High-Gain Millimeter-Wave Patch Array Antenna for Unmanned Aerial Vehicle Application

sensors

Communication High-Gain Millimeter-Wave Patch Array Antenna for Unmanned Aerial Vehicle Application

Kyei Anim 1 , Jung-Nam Lee 2 and Young-Bae Jung 1,*

1 Electronics Engineering Department, Hanbat National University, Daejeon 34158, Korea; [email protected] 2 Electronics and Telecommunications Research Institute (ETRI), Daejeon 34129, Korea; [email protected] * Correspondence: [email protected]; Tel.: +82-42-821-1136

Abstract: A high-gain millimeter-wave patch array antenna is presented for unmanned aerial vehicles (UAVs). For the large-scale patch array antenna, microstrip lines and higher-mode surface wave radiations contribute enormously to the antenna loss, especially at the millimeter-wave band. Here, the element of a large patch array antenna is implemented with a substrate integrated waveguide (SIW) cavity-backed patch fed by the aperture-coupled feeding (ACF) structure. However, in this case, a large coupling aperture is used to create strongly bound waves, which maximizes the coupling level between the patch and the feedline. This approach helps to improve antenna gain, but at the same time leads to a signiﬁcant level of back radiation due to the microstrip feedline and unwanted surface-wave radiation, especially for the large patch arrays. Using the SIW cavity-backed patch and stripline feedline of the ACF in the element design, therefore, provides a solution to this problem. Thus, a full-corporate feed 32 × 32 array antenna achieves realized gain of 30.71–32.8 dBi with radiation efﬁciency above 52% within the operational band of 25.43–26.91 GHz. The fabricated antenna also retains being lightweight, which is desirable for UAVs, because it has no metal plate at Citation: Anim, K.; Lee, J.-N.; Jung, the backside to support the antenna. Y.-B. High-Gain Millimeter-Wave Patch Array Antenna for Unmanned Keywords: aperture-coupled feed; cavity backed patch; gain; lightweight; millimeter-wave; stripline; Aerial Vehicle Application. Sensors substrate integrated waveguide; surface wave; unmanned aerial vehicle 2021, 21, 3914. https://doi.org/ 10.3390/s21113914

Academic Editors: Youchung Chung 1. Introduction and Cynthia M. Furse Unmanned aerial vehicles (UAVs) and drone technologies have seen a rapid growth Received: 7 May 2021 of interest due to the breakthroughs in microprocessors and artificial intelligence (AI) [1,2] Accepted: 3 June 2021 to make them more accessible and affordable. UAVs can operate either autonomously Published: 6 June 2021 or remotely controlled [3,4] to carry out tasks such as search and rescue operations, security patrol, cargo transportation, and provide seamless wireless coverage where no Publisher’s Note: MDPI stays neutral communication infrastructure is available [5–7]. with regard to jurisdictional claims in UAV antennas must, therefore, achieve a good trade-off between the commercial published maps and institutional affil- criteria and technical design issues. This includes high performances (i.e., gain), low iations. power consumption, restrictive physical properties (e.g., size and weight), and low cost [8]. Owing to the trend of miniaturized electronic devices and increasing demands for high data rates, microstrip patch antennas (MPAs) are better suited for UAV technologies than other antenna types, especially at the millimeter-wave (mm-wave) band. This is because, Copyright: © 2021 by the authors. at mm-wave frequencies, the MPAs become small and lightweight. However, the standard Licensee MDPI, Basel, Switzerland. MPAs having huge radiation aperture contradict the high-gain requirement for the UAV This article is an open access article applications due to the lower radiation efficiency. This is because of the significant losses distributed under the terms and attributable to higher-mode surface wave radiation, leaky-wave radiation from the feeding conditions of the Creative Commons network, and ohmic loss by the dielectric substrate, especially in higher frequencies. Thus, Attribution (CC BY) license (https:// a large-scale array of standard MPA cannot achieve superior gain results in the millimeter- creativecommons.org/licenses/by/ wave band. A more innovative design must be investigated. 4.0/).

Sensors 2021, 21, 3914. https://doi.org/10.3390/s21113914 https://www.mdpi.com/journal/sensors Sensors 2021, 21, 3914 2 of 11

Several high-gain horn antennas have been reported in the literature with promising electrical and physical features [9–11]. For example, they have a compact size for flush- mounted installation on the UAV platform to improve aerodynamics and broadband for increased capacity. The horn antenna gain ranges up to 25 dBi, with 10~20 dBi being typical. Nonetheless, some UAV applications demand extremely high antenna gain above 25 dBi in order to have a reliable communication link with the ground control stations in higher frequency bands. This is because the path losses increase with the increase of operating frequency and distance. In such a situation, the horn antennas cannot be chosen over the patch array antenna with large radiation aperture. In this paper, a corporate-fed 32 × 32 patch array antenna is developed at the millimeter-wave band for UAV applications. To achieve higher gain, the element’s aperture- coupled feed (ACF) is modified by increasing the coupling slot size. This approach increases the coupling levels between the patch and the feedline to improve radiation efficiency. However, at the same time, strong surface waves and back radiation due to the feedline oc- cur, especially in the large-scale patch arrays in higher frequency bands. This phenomenon causes significant antenna losses to undermine the gain improvement. Hence, the element antenna is implemented with SIW cavity-backed patch to create strongly bound input power and suppress any undesired higher-mode surface wave. Also, a stripline feedline is used in the ACF scheme to minimize back radiation level. The experimental results indicated a high gain of 30.71–32.8 dBi with the radiation efficiency of at least 52% within operational band of 25.43–26.91 GHz. The fabricated antenna retains low weight because it has no metal plate to back the antenna’s bottom ground plane. Thus, the external WR-28 waveguide-to-coax adapter is mounted directly at the antenna’s backside for measurement.

