Differentially Fed Dual-Band Base Station Antenna With Multimode Resonance and High Selectivity for 5G Applications

A dual-polarized antenna with multimode resonance and high selectivity is proposed in this article to cover the 5G sub-6 GHz bands. The proposed antenna achieves dual wide impedance bandwidth characteristics by incorporating a dual mode coupled patch and four planar coupled strips around the driven patch. Through the four resonant modes of these structures, the antenna effectively covers the two desired frequency bands. Moreover, the electric/magnetic coupling between the driven patch, dual mode coupled patch, and planar coupled strips enables the creation of three radiation nulls that suppress unwanted radiation. To further improve the out-of-band rejection level and half power beamwidth (HPBW), four shorted strips are introduced around the radiator. The introduction of these strips results in a fourth radiation null at higher out-of-band frequencies and expands the antenna’s HPBW from 52° to 62°. To demonstrate the feasibility of the design, both the proposed antenna and its array were manufactured and tested. Measured results show that the filtering element was able to operate within frequency bands of 3.24–3.83 GHz (16.7%) and 4.74–5.30 GHz (11.2%) with a reference of $\vert \text{S}_{\text {dd11}} \vert < -14$ dB. The input ports exhibited a high level of isolation, measuring 40 dB. Furthermore, the four radiation nulls proved effective in suppressing out-of-band radiation.

Abstract-A dual-polarized antenna with multimode resonance and high selectivity is proposed in this article to cover the 5G sub-6 GHz bands.The proposed antenna achieves dual wide impedance bandwidth characteristics by incorporating a dual mode coupled patch and four planar coupled strips around the driven patch.Through the four resonant modes of these structures, the antenna effectively covers the two desired frequency bands.Moreover, the electric/magnetic coupling between the driven patch, dual mode coupled patch, and planar coupled strips enables the creation of three radiation nulls that suppress unwanted radiation.To further improve the out-of-band rejection level and half power beamwidth (HPBW), four shorted strips are introduced around the radiator.The introduction of these strips results in a fourth radiation null at higher out-of-band frequencies and expands the antenna's HPBW from 52 • to 62 • .To demonstrate the feasibility of the design, both the proposed antenna and its array were manufactured and tested.Measured results show that the filtering element was able to operate within frequency bands of 3.24-3.83GHz (16.7%) and 4.74-5.30GHz (11.2%) with a reference of |S dd11 | < −14 dB.The input ports exhibited a high level of isolation, measuring 40 dB.Furthermore, the four radiation nulls proved effective in suppressing out-ofband radiation.

