 -30 0 30 60 90
two simultaneously excited ports of equal magnitude and in-phase, located Theta ()
between different elements. Fig. 14. Calculated radiation patterns of dual-feed antenna for difference beam
with two simultaneously excited ports of equal magnitude and out-of-phase,
20 located between different elements.
element 1/2 element 3/4
element 2/3 element 1/2 element 3/4
element 4/5
element 2/3 element 4/5
10 20

0 10
Gain (dBi)

0
-10
Gain (dBi)

-10
-20

-20
-30
-90 -60 -30 0 30 60 90
Theta () -30
-90 -60 -30 0 30 60 90
Theta ()
Fig. 12. EM-simulated radiation patterns of dual-feed antenna for sum beam Fig. 15. EM-simulated radiation patterns of dual-feed antenna for difference
with two simultaneously excited ports of equal magnitude and in-phase, located beam with two simultaneously excited ports of equal magnitude and out-of-
between different elements. phase, located between different elements.

Authorized licensed use limited to: Onur Akarsu. Downloaded on October 06,2025 at 10:16:24 UTC from IEEE Xplore. Restrictions apply.
© 2025 IEEE. All rights reserved, including rights for text and data mining and training of artificial intelligence and similar technologies. Personal use is permitted,
but republication/redistribution requires IEEE permission. See https://www.ieee.org/publications/rights/index.html for more information.
 This article has been accepted for publication in IEEE Transactions on Antennas and Propagation. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TAP.2025.3613531

1 2 3 4 5 6 7 8 9 10 larger magnitudes are. As illustrated in Fig. 16, the current

distributions of patch elements i (i = 1, 2…N/2) are of equal
Port 1 Port 2 amplitude and out of phase with the those of patch element
(a) N+1-i, thereby enabling the generation of a difference beam. A
trade-off in the selection of feeding positions is made by
choosing two excited ports located between elements 3 and 4
Port 1 Port 2 based on a comprehensive consideration of SLL of the sum in
(b) Fig. 12 and difference beam in Fig. 15.

Active Reflectional Coefficients (dB)

Port 1 Port 2
(c)
-5

Port 1 Port 2
(d) -10

Fig. 16. Current vector distributions of dual-feed antenna for difference beam
with two simultaneously excited ports of equal magnitude and out-of-phase,
located between the elements (a) 1 and 2 (9 and 10), (b) 2 and 3 (8 and 9), (c) 3 -15 Active S(1:1)
and 4 (7 and 8), (d) 4 and 5 (6 and 7).
Active S(2:2)
Active S(3:3)
D. Dual-Feed Patch Antenna Array for Difference Beam -20
Upon completing the design of the sum beam, the realization 23.0 23.5 24.0 24.5 25.0
of difference beam is possible by exciting two ports with equal Frequency (GHz)
amplitude and out-of-phase signals. As shown in Figs. 9(c)-(e),
the magnitudes of patch elements exhibit a decrease with the Fig. 17. Simulated active reflectional coefficients for triple-feed antenna array.
increase in distance from the source terminal. Due to two ports
being symmetrical about the center of array, the array factor 20
single-feed
fdd(, ) for the difference beam is derived as follows:
dual-feed
10 triple-feed
a p1_ i  a p 2 _ N 1i , i  1, 2... N 2 (12)

0
a p1_ i  a p 2 _ i  (a p1_ N 1i  a p 2 _ N 1i ), i  1, 2... N 2 (13)
Gain (dBi)

N /2 -10
f dd ( , )=  2  a p1_ i  a p 2 _ i   e  j  d (sin  cos  ) (14)
i 1
-20
According to the equations (12)-(14), the synthesized
radiation patterns of a dual-feed antenna for the difference beam, port1 port2 port3
as shown in Fig. 14, are investigated under different feeding -30
-90 -60 -30 0 30 60 90
positions of two excited ports. It can be observed that as the Theta ()
distance between the two excited ports decreases, the side lobe
level (SLL) of the dual-feed antenna array reduces, while the Fig. 18. Comparison of simulated patterns of single-dual-and triple-feed
beamwidth increases. Figs. 15 and 16 show EM-simulated antenna array for the sum beam with all simultaneously excited ports of equal
radiation patterns and current distributions, respectively. The magnitude and in-phase.
simulated radiation pattern in Fig. 15 is overall consistent with
the calculated one depicted in Fig. 14. The underlying reason E. Triple-Feed Monopulse Antenna Array
for this phenomenon is that the decrease of the distance between To further enhance the EIRP of system while maintaining the
two simultaneously excited ports results in an increase in the same HPBW as those of the single- and dual-feed antenna
number of patch elements with relatively larger magnitudes. As arrays, a three-feed configuration is implemented by adding one
illustrated in Figs. 16(a)-(d), it can be further verified by a extra feeding port onto the center of array, based on dual-feed
comparative analysis of the current distributions with varying antenna in Fig. 5(a). By turning the parameter r of the
feeding positions. Magnitudes of patch elements located transformer, three simultaneously excited ports are well
between two excited ports, suffering from impedance matched as evidenced by the simulated active reflectional
mismatching, are significantly smaller than that of patch coefficients in Fig. 17. Fig. 18 compares the simulated gain of
elements located outside. It is clear that the closer two excited three-feed antenna with those of single-and dual-feed
ports are, the more the number of patch elements with relatively counterparts for the sum beam. It can be found that small