2. Radiating Element Design The geometry of the proposed array element is shown in Figure1, which is a multilayer structure with three dielectric substrates made of Taconic TLY-5 (i.e., Substrate #1, Substrate #2, and Substrate #3, εr = 2.2, tanδ = 0.0009, height h = 0.508 mm). It also consists of three copper layers, i.e., top layer GND #1, middle or isolating layer GND #2, and bottom layer GND #3. The radiating patch resides on substrate 1, as shown in Figure1a. Several metal vias spaced along the rectangular opening are installed in the substrate and connected to the top and middle ground planes (GND #1 and GND #2), thus forming the SIW cavity that is the backing of the patch. The metal vias’ diameter dv, and spacing sv are carefully optimized to reduce the leakage loss and provide feasible fabrication. The rectangular patch on the top surface of substrate #1 is fed by a stripline feedline integrated on substrate #2 and substrate #3 through the coupling slot or aperture in the isolating ground plane GND #2, thus constituting the ACF scheme (see Figure1a). The coupling slot is centrally positioned under the patch to produce lower cross-polarization due to the symmetry of the configuration [9] according to the X- and Y-axis, as illustrated in Figure1b. The implemented feeding configuration in this element design is similar to the conventional ACF. The only difference is at the coupling-slot size. In contrast to the conventional ACF that contains a small non-resonant coupling slot, the proposed element antenna has a relatively large coupling aperture whose width ws is about seven times larger than that of the conventional coupling slot. This modification shows a significant improvement of the coupling levels (see Figure2a) as the magnetic polarization of the slot (a dominant mechanism for the coupling) is highly dependent on the size and shape of the aperture or slot [12]. Although maximum coupling between the feedline and the patch is obtained in Figure2a with larger coupling-aperture size, a smaller non-resonant aperture produces lower back radiation with a better impedance matching to result in less spurious radiation and better efficiency. Thus, a large coupling aperture utilized in this element antenna causes significant spurious radiation due to the strong higher-mode surface wave and transmission line radiations to lower efficiency, especially in large-scale patch array in higher frequency bands. Also, the microstrip line of the ACF underneath the patch causes undesired reactive Sensors 2021, 21, x FOR PEER REVIEW 3 of 12

Sensors 2021, 21, 3914 3 of 11

loading effects at the element antenna’s input port because of the distortion in the electric Sensors 2021, 21, x FOR PEER REVIEW 3 of 12

ﬁeld( betweena) the isolating ground plane and the patch [13(].b) Figure 1. The geometry of element antenna: (a) 3D exploded view and (b) top view.

The implemented feeding configuration in this element design is similar to the conventional ACF. The only difference is at the coupling-slot size. In contrast to the conventional ACF that contains a small non-resonant coupling slot, the proposed element antenna has a relatively large coupling aperture whose width ws is about seven times larger than that of the conventional coupling slot. This modification shows a significant improvement of the coupling levels (see Figure 2a) as the magnetic polarization of the slot (a dominant mechanism for the coupling) is highly dependent on the size and shape of the aperture or slot [12]. Although maximum coupling between the feedline and the patch is obtained in Figure 2a with larger coupling-aperture size, a smaller non-resonant aperture produces lower back radiation with a better impedance matching to result in less spurious radiation and better efficiency. Thus, a large coupling aperture utilized in this element antenna causes significant spurious radiation due to the strong higher-mode surface wave and transmission line radiations to lower efficiency, especially in large-scale patch array in higher frequency bands. Also, the microstrip line of the ACF underneath the patch causes undesired(a) reactive loading effects at the element antenna’s(b) input port because of the Figuredistortion 1. The geometryin the electricof element field antenna: between (a) 3D exploded the isolating view and ( bground) top view. plane and the Figure 1. The geometry of element antenna: (a) 3D exploded view and (b) top view. patch [13]. The implemented feeding configuration in this element design is similar to the conventional ACF. The only difference is at the coupling-slot size. In contrast to the conventional ACF that contains a small non-resonant coupling slot, the proposed element antenna has a relatively large coupling aperture whose width ws is about seven times larger than that of the conventional coupling slot. This modification shows a significant improvement of the coupling levels (see Figure 2a) as the magnetic polarization of the slot (a dominant mechanism for the coupling) is highly dependent on the size and shape of the aperture or slot [12]. Although maximum coupling between the feedline and the patch is obtained in Figure 2a with larger coupling-aperture size, a smaller non-resonant aperture produces lower back radiation with a better impedance matching to result in less spurious radiation and better efficiency. Thus, a large coupling aperture utilized in this element antenna causes significant spurious radiation due to the strong higher-mode sur- Sensors 2021, 21, x FOR PEER REVIEW 4 of 12 face wave and transmission line radiations to lower efficiency, especially in large-scale patch array in higher frequency bands. Also, the microstrip line of the ACF underneath the patch causes undesired(a) reactive loading effects at the element antenna’s input port because of the distortion in the electric field between the isolating ground plane and the patch [13].