I. INTRODUCTION
W ITH the growing demand for high-speed data transfer rates and increasing numbers of connected devices, 5G base station antennas (from sub-6 GHz to millimeter-wave frequencies) have become an essential component of modern communication systems.A 5G sub-6 GHz base station antenna is required to support the sub-6 GHz frequency bands that have been allocated for 5G wireless communication by regulatory bodies around the world.These lower frequency band antennas are important for 5G because they can provide better coverage and penetration through obstacles, such as walls and buildings, and support for Internet of Things (IoTs) devices.
The frequency bands n78 (3.3-3.8GHz) and n79 (4.8-5.0GHz) are crucial sub-6 GHz frequency bands for 5G communication and have been adopted by numerous countries worldwide.Recently, some innovative antennas were designed in [1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], and [13] to cover part or single of these frequency bands.However, in integrated base station systems, single-band antennas are not sufficient for providing extensive frequency range coverage within the limited space.Thus, there is a growing interest in the research of dual/multiband antennas.In addition to the dual/multiband characteristics, filtering performance is also crucial for base station antennas to reduce the mutual coupling with other antennas.Therefore, the design of a dual-polarized antenna that possesses dual-frequency bands and filtering capability is of great significance for base station systems.
Traditionally, the filter and antenna were designed independently and then connected [14].While this approach could yield a good filtering response, the resulting combination tended to be bulky, making it unsuitable for the highly integrated communication systems that are prevalent today.In [15], by reshaping the final resonator of a traditional bandpass filter, a filtering antenna covering 2.26-2.66GHz was realized.In this design, the final resonator acts not only as a resonator of the bandpass filter, but also as the radiator of the antenna.By integrating two bandpass filters into an ultrawideband antenna, a novel reconfigurable filtering antenna with three switchable states was presented in [16].Similarly, by introducing filtering circuits on the feeding structures, some filtering antennas with good gain suppression levels were proposed in [17] and [18].However, the bandwidths of these antennas are relativity narrow.
Apart from the methods mentioned above, filtering antennae can usually be realized by introducing radiation nulls in unwanted frequency bands to suppress the radiation [19], [20], [21], [22], [23], [24].By introducing parasitic loops above and below the dipole arms, a dual-polarized filtering antenna with two radiation nulls was achieved in [19].This antenna can cover a wide frequency range from 1.66 to 2.73 GHz with a stable gain.In [21], a dual-polarized filtering antenna covering 2.3-2.9GHz was realized by using defected ground structures (DGSs).However, the radiation patterns of this antenna are not stable within the working band.By exciting the out-ofband resonant mode of the feeding structure and radiator, a novel filtering antenna without additional filtering structure was presented in [23].
Some dual-band filtering antennas were achieved in [25], [26], and [27] to work at the target 5G sub-6G frequency bands.However, all these antennas have single polarization.They are not suitable for base stations.In [28] and [29], two dual-band dual-polarized filtering antennas with band-notch characteristics were presented to cover the target frequency bands.The impedance bandwidths of them are 11.4% (lower band), 5.1% (higher band) and 19.8% (lower band), 13.2% (higher band).However, the out-ofband rejection levels of these antennas are relatively low.By introducing a filtering circuit on the feeding structure and etching slots on the radiator, a dual-band dual-polarized filtering antenna was achieved in [30].This antenna has a good gain suppression level at out-of-bands.However, the impedance bandwidth of its lower band is relatively narrow to cover the 3.3-3.8GHz band.
In this article, by integrating a dual mode coupled patch, four planar coupled strips, and four shorted strips around the driven patch, four resonant modes and four radiation nulls can be successfully realized to cover the n78 (3.3-3.8GHz) and n79 (4.8-5.0GHz) bands.Different from other parasitic patches, the coupled patch in this design has two resonant modes within the frequency band from 2.5 to 6.5 GHz.The additional resonant mode is obtained by simply etching a pair of crossing open slots on the patch.The resonant frequency of this mode can be modified by manually adjusting the parasitic loop located above the coupled patch.Four shorted strips are introduced to improve the out-of-band rejection level and half-power beamwidth (HPBW) of the antenna.This results in the appearance of a fourth radiation null, and an improvement in the HPBW from 52 • to 62 • .To confirm the design concept, a four-element array was also developed, produced, and evaluated.The experimental outcomes are consistent with the simulated data.The Ansys HFSS software was utilized to generate all simulated results.

II. CONFIGURATION OF THE ANTENNA ELEMENT
Fig. 1 provides a visual representation of the configuration for the proposed dual-band, dual-polarized filtering antenna element.The antenna is composed of a driven patch, four planar coupled strips, a coupled patch, a parasitic loop, and four shorted strips.All these structures are printed on Rogers 4003 substrates, as depicted in Fig. 1(a).The thickness of substrates 1 and 2 is 0.305 mm, while substrate 3 has a thickness of 0.813 mm.Fig. 1(b) shows the lateral view of the proposed antenna.The supports and superglue play a crucial role in ensuring the stability of each layer in this design.The supports serve two primary purposes.First, they enable the realization of the shorting strip through PCB technology.Second, they maintain the proper spacing between the three substrates.By appropriately sizing the square grooves on the substrates, the supports securely clamp the substrates in place.In addition, applying superglue at the junction of the square groove and the support eliminates the need for screws, providing a stable structure for the antenna.Fig. 1(c) and (d) shows the detailed dimensions for each component.By feeding the proposed two-layer radiator with four 50 microstrip lines, a dual-band dual-polarized antenna element with filter response is presented.The detailed design parameters of the microstrip lines are shown in Fig. 1(e).The S-parameters of the differential ports can be obtained by using the following equations: The excitation of the multiresonant modes and the introduction of the radiation nulls will be interpreted in Section III.