Authorized licensed use limited to: Onur Akarsu. Downloaded on October 06,2025 at 10:16:24 UTC from IEEE Xplore. Restrictions apply.
© 2025 IEEE. All rights reserved, including rights for text and data mining and training of artificial intelligence and similar technologies. Personal use is permitted,
but republication/redistribution requires IEEE permission. See https://www.ieee.org/publications/rights/index.html for more information.
 This article has been accepted for publication in IEEE Transactions on Antennas and Propagation. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TAP.2025.3613531

difference between their beamwidths and gains is observed, Fig. 20 compares the performances of series-fed uniform
indicating the function of on-antenna power combining is arrays scaled from 1  8 to 1  32. Return losses in these antenna
successfully realized in three-feed configuration. Therefore, arrays varies little whereas the gains don’t exhibit the expected
compared with the transmitter system equipped with a single- theoretical improvement. This limitation in gain can be
and dual-feed antenna array, the EIRP of the transmitter system attributed to two primary factors: first, a substantial increase in
with a triple-feed antenna array exhibits an increase by 10  insertion loss of feeding line [29], and second, patch elements
Log10(3) ~ 4.8 dB and 10  Log10(3/2) ~ 1.8 dB, respectively, located at the array sides contribute less effectively to radiation
while maintaining the same beamwidth. In comparison with the compared with the central positioned elements, particularly in
SLL of single-and dual-feed antenna array, that of triple-feed the case of the series-fed tapered antenna array.
counterpart is improved because the magnitude difference
between the center and side portions of patch elements is 0
enlarged by the additional port placed at the center of the array.

sum beam1 sum beam2 difference beam -10 18

20 116

Return Loss (dB)

132
10 -20
18

116
0 -30
Gain (dBi)

132
-10
-40
23.0 23.5 24.0 24.5 25.0
Frequency (GHz)
-20
(a)
20
-30 18
-90 -60 -30 0 30 60 90
116
Theta () 10
132
Fig. 19. Simulated radiation patterns of triple-feed antenna array under three
Gain (dBi)

different operated modes. 0

Three different beams of three-feed antenna array are

generated by controlling the magnitudes and phases of three
-10
excited ports. As summarized in Table I, sum beam 1, which is
advantageous for long-range monopulse radar, is obtained by
simultaneously exciting all three ports with equal magnitude
and in-phase signals. Sum beam 2, suitable for middle-range -20
-90 -60 -30 0 30 60 90
monopulse radar, is achieved only by port 1 and 3 with equal
magnitude and in-phase signals. The discrepancy in the distance Theta ()
(b)
detection range of beams 1 and 2 is attributed to the different 21
EIRPs of the respective beams. As shown in Fig. 19, compared Theoretical gain
with SLL (depicted as the bule dash curve in Fig. 18) of dual- 20 Practical gain
feed antenna array, the one of sum beam 2 is deteriorated by 6.2 19
dB due to the parasitic radiation introduced by the newly added
bent g/2 TL located at port 2 [21]. The difference beam in Fig. 18
Gain (dBi)

19 is generated by simultaneously exciting port 1 and 3 with 17

equal magnitude and out-of-phase signals.
16
TABLE I
THREE DIFFERENT BEAMS OF THREE-FEED ANTENNA ARRAY UNDER 15
DIFFERENT EXCITED PORTS
14
Port Sum beam 1 Sum beam 2 Difference beam 13
Port 1 +1 +1 +1
18 116 124 132
Array Scale
Port 2 +1 0 0 (c)
Port 3 +1 +1 -1 Fig. 20. Simulated (a) return loss, (b) radiation pattern in xoz plane, (c) gain
against array scale for multi-feed uniform antenna array.

Authorized licensed use limited to: Onur Akarsu. Downloaded on October 06,2025 at 10:16:24 UTC from IEEE Xplore. Restrictions apply.
© 2025 IEEE. All rights reserved, including rights for text and data mining and training of artificial intelligence and similar technologies. Personal use is permitted,
but republication/redistribution requires IEEE permission. See https://www.ieee.org/publications/rights/index.html for more information.
 This article has been accepted for publication in IEEE Transactions on Antennas and Propagation. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TAP.2025.3613531

0 0.10 1.0 0
 = 0.05 mm  = +0.01 mm

Magnitude Difference (dB)

Magnitude difference
-5 Phase difference
 = 0.03 mm  = +0.03 mm

Phase difference ()

0.05 0.5
S parameters (dB)

-10
 = 0.01 mm  = +0.05 mm

Reflection coefficients (dB)

-15 0.00 0.0
-5
-20
S11 S22 -0.05 -0.5
-25 S21 S23
S31 S33
-30 -0.10 -1.0
23.0 23.5 24.0 24.5 25.0 23.0 23.5 24.0 24.5 25.0 -10
Frequency (GHz) Frequency (GHz)
(a)
0 1.0 -160
Magnitude Difference (dB)

Magnitude difference
-5 Phase difference

Phase difference ()

0.5 -170 -15
S parameters (dB)

-10 S11 S22

S21 S23
-15 S31 S33 0.0 -180
-20
-0.5 -190
-25 -20
23.0 23.5 24.0 24.5 25.0
-30 -1.0 -200
23.0 23.5 24.0 24.5 25.0 23.0 23.5 24.0 24.5 25.0 Frequency (GHz)
Frequency (GHz) Frequency (GHz)
(b)
(a)
0 0.6
0
Magnitude Difference (dB)

12
-5 0.0 S21S31
 S21S41

Phase difference ()  = 0.05 mm  = +0.01 mm
S parameters (dB)

-10 -0.6 2131 2141 9  = 0.03 mm  = +0.03 mm

-15 S11
 S31
 S22
 S24

-1.2
6  = 0.01 mm  = +0.05 mm
S21
 S41
 S23

3
Normalized Gain (dB)
-10
-20 -1.8
0
-25 -2.4
-3
-30 -3.0
23.0 23.5 24.0 24.5
Frequency (GHz)
25.0 23.0 23.5 24.0 24.5 25.0 -20
Frequency (GHz)
(c)
Fig. 21. Simulated S-parameters, magnitude and phase difference of (a) two- Semi-coax
way in-phase PD, (b) two-way out-of-phase PD, (c) three-way PD.
-30 
Signal via
-40
-90 -60 -30 0 30 60 90
20.25mm Theta ()
(b)
92.13mm
(a) Fig. 23. (a) simulated return loss, (b) radiation pattern against the offset  of
signal through via in the proposed triple-feed antenna.