(a) (b)

Figure 2. 2. SimulatedSimulated current current distributions distributions on onelement element antenna antenna at 26.2 at 26.2GHz. GHz. (a) Conventional (a) Conventional versus versus proposed slot size. (b) Conventional versus SIW cavity-backed microstrip patch. proposed slot size. (b) Conventional versus SIW cavity-backed microstrip patch.

ToTo overcome overcome these these drawbacks drawbacks of ofthe the ACF ACF with with large large aperture aperture size, size,SIW SIWcavity-backed cavity-backed patch has been combined with with a a stripline stripline feeding feeding line line to to implement implement the the radiating radiating struc- structure, ture,as shown as shown in Figure in Figure1. This 1. This configuration configuration significantly significantly suppresses suppresses undesired undesired higher-mode higher- modesurface surface waves, waves, as illustrated as illustrated in Figure in Figure2b, as 2b, well as as well back as radiationback radiation to maximize to maximize the radiation the radiation efficiency. Figure 3a indicates that increasing the slot width ws leads to a corre- efficiency. Figure3a indicates that increasing the slot width ws leads to a corresponding sponding increase in radiation efficiency. This validates this design concept since the ra- increase in radiation efficiency. This validates this design concept since the radiation diation efficiency peaks at ws = 0.146λ, which is better than conventional ACF. Hence, the efficiency peaks at w = 0.146λ, which is better than conventional ACF. Hence, the element element antenna exhibitss gain variation less than 1 dBi from 7.82 to 6.89 dBi within the antenna exhibits gain variation less than 1 dBi from 7.82 to 6.89 dBi within the band of band of interest, as shown in Figure 3b. A peak gain of 7.82 dB occurs at 26.4 GHz with a radiation efficiency of more than 89%. See Figure 4a, the simulated model of element antenna operates from 25.58 to 27.04 GHz for |S11| < −10 dB (5.2%). The simulation results in Figure 4b indicates half power beamwidth (HPBW) of 70° and 73° in the XOZ- and YOZ-plane, respectively. The level of cross-polarization in the boresight direction is less than −20 dB. The geometric dimensions of the radiating element can be found in Table 1.

(a) (b) Figure 3. Simulation results of the element antenna: (a) Radiation efficiency in variation of coupling slot width, ws. (b) Gain and radiation efficiency.

Table 1. Dimensions of the Element Patch Antenna.

Parameters Value Parameters Value dv 0.50 mm wf 0.50 mm xs 0.50 mm wp 4.50 mm ys 0.50 mm ws 5.50 mm sv 0.75 mm ls 3.58 mm Sensors 2021, 21, x FOR PEER REVIEW 4 of 12

(b) Figure 2. Simulated current distributions on element antenna at 26.2 GHz. (a) Conventional versus proposed slot size. (b) Conventional versus SIW cavity-backed microstrip patch.

To overcome these drawbacks of the ACF with large aperture size, SIW cavity-backed patch has been combined with a stripline feeding line to implement the radiating structure, as shown in Figure 1. This configuration significantly suppresses undesired higher- mode surface waves, as illustrated in Figure 2b, as well as back radiation to maximize the radiation efficiency. Figure 3a indicates that increasing the slot width ws leads to a corre- Sensors 2021, 21, 3914 4 of 11 sponding increase in radiation efficiency. This validates this design concept since the radiation efficiency peaks at ws = 0.146λ, which is better than conventional ACF. Hence, the interest,element as antenna shown exhibits in Figure gain3b. Avariation peak gain less of than 7.82 1 dBdBi occurs from 7.82 at 26.4 to 6.89 GHz dBi with within a radiation the band of interest, as shown in Figure 3b. A peak gain of 7.82 dB occurs at 26.4 GHz with a efﬁciency of more than 89%. See Figure4a, the simulated model of element antenna radiation efficiency of more than 89%. See Figure 4a, the simulated model of element an- operates from 25.58 to 27.04 GHz for |S11| < −10 dB (5.2%). The simulation results in tenna operates from 25.58 to 27.04 GHz for |S11| < −10 dB (5.2%). The simulation results Figure4b indicates half power beamwidth (HPBW) of 70 ◦ and 73◦ in the XOZ- and YOZ- in Figure 4b indicates half power beamwidth (HPBW) of 70° and 73° in the XOZ- and plane, respectively. The level of cross-polarization in the boresight direction is less than YOZ-plane, respectively. The level of cross-polarization in the boresight direction is less − than20 dB. −20 The dB. geometric The geometric dimensions dimensions of the of the radiating radiating element element can can be be found found in in Table Table1. 1.