III. RADIATION NULLS AND MULTIRESONANCE
The simulated reflection coefficient and boresight realized gain of the proposed antenna could be observed in Fig. 2. As depicted, the proposed antenna can cover the desired n78 (3.3-3.8GHz) and n79 (4.8-5.0GHz) bands with four resonances.The simulated gains within the operating bands can achieve 8.8 and 9.6 dBi.Besides, this antenna has four radiation nulls within the frequency band 2.5-6.5 GHz.Two of them are between the two target 5G Sub-6 GHz bands.The other two are at higher out-of-band.The first, second, and third radiation nulls in this design are realized by introducing the coupled resonators, while the fourth radiation null is introduced by the shorted strips.To explain the working principles of the proposed antenna, coupled resonator filter topology will be introduced first.
Coupled resonator topology is a grouping of resonant circuits that are interconnected in a way that enables energy exchange and interaction.Multiple resonators are connected to one another in this topology via a common coupling element, such as a capacitor or an inductor.The coupling component enables energy exchange and mutual behavior modification between the resonators.In coupled resonator topology, there are several paths from the source to the load.Transmission zeros can be obtained if the signals in the paths are equal in amplitude but opposite in phase [31].Undoubtedly, this technique proves to be effective in introducing resonant modes and radiation nulls for the design of broadband/multiband filtering antennas.Nevertheless, the challenges lie in the introduction and allocation of the coupled resonators.
It can be observed from Fig. 2 that there are four resonant modes in this design.The first (M 1 ) resonant mode is the TM 10 mode of the driven patch.The second (M 2 ) and fourth (M 4 ) resonant modes are TM 10 and open slot mode of the coupled dual-mode patch.The third (M 3 ) resonant mode is the resonant mode of the planar coupled strip.All these resonant modes can be equivalent to parallel LC resonators.At their resonant frequencies, the phase of the transmission coefficient will be 0, and the RF signal will be transmitted to the free space.However, this shunt LC pair will show different properties at off-resonance frequencies.If the RF signal is above the resonant frequency, the resonator will provide a −90 • phase shift.If the RF signal is below the resonant frequency, the phase shift will be +90 • .It can be concluded as follows [31]: Since all these coupled resonators (coupled structures) are coupled to the main resonator (driven patch), the coupling between the driven patch and the coupled structures is assumed as the main part when calculating the phase shift.The interference among coupled resonators will be ignored.Based on the above information, the radiation nulls of this antenna will be analyzed using multipath topology.

A. First Radiation Null
The first radiation null in this design is caused by the coupled patch.Due to the coupling between the driven patch and the coupled patch, the TM 10 mode (M 2 ) of the coupled patch can be excited.Thus, here are two paths from the input port to the output port around f M2 .They are P i -M 1 -M 2 -P o and P i -M 1 -P o .Since both paths contain the same part, P i -M 1 , it will not be included when calculating the phase difference between the main path and the coupled path.In addition, the path from the resonators to the output port will not contribute to the phase shift [4].Thus, we only need to take the phase shift of the coupling and the phase shift provided by the coupled resonator into consideration.
The simulated electric and magnetic field distributions of the proposed antenna on the reference planes at f M2 are shown in Fig. 3. Since the E-field vector does not extend from the driven patch to the coupled patch, there is no electric coupling between them.In contrast, the coupling between the driven patch and the coupled patch can be classified as a magnetic coupling, as the H -field forms a closed loop.
The phase shift of the two dominant paths is given in Table I.Without loss of generality, the phase shift of the parasitic path (P i -M 1 -P o ) is assumed to be 0 • .Due to the magnetic coupling between the driven patch and coupled patch, a −90 • is introduced in the main path (P i -M 1 -M 2 -P o ).Then, considering the phase shift that the second resonant mode provided, a transmission zero will appear at the frequencies above f M2 .The simulated boresight realized gain of the proposed antenna under different L 1 is shown in Fig. 4. As can be observed, the first radiation null can be adjusted by changing the length of the L 1 with nearly no effect on other radiation nulls.

B. Second Radiation Null
The second radiation null in this design is obtained by the planar coupled strip.These coupled strips can be interpreted as resonators with a length of λ g /2 (where λ g represents the guided wavelength) [32].Fig. 5 illustrates their electromagnetic characteristics, which resemble those of printed dipole antennas.When the parasitic strip resonates, the E-field reaches its maximum value at both ends of the strip.Conversely, there is an E-field null in the middle of the strip.However, the current distribution exhibits the opposite pattern, with the minimum current occurring at the strip's ends and the maximum current occurring at its center.The coupling between the driven patch and the parasitic strip can excite a new resonant mode (M 3 ).Thus, a new path (P i -M 1 -M 3 -P o ) for the RF signal can be achieved around f M3 .Since the phase shifts provided by the coupled resonator are fixed in the lowerand upper-frequency regions.The phase difference between