III. MEASUREMENT AND DISCUSSION

The proposed multi-feed monopulse antenna array is
fabricated based on two Rogers 5880 PCB laminates, which are
(b) bonded together with Rogers 4450f bonding film. In fact, the
measurement of the performance of the proposed antenna array
is conducted by using a multi-way power divider rather than a
chip which is temporarily unavailable. The design of power
divider (PD) is detailed in [49], [50], and therefore not
elaborated upon here. Figs. 21 (a)-(c) show the simulated
performance of three kinds of PDs. The simulated magnitude
(c)
and phase difference of 2-way in-phase T-type PD are less than
0.05 dB and 0.5, respectively. The magnitude and phase
difference of 2-way out-of-phase T-type PD are less than 0.85
dB and (1805), respectively, while the maximum ones of 3-
way PD are about 1.3 dB and 8.4, respectively. The difference
of magnitude and phase has minimal impact on the performance
(d)
of antenna array. The dual-feed antenna array in Fig. 5(a) has
Fig. 22. Photographs of fabricated prototypes of proposed multi-feed antenna not been fabricated as its functionality is incorporated within
array. (a) Top view. Bottom view of (b) sum beam 1, (c) sum beam 2, (d) the triple-feed antenna in Fig. 18. The photographs of fabricated
difference beam. antenna prototypes are shown in Fig. 22. Triple-feed 1  10

Authorized licensed use limited to: Onur Akarsu. Downloaded on October 06,2025 at 10:16:24 UTC from IEEE Xplore. Restrictions apply.
© 2025 IEEE. All rights reserved, including rights for text and data mining and training of artificial intelligence and similar technologies. Personal use is permitted,
but republication/redistribution requires IEEE permission. See https://www.ieee.org/publications/rights/index.html for more information.
 This article has been accepted for publication in IEEE Transactions on Antennas and Propagation. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TAP.2025.3613531

10

antenna array is presented in Fig. 22(a). A three-way in-phase 0 20

power divider situated at the bottom of multi-layer PCB in Fig.
22(b) is utilized to achieve equal-amplitude and in-phase -5 10
conditions necessary for the operation of sum beam 1. As
shown in Figs. 22(c) and (d), two-way in-phase and out-of- -10 0
phase power dividers are essential for the operation of sum

S11 (dB)

Gain (dBi)
beam 2 and difference beam, respectively. -15 -10
Fig. 23 provides the simulated return loss and radiation
-20 -20
pattern against the offset  of signal via, which shows the
proposed structure is robust. The radiation pattern of the S11
-25 -30
fabricated antenna array is tested in compact range anechoic Gain
champers of size 20 m  10 m  10 m in Fig. 24. The antenna -30 -40
gain is obtained by comparing with a K-band standard horn [48]. 23.0 23.5 24.0 24.5 25.0
Frequency (GHz)
(a)
0 20

-5 10

-10 0

Gain (dBi)
S11 (dB)

-15 -10

-20 -20
Test antenna S11
-25 -30
Gain
-30 -40
23.0 23.5 24.0 24.5 25.0
Frequency (GHz)
(b)
0 20

-5 10
Fig. 24. Measurement setup in compact range anechoic champers of size 20 m
 10 m  10 m.
-10 0

Gain (dBi)
S11 (dB)

A. Impedance Bandwidth and Broadside Gain -15 -10

Fig. 25 shows the simulated and measured RLs and peak gain.
As illustrated in Fig. 25(a), the measured return loss S11 of the -20 -20
sum beam 1 in Fig. 22(b) is better than 10 dB with an impedance
-25 S11 -30
bandwidth exceeding 4.6% (ranging from 23.4 to 24.5 GHz).
The measured peak gain is 15.7 dBi at 24 GHz with 3-dB gain Gain
-30 -40
bandwidth wider than 7.5% (from 23.2 to 25 GHz). Compared 23.0 23.5 24.0 24.5 25.0
with simulated gains in Fig. 19, the measured gains are reduced Frequency (GHz)
due to the introduction of the power divider and semi-coaxial (c)
transitions. As shown in Fig. 25(b), the measured return loss Fig. 25. Simulated and measured return losses and gain of (a) sum beam 1, (b)
S11 of the sum beam 2 is better than 10 dB with impedance sum beam 2, and (c) difference beam. (simulation: dash, measurement: solid).
bandwidth wider than 7.5% (from 23.2 to 25.0 GHz). The
measured peak gain is 14.5 dBi at 24 GHz with 3-dB gain B. Radiation Pattern
bandwidth of over 8.3%. As illustrated in Fig. 25(c), the The simulated and measured radiation patterns of three
measured return loss S11 of the difference beam in Fig. 22(d) beams at 24 GHz are shown in Fig. 26. Compared with the
is better than 10 dB with impedance bandwidth wider than 6.2%. simulated radiation patterns of sum beam 1 and 2 in Fig. 19,
The measured peak gain is 12.9 dBi at 24 GHz with 3-dB gain SLL in Fig. 26 deteriorates somewhat due to some parasitic
bandwidth wider than 7.9%. The simulated radiation radiation introduced by the semi-coaxial structure. As
efficiencies of sum beam 1, sum beam 2, and difference beam illustrated in Figs. 26(a) and (b), the measured HPBW in xoz
are 86%, 62%, and 82%, respectively. Compared with the sum plane is 10.2 while the measured one in yoz plane is 67.0. The
beam 1, the radiation efficiency of the sum beam 2 decreases measured SLL of the sum beam 1 in xoz plane is 16.7 dB. For
due to the 50  resistor load. the sum beam 2, the measured HPBW in Fig. 26(c) is 7.4 while

