Directivity Improved Antenna with a Planar Dielectric Lens for Reducing the Physical Size of the On-Vehicle Communication System

Abstract

As the physical size of a communication system for satellites or unmanned aerial vehicles demands to be reduced, a compact antenna with high directivity is proposed as a core element essential to the wireless device. Instead of using a horn or an array antenna, a unit planar antenna is combined with a surface-modulated lens to convert a low antenna gain to a high antenna gain. The lens is not a metal-patterned PCB but is dielectric, which is neither curved nor very wide. This palm-sized lens comprises pixels with different heights from the backside of PolyPhenylene Sulfide (PPS) as the dielectric base. The antenna gain from the unit antenna of 4.5 cm × 4.5 cm is enhanced by 10 dB with the help of a compact dielectric lens of 7.5 cm × 7.5 cm at 24.5 GHz as the frequency of interest. The antenna design is verified by far-field measurement as well as near-field observation, including sensing a metal object behind a blocking wall by using an RF test bench. Moreover, antenna performance is understood by making a comparison with conventional designs of antennas in terms of directivity and physical sizes.

1. Introduction

Connecting one node to another on a communication network wirelessly is not a surprise any longer. As the level of the technology has advanced to suit user requirements by introducing more complicated architectures, algorithms and circuitry than before, and some elements are excluded from the system and new elements are brought in, there are elements that are indispensable to the wireless communication system. One of them is frequency. Working as the primary element to form a channel between a transmitting node and a receiving node, the frequency of one node should be tuned to that of another. The frequency has been driven to go up through GHz-mobile services like UMTS, WiFi and LTE-A to X-band, Ku-band and higher. Increasing frequency comes from the goal of securing a high data transmission rate and targeting a small number of receivers instead of many distributed over a wide angle as elaborated in [1,2,3]. Real-time interaction added video streaming and inter-satellite operation in an LEO constellation, which needs much faster communication than that. For this purpose, frequency is set at Ku-band above 12 GHz or K-band above 18 GHz. It is suggested to exceed 20 GHz for broader bands, and when it is near 30 GHz, they call it the millimeter-wave band. This takes advantage of the increased data transfer rate, but on the flip side, the signal is sensitive to the environment and ends up with path loss.

Wireless device developers have built systems working in the high frequency band by employing high-gain antennas; they look for solutions to compensate for path loss like array antennas [4]. When it comes to high-gain antennas, a horn or the parabolic reflector can usually be considered. However, these volumetric structures take up a relatively large space and make the whole system heavy. Hence, the microstrip patch array antenna is preferred as a planar structure [5,6]. Whether military or commercial, use-cases demanding increased antenna gain see the array antennas placed in the transceiver. Nonetheless, when entering the millimeter-wave band, drawbacks of the array antenna are addressed. The RF signal undergoes an increasing amount of loss as it travels from the input port to the far-end in the power divider as a result of high sensitivity to the loss factors of the dielectric and conductor. When the array antenna should be larger to catch up with a necessary gain against attenuation along the signal path, the growth in loss overshadows the antenna gain. This causes heating from beamforming chipsets.

Given that the use of patch array antennas runs into such problems, a transmitarray or superstrate is worthy of being tried to tackle them. Placed above the antenna as the source of radiation, the superstrate is introduced to enhance the strength of far-field radiation, and it is also called a metasurface lens, which is analogous to lenses in optics.

A diverging wave from the source antenna can be changed to high directivity at the far-zone field by way of a superstrate as described in Figure 1. A conventional lens could generate phase shifts for the incoming rays by bending the surfaces of the dielectric material. As the gap between the source antenna and the superstrate needs to be shorter than the conventional technique where the plane of the incidence is far enough from the source antenna, the phase distribution of the superstrate is implemented by using the idea of new materials. There are representative methods such as the phase compensation approach, graded index approach, double negative parameters, etc. Some examples of them are briefly mentioned. A multi-layered PCB as the superstrate was added to the source antenna by Datthanasombat et al. to change the quasi-spherical wavefronts into parallel wavefronts [7]. Raising the frequency to a higher channel in the mm wave band, a Q-band transmitarray was presented by Kaouach et al. using a layered PCB for increasing the gain [8]. Dussoptia et al. took a similar step to the previous references to have a high antenna gain over an expanded bandwidth [9,10]. Ray optics were employed at the initial stage of the design by Bai et al. to make the radiation aperture wider for increasing directivity [11]. Wang and Liang had more layers in the superstrate with gaps and trimmed the gaps and heights there [12,13]. Slits were introduced to the metal plane to couple the fields from the source with the top layer [14]. Their geometries could generate beams of improved directivity.

In this paper, different from the PCB-type superstrates that conventional methods use, a dielectric slab is employed as a small footprint lens for the source of radiation. As pictorially explained below, the proposed structure gives wireless systems of aerospace communication positive effects.

Aircraft and satellites located at high altitudes can communicate with ground stations on which parabolic reflectors as well as horn antennas are mounted. The horn antenna has a reputation for directional radiated fields. Its antenna gain is such a merit that alone or combined with a reflector, it can generate high directivity as much as a large patch array antenna and relieve the active RF block of necessitating multiple-stage power amplifiers. As is shown in Figure 2, the horn antenna is long while the other antennas of the same antenna gain are relatively short. This metallic waveguide antenna can be replaced by the proposed antenna. As is given on the right side of Figure 2, a dielectric slab is placed over the source antenna. As conventional metasurface lenses take the form of multi-layered and wide PCBs, there are air gaps or bonding films between one wide layer and another. On the contrary, the contribution is given by making a flat lens out of one homogeneous dielectric slab PPS, which is better than multi-layered PCB geometry in terms of errors from vertical and horizontal misalignment of stacking and uncertainty of materials and thickness of adhesive films. These errors will make the electromagnetic properties of the antenna of interest worse in K- or Ka-band and above. Therefore, a single layer material is good for the lens, but for it to work up to expectations, it requires a sufficient thickness that a single PCB substrate in the market cannot satisfy. This is why PPS is chosen. In order to make the single dielectric slab play the role of lensing, the required phase distribution of the structure is implemented with certain heights in pixels. Drilling and grinding in the fabrication process are adopted to make the surface of the lens. This increases the antenna gain by around 10 dB with a lens of 7.5 cm × 7.5 cm at 24.5 GHz, with reference to the source antenna. The design is validated by far-field and near-field tests. Also, it is proven that the lens can improve the function of sensing a metal object behind a blocking wall. This is tested by using an RF test bench comprising a two-port vector network.

1.1. Design of the Source Antenna

The proposed structure comprises two parts. One is the source antenna and the other is the lens. Firstly, the design of the source antenna is addressed. While most of the literature on planar antennas have a horn antenna or a single patch antenna as the primary source, we do not follow suit. When a horn antenna is used to excite the superstrate, the gap between the two parts is enormous. The superstrate becomes very large as the source antenna is located far away, which is improper for the objective of size reduction. As for a single patch antenna, since the initial gain is small, to eventually have a large antenna gain, it is unavoidable to make the geometry large.

The source antenna is a 2 × 2 array as in Figure 3a,b. Generating an antenna gain higher than a single patch helps it achieve high directivity in the final structure. Each patch has an area of 2.9 mm × 2.9 mm on a 45 mm × 45 mm substrate of 4350B. The reflection coefficient in Figure 3c,d shows the operating frequency as desired and the beam pattern with an antenna gain of 9 dBi. Next, it is combined with the lens.

1.2. Design of the Lens

The lens is presented as follows.

A dielectric slab sized 75 mm × 75 mm is placed over the source antenna with a gap of 4.7 cm as given in Figure 4a. On a very compact plane, the surface of the dielectric slab is formed on the basis of the phase distribution of Figure 4b, which is the result of a calculation for converting the phase distribution of the incoming wave from the source to the one required for the beam pointed at the zenith. To implement the phases of the pixels of the surface, which correspond to the tiles of its discretized version, either different permittivity dielectric cubes of the same height or the same permittivity of different heights is possible to think of. Realization begins by relating the height of the design parameter of the dielectric unit with the phase as in Figure 4c. This is applied to expressing the phases with the heights of all the pixels as in Figure 4d working for the incoming wavefront from the 2 × 2 array antenna. When phases of 100° and −170° are required, Figure 4c maps them to 1.41 mm and 6.4 mm, respectively, in height. This mapping applies to other phases and pixels. Through this process which is explained again in Appendix A, the antenna is realized. The simulated far-field patterns of the source antenna alone and the lens-combined antenna are plotted based on this implementation. It is observed that there is a distinct enhancement in the view-point of antenna gain from the source antenna to the lens-added patch antenna. About a 10 dB increment is noted by looking at Figure 4e. It seems equivalent to the effect of a 10 cm long horn antenna in that it can generate a far-field pattern of almost the same gain as the proposed structure.

2. Characterizing the Antennas Through RF Measurement

The proposed design scheme to substantially enhance the antenna gain of a small-sized array antenna is validated by fabrication of the prototypes and experiments with them. It has procedural steps from the primary source antenna to the lens-combined geometry. As steps 1 and 2, the primary source and its modified version with a fixture are electromagnetically characterized.

In step 1, the electrical property of the source antenna, which is prototyped as in Figure 5a, is checked. The 2 × 2 array’s metal patches on 4350B as the substrate mentioned in Figure 3 are fed by a power divider whose port is connected to the port coming from the vector network analyzer. Figure 5b shows that it meets the resonance condition in that S₁₁ as the reflection coefficient at the port becomes lower than −10 dB at 24.5 GHz. The bare board as above is handled in just the initial phase of a series of tests, and should be put into a fixture for experimentally investigating far-field radiation as in step 2. In the antenna test setup, it is necessary to place the antenna on a feeder or turn-table with the fixture, and this harnessing may affect the electrical and electromagnetic properties of the antenna because it is not the material of air but metal or dielectrics. The effect of the fixture on the antenna is considered through modeling and physical implementation as in Figure 5c,d. Using a 3D-printed fixture of PLA, the simulated and measured S₁₁ curves agree with each other, showing resonance at the target frequency. Figure 5f of the AUT in the anechoic chamber to (d) presents the simulated and measured beam patterns as the same as Figure 3d as a broad beam.

The lens is manufactured as in Figure 6a, which is a homogenous dielectric slab. The model in Figure 6b implies that the lens goes into the unoccupied area of the jig of Figure 5c. This model is simulated, and its far-field pattern is obtained. Figure 6c plots the simulated beam patterns of the antenna without and with the jig, and they have no difference except for roughly 1-dB loss as in 19 dBi to 18 dBi. Measurement that verifies the design and simulation like Figure 6d is conducted by employing a full setup of Figure 6e. The objective of the proposed method is confirmed by observing the remarkably enhanced gain as in Figure 6e though both the beam patterns experience loss due to loss of the cable, its torsion, uncertainty from mechanical contacts, etc. [15,16,17,18,19].

3. Observation of the Strengths of the Lens in View of Wireless Sensing

So far, the proposed structure has been dealt with primarily from the standpoint of antenna functions and geometrical characteristics. Meanwhile, secondary aspects of the antenna such as playing a core role in wireless sensing are addressed. The line of sight (LoS) link, as in free-space communication, is dubbed into reflection by a metal plane. In addition, the antenna is used to detect hidden objects as in radars.

Far-field measurement showing the EM power distribution on the spherical space at a long distance from the AUTs requires relatively high input power. With low input power, near-zone experiments using the VNA with two ports can characterize the AUTs. Figure 7b,c show the reflected signals. They have 50 cm as the traveling distance. Port 2 adopts an 8 × 8 array to amplify the signal strength in an RF-passive manner, considering the low power.

Hands-on experiences of wireless links from the TX to the RX are acquired through checking the transmission coefficient from one antenna at port 1 to that at port 2 of the VNA. As mentioned previously, contrary to the far-field pattern tests, since the VNA has a limited level of input power, the test setups in Figure 8 take 50 cm as the distance for both the straight and reflected paths. At first sight, the distance might look very short, but it amounts to 40.8 wavelengths, which seems like quasi-far-field and makes the millimeter-wave signal become very weak at the RX. In order to increase the sensitivity of the signal reception, instead of a simple antenna, an 8 × 8 array is connected to the RX side. On the TX side, the tests take turns from the source antenna to the flat lens-combined structure. The greater antenna gains on both sides lead to sensible results.

$$P_{RF\_RX}{|}_{{Ant.}_{Ref}}=\zeta \times {\displaystyle \frac{{{|E_0|}_{Freq.}}^2}{{{Dist.}_{Ref}}^2}}\times G_{TX\_Ref}\times G_{RX\_Ref}\times {\zeta }_{RX}$$