Sensors 2021, 21, x FOR PEER REVIEW 5 of 12

(a) lfx 1.25 mm (b) lp 2.58 mm

FigureFigure 3. Simulation 3. Simulation results results of theof the element element antenna:lf antenna: (a) ( Radiationa) Radiation4.00 efﬁciency efficiency mm in in variation variation of of coupling coupling slot slot width, width,ws ws.(b. ()b Gain) Gain and radiation efficiency. and radiation efﬁciency. Table 1. Dimensions of the Element Patch Antenna.

Parameters Value Parameters Value dv 0.50 mm wf 0.50 mm xs 0.50 mm wp 4.50 mm ys 0.50 mm ws 5.50 mm sv 0.75 mm ls 3.58 mm

(a) (b)

FigureFigure 4. 4.Simulated Simulated results results of of the the proposed proposed element element antenna. antenna. (a) ( |S11|.a) |S11|. (b) ( Radiationb) Radiation patterns patterns at 26.2 at 26.2 GHz. GHz.

Table 1.TheDimensions geometry of of the the Element element Patch antenna Antenna. in Figure 1, however, can lead to power leakage to the parallel plate mode due to the use of the stripline feed to cause a decrease in radiation efficiency.Parameters Figure 5 shows Value the parallel plate Parametersmode of the stripline, indicating Value the electric field dlinesv due to the voltage0.50 mm |Vs| applied to thewf coupling slot. Thus,0.50 mmthe amount of parallel platexs mode is contingent0.50 mm on the amplitude wofp |Vs|. The impedance4.50 mm seen looking from the yslots toward the patch0.50 mmis purely real (Ra) atw sresonance for the5.50 aperture-coupled mm sv 0.75 mm ls 3.58 mm antenna (ACA), as illustrated in the circuit model in Figure 6a [14]. As seen from Figure lfx 1.25 mm lp 2.58 mm 6b, Ra is verylf low compared4.00 to mmthe 60-Ω input impedance of the element antenna within the operating frequency band. The patch, therefore, appears to behave like a circuit that shorts out the slot at these resonant frequencies, resulting in the parallel plate mode sup- The geometry of the element antenna in Figure1, however, can lead to power leakage pression. It should be noted that Lf and Cf represent the feedline, whereas the patch is to the parallel plate mode due to the use of the stripline feed to cause a decrease in radiation characterized by Lp, Cp, and Rp. efﬁciency. Figure5 shows the parallel plate mode of the stripline, indicating the electric

Figure 5. Parallel plate waveguide developed from the ground planes of the stripline of element antenna. Sensors 2021, 21, x FOR PEER REVIEW 5 of 12

lfx 1.25 mm lp 2.58 mm lf 4.00 mm

(a) (b) Figure 4. Simulated results of the proposed element antenna. (a) |S11|. (b) Radiation patterns at 26.2 GHz. Sensors 2021, 21, 3914 5 of 11 The geometry of the element antenna in Figure 1, however, can lead to power leakage to the parallel plate mode due to the use of the stripline feed to cause a decrease in radia- ﬁeldtion efficiency. lines due toFigure the voltage 5 shows |Vs| the parallel applied plate to the mode coupling of the slot.stripline, Thus, indicating the amount the ofelec- parallel platetric field mode lines is due contingent to the voltage on the |Vs| amplitude applied ofto |Vs|.the coupling The impedance slot. Thus, the seen amount looking of from theparallel slot towardplate mode the is patch contingent is purely on realthe amplitude (Ra) at resonance of |Vs|. forThe the impedance aperture-coupled seen looking antenna (ACA),from the as slot illustrated toward the in patch the circuit is purely model real in(Ra Figure) at resonance6a [ 14]. for As the seen aperture-coupled from Figure6b, Ra antenna (ACA), as illustrated in the circuit model in Figure 6a [14]. As seen from Figure is very low compared to the 60-Ω input impedance of the element antenna within the 6b, Ra is very low compared to the 60-Ω input impedance of the element antenna within operating frequency band. The patch, therefore, appears to behave like a circuit that shorts the operating frequency band. The patch, therefore, appears to behave like a circuit that out the slot at these resonant frequencies, resulting in the parallel plate mode suppression. shorts out the slot at these resonant frequencies, resulting in the parallel plate mode sup- It should be noted that Lf and Cf represent the feedline, whereas the patch is characterized pression. It should be noted that Lf and Cf represent the feedline, whereas the patch is by Lp, Cp, and Rp. characterized by Lp, Cp, and Rp.

Sensors 2021, 21, x FOR PEER REVIEWFigure 5. 5. ParallelParallel plate plate waveguide waveguide developed developed from fromthe ground the ground planes of planes the stripline of the of stripline element 6 of 12 of ele- antenna. ment antenna.