Authorized licensed use limited to: Onur Akarsu. Downloaded on October 06,2025 at 10:16:24 UTC from IEEE Xplore. Restrictions apply.
© 2025 IEEE. All rights reserved, including rights for text and data mining and training of artificial intelligence and similar technologies. Personal use is permitted,
but republication/redistribution requires IEEE permission. See https://www.ieee.org/publications/rights/index.html for more information.
 This article has been accepted for publication in IEEE Transactions on Antennas and Propagation. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TAP.2025.3613531

11

0
the measured one in Fig. 26(d) is 64.6. The measured SLL of Simulated: Measured:
the sum beam 2 is only 8.4 dB, which is caused by the parasitic Co-pol Co-pol
X-Pol X-Pol

Normalized Gain (dB)

radiation of the semi-coaxial structure and bent g/2 TL located -10
at port 2. The measured SLL in Fig. 26(c) exhibits a slight
improvement over the simulated one due to the fabrication and -20
assembly tolerances inherent in the semi-coaxial structure. As
shown in Fig. 26(e), the measured null depth of the difference -30
beam is about 33.3 dB. It is found the difference beam is not
strictly symmetric about theta = 0, which is due to the errors in -40
-90 -60 -30 0 30 60 90
magnitude and phase of two-way in-phase power divider, Theta ()
leading to an amplitude imbalance between the two peaks in the (a)
0
difference beam [29]. The measured SLL in the difference
beam is about 10.7 dB. The measured cross polarizations of -10

Normalized Gain (dB)

three beams are less than 20 dB.
C. Comparison and Discussion -20

The important measured results and geometry characteristics

of the reported monopulse antenna arrays, along with the -30 Simulated： Measured：
Co-pol Co-pol
proposed multi-feed ones, are summarized in Table II. X-Pol X-Pol
Compared with waveguide or GWG-based monopulse antenna -40
-90 -60 -30 0 30 60 90
arrays [2], [4], [5], [8], [11], the proposed counterpart offers Theta ()
several advantages, including low cost, compact size, low 0
(b)
profile, relaxed manufacturing tolerance requirements, Simulated: Measured:
Co-pol Co-pol
seamless integration with front-end components. Multiple SIW X-Pol X-Pol
Normalized Gain (dB)
-10
subarrays,which constitute the monopulse array [12], [13], [20],
occupy relatively large circuit sizes. Additionally, an
-20
impedance transformer is typically required in order to interface
a chip with SIW components, resulting in increased size and
-30
design complexity. Co-layer monopulse comparator network is
utilized in antenna arrays [21], [22], [28], where parasitic
radiation of the comparator network could compromise the -40
-90 -60 -30 0 30 60 90
performance of the antenna to a greater or lesser extent. The Theta ()
(c)
scale of monopulse antenna arrays [22], [23], [25], [27], [28] 0
are limited to 2  2 with limited gain, and it scale is found to be
proportional to the corresponding comparator feeding network.
Normalized Gain (dB)

-10
Therefore, when it scales up to a large size, for example 2  8
array [27], 3232 array [29], the number of comparator feeding -20
networks also increase linearly, thereby increasing both the
circuit size and the design complexity. To enhance the gain of -30 Simulated： Measured：
antenna array, a patch element is replaced by 8  8 subarray [21] Co-pol Co-pol
and 10  8 subarray [24]. However, most monopulse antenna X-Pol X-Pol
-40
arrays [2]-[29] adopt a four-quadrant symmetrical layout, -90 -60 -30 0 30 60 90
Theta ()
typically incorporating at least two subarrays. Under the way of (d)
array-based spatial power combining, the beamwidth of the 0
Simulated: Measured:
monopulse array [2]-[29] for the sum beam is reduced by half Co-pol Co-pol
Normalized Gain (dB)

X-Pol X-Pol
compared with single element or subarray. However, without -10
increasing the antenna panel size, the proposed multi-feed
monopulse antenna array realizes low-loss on-antenna power -20
combining and maintains the same beamwidth with the single-
feed counterpart. Meanwhile, the need for a power combiner -30
and passive impedance matching network between antenna and
power amplifier is eliminated. Moreover, EIRP of transmitter -40
system integrated with the proposed multi-feed antenna array -90 -60 -30 0 30 60 90
Theta ()
exhibits an increase by 10  Log10(N) (N 3) dB whereas its (e)
single-feed counterpart achieves only 10  Log10(2) dB increase Fig. 26. Simulated and measured radiation patterns at 24 GHz. (a) xoz plane, (b)
yoz plane of sum beam 1, (c) xoz plane, (d) yoz plane of sum beam 2, (d) xoz
at the expense of larger circuit size and reduced beamwidth. plane of difference beam.