TABLE II PHASE SHIFT BETWEEN THE MAIN AND COUPLED PATHS
the main path (P i -M 1 -M 3 -P o ) and the parasitic path (P i -M 1 -P o ) is decided by the coupling method.
As mentioned above, when the parasitic strip resonates, the maximum E-field value appears at its two open ends.Therefore, electric coupling can be easily realized by placing one of the ends of the planar coupled strip close to the driven patch.The simulated E-field distribution on reference plane 3 at f M3 is shown in Fig. 6.The electric coupling between the driven patch and planar coupled strip can be confirmed since the E-field vector points from the driven patch to the parasitic strips.Fig. 7 shows the simulated boresight realized gain of the proposed antenna under different L 2 .By increasing the length of the planar coupled strips, the second radiation null can be easily shifted to a lower frequency band with little influence on other radiation nulls.

C. Third Radiation Null and Open Slot Mode
The third radiation null in this design is also introduced by the coupling between the driven patch and the coupled patch.Different from traditional dual-layer microstrip patch antenna, the coupled patch in this design has two resonant modes (TM 10 mode and open slot mode).Thus, two radiation nulls can be introduced by this single coupled patch.
To explain the working principles of the coupled dual-mode patch more effectively, two reference antennas are designed and simulated.As shown in Fig. 8, the open slots and parasitic polygon rings were removed in reference antenna 1 (Ant-1).At the same time, only the parasitic polygon ring was removed in reference antenna 2 (Ant-2).When the input impedance of the reference antennas in Fig. 8   the coupled patch of reference antenna 2 is shown in Fig. 9(a).As can be seen, the E-field density concentrates at the center part of the coupled patch.There are two E-field nulls along the slots.
The introduction of the resonant mode at 6 GHz did not lead to a wide impedance bandwidth due to a significant frequency ratio (FR) between the strip mode and the new mode.To address this issue, a parasitic loop was added above the slot-loaded coupled patch to reduce the FR.The impact of the polygon loop is illustrated in Fig. 8(c), showing a shift in the resonant frequency of the open slot mode from 6 to 5.2 GHz.Furthermore, Fig. 9(b) indicates that the antiphase point of the electric field moved toward the edges, resulting in a resonant frequency shift toward a lower frequency.
Since the maximum E-field of the driven patch is at both ends, and the slot is in the center part.The coupling between the driven patch and the slots is magnetic coupling.This can be confirmed by the simulated H -field distribution on reference plane 2 in Fig. 10.Then, the phase shift of the main path around f M4 can be calculated.Table III indicates that the main and secondary paths are in phase below f M4 but become out-of-phase above f M4 .This indicates the presence of a transmission zero above f M4 .The boresight realized gain   of the proposed antenna with varying gap sizes (G) is shown in Fig. 11.As the gap size increases, the third radiation null shifts toward a higher frequency band while the other radiation nulls remain unaffected.

D. Fourth Radiation Null and Enhancement of HPBW
The fourth radiation null in this design is introduced by the shorted strips.However, different from other radiation nulls, this radiation null is not caused by the coupled resonator.The shorted strip can be equivalent to an inductor with one end connected to the ground.As shown in Fig. 12, the open end of the shorted strip exhibits the highest electric field intensity and is in close proximity to the open ends of the driven patch, where the maximum electric field also occurs.Thus, there will be electric coupling between the driven patch and shorted strips.The equivalent circuit of the shorted strip is shown in Fig. 12(c).At resonance, the series LC circuit acts as a short circuit, causing all power to be reflected to the source.Thus, a radiation null can be obtained.Fig. 13 displays  the relationship between the height of the strip (H s ) and the frequency of the fourth radiation null.As can be seen, when H s increases, the fourth radiation null shifts toward the lower frequency band without influencing other radiation nulls.
Apart from filtering performance and impedance bandwidth, the HPBW plays a crucial role in evaluating the performance of a dual-band, dual-polarized filtering base station antenna.This parameter is closely associated with the coverage of the base station antenna.The incorporation of these shorted strips has a dual benefit of not only introducing a radiation null but also significantly enhancing the HPBW of the proposed antenna.Specifically, Fig. 14 demonstrates that the HPBW at the higher band can be effectively increased to approximately 62 • while having minimal impact on the HPBW at the lower band.