(a) (b)

FigureFigure 6. (a) Equivalent6. (a) Equivalent circuit circuit of aperture-coupled of aperture-coupled antenna. antenna. ((bb) Impedance seen seen looking looking from from the theslot slottoward toward the patch the patch within the operating band of the element antenna. within the operating band of the element antenna. 3. Array Antenna Design 3. Array Antenna Design The structure of the 32 × 32 array based on the proposed element antenna is shown inThe Figure structure 7a. The ofantenna’s the 32 × geometry32 array comprises based on three theproposed substrate layers element with antenna a corporate is shown in Figurefeeding7 a.network The antenna’s to distribute geometry power among comprises the constituent three substrate elements. layers Its feeding with network a corporate feedinghas Tee-junctions, network to distributewhich are powertwo-way among equal power the constituent dividers to elements. ensure uniform Its feeding distribu- network hastion Tee-junctions, of power. Some which sections are two-way of the feeding equal lines power are dividers bent to achieve to ensure 180° uniform phase shift distribution be- of power.tween adjacent Some sections elements of in the the feeding x-axis for lines in-phase are bent excitation, to achieve as illustrated 180◦ phase in Figure shift between7b. adjacentThe center-to-center elements in thedistancex-axis between for in-phase adjacent excitation,radiating patches as illustrated is 7.2 mm inin the Figure x-direc-7b. The center-to-centertion and 8 mmdistance in the y-direction. between The adjacent entire radiatingradiation aperture patches of is the 7.2 full-corporate mm in the x -directionfeed 32 × 32 array antenna covers an area of 232.6 mm (20λ0) × 257 mm (22.5λ0) (x × y). The and 8 mm in the y-direction. The entire radiation aperture of the full-corporate feed 32 × 32 height of the antenna is 1.664 mm (0.14 λ0). The paths of the input signals to the radiating λ × λ × arraypatches antenna are detailed covers anin Figure area of 5 232.6as Block mm III (20→ Block0) I.257 mm (22.5 0)(x y). The height of the antennaBlock isIII: 1.664 Waveguide-to-stripline mm (0.14 λ0). The Transition-The paths of the input launched signals signal to in the the radiating waveguide patches areexcites detailed the in matching Figure5 patch as Block of the III transition→ Block at I. 26.2 GHz. The matching patch is located in theBlock feed aperture III: Waveguide-to-stripline in the bottom ground Transition-Theplane of substrate launched #3. signal in the waveguide excites theBlock matching II: Feed network-The patch of the input transition signals at are 26.2 electromagnetically GHz. The matching coupled patch to the is feed- located in theline feed on aperture the top surface in the of bottom substrate ground #3 from plane the matching of substrate patch. #3. The signals are then transferred to and evenly distributed in the feeding network. Block I: Radiating part-The SIW cavities on substrate #1, comprising coupling apertures and metal vias, are designed to transfer power between the substrate layers. Thus, the uniformly distributed signals in the feeding network are also coupled electromagnetically through the apertures or slots, which are etched in the middle ground plane, to the patches. The signals are subsequently re-radiated by the patches without the influence of the higher-mode surface waves due to the cavity-backed technique. Sensors 2021, 21, 3914 6 of 11

Block II: Feed network-The input signals are electromagnetically coupled to the feedline on the top surface of substrate #3 from the matching patch. The signals are then transferred to and evenly distributed in the feeding network. Block I: Radiating part-The SIW cavities on substrate #1, comprising coupling apertures and metal vias, are designed to transfer power between the substrate layers. Thus, the uniformly distributed signals in the feeding network are also coupled electromagnetically through the apertures or slots, which are etched in the middle ground plane, to the patches. The signals are subsequently re-radiated by the patches without the inﬂuence of Sensors 2021, 21, x FOR PEER REVIEW 7 of 12 the higher-mode surface waves due to the cavity-backed technique.

(a)

(b) (c) Figure 7. (a) Configuration of the proposed array antenna. (b) Detailed 3-D view of the corporate-feed network. (c) De- Figuretailed 7.3-D( viewa) Conﬁguration of the waveguide-to-stripline of the proposed transition. array antenna. (b) Detailed 3-D view of the corporate-feed

network. (c) Detailed3.1. 3-D Block view II: Feed of Network the waveguide-to-stripline transition. As pointed out previously, the feed network (see Figure 7b) adopts a Tee-junction for 3.1. Block II: Feed Networksignal division. It also contains a microstrip bend (i.e., delay line), which follows the Tee- junction to achieve a 180-degree phase difference between the adjacent elements because As pointed outthe previously, delay line length theis adjusted feed to networkλg/2 (where λg (see is the Figure guided wavelength7b) adopts at the center a Tee-junction frequency. With reference to Figure 7b, no additional metal vias were installed along the for signal division.feedlines It also of the contains feed network a microstrip to suppress the bend surface waves(i.e., from delay feed line), structure which and follows the Tee-junction to achieve a 180-degree phase difference between the adjacent elements because the delay line length is adjusted to λg/2 (where λg is the guided wavelength at the center frequency. With reference to Figure7b, no additional metal vias were installed along the feedlines of the feed network to suppress the surface waves from feed structure and ground planes (i.e., stripline shielding). This is because it will further increase the complexity of the full array structure. The prototyping of the antenna, thus, becomes more difﬁcult to perform, making the antenna more susceptible to fabrication errors, which eventually deteriorate the measured antenna performance. In addition, the metal via itself could also radiate, especially at a higher frequency, as the size is comparable to the Sensors 2021, 21, 3914 7 of 11

operating wavelength to result in radiation losses. It is, therefore, imperative to take into consideration the coupling between the feedlines without the stripline shielding. Figure8 a shows the coupling coefficients between the feedlines in terms of |S12|, |S21| and |S23|, and |S32|. It can be seen that the coupling coefficient is lower than −30 dB over the frequency range of 24.5–27.5 GHz. This suggests that less than 0.1% of the input power contributes to the coupling between feedlines, which is negligible to reduce efficiency. Figure8a also indicates that a 180-degree phase difference is achieved by the delay line.

3.2. Block III: Waveguide-to-Stripline Transition Structure The waveguide-to-stripline transition of the array antenna is shown in Figure7c. This transition consists of a rectangular waveguide, feedline, waveguide-short pattern, metal vias, and matching patch. The top substrate layer (i.e., antenna’s substrate #2) is cladded with copper layer as the upper ground plane to achieve the stripline configuration. Thus, Sensors 2021, 21, x FOR PEER REVIEWany undesired transmission line radiation is suppressed to improve the radiation efficiency, 8 of 12 especially for the large-scale patch array. On top of the lower substrate (i.e., substrate #3) resides the feeding line. The matching patch is etched on the substrate’s backside with a short-terminatedground planes (i.e., waveguide stripline shielding). using the This metal is because vias. WR-28 it will further waveguide increase was the used com- with dimensionsplexity of the of afull = 7.112 array mm structure. and b =The 3.556 prototyping mm. of the antenna, thus, becomes more difficultReferring to perform, to Figure making7c, the the waveguide’s antenna more dominant susceptible TE10 to fabrication mode transforms errors, which to the feedline’seventually quasi-TEM deteriorate mode the measured through theantenna radiation performance. mode TM01 In addition, of the matching the metal patch. via itself Thus, bycould adjusting also radiate, the length especially and width at a higher of the frequency, matching patch,as the size the desiredis comparable operating to the frequency oper- andating good wavelength impedance to matchingresult in radiation can be realized losses. It to is, lower therefore, insertion imperative loss. Figure to take8b into shows con- the simulatedsideration reflection the coupling |S11| between and transition the feedlines |S21| without characteristics the stripline of the shielding. optimized Figure waveguide- 8a to-striplineshows the transitioncoupling coefficients in a back-to-back between configuration. the feedlines in The terms transition of |S12|, operates |S21| and from |S23|, 24.86 to 28.24and GHz|S32|. for It |S11|can be

(a) (b)

FigureFigure 8. (8.a )(a) Coupling Coupling coefﬁcientcoefficient and and phase phase of offeed feed network. network. (b) Simulated (b) Simulated S11, S21, S11, and S21, phase and of waveguide-to-stripline phase of waveguide-to- transition. stripline transition. 3.2. Block III: Waveguide-to-Stripline Transition Structure The waveguide-to-stripline transition of the array antenna is shown in Figure 7c. This transition consists of a rectangular waveguide, feedline, waveguide-short pattern, metal vias, and matching patch. The top substrate layer (i.e., antenna’s substrate #2) is cladded with copper layer as the upper ground plane to achieve the stripline configuration. Thus, any undesired transmission line radiation is suppressed to improve the radiation efficiency, especially for the large-scale patch array. On top of the lower substrate (i.e., substrate #3) resides the feeding line. The matching patch is etched on the substrate’s backside with a short-terminated waveguide using the metal vias. WR-28 waveguide was used with dimensions of a = 7.112 mm and b = 3.556 mm. Referring to Figure 7c, the waveguide’s dominant TE10 mode transforms to the feedline’s quasi-TEM mode through the radiation mode TM01 of the matching patch. Thus, by adjusting the length and width of the matching patch, the desired operating frequency and good impedance matching can be realized to lower insertion loss. Figure 8b shows Sensors 2021, 21, x FOR PEER REVIEW 9 of 12

the simulated reflection |S11| and transition |S21| characteristics of the optimized waveguide-to-stripline transition in a back-to-back configuration. The transition operates from 24.86 to 28.24 GHz for |S11| < −10 dB (12.7%) with an insertion loss |S21| beer than −0.6 Sensors 2021, 21, 3914 dB from 25 to 27.95 GHz. The insertion loss is due to the transmission line (loss8 of = 11 0.012 dB/mm) and the two transition structures (0.1 dB each) for the back-to-back arrangement. Figure 8b also shows the phase response of the waveguide-to-stripline transition.

4. Results and Discussion4. Results and Discussion The photographsThe of photographs the fabricated of the 32 fabricated× 32 array 32 × antenna 32 array areantenna shown are shown in Figure in Figure9. It is9. It is observed that theobserved backside that of the the antenna backside has of the no antenna metal plate has no attached metal plate to the attached bottom to ground the bottom plane for lightweightground purposes, plane for which lightweight is desirable purposes, for which UAVs. is desirable Thus, the for WR-28 UAVs. waveguideThus, the WR-28 to coax adapter iswaveguide directly connected to coax adapter to the is directly antenna’s connected feed pointto the antenna’s for measurement. feed point for measurement.