Authorized licensed use limited to: Onur Akarsu. Downloaded on October 06,2025 at 10:16:24 UTC from IEEE Xplore. Restrictions apply.
© 2025 IEEE. All rights reserved, including rights for text and data mining and training of artificial intelligence and similar technologies. Personal use is permitted,
but republication/redistribution requires IEEE permission. See https://www.ieee.org/publications/rights/index.html for more information.
 This article has been accepted for publication in IEEE Transactions on Antennas and Propagation. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TAP.2025.3613531

12

TABLE II
COMPARISON BETWEEN PROPOSED AND REPORTED MONOPULSE ANTENNA ARRAYS

Null
FBW Feeding Planar Array Size Gain SLL X-pol Power
Ref depth Dim. BW* EIRP
(%) network integration scale (000) (dBi) (dB) (dB) combinner
(dB)

[2] 6.5 waveguide difficult 32# 16.112.78.7 24 19 >19 30 2-D array-based half 10Log10(2)

[4] 8.7 waveguide difficult 1616 14.814.82.2 32.3 14 30 24 2-D array-based half 10Log10(2)

[5] 21 GWG difficult 1616 17.417.43 30 16 27 38 2-D array-based half 10Log10(2)

[8] 0.83 GWG difficult  4718.81.5 27.8 14 30.8 50 2-D array-based half 10Log10(2)

[11] 10 GWG difficult 2016 16162.4 >30.6 22 22 37 2-D array-based half 10Log10(2)

[12] 23 SIW difficult 44  19.4 15 30 -31.1 2-D array-based half 10Log10(2)

[13] 13.1 SIW difficult 44 9.310.60.3 15.8 >10 25 31 2-D array-based half 10Log10(2)

[20] 1.6 SIW difficult 44 3.14.40.2 16.0 >15 30 33 1-D array-based half 10Log10(2)

[21] 5.6 microstrip easy 1616 13.312.40.02 24.5 17  30 2-D array-based half 10Log10(2)

[22] 1.9 microstrip easy 22 3.23.20.03 10.6   30 2-D array-based half 10Log10(2)

2-layer
[23] 2.1 easy 22 1.61.80.04 12.5 19 20 27 2-D array-based half 10Log10(2)
microstrip
5-layer
[24] 2.9 easy 804 16.915.80.76 25.3 24 23 30 2-D array-based half 10Log10(2)
stripline
3-layer
[25] 13.1 easy 22  11.2 >13 19 <25 2-D array-based half 10Log10(2)
microstrip
4-layer
[26] 50 easy 44  ~13   48 2-D array-based half 10Log10(2)
microstrip
4-layer
[27]  easy 82  11.8 12.5 18 <17.6 2-D array-based half 10Log10(2)
microstrip

[28] 3.3 microstrip easy 22 1.92.30.03 8.5   18 2-D array-based half 10Log10(2)

8-layer
[29] 28.8 easy 3232 21.821.80.3 28.6 12 30 56.5 2-D array-based half 10Log10(2)
SICL
This 4-layer
>4.6 easy 110 1.67.40.07 15.7 16.7# 20 33.3 1-D on-antenna same 10Log10(3)
work SICL
#
element number of antenna array; *beamwidth in comparison with single subarray or element; #sum beam1

IV. CONCLUSION antenna array has the potential to be a valuable asset for
miniaturized mid-and long-range monopulse radar systems.
A multi-feed monopulse antenna array with compact size has
been investigated in this paper. An equivalent circuit and a REFERENCES
theory of power control are developed to analyze the array
[1] D. Barton and S. Sherman, Monopulse Principles and Techniques, 2nd
factor of the sum and difference beam. A good agreement ed., Norwood, MA, USA: Artech House, 2011.
between the calculated and EM-simulated pattern is observed. [2] Z. Cao, Y. Chen, and H. Meng, “A W-band two-dimensional monopulse
The desired sum and difference beams is achieved by sparse array antenna,” IEEE Trans. Antennas Propag., vol. 70, no. 10, pp.
controlling the magnitudes and phases of multiple excited ports. 9260-9269, Oct. 2022.
[3] Y. Chen, Z. Cao, L. Zhang, H. Guo, and H. Meng, “Bandwidth
Compared with conventional monopulse antenna array based enhancement of comparator applied in W-band circularly polarized
on a four-quadrant symmetrical layout, the proposed multi-feed monopulse antenna,” IEEE Antennas Wireless Propag. Lett., vol. 22, no.
antenna array achieves a sum beam with beamwidth that is 9, pp. 2175-2179, Sep. 2023.
twice as wide. Additionally, the circuit size is reduced by half, [4] S. Wu, J. Li, J. Yao, Y. Cao, Y. Li, and X. Y. Zhang, “3D-printed Ka-
band circularly-polarized monopulse antenna array with high gain and
as only one subarray is employed in this work. The results show low axial ratio,” IEEE Trans. Antennas Propag., vol. 72, no. 11, pp. 8282-
that on-antenna power combining is effectively achieved in the 8293, Nov. 2024.
proposed multi-feed antenna array without any power combiner. [5] A. Vosoogh, A. Haddadi, A. U. Zaman, J. Yang, H. Zirath, and A. A.
As a result, EIRP of transmitter system equipped with the Kishk, “W-band low profile monopulse slot array antenna based on gap
waveguide corporate-feed networks,” IEEE Trans. Antennas Propag., vol.
proposed multi-feed antenna array exhibits an enhancement of 66, no. 12, pp. 6997-7009, Dec. 2018.
10  Log10(N) (N  3) dB. The proposed multi-feed monopulse [6] J. L. Vazquez-Roy, A. Tamayo-Dominguez, E. Rajo-Iglesias, and M.
Sierra-Castaner, “Radial line slot antenna design with groove gap

Authorized licensed use limited to: Onur Akarsu. Downloaded on October 06,2025 at 10:16:24 UTC from IEEE Xplore. Restrictions apply.
© 2025 IEEE. All rights reserved, including rights for text and data mining and training of artificial intelligence and similar technologies. Personal use is permitted,
but republication/redistribution requires IEEE permission. See https://www.ieee.org/publications/rights/index.html for more information.
 This article has been accepted for publication in IEEE Transactions on Antennas and Propagation. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TAP.2025.3613531