E. Results and Comparison
Fig. 15 shows the fabricated prototype of the proposed antenna.The performance of this fabricated antenna element was evaluated through S-parameter and far-field measurements conducted in the Antenna Laboratory at the University of Kent, using the Keysight P9377B Vector Network Analyzer  and anechoic chamber.Fig. 16 illustrates the comparison between the measured and simulated S-parameters, revealing a close agreement between the two sets of results.From the measured results, it can be observed that the fabricated antenna operates effectively in two distinct frequency bands, spanning 3.24-3.83GHz and 4.74-5.30GHz, respectively.The two input ports of the antenna are well isolated with a measured isolation value of higher than 40 dB within the operating frequency bands.The proposed antenna achieves a maximum realized gain of 8.1 and 8.8 dBi in the lower band and higher band, respectively.Furthermore, the realized gain outside of the two desired frequency bands can be effectively suppressed through the four radiation nulls.
The radiation patterns of the fabricated antenna in the horizontal plane at various frequencies are displayed in Fig. 17  Table IV provides a comparison of the key factors of the proposed antenna with those of previously published reference antennas.The antennas in [25] and [27] can cover two separate frequency bands with good out-of-band rejection levels.However, both of these antennas are single-polarized antennas, which are not suitable for base station applications in 5G/6G.A dual-polarized antenna is not a simple combination of two single-polarized antennae with two orthogonal polarization ports.The design of a dual-polarized antenna is more difficult than the design of a single-polarized one.The antenna in [28] realizes a higher gain than the proposed antenna with a larger radiator size.However, the impedance bandwidths of this antenna are 11.4% and 5.1% with a reference level of VSWR < 2 (it should be |S 11 | < −14 dB in base station applications).It is worth noting that this design can not fully cover the target 5G sub-6GHz bands (3.3-3.8 and 4.8-5.0GHz).Besides, our design obtains three additional radiating nulls compared to this antenna, which only exhibits a single radiating null.The design in [29] obtains a smaller radiator size than the proposed antenna.However, the isolation between the input ports is only 20 dB.When compared with the design presented in [30], our antenna exhibits a 35.7% wider impedance bandwidth in the lower band and a 47.4% wider impedance bandwidth in the higher band.The aperture of our antenna is only 81.8% of the size of the design in [30].Moreover, the higher band HPBW of the antenna in reference [30] is only 52 • , which is significantly narrower than that of the proposed antenna.Its narrow HPBW in the higher band limits its practical use in base station systems.Furthermore, the proposed design has more independently adjustable radiation nulls than the antenna in [30] without introducing an extra circuit.
After a thorough comparison with other designs in the table, it can be concluded that the proposed antenna has superior overall performance and exhibits significant potential for employment in 5G base stations.

IV. ARRAY ANTENNA
As shown in Fig. 18, by arranging four proposed antenna elements in a row, a 1 × 4 linear array antenna is realized.
Authorized licensed use limited to the terms of the applicable license agreement with IEEE.Restrictions apply.The spacing between these antenna elements is 60 mm.The entire array antenna's footprint is 240 × 65 mm.As depicted in Fig. 19, a prototype of the proposed array antenna was fabricated for the purpose of confirming the design concept.
The measured results in Fig. 20 show that the proposed array antenna can cover two wide frequency ranges of 3.26-3.82and 4.72-5.31GHz with a reference reflection coefficient level of −14 dB.Within these operation bands, the proposed array antenna realizes a peak gain of 13.2 and 14.6 dBi.The variation between the measured gains and simulated gains is primarily attributed to measurement error and additional insertion loss introduced by the feeding network, coaxial cables, and SMA connectors.Besides, the radiation outside the two operation bands is successfully suppressed by the radiation nulls.
The normalized radiation patterns of the proposed array antenna were also tested and compared with the simulated results.The two planes of most interest for ±45 • dualpolarized base station antennas are the horizontal plane and vertical plane; hence, only these two planes were selected during the test.The measured radiation patterns of the fabricated array antenna agree well with the simulated results.However, there are some minor differences in the crosspolarization level, which could be caused by radiation from the cables and connectors.At the center frequency of the lower band, the proposed antenna exhibits an HPBW of 69 • in the horizontal plane and 18 • in the vertical plane, as depicted in Fig. 21(a) and (b).Fig. 21(c) and (d) illustrates that the proposed antenna has a horizontal plane HPBW of 62 • and a vertical plane HPBW of 12 • at the center frequency of the higher band.

V. CONCLUSION
The main emphasis of this article is the development of a dual-band, dual-polarized filtering antenna utilizing coupled structures.The integration of a dual-mode coupled patch and four planar coupled strips surrounding the driven patch enables the attainment of four resonant modes and three radiation nulls.Further enhancement of the antenna's out-of-band rejection level and HPBW is achieved by introducing four shorted strips.As a result, a fourth radiation null is observed, and the HPBW is improved from 52 • to 62 • .The measured results for both the antenna element and its array were found to be consistent with the simulated results, indicating the accuracy of the design.Results from the testing also revealed that the proposed designs could effectively cover the targeted 5G spectrum for numerous countries.With its dual-wideband, dual-polarized filtering capabilities, this antenna is a promising candidate for use in 5G base-station applications.

Fig. 1 .
Fig. 1.Structure of the proposed design.(a) Component breakdown view.(b) Lateral view.(c) Overhead view of the top substrate.(d) Overhead view of the middle substrate.(e) Transmission line layer.

Fig. 3 .
Fig. 3. (a) Configuration of the reference planes.(b) Simulated E-field distribution on reference plane 1. (c) Simulated H -field distribution on reference plane 2 at f M2 .

Fig. 5 .
Fig. 5. Simulated (a) E-field distribution and (b) current distribution on the parasitic strips at 5.2 GHz.
(a) and (b) is compared, it becomes apparent that the etched slots on the coupled patch have caused a new resonant mode to be excited at approximately 6 GHz.The simulated E-field distribution of

Fig. 9 .
Fig. 9. Simulated E-field distributions on the coupled patch of (a) reference antenna 2 and (b) proposed antenna.

Fig. 12 .
Fig. 12.(a) Configuration of the reference plane 4. (b) Simulated E-field distribution on reference plane 4. (c) Equivalent circuit of the shorted strip.

Fig. 16 .
Fig. 16.Measured and simulated S-parameters and boresight realized gain of the proposed antenna element.
, revealing stable radiation patterns within the operating frequency bands.The measured HPBW of the antenna is around 68 • (±4 • ) in the lower band and approximately 62 • (±3 • ) in

Fig. 18 .
Fig. 18.Configuration of the proposed array antenna.thehigher band.The cross-polarization levels of the measured antenna are found to be less than −17 dB over a range of ±90 • and less than −34 dB at the boresight direction.TableIVprovides a comparison of the key factors of the proposed antenna with those of previously published reference antennas.The antennas in[25] and[27] can cover two separate frequency bands with good out-of-band rejection levels.However, both of these antennas are single-polarized antennas, which are not suitable for base station applications in 5G/6G.A dual-polarized antenna is not a simple combination of two single-polarized antennae with two orthogonal polarization ports.The design of a dual-polarized antenna is more difficult than the design of a single-polarized one.The antenna in[28] realizes a higher gain than the proposed antenna with a larger radiator size.However, the impedance bandwidths of this antenna are 11.4% and 5.1% with a reference level of VSWR < 2 (it should be |S 11 | < −14 dB in base station applications).It is worth noting that this design can not fully cover the target 5G sub-6GHz bands (3.3-3.8 and 4.8-5.0GHz).Besides, our design obtains three additional radiating nulls compared to this antenna, which only exhibits a single radiating null.The design in[29] obtains a smaller radiator size than the proposed antenna.However, the isolation between the input ports is only 20 dB.When compared with the design presented in[30], our antenna exhibits a 35.7% wider impedance bandwidth in the lower band and a 47.4% wider impedance bandwidth in the higher band.The aperture of our antenna is only 81.8% of the size of the design in[30].Moreover, the higher band HPBW of the antenna in reference[30] is only 52 • , which is significantly narrower than that of the proposed antenna.Its narrow HPBW in the higher band limits its practical use in base station systems.Furthermore, the proposed design has more independently adjustable radiation nulls than the antenna in[30] without introducing an extra circuit.After a thorough comparison with other designs in the table, it can be concluded that the proposed antenna has superior overall performance and exhibits significant potential for employment in 5G base stations.

Fig. 21 .
Fig. 21.Measured and simulated normalized radiation patterns of the proposed array antenna at (a) 3.6 GHz in horizontal plane, (b) 3.6 GHz in vertical plane, (c) 5.0 GHz in horizontal plane, and (d) 5.0 GHz in vertical plane.

TABLE III PHASE
SHIFT BETWEEN THE MAIN AND COUPLED PATHS

TABLE IV COMPARISON
OF THE PREVIOUSLY PRESENTED DUAL-BAND ANTENNAS