(a) (b)

Figure 9. FigurePhotograph 9. Photograph of the of fabricated the fabricated 32 ×32 32× 32 array array antenna.antenna. (a ()a Top) Top view. view. (b) (Backb) Back view. view. Figure 10a,b exhibit the experimental results of |S11|, gain, and radiation efficiency. Figure 10a,b exhibit the experimental results of |S11|, gain, and radiation efficiency. There is a good agreement between the simulated and measured results. The measured There is a good agreement|S11| < −10 dB between bandwidth the is simulated 5.65%, i.e., from and 25.43 measured to 26.91 results.GHz, while The the measured measured gain |S11| < −10 dB bandwidthvaries between is 5.65%,30.71 and i.e., 32.8 from dBi 25.43within to this 26.91 band. GHz, A peak while gain theof 32.8 measured dBi is observed gain at varies between 30.7126.2 andGHz. 32.8 On the dBi other within hand, this the band. simulated A peak gain gain varies of between32.8 dBi 33.9 is observed and 35 dBi. at The 26.2 GHz. On the otheroverall hand, disparity the between simulated the gainsimulated varies and between measured 33.9 gain and is generally 35 dBi. Thedue overallto the inser- disparity betweention the loss simulated of 0.3 dB of and themeasured coaxial waveguide gain is adapter generally or transition. due to the Other insertion contributing loss factors include copper surface roughness, substrate shrinkage, and assembly tolerance, of 0.3 dB of the coaxial waveguide adapter or transition. Other contributing factors include which are difficult to account for during the full-wave analysis. They significantly affect copper surface roughness,the performances substrate of antennas shrinkage, in millimeter-wave and assembly bands, tolerance, especially which in multilayered are difficult struc- to account for duringtures. the The full-wave radiation efficiency analysis. plotted They in significantly Figure 10b is affectabove the52% performanceswithin the band ofof op- antennas in millimeter-waveeration. The simulated bands, especiallyand measured in far-field multilayered radiation structures. patterns at 25.43, The radiation 25.9, 26.2, and efficiency plotted26.9 in Figure GHz are 10 shownb is above in Figure 52% 11, within respectively. the band Well-maintained of operation. beam The simulated shapes at these and measured far-fieldthree frequencies radiation can patterns be observed. at 25.43, The 25.9, measured 26.2, and HPBWs 26.9 for GHz all the are three shown frequency in Figure 11, respectively.points are Well-maintained 2° and 2.2° in XOZ- beam and YOZ-plane, shapes at respectively. these three frequencies can be ◦ ◦ Sensors 2021, 21, x FOR PEER REVIEWobserved. The measured HPBWs for all the three frequency points are 2 and 2.2 10 ofin 12 XOZ-

and YOZ-plane, respectively.

(a) (b)

FigureFigure 10.10. ((aa)) ReflectionReflection coefficient. coefficient. (b (b) )Gain Gain and and efficiency. efficiency.

(a)

(b) Sensors 2021, 21, x FOR PEER REVIEW 10 of 12

Sensors 2021, 21, 3914 (a) (b) 9 of 11 Figure 10. (a) Reflection coefficient. (b) Gain and efficiency.

(a)

Sensors 2021, 21, x FOR PEER REVIEW 11 of 12

(b)

(c)

(d)

Figure 11.FigureSimulated 11. Simulated and measuredand measured radiation radiation patterns patterns ofof thethe 32 ×× 3232 planar planar array array antenna antenna at different at different frequency frequency points: points: (a) 25.43, (b) 25.9, (c) 26.2, and (d) 26.9 GHz. (a) 25.43, (b) 25.9, (c) 26.2, and (d) 26.9 GHz. 5. Conclusions In summary, this paper presents a millimeter-wave array antenna for unmanned aerial vehicles (UAVs). The array element utilizes a large coupling aperture (compared with the conventional ACF slot size) to enhance coupling power between the feeding line and patch. However, with this design concept, the large-scale patch array results in significant transmission lines and higher-mode surface wave radiations to increase antenna loss. The radiating element is therefore implemented with the SIW cavity-backed patch and stripline feeding line of ACF. Thus, a full-corporate feed 32 × 32 array antenna including the transition achieves gain between 30.7 and 32.8 dBi with radiation efficiency above 52% within the band of interest. Also, the fabricated antenna has no metal plate at the backside to maintain its lightweight attribute, essential for UAV applications.

Author Contributions: K.A. designed, simulated, and optimized the 32×32 planar array antenna. J.- N.L. contributed to the measurement and the preparation of this article. K.A. prepared this article, including its revision. Y.-B.J. provided important suggestions regarding the design, measurement, and preparation of this article and is the supervisor of the research group. All authors have read and agreed to the published version of the manuscript. Funding: This work was supported by Institute of Information & communications Technology Plan- ning & Evaluation (IITP) grant funded by the Korea government (MSIT) (No.2021-0-00045, Devel- opment of movable high-capacity mobile communication infra for telecommunication disaster and rescue). Institutional Review Board Statement: Not applicable. Sensors 2021, 21, 3914 10 of 11

5. Conclusions In summary, this paper presents a millimeter-wave array antenna for unmanned aerial vehicles (UAVs). The array element utilizes a large coupling aperture (compared with the conventional ACF slot size) to enhance coupling power between the feeding line and patch. However, with this design concept, the large-scale patch array results in signiﬁcant transmission lines and higher-mode surface wave radiations to increase antenna loss. The radiating element is therefore implemented with the SIW cavity-backed patch and stripline feeding line of ACF. Thus, a full-corporate feed 32 × 32 array antenna including the transition achieves gain between 30.7 and 32.8 dBi with radiation efﬁciency above 52% within the band of interest. Also, the fabricated antenna has no metal plate at the backside to maintain its lightweight attribute, essential for UAV applications.