13

waveguide feed for monopulse radar systems,” IEEE Trans. Antennas Monopulse Array Antenna with Compact Comparator Circuit,” IEEE
Propag., vol. 67, no. 10, pp. 6317-6324, Oct. 2019. Access, vol. 13, pp. 48359–48366, Mar. 2025.
[7] R. Shen, X. Ye, J. Xie, Z. Chen, and C. Jin, “A W-band circular box-horn [29] Z. Wang, Y. Hu, L. Xiang, J. Xu, and W. Hong, “A wideband high-gain
antenna array radiating sum and difference beams with suppressed planar monopulse array antenna for Ka-band radar applications,” IEEE
sidelobe,” IEEE Trans. Antennas Propag., vol. 67, no. 9, pp. 5934-5942, Trans. Antennas Propag., vol. 71, no. 11, pp. 8739–8752, Nov. 2023.
Sep. 2019. [30] K. K. Katare, S. Chandravanshi, A. Sharma, A. Biswas, and M. J. Akhtar,
[8] A. Tamayo-Dominguez, J.-M. Fernandez-Gonzalez, and M. Sierra- “Anisotropic metasurface-based beam-scanning dual-polarizated fan-
Castaner, “Monopulse radial line slot array antenna fed by a 3-D-printed beam integrated antenna system,” IEEE Trans. Antennas Propag., vol. 67,
cavity-ended modified Butler matrix based on gap waveguide at 94 GHz,” no. 12, pp. 7204–7215, Dec. 2019.
IEEE Trans. Antennas Propag., vol. 69, no. 8, pp. 4558-4568, Aug. 2021. [31] S. Li, T. Chi, J. S. Park, and H. Wang, “A multi-feed antenna for antenna
[9] X. Jiang, Y. Shi, W. Yu, F. Jia, X. Wang, and Q. Wu, “A compact Ka- level power combining,” in Proc. IEEE APS/URSI, Jun./Jul. 2016, pp.
band low-sidelobe monopulse antenna array based on mixed gap 1589–1590.
waveguide and hollow waveguide multilayer feeding network,” IEEE [32] B. Goettel, P. Pahl, C. Kutschker, S. Malz, U. R. Pfeiffer, and T. Zwick,
Trans. Antennas Propag., vol. 71, no. 11, pp. 8714-8725, Nov. 2023. “Active multiple feed on-chip antennas with efficient in-antenna power
[10] M. Ferrando-Rocher, J. I. Herranz-Herruzo, A. Valero-Nogueira, and B. combining operating at 200–320 GHz,” IEEE Trans. Antennas Propag.,
Bernardo-Clemente, “All-metal monopulse antenna array in the Ka-band vol. 65, no. 2, pp. 416–423, Feb. 2017.
with a comparator network combining ridge and groove gap waveguides,” [33] T. Chi, S. Li, J. S. Park, and H. Wang, “A multifeed antenna for high-
IEEE Antennas Wireless Propag. Lett., vol. 22, no. 6, pp. 1–5, Apr. 2023. efficiency on-antenna power combining,” IEEE Trans. Antennas Propag.,
[11] X. Cheng, X. Chen, J. Liu, X. Liu, J. Li, S. Gao, and A. A. Kishk, vol. 65, no. 12, pp. 6937–6951, Dec. 2017.
“Compact Dual-Polarized Low Sidelobe Monopulse Slot Antenna Array [34] W. Duan, X. Y. Zhang, S. Liao, K. X. Wang, and Q. Xue, “Multiport
Based on Gap Waveguide Technology,” IEEE Trans. Antennas Propag., power combining patch antenna with stable reflection coefficient and
vol. 73, no. 2, pp. 1215-1220, Feb. 2025. radiation pattern in six polarization states,” IEEE Trans. Antennas
[12] K. Han, Z.-X. Jiang, W. Liu, Z.-L. Deng, and G.-L. Huang, “Dual- Propag., vol. 67, no. 2, pp. 719–729, Feb. 2019.
Polarized SIW-Based Scalable Antenna Array and 2-D Monopulse [35] S. Bahrami, D. Lee, W. Hong, and H.-J. Song, “Excitation flexibility of
Beamforming Applications,” IEEE Antennas Wireless Propag. Lett., vol. on-antenna power combining,” IEEE Antennas Wireless Propag. Lett.,
24, no. 6, pp. 1392-1396, Jun. 2025. vol. 22, no. 6, pp. 1376–1380, Jun. 2023.
[13] J. Zhu, S. Liao, S. Li, and Q. Xue, “60 GHz substrate-integrated [36] T. Le Gall, A. Ghiotto, S. Varault, and B. Louis, “16-Port cavity-backed
waveguide-based monopulse slot antenna arrays,” IEEE Trans. Antennas aperture-coupled patch antenna for on-antenna power combining,” IEEE
Propag., vol. 66, no. 9, pp. 4860–4865, Sep. 2018. Trans. Antennas Propag., vol. 72, no. 3, pp. 2124–2134, Mar. 2024.
[14] B. Liu et al., “Substrate integrated waveguide (SIW) monopulse slot [37] S. Bahrami, D. Lee, W. Hong, and H.-J. Song, “Analysis of on-antenna
antenna array,” IEEE Trans. Antennas Propag., vol. 57, no. 1, pp. 275– power combining for realization of pattern-reconfigurable multifeed
279, Jan. 2009. microstrip antennas,” IEEE Trans. Antennas Propag., vol. 71, no. 4, pp.
[15] Y. J. Cheng, W. Hong, and K. Wu, “94 GHz substrate integrated 3091–3103, Apr. 2023.
monopulse antenna array,” IEEE Trans. Antennas Propag., vol. 60, no. 1, [38] J. Zhang, G. Xu, K. Jin, K. Kang, and W.-Y. Yin, “A generalized
pp. 121–129, Jan. 2012. scattering parameters analysis method for multi-port fed antenna and
[16] K. Tekkouk, M. Ettorre, and R. Sauleau, “Multibeam pillbox antenna amplitude-phase controlled arrays,” IEEE Trans. Antennas Propag. Early
integrating amplitude-comparison monopulse technique in the 24 GHz Access.
band for tracking applications,” IEEE Trans. Antennas Propag., vol. 66, [39] Z. Chen, X. Zhu, and L. Zhong, “Multi-port power combining grid array
no. 5, pp. 2616-2621, May 2018. antenna on fan-out wafer level packaging,” in Proc. IEEE MTT-S Int.
[17] M. Kuznetcov, S. K. Podilchak, and M. Sellathurai, “SIW sub-array Microw. Workshop Ser. Adv. Mater. Processes RF THz Appl. (IMWS-
antenna with high isolation offering dual-polarized monopulse patterns,” AMP), Jul. 2020, pp. 1-3.
IEEE Open J. Antennas Propag., vol. 5, no. 1, pp. 73–81, Feb. 2024. [40] A. G. Derneryd, “Linearly polarized microstrip antennas,” IEEE Trans.
[18] Z. Sun et al., “A dual-polarized 2-D monopulse antenna array based on Antennas Propag., vol. AP-24, no. 6, pp. 846–851, Nov. 1976.
substrate integrated waveguide,” IEEE Trans. Antennas Propag., vol. 70, [41] D. G. Babas and J. N. Sahalos, “Synthesis method of series-fed microstrip
no. 12, pp. 11771–11778, Dec. 2022. antenna arrays,” Electron. Lett., vol. 43, no. 2, pp. 78-80, Jan. 2007.
[19] W. Li, S. Liu, J. Deng, Z. Hu, and Z. Zhou, “A compact SIW monopulse [42] H. halili, K. Mohammadpour-Aghdam, S. Alamdar, and M. ammad-
antenna array based on microstrip feed,” IEEE Antennas Wireless Propag. Taheri, “Low-cost series-fed microstrip antenna arrays with extremely
Lett., vol. 20, pp. 93–97, 2021. low sidelobe levels,” IEEE Trans. Antennas Propag., vol. 66, no. 9, pp.
[20] Z. Sun, S. Liu, Z. Hu, W. He, X. Zheng, and Y. Yang, “A compact dual 4606–4612, Sep. 2018.
linearly-polarized single-plane monopulse antenna array based on SIW [43] H. Yi, L. Li, J. Han, and Y. Shi, “Traveling-wave series-fed patch array
and strip-line feed,” IEEE Trans. Antennas Propag., vol. 70, no. 8, pp. antenna using novel reflection-canceling elements for flexible beam,”
6732–6739, Aug. 2022. IEEE Access, vol. 7, pp. 111466–111476, Aug. 2019.
[21] H. Wang, D.-G. Fang, and X. G. Chen, “A compact single layer [44] A. Vallecchi and G. B. Gentili, “Design of dual-polarized series-fed
monopulse microstrip antenna array,” IEEE Trans. Antennas Propag., vol. microstrip arrays with low losses and high polarization purity,” IEEE
54, no. 2, pp. 503–509, Feb. 2006. Trans. Antennas Propag., vol. 53, no. 5, pp. 1791–1798, May 2005.
[22] N. Ahmad, N. Shoaib, H. Nawaz, Q. H. Abbasi, and S. Nikolaou, “Dual- [45] C. Min and C. E. Free, “Analysis of traveling-wave-fed patch arrays,”
polarized (CP, LP, and Slant) series-fed, monopulse planar array antenna,” IEEE Trans. Antennas Propag., vol. 57, no. 3, pp. 664–670, Mar. 2009.
IEEE Trans. Antennas Propag., vol. 72, no. 10, pp. 8070–8075, Oct. 2024. [46] Y.-H. Yang, J.-L. Guo, B.-H. Bao, Y.-M. Cai, and G.-N. Zhou, “The
[23] S. A. Khatami, J. Meiguni, A. Amn-e Elahi, and P. Rezaei, “Compact via- design of dual circularly polarized series-fed arrays,” IEEE Trans.
coupling fed monopulse antenna with orthogonal tracking capability in Antennas Propag., vol. 67, no. 1, pp. 574–581, Jan. 2019.
radiation pattern,” IEEE Antennas Wireless Propag. Lett., vol. 19, no. 8, [47] Y. J. Cheng, J. Wang, and X. L. Liu, “94 GHz substrate integrated
pp. 1443–1446, Aug. 2020. waveguide dual-circular-polarization shared-aperture parallel-plate long-
[24] L. Zou, X. Wang, and J. Zang, “Series-fed monopulse microstrip array slot array antenna with low sidelobe level,” IEEE Trans. Antennas
antenna with stripline quadrature hybrid comparator network,” IEEE Propag., vol. 65, no. 11, pp. 5855–5861, Nov. 2017.
Access, vol. 9, pp. 169177–169192, Dec. 2021. [48] C. A. Balanis, Antenna Theory Analysis and Design, 4th ed., Hoboken,
[25] H. Sun, X. Ge, W. He, and L. Zhao, “A reconfigurable antenna with sum NJ, USA: Wiley, 2016.
and difference patterns for WLAN access points,” IEEE Antennas [49] D. M. Pozar, Microwave Engineering, 4th ed. New York, NY, USA:
Wireless Propag. Lett., vol. 19, pp. 1073–1077, Jul. 2020. Wiley, 2011.
[26] S. Karamzadeh, “Polarization diversity controllable monopulse antenna [50] J. C. Chiu, J. M. Lin, and Y. H. Wang, “A novel planar three-way power
array design,” IEEE Access, vol. 11, pp. 7078–7084, Jan. 2023. divider,” IEEE Microw. Wireless Compon. Lett., vol. 16, no. 8, pp. 449–
[27] M. Qu, Z. Chen, J. Su, and Z. Li, “Design of polarization reconfigurable 451, Aug. 2006.
beam-scanning monopulse phased array,” IEEE Trans Antennas Propag.,
vol. 72, no. 3, pp. 2879–2884, Mar. 2024.
[28] N. Ahmad, M. Waseem, N. Shoaib, H. Nawaz, S. Shoaib, K. A. Aljaloud,
and R. Hussain, “Polarization Reconfigurable, Single-Layer, 2-D,

