A High-Gain Filtering Fabry-Perot Antenna With RCS Reduction

A high-gain filtering Fabry-Perot antenna (FPA) with radar cross Section (RCS) reduction is proposed. The design utilizes a partially reflecting surface (PRS) which is integrated with a chessboard arranged metasurface (CAMS) in a single-layered substrate. The CAMS is composed of periodic strip arrays to reduce monostatic RCS based on the phase cancellation principle. The PRS structure is a cross-slot array which is designed to work with both orthogonal orientations of the CAMS elements while yielding a high reflection magnitude. Accordingly, the proposed FPA achieves a high-gain broadside radiation and wideband RCS reduction. The filtering functionality is achieved from a T-shaped feed patch, which is, interestingly, enhanced by the integration of CAMS in the PRS. For validation, a prototype operating at 6.0 GHz has been fabricated and measured. The antenna with an overall size of $3.2\lambda _{{0}} \times 3.2\lambda _{{0}} \times 0.57\lambda _{{0}}$ at 6.0 GHz exhibits an impedance matching bandwidth (BW) of 5.73 – 6.23 GHz, a 3-dB gain BW of 5.76 – 6.17 GHz with the peak gain of 14.63 dBi, and an out-of-band suppression of ≥20 dB. Meanwhile, it achieves an RCS-reduction bandwidth of about 4.0 to 9.0 GHz (76.9%).

SECTION I. Introduction

Fabry-Perot antennas (FPAs) [1], also named as resonant cavity antennas, basically consist of a simple source in a parallel-plate cavity constructed with a ground plane (GND) at the bottom and a partially reflecting surface (PRS) at the top. Owing to the features of high gain with simple structure, the FPAs are attractive in various applications such as onboard satellites, air-craft, and unmanned aerial vehicles. For safety in these applications, stealth characteristic or low radar cross Section (RCS) is critically required. Generally, the antenna, which contains several metallic components, has a considerable contribution to the overall RCS of the platform. Accordingly, FPAs with RCS reduction have been received substantial attention in the last decade. Different methods have been presented for reducing the RCS of FPAs, which are summarized as follows:

1. An effective method for wideband operation is adding absorbing structure to the PRS, which was first presented in [2]. For easy implementation, the absorber-integrated PRS is built on a single-layered substrate [2], [3], [4], [5], [6], i.e., the PRS is patterned on the bottom side to ensure a high reflectivity for the enhanced gain of FPA, whereas the absorber structure is placed on the top side to absorb incoming waves. Typically, metallic square loops with mounted resistors are employed, however, they may degrade the antenna efficiency. Recently, multi-layered PRSs have been presented to improve the performances of RCS-reduction FPAs such as enhancing bandpass response [7], radiation pattern [8], and reducing RCS for both co- and cross-polarized incident waves [9]. The multi-layered configurations, however, increase the complexity and realization cost of the antenna.

2. Another common approach is based on the phase cancellation principle of chessboard arranged metasurface (CAMS) [10] which is a combination of artificial magnetic conductor (AMC) and perfect electric conductor (PEC) cells. Since the AMC and PEC cells have opposite reflection phases, for any normal incident wave, the reflections from CAMS cancel out, and consequently, the monostatic RCS is reduced. Following this approach, many polarization conversion surfaces (PCS), which yield $180\pm 30^{\circ}$ reflection phase variations between two neighboring cells, have been used in FPAs to reduce RCS within a wide frequency range [11], [12], [13], [14], [15], [16], [17], [18], [19].

3. Several other techniques were also applied in FPAs for RCS reduction, which include coding metasurface [20], [21], phase gradient metasurface [22], [23], receiver-transmitter metasurface [24], phase delay line [25], asymmetric electromagnetic absorption [26]. These techniques are effective only in a typically narrow bandwidth.

Besides the desired operational bandwidth, the aforementioned FPAs commonly yield undesired radiations, e.g., high-order resonant modes, which could degrade the stealth characteristic. To suppress the undesired radiations, filtering antennas, which are designed for achieving both antenna and filter function in a single structure, have been considered as an efficient solution [27]. Also, the filtering antennas contribute to reduce losses, complexity, size, and cost in microwave devices. For higher gain, the filtering antennas are traditionally arrayed in the previous works, such as [28], [29], [30], and [31], which require complicated feeding networks and design configurations. Recently, a FPA with simple structure has been reported for both filtering and high-gain features [32]. However, this filtering FPA did not yield the scattering reduction.

This paper presents an FPA concept that can improve the filtering characteristics while still providing RCS reduction and keeping a simple configuration, i.e., with a single-layered substrate. The design employs a PRS integrated with a CAMS. The PRS unit cell is carefully designed to work in two orthogonal orientations of the CAMS elements, thus enabling the use of chessboard arrangement to reduce RCS while keeping the PRS functionality. Interestingly, the CAMS-integrated PRS also exhibits filtering characteristics, which is used together with the feeding patch to obtain the overall filtering performance of the design.

SECTION II. Antenna Design and Characteristics

A. CAMS-Integrated FPA Design

Fig. 1 shows the schematic model of the proposed FPA which includes a feed source in a parallel-plate cavity formed by a conducting GND and a CAMS-integrated PRS.

FIGURE 1.

Schematic model of the proposed FPA.

For completion, we briefly summarize the classical theory of FPA in [33] in the following. For a FPA, a radiation at the broadside is obtained when the reflection phase of the PRS ($\Phi _{PRS}$ ) is theoretically the same as

$$\begin{equation*} \Phi _{1}(f) = \frac {4\pi f} {c} H_{c} + (2N-1)\pi, N = 0, 1, 2, \ldots \tag{1}\end{equation*}$$ where $f$ is the frequency, $c$ is the speed of light in free space, $H_{c}$ is the cavity height and is chosen a value of approximately $\lambda _{0}/2$ ($\lambda _{0}$ is the wavelength in free-space at the resonance frequency). When (1) is satisfied and the PRS size is sufficiently large, the total directivity of the FPA ($D_{total}$ in dB scale) can be given by [34] $$\begin{equation*} D_{total} = D_{source} + D_{PRS} \tag{2}\end{equation*}$$ where $D_{source}$ is directivity of the feed source, $D_{PRS}$ is the enhanced-directivity due to the PRS. $D_{PRS}$ is depended on the reflection magnitude of the PRS, $|\Gamma _{PRS}|$ , as follows $$\begin{equation*} D_{PRS} = 10\log _{10} \frac {1+|\Gamma _{PRS}|}{1-|\Gamma _{PRS}|} \tag{3}\end{equation*}$$ Equation (3) indicates that the PRS needs to yield a large value of $|\Gamma _{PRS}|$ , i.e., close to 1 (or 0 dB), to obtain a high value of $D_{total}$ .

The CAMS is designed utilizing the top conductive layer of the PRS (Fig. 1). It is implemented based on the phase cancellation principle [10] to reduce RCS within a wