A Compact Planar Wi-Fi Antenna with Optimized Radiation Patterns for Small UAV Applications

1. Introduction

With the rapid growth of the number of small unmanned aircraft vehicles (UAVs), their antenna design is becoming more and more important to ensure a stable link [1]. This study addresses an antenna design for light-and-small civil UAV applications in urban areas. Small UAVs tend to be light in weight and small in size for portability, so they are not suitable for mounting turntables or large phased-array antennas. Figure 1 shows the common scenarios of light-and-small civil UAV applications in urban areas, in which a person holding a remote control operates a quadcopter.

The proposed antenna is mounted below the propeller on the outermost side of the wing of a small UAV, away from large metal conductors such as printed circuit boards (PCBs) and batteries in the UAV body. The UAV takes off from the ground, flies vertically above ground objects, and then flies forward to the farthest position. The entire flight must avoid hitting ground objects such as people, trees, and buildings, so the UAV must keep flying higher than them. The distances between the UAV and the remote control are in order of taking off from the ground, flying vertically above ground objects, and flying forward to the farthest position.

For ensuring stable transmission quality and stability, the gain and polarization characteristics of the antenna are determined based on the Friis Transmission Equation [2]. The antenna must meet the following requirements to reduce interference and provide more antenna gain when the small UAV flies farther in urban areas. First, its $\left|{\mathrm{S}}_{11}\right|$ must be equal to or smaller than $-10 \mathrm{d}\mathrm{B}$ to ensure that the power can be delivered to the antenna efficiently. Second, it must have a small size and low profile to meet the portability and aerodynamic requirements for small UAV applications. Third, it must have stronger radiation below the UAV to reduce interference from rare-use directions. Because the UAV is always above the remote control as it flies away, it is farther away from the remote control than when it was just taking off. Fourth, its maximum radiations, the directions of the maximum radiation in each elevation plane, must be below the UAV and between 10 and 30 degrees from the horizontal plane to provide more antenna gain when the small UAV flies farther in urban areas. Finally, due to the boundary condition of the earth's ground plane, the antenna must have good vertically polarized radiation for long-distance applications.