(1)

$$P_{RF\_RX}{|}_{{Ant.}_{New}}=\zeta \times {\displaystyle \frac{{{|E_0|}_{Freq.}}^2}{{{Dist.}_{Ref}}^2}}\times G_{TX\_New}\times G_{RX\_New}\times {\zeta }_{RX}$$

(2)

$${\mathsf{\Delta }P}_{RF\_RX}\left[dB\right]=P_{RF\_RX}{|}_{Ant{.}_{New}}\left[dB\right]-P_{RF\_RX}{|}_{{Ant.}_{Ref}}\left[dB\right]$$

(3)

where P_{RF_RX}, ${{|E_0|}_{Freq.}}^2$, G and $\zeta$ imply the RF power of the received signal expressed in S₂₁, initial RF power, antenna gain and coefficient of the electromagnetic radiated power, respectively, at the frequency of interest. These findings are verified through the following tests. Firstly, LoS signals are observed as S₂₁ regarding the source antenna as the TX or the proposed lens as the TX as in Figure 8a. Between the source antenna and the 8 × 8 array, S₂₁ is −31 dB, which jumps to −23 dB as in Figure 8b by substituting the flat lens for the source antenna. Comparing the values of S₂₁ at the target frequency, −30 dB without the lens becomes −21 dB, meaning an improvement in the transmission. The direct path is changed to the V-shaped reflected path to investigate the capability of RF sensing on a target placed behind a screen. Figure 8c is the photograph of the reflected paths for two cases (without or with the lens), which has nothing behind the foam screen in common. Foam is reckoned to be a type of air for permittivity. Since there is no reflection due to nothing being hidden, the port receives −55 dB as in Figure 8d, which is understood as empty. The curves are compared and show both of them having no reflected signal of a measurable value. Figure 8e involves the sensing of a glass cup behind the screen with the proposed antenna. Seeing Figure 8f, S₂₁ increases to −47 dB from the cases that hide nothing. The lens results in −44 dB. Compared to the case without lens, the signal level becomes larger by 5 dB; when referring to the LoS case, the increment rate is lower due to wave scattering by the round shape and material loss of the cup. Lastly, a palm-sized metal pad is introduced to the backside of the screen as in Figure 8g. The object is made out of copper. Figure 8h presents that the S₂₁ values become −24 dB and −31 dB, respectively, for the source antenna alone and the flat lens-combined geometry. An improvement in object sensing is about 10 dB from −31 dB to −22 dB. This effect is compared to the LoS cases, and reflection from the metal has a similar behavior to the direct path. The proposed structure enables the beam to be narrow enough to sense a metal object, which is a few centimeters in width and strong enough to make the reflected signal strengths similar to LoS cases.

4. Conclusions

An antenna comes to have a notably increased antenna gain by making a contribution of developing a compact and flat lens for the usage of 24.5 GHz, representing millimeter-wave bands, and necessitating solutions to compensate for fast attenuation in propagation. The proposed lens is a 75 mm × 75 mm sized homogeneous dielectric slab; the surface of which has different heights on the pixels corresponding to the desired phases for the purpose of changing the ones of the incident wave to those of the transmitted plane wave. As a superstrate whose footprint is not large over the 2 × 2 array as the source antenna, the flat lens enhances the antenna gain of the source antenna by about 10 dB. This effect from this contribution is equivalent to what a 10 cm long horn or 8 × 8 patch array can generate, which means effective miniaturization is obtained. The proposed design is validated by prototyping and far-field measurement. In addition, near-zone signal strength observation using the antennas fed by ports of the VNA as the test bench has been conducted to figure out advantages in the lens-combined structure. The signal strengths are measured on the straight and bouncing paths between the source antenna alone or planar lens-combined geometry at port 1 and the 8 × 8 array for sensitivity increment at port 2. When the TX employs the lens-combined antenna, the signal strength improves by nearly 9 dB. Additionally, this improvement helps sensing of the hidden object and has a better performance in that the level of detection of a small metal object goes up by almost 9 dB. The proposed design can solve the problems of conventional antennas mounted on a satellite payload like patch arrays and horns or parabolic reflector antennas, which suffer from conductor and dielectric loss and large volumes, respectively. In addition, this will be potentially used as a lighter and smaller device in communication and sensing as well.