Authorized licensed use limited to: Onur Akarsu. Downloaded on October 06,2025 at 10:16:24 UTC from IEEE Xplore. Restrictions apply.
© 2025 IEEE. All rights reserved, including rights for text and data mining and training of artificial intelligence and similar technologies. Personal use is permitted,
but republication/redistribution requires IEEE permission. See https://www.ieee.org/publications/rights/index.html for more information.
 This article has been accepted for publication in IEEE Transactions on Antennas and Propagation. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TAP.2025.3613531

14

Chuanming Zhu received the Ph.D. degree Distinguished Design Award, and the IEEE/ACM ASP-DAC
in information and communication Best Design Award. He has been an Associate Editor and a
engineering from the Nanjing University of Guest Editor of the IEEE Solid–State Circuits Letters (SSCL),
Science and Technology, Nanjing, China, in a Guest Editor of the IEEE Journal of Solid–State Circuits
2018. He is currently working with Tsinghua (JSSC), and a Distinguished Lecturer of the IEEE Solid-State
University. His current research interests Circuits Society (SSCS).
include chip-in-package, and millimeter-
wave heterogeneous integration, antenna
array, and millimeter-wave CMOS Baoyong Chi (Senior Member, IEEE)
integrated circuits. received the B.S. degree in microelectronics
from Peking University, Beijing, China, in
1998, and the Ph.D. degree from Tsinghua
Haikun Jia (Member, IEEE) received the University, Beijing, in 2003. From 2006 to
B.S. and Ph.D. degrees in electronics 2007, he was a Visiting Assistant Professor
engineering from Tsinghua University, with Stanford University, Stanford, CA,
Beijing, China, in 2009 and 2015, USA. He is currently a Full Professor and
respectively. He is currently an Associate the Deputy Director of the School of Integrated Circuits,
Professor with the School of Integrated Tsinghua University. He has authored over 140 academic
Circuits, Tsinghua University. His research articles and two books. He holds more than 20 patents. His
interests are in the field of millimeter-wave current research interests include radio frequency
and high-speed circuit and system design, including power (RF)/millimeter-wave integrated circuit design, analog
amplifiers (PAs), voltage-controlled oscillators (VCOs), and integrated circuit design, and monolithic wireless transceiver
frequency modulated continuous-wave (FMCW) radars. chips for radar and communication.
Dr. Chi has been a Technical Program Committee (TPC)
Member of the IEEE Asian Solid-State Circuits Conference (A-
Wei Deng (Senior Member, IEEE) received SSCC) since 2005.
the B.S. and M.S. degrees in electronic
engineering from the University of
Electronic Science and Technology of China
(UESTC), Chengdu, China, in 2006 and
2009, respectively, and the Ph.D. degree in
electronic engineering from Tokyo Institute
of Technology, Tokyo, Japan, in 2013.
From 2013 to 2014, he was a Post-Doctoral Researcher with
Tokyo Institute of Technology. From 2015 to 2019, he was with
Apple Inc., Cupertino, CA, USA, working on radio frequency
(RF), millimeter-wave (mm-wave), and mixed-signal
integrated circuit (IC) design for wireless transceivers and
Apple A-series processors. Since 2019, he has been with the
School of Integrated Circuits, Tsinghua University, Beijing,
China, where he is currently an Associate Professor. He has
authored or co-authored more than 160 IEEE journal articles
and conference papers. His research interests include RF, mm-
wave, terahertz, and mixed-signal integrated circuits and
systems for wireless communications, radars, and imaging
systems.
Dr. Deng is a Technical Program Committee (TPC) Member
of the IEEE International Solid-State Circuits Conference
(ISSCC), the IEEE VLSI Symposium On Technology and
Circuits (VLSI), the IEEE Custom Integrated Circuits
Conference (CICC), the IEEE Asian Solid-State Circuits
Conference (A-SSCC), and the IEEE European Solid-State
Circuits Conference (ESSCIRC). He was a recipient or co-
Recipient of several national and international awards,
Including China Youth Science and Technology Innovation
Award, the IEEE SSCS Predoctoral Achievement Award,
Chinese Government Award For Outstanding Self-Financed
(nongovernment sponsored) Students Abroad, the Tejima
Research Award, the IEEE Radio Frequency Integrated Circuits
(RFIC) Symposium Best Student Paper Award, the A-SSCC

Authorized licensed use limited to: Onur Akarsu. Downloaded on October 06,2025 at 10:16:24 UTC from IEEE Xplore. Restrictions apply.
© 2025 IEEE. All rights reserved, including rights for text and data mining and training of artificial intelligence and similar technologies. Personal use is permitted,
but republication/redistribution requires IEEE permission. See https://www.ieee.org/publications/rights/index.html for more information.