Although many antenna architectures have been proposed [3,4,5,6,7,8,9,10,11,12,13,14] to meet the requirements for UAV applications, there is not yet a suitable one that can fully meet the requirements for small UAV applications in urban areas. Due to their small size and lightweight nature, folded dipole antennas have been proposed [3,4] for small and compact UAV applications, respectively. The fragmented antenna was proposed [3], and its overall dimensions and total mass are $0.96{\mathsf{\lambda }}_0\times 0.8{\mathsf{\lambda }}_0\times {0.8\mathsf{\lambda }}_0$ at 240 MHz and 18 g (including the matching circuit). A compact electrically tunable VHF antenna [4] was integrated into the landing gear of a compact UAV and has good vertically polarized radiation. However, neither of them has stronger radiation below the UAV, and their maximum radiations are near the horizontal plane. In addition, both of them need an extra matching circuit to match their characteristic impedance to 50 $\mathsf{\Omega }$ for a better reflection coefficient at the operation frequency. The low-profile, quasi-omnidirectional Substrate Integrated Waveguide (SIW) Multihorn Antenna was proposed [5] for small nonmetallic UAV applications. It has good vertically polarized radiation and a low profile of 0.028 ${\mathsf{\lambda }}_0$ at 2.4 GHz. However, it does not have stronger radiation below the UAV, and its lateral dimensions of $0.76{\mathsf{\lambda }}_0\times 0.76{\mathsf{\lambda }}_0$ are too large for small UAVs. Due to their low profile and ability to conform to planar and nonplanar surfaces, patch antennas were proposed [6,7] for UAV applications. The high-gain dual-mode cylindrical conformal rectangular patch antenna [6] has a low profile of 0.508 mm (0.0175 ${\mathsf{\lambda }}_0$ at 10.323 GHz), stronger radiation below the quadcopter UAV, and good vertically polarized radiation. The required radius of the arm of the UAV is 25 mm at the lowest operating frequency of 10.288 GHz, and it has to mount 3 patch antennas on the UAV arm to cover the hemisphere below the UAV. The conformal patch antennas [7] have a low profile of 3 mm (0.058 ${\mathsf{\lambda }}_0$ at 5.8 GHz) and stronger radiation below the UAV. However, their maximum radiations are directly above each patch. Both of them proposed assembling multiple patch antennas on the UAV to ensure a stable link in the required directions. It takes up too much surface area, so they are not suitable for small UAVs. The design of a combined printed helical spiral antenna and helical inverted-F antenna [8] was miniaturized by co-winding the radiation elements on the same ceramic rod with a height of 24.7 mm (0.129 ${\mathsf{\lambda }}_0$ at 1.57 GHz) and a radius of 6 mm. It was proposed for UAVs with global positioning system (GPS) L1 band and telemetry communication frequency band (2.33 GHz) applications. However, it does not have stronger radiation below the UAV at 2.33 GHz. The low-profile broadband plasma antenna [9] has different radiation patterns at different operating frequencies, and it also has a low profile of 0.105 ${\mathsf{\lambda }}_0$ at 30 MHz. Its radiation pattern is omnidirectional at 30 MHz, and it has stronger radiation below the UAV at 300 and 500 MHz. However, it must be matched with an extra Butterworth low-pass filter and high-pass filter to isolate the 13.56 MHz radio frequency signal that is used to activate plasma. Its complex structure and the need to use special materials such as plasma will result in a heavier weight and higher cost, so it is not suitable for small UAVs. The broadband slotted blade dipole antenna [10] has different radiation patterns at different operating frequencies. However, it must be matched with a special matching circuit, and it does not have stronger radiation below the UAV. In addition, its radiation patterns at all operating frequencies have nulls near the horizontal plane, which will cause the smallest gain when the UAV flies to the farthest position. Due to the advantages of its simple structure and low manufacturing cost using modern printed circuit technology, printed planar antennas were proposed [11,12,13] for UAV applications. The planar dual-mode dipole antenna [11] has a short electrical length of 0.3297 ${\mathsf{\lambda }}_0$ at 0.86 GHz and good vertically polarized radiation. The omnidirectional vertically polarized antenna [12] also has good vertically polarized radiation, and it has different radiation patterns at different operating frequencies. However, neither of them has stronger radiation below the UAV, and their maximum radiations are near the horizontal plane. The wideband single-sided folded-off-center-fed dipole antenna [13] has good vertically polarized radiation and different radiation patterns at different operating frequencies. At 5 and 5.5 GHz, it has stronger radiation below the small UAV, and its maximum radiations are below the UAV and between 10 and 30 degrees from the horizontal plane. At 3.5 GHz, it has quasi-omnidirectional radiation, and its maximum radiations are near the horizontal plane. However, its electrical length of 0.469 ${\mathsf{\lambda }}_0$ is too long for small UAV applications. A compact slot antenna with coplanar waveguide-fed [14] was not proposed for small UAV applications, and it has a very short electrical length of $0.16{\mathsf{\lambda }}_0\left(\mathrm{L}\right)$ and a low profile of $0.0128{\mathsf{\lambda }}_0\left(\mathrm{h}\right)$ to meet the portability and aerodynamic requirements for the small UAV applications. However, it does not have stronger radiation below the UAV, and its maximum radiations are near the horizontal plane.

Summarizing the above literature reviews, most proposed antennas have a radiation pattern with maximum radiations near the horizontal plane for UAV applications. This feature can meet the requirements for UAV applications in open areas and provide more antenna gain when the small UAV flies farther. The wideband single-sided folded-off-center-fed dipole antenna [13] is the best one; it has suitable radiation patterns to meet the requirements for small UAV applications in both urban and open areas. In addition, it also has the advantage of a simple and cost-efficient structure, which can make light-and-small civil UAVs more affordable. However, its electrical length of 0.469 ${\mathsf{\lambda }}_0$ at 3.5 GHz is too long to meet the portability and aerodynamic requirements for small UAV applications.

Based on the structure with a short electrical length [14], this paper proposes a dual-band antenna with optimized radiation patterns for small UAV applications in both urban and open areas, respectively. It has a shorter electrical length of 0.132 ${\mathsf{\lambda }}_0$ at 2.4 GHz than the best one that has been proposed [13]. First, this study considers that the small UAV has a camera function and needs to transmit real-time images to the remote control for the user to preview or store. To meet this application requirement, Wi-Fi 2.4 GHz and 5.8 GHz were selected as the operating frequencies. Secondly, since many technologies, such as Zigbee and Bluetooth, use 2.4 GHz as the operating frequency [15], there will be more interference at this frequency. Therefore, a quasi-directional radiation pattern was designed at 2.4 GHz, and its maximum radiations are near the horizontal plane for small UAV applications in open areas with less interference, such as mountains, oceans, or farmland. Third, an end-fire radiation pattern was designed at 5.8 GHz for small UAV applications in urban areas. It has stronger radiation below the UAV, and its maximum radiations are below the UAV and between 14 and 29 degrees from the horizontal plane. Finally, to ensure that the characteristics of this antenna will not be significantly affected when it is applied to small UAVs, the following assumptions are made in this study: First, as shown in Figure 1, the objects around the antenna, such as the UAV wing and propeller, must be made of non-metallic material. Second, the large metal conductors like PCBs and batteries in the UAV body must be located in the far-field region of the proposed antenna. Based on the definition of the far-field region [16], the distance between the small UAV body and the proposed antenna must be more than 46 mm.

By performing the theoretical analysis, full-wave simulation, and prototype measurement, we verified that the proposed antenna can meet the requirements for light-and-small civil UAV applications in both urban and open areas. The rest of this paper is as follows: Section II presents the antenna design methodology, which includes the proposed antenna's structural topology and theoretical analysis. Section III presents the full-wave simulations for validating the design theory. Section IV presents the antenna prototype, the measurement setup, the comparisons of the simulated data and measured results of the prototype, and discussions about the results. Finally, Section V summarizes the results and future work.

2. Antenna Design and Analysis

In this study, a compact slot antenna with coplanar waveguide (CPW)-fed was designed that supports Wi-Fi dual-band communication and has a single-sided planar structure.

2.1. Antenna Structure

Figure 2 shows the configuration of the proposed antenna. At 5.8 GHz, the four corners of its outer conductor are the main radiation sources, forming end-fire radiation. At 2.4 GHz, the two sides of the outer conductor are the main effective radiation sources, forming quasi-omnidirectional radiation.

Table 1 displays the values of geometrical parameters for the proposed antenna. The thickness of the metal layer (${\mathrm{M}}_{\mathrm{t}}$) is 0.035 mm, of which 0.031 mm is copper, 0.003 mm is nickel, and 0.001 mm is gold. It was printed on the FR4 substrate with a relative permittivity of 4.4, a loss tangent of 0.02, a thickness of 1.6 mm (h), and planar dimensions of 16.5 mm (L) by 30.3 mm (W). At the fed port, the width of the central conducting track (${\mathrm{W}}_{\mathrm{f}}$) is 2.5 mm, and a pair of outer conducting tracks are separated from the central conducting track by a small gap (g) of 0.3 mm to achieve a characteristic impedance of 50 $\mathsf{\Omega }$. The antenna is symmetrical on both sides of the x-z plane, so the radiation generated by the current components in the +y and -y directions will cancel each other out. Therefore, the currents in the +z and -z directions are the main radiation sources, making this antenna have good vertically polarized radiation.

2.2. Antenna Design for 5.8 GHz

Figure 3 shows the simulated surface current density at 5.8 GHz, where the areas with higher surface current densities are the main radiation areas. The outer conductor forms four monopole antennas from each triangle vertex to its two adjacent corners. These four monopoles have the same wavelength, ${\mathsf{\lambda }}_{\mathrm{H}}$, and the upper resonance frequency, ${\mathrm{f}}_{\mathrm{H}}$, can be approximated by using (1), where c is the speed of light in vacuum ($3\times {10}^8$ m/s).

$${\mathrm{f}}_{\mathrm{H}}\cong \frac{\mathrm{c}}{{\mathsf{\lambda }}_{\mathrm{H}}}=\frac{\mathrm{c}}{\left({\mathrm{L}}_2+{\mathrm{L}}_3+{\mathrm{L}}_1/2\right)/0.23}=5.63 \mathrm{G}\mathrm{H}\mathrm{z},$$

(1)

There are higher surface current densities at the four corners inside the outer conductor, which approximate four point-sources, named Element A, Element B, Element C, and Element D. The radiation pattern is mainly synthesized from these four point-sources. Because of the electric field that radiates between the inner and outer conductors, the proposed antenna approximates a two-element array of two point-sources, Element 1 and 2, as shown in Figure 4.

Where the radiation patterns of Elements 1 and 2 are equal to the sum of Elements A and C and the sum of Elements B and D, respectively. Furthermore, since the cross-sectional area of the inner conductor used as the reference plane is very small, the radiation patterns of both Element 1 and 2 approximate a vertical infinitesimal dipole. Based on the directivity of the vertical infinitesimal (Hertzian) dipole [17], the normalized pattern function of the vertical infinitesimal dipole is ${\mathrm{s}\mathrm{i}\mathrm{n}}^2\mathsf{\theta }$. The normalized array factor of Element 1 and 2 is adapted from (2) [18]. Where k is the wave number equal to $2\mathsf{\pi }/{\mathsf{\lambda }}_{5.8\mathrm{G}}$, and d is the distance between Element 1 and 2, approximately equal to ${\mathrm{L}}_{\mathrm{i}}+\left({\mathrm{L}}_1/2\right)\times 2$.

$${\left({\mathrm{A}\mathrm{F}}_{12}\right)}_{\mathrm{n}}=\mathrm{c}\mathrm{o}\mathrm{s}\left[\frac{1}{2}\left(\mathrm{k}\mathrm{d}\mathrm{c}\mathrm{o}\mathrm{s}\mathsf{\theta }+{\mathsf{\beta }}_{\mathrm{d}}\right)\right]=\mathrm{c}\mathrm{o}\mathrm{s}\left\{\frac{1}{2}\left[\frac{2\mathsf{\pi }}{{\lambda }_{5.8G}}\left({\mathrm{L}}_{\mathrm{i}}+\left({\mathrm{L}}_1/2\right)\times 2\right)\mathrm{c}\mathrm{o}\mathrm{s}\mathsf{\theta }+{\mathsf{\beta }}_{\mathrm{d}}\right]\right\}=\mathrm{c}\mathrm{o}\mathrm{s}\left[\frac{1}{2}\left({100.92}^{°}\times \mathrm{c}\mathrm{o}\mathrm{s}\mathsf{\theta }+{\mathsf{\beta }}_{\mathrm{d}}\right)\right],$$

(2)

At 5.8 GHz, the difference in phase excitation, ${\mathsf{\beta }}_{\mathrm{d}}={\mathsf{\beta }}_{\mathrm{A}}-{\mathsf{\beta }}_{\mathrm{B}}={\mathsf{\beta }}_{\mathrm{C}}-{\mathsf{\beta }}_{\mathrm{D}}$, can be derived by (3). Where ${\mathsf{\beta }}_{\mathrm{A}}$, ${\mathsf{\beta }}_{\mathrm{B}}$, ${\mathsf{\beta }}_{\mathrm{C}}$, and ${\mathsf{\beta }}_{\mathrm{D}}$ are the phase excitations of the Element A, B, C, and D, respectively. In addition, ${\mathsf{\beta }}_{\mathrm{A}}$ is equal to ${\mathsf{\beta }}_{\mathrm{C}}$, and the negative sign indicates the phase lag.

$${\mathsf{\beta }}_{\mathrm{d}}=-\frac{2\times \mathsf{\pi }}{{\mathsf{\lambda }}_{5.8\mathrm{G}}}\times \left({\mathrm{L}}_{\mathrm{i}}+\left({\mathrm{L}}_1/2\right)\times 2-{\mathrm{T}}_{\mathrm{w}}\right)=-{90.48}^{°},$$

(3)

In (2), replace ${\mathsf{\beta }}_{\mathrm{d}}$ with $-{90.48}^{°}$, the normalized array factor of Element 1 and 2 can be written as (4). It has two maximums at $\mathsf{\theta }=+26°$ and $\mathsf{\theta }=-26°$ and only one minimum at $\mathsf{\theta }=+180°$, and it forms an end-fire array toward $0°$.

$${\left({\mathrm{A}\mathrm{F}}_{12}\right)}_{\mathrm{n}}=\mathrm{c}\mathrm{o}\mathrm{s}\left[\frac{1}{2}\left({100.92}^{°}\times \mathrm{c}\mathrm{o}\mathrm{s}\mathsf{\theta }-{90.48}^{°}\right)\right],$$

(4)

The normalized total field of the two-element array antenna is written as (5). It has two maximums at $\mathsf{\theta }=+60°$ and $\mathsf{\theta }=-60°$ and only one minimum at $\mathsf{\theta }=+180°$, and it forms an end-fire radiation toward $0°$.

$${\left({\mathrm{E}}_{\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}\right)}_{\mathrm{n}}={\left\{{\mathrm{s}\mathrm{i}\mathrm{n}}^2\mathsf{\theta }\times {\left({\mathrm{A}\mathrm{F}}_{12}\right)}_{\mathrm{n}}\right\}}_{\mathrm{n}},$$

(5)

Figure 5 shows the relative power patterns of the vertical infinitesimal dipole, the array factor of Element 1 and 2, and the total field of the two-element array antenna.

2.3. Antenna Design for 2.4 GHz

Figure 6 shows the simulated surface current densities at 2.4 GHz, where the areas with higher surface current densities are the main radiation areas. The two sides of the outer conductor form a pair of L-shaped monopoles, Monopole E and Monopole F. These two monopoles have the same wavelength, ${\mathsf{\lambda }}_{\mathrm{L}}$, and the lower resonance frequency, ${\mathrm{f}}_{\mathrm{L}}$, can be approximated by using (6).

$${\mathrm{f}}_{\mathrm{L}}\cong \frac{\mathrm{c}}{{\mathsf{\lambda }}_{\mathrm{L}}}=\frac{\mathrm{c}}{\left\{\left[{\mathrm{W}}_1+\left({\mathrm{W}}_2/2\right)\right]+\left[{\mathrm{L}}_{\mathrm{i}}+\left({\mathrm{L}}_1/2\right)\times 2\right]\right\}/0.23}=2.52 \mathrm{G}\mathrm{H}\mathrm{z},$$

(6)

Due to the symmetrical design of the architecture, the radiations from the two bottom paths will cancel each other out. Because of the electric field that radiates between the inner and outer conductors, the proposed antenna approximates a vertical monopole antenna. In addition, since the cross-sectional area of the inner conductor used as the reference plane is very small, the radiation pattern of the vertical monopole antenna approximates that of a vertical dipole with an effective length of ${\mathrm{L}}_{\mathrm{i}}+\left({\mathrm{L}}_1/2\right)\times 2$, equivalent to 0.122 ${\mathsf{\lambda }}_{\mathrm{L}}$. Its maximum radiations are near the horizontal plane.

3. Full-Wave Simulation

The proposed antenna was simulated by the ANSYS High-Frequency Structure Simulator (HFSS) [19]. From (1), ${\mathrm{f}}_{\mathrm{H}}$ decreases as ${\mathrm{L}}_2$ increases, and ${\mathrm{L}}_2$ can be derived from (7), ${\mathrm{L}}_2$ increases as ${\mathrm{T}}_{\mathrm{h}}$ increases. Hence, ${\mathrm{f}}_{\mathrm{H}}$ decreases as ${\mathrm{T}}_{\mathrm{h}}$ increases.

$${\mathrm{L}}_2=\sqrt{{{\mathrm{T}}_{\mathrm{h}}}^2+{\left(\frac{{\mathrm{T}}_{\mathrm{w}}}{2}\right)}^2}=\sqrt{{5.7}^2+{\left(\frac{1.5}{2}\right)}^2}=5.75,$$

(7)

Figure 7 shows the simulated reflection coefficients for different values of ${\mathrm{T}}_{\mathrm{h}}$. The simulated data are in agreement with (1) and (7), ${\mathrm{f}}_{\mathrm{H}}$ decreases as ${\mathrm{T}}_{\mathrm{h}}$ increases, and ${\mathrm{f}}_{\mathrm{H}}$ is about 5.63 GHz when ${\mathrm{T}}_{\mathrm{h}}$ is equal to 5.7 mm.

From (6), ${\mathrm{f}}_{\mathrm{L}}$ decreases as ${\mathrm{W}}_1$ increases, and Figure 8 shows the simulated reflection coefficients for different values of ${\mathrm{W}}_1$. The simulated data are in agreement with (6),${\mathrm{f}}_{\mathrm{L}}$ decreases as ${\mathrm{W}}_1$ increases.

From (2) and (5), the normalized total field at 5.8 GHz changes as ${\mathsf{\beta }}_{\mathrm{d}}$ changes, and from (3), ${\mathsf{\beta }}_{\mathrm{d}}$ gets closer to $-{100.92}^{°}$ as ${\mathrm{T}}_{\mathrm{w}}$ decreases closer to zero. Hence, the end-fire radiation becomes more significant as ${\mathrm{T}}_{\mathrm{w}}$ decreases. Figure 9 shows the simulated radiation patterns at 5.8 GHz in the y-z plane for different values of ${\mathrm{T}}_{\mathrm{w}}$. The simulated data are in agreement with (2), (3), and (5), and the end-fire radiation becomes more significant as ${\mathrm{T}}_{\mathrm{w}}$ decreases.

4. The Comparisons of Simulated Data and Measured Results of The Prototype

Figure 10 shows the prototype of the proposed antenna with an SMA connector [20] soldered to it, where the hand-soldered tin had not been incorporated in the simulation. The antenna had been measured by the Atenlab A3 compact type OTA measurement system [21] and compared with its simulated data.

Figure 11a–c show the measurement setup for radiation patterns in the x-y, x-z, and y-z planes, respectively. The proposed antenna was placed on a turntable and connected to the coaxial cable of the measurement system which had not been incorporated in the simulation. Figure 11a shows the proposed antenna was connected directly to the coaxial cable of the measurement system, and the cable was parallel to the z-axis of the proposed antenna. Figure 11b,c show the proposed antennas were connected to the coaxial cable of the measurement system via a right-angle SMA adapter [22], and the cable was perpendicular to the z-axis of the proposed antennas.

4.1. Reflection Coefficients

Figure 12 shows the comparison of the simulated and measured reflection coefficients of the proposed antenna. The simulated curve shows that the proposed antenna can meet $\left|S_{11}\right|\le -10\mathrm{d}\mathrm{B}$ from 2.11 to 2.58 GHz and from 5.06 to 7.5 GHz, where the fractional bandwidths are 20.04 % and 38.05 %, respectively. The measured result is in agreement with the simulated one, and it can meet Wi-Fi 2.4 and 5.8 GHz requirements for small UAV applications.

4.2. Radiation Patterns

Figure 13, Figure 14, and Figure 15 show the comparisons of the simulated and measured radiation patterns of the proposed antenna in the x-y, x-z, and y-z planes, respectively. The simulated data show the proposed antenna has good vertically polarized radiation. Figure 13a,b show the radiation patterns of the proposed antenna in the x-y plane at 2.4 and 5.8 GHz, respectively, and the vertically polarized radiations are quasi-omnidirectional. Due to the hand-soldered tins on the main radiation areas of 2.4 GHz, the measured horizontally polarized radiation is larger than the simulated one, as shown in Figure 13a.

In the x-z and y-z planes, due to the cable effect, the measured horizontally polarized radiations are much larger than the simulated data, as shown in Figure 14 and Figure 15. In addition, due to the cable effect, the measured vertically polarized radiation is larger than the simulated one at −120° to +120° in the y-z plane at 5.8 GHz, as shown in Figure 15b.

The simulated vertically polarized radiation patterns at 2.4 GHz are quasi-omnidirectional, and the maximum radiations in the x-z and y-z planes are near the horizontal plane, as shown in Figure 13a, Figure 14a, and Figure 15a. The simulated vertically polarized radiation patterns at 5.8 GHz have stronger radiation in the upper hemisphere (below the UAV), as shown in Figure 14b and Figure 15b. Its maximum radiations in the x-z and y-z planes are in the upper hemisphere and between 14 and 29 degrees from the horizontal plane.

Table 2 displays the characteristics of the simulated vertically polarized radiation patterns. Ignoring the influence caused by the hand-soldered tin and the cable effect of the measurement system, all measured results of radiation patterns are in agreement with simulated data. In E-planes, the beamwidths of relative -10 dB at 2.4 GHz are no less than 142 degrees. Moreover, at 5.8 GHz, maximum gain angles are between 17 and 29 degrees off the horizontal.

Table 3 displays the comparison of the proposed antenna and published works on planar antennas. The proposed planar antenna has the following advantages. First, it has a single-sided structure, which has a relatively low cost to make small UAVs more affordable. Second, it has a minimum electrical length (L) of 0.132 ${\mathsf{\lambda }}_0$ and a low profile (h) of 0.0128 ${\mathsf{\lambda }}_0$ to meet the portability and aerodynamic requirements for small UAV applications. Third, at 5.8 GHz, it has stronger radiation in the upper hemisphere (below the small UAV) to reduce interference from rare-use directions. Its maximum radiations are below the UAV and between 14 and 29 degrees from the horizontal plane to provide more antenna gain when the small UAV flies farther in urban areas. At 2.4 GHz, it has quasi-omnidirectional radiation to ensure a stable link in all directions. Its maximum radiations are near the horizontal plane to provide more antenna gain when the small UAV flies farther in open areas. Finally, it has good vertically polarized radiation at both 2.4 and 5.8 GHz for long-distance applications.

5. Conclusions

In summary, we report a compact planar Wi-Fi antenna with optimized radiation patterns for small UAV applications in urban and open areas, respectively. The proposed antenna had been prototyped and measured, and the measured results are in agreement with the simulated data. It has end-fire radiation to reduce the interference from rare-use directions at 5.8 GHz and quasi-omnidirectional radiation to ensure a stable link in all directions at 2.4 GHz. The directions of the maximum radiation in each elevation plane at 5.8 and 2.4 GHz provide more antenna gain when the small UAV flies farther in urban and open areas, respectively. In future work, the proposed antenna will be mounted on a small UAV for field experiments.