Author Contributions: K.A. designed, simulated, and optimized the 32 × 32 planar array antenna. J.-N.L. contributed to the measurement and the preparation of this article. K.A. prepared this article, including its revision. Y.-B.J. provided important suggestions regarding the design, measurement, and preparation of this article and is the supervisor of the research group. All authors have read and agreed to the published version of the manuscript. Funding: This work was supported by Institute of Information & communications Technology Planning & Evaluation (IITP) grant funded by the Korea government (MSIT) (No.2021-0-00045, Development of movable high-capacity mobile communication infra for telecommunication disaster and rescue). Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: Not applicable. Conﬂicts of Interest: The authors declare no conﬂict of interest.

References 1. Adnan, N.; Nordin, S.; bin Bahruddin, M.A. Sustainable interdependent networks from smart autonomous vehicle to intelligent transportation networks. In Sustainable Interdependent Networks II; Springer: Cham, Switzerland, 2019; pp. 121–134. 2. Huo, Y.; Lu, F.; Wu, F.; Dong, X. Multi-Beam Multi-Stream Communications for 5G and beyond Mobile User Equipment and UAV Proof of Concept Designs. In Proceedings of the 2019 IEEE 90th Vehicular Technology Conference (VTC2019-Fall), Honolulu, HI, USA, 22–25 September 2019; pp. 1–5. 3. Bento, M.D.F. (Ed.) Unmanned Aerial Vehicles: An Overview; Inside GNSS: Hoboken, NJ, USA, 2008; pp. 54–61. 4. Gupta, S.G.; Ghonge, M.M.; Jawandhiya, A.P. Review of unmanned aircraft system (UAS). Int. J. Adv. Res. Comput. Eng. Technol. 2013, 2, 1646–1658. [CrossRef] 5. Huo, Y.; Dong, X. Millimeter-Wave for Unmanned Aerial Vehicles Networks: Enabling Multi-Beam Multi-Stream Communications. arXiv Prepr. 2018, arXiv:1810.06923. 6. Franke, U.E. Civilian drones: Fixing an image problem? In ISN Blog; International Relations and Security Network: Zurich, Switzerland, 2015; Volume 5. 7. Zeng, Y.; Zhang, R.; Lim, T.J. Wireless communications with unmanned aerial vehicles: Opportunities and challenges. IEEE Commun. Mag. 2016, 54, 36–42. [CrossRef] 8. Arpaio, M.J.; Fuschini, F.; Vitucci, E.M.; Degli Esposti, V.; Barbiroli, M.; Masotti, D. Lightweight Microstrip Patch Array for Broadband UAV Applications over 5G networks. In Proceedings of the 2019 Conference on Microwave Techniques (COMITE), Pardubice, Czech Republic, 16–18 April 2019. 9. Teniente, J.; Gonzalo, R.; del-Rio, C. Innovative High-Gain Corrugated Horn Antenna Combining Horizontal and Vertical Corrugations. In IEEE Antennas and Wireless Propagation Letters; IEEE: Greenvile, SC, USA, 2006; Volume 5, pp. 380–383. 10. Al-Nuaimi, M.K.T.; Hong, W.; Zhang, Y. Design of High-Directivity Compact-Size Conical Horn Lens Antenna. In IEEE Antennas and Wireless Propagation Letters; IEEE: Greenvile, SC, USA, 2014; Volume 13, pp. 467–470. 11. Baldazzi, E.; Al-Rawi, A.; Cicchetti, R.; Smolders, A.B.; Testa, O.; van Coevorden Moreno, C.D.J.; Caratelli, D. A High-Gain Dielectric Resonator Antenna with Plastic-Based Conical Horn for Millimeter-Wave Applications. In IEEE Antennas and Wireless Propagation Letters; IEEE: Greenvile, SC, USA, 2020; Volume 19, pp. 949–953. 12. Zarreen, A.A.S. An Introduction of Aperture Coupled Microstrip Slot Antenna. Int. J. Eng. Sci. Technol. 2010, 2, 36–39. Sensors 2021, 21, 3914 11 of 11

13. Satyanarayana, D.S.S.; Bathula, A. Aperture Coupled Microstrip Antenna Design and Analysis using MATLAB. Master’s Thesis, California Polytechnic State University, San Luis Obispo, CA, USA, June 2019. 14. Yamamoto, M.; Itoh, K. Behavior of parallel plate mode in a slot-coupled patch antenna with a stripline feed. In Proceedings of the IEEE Antennas and Propagation Society International Symposium. 1998 Digest. Antennas: Gateways to the Global Network. Held in conjunction with: USNC/URSI National Radio Science Meeting (Cat. No.98CH36194), Atlanta, GA, USA, 21–26 June 1998; Volume 2, pp. 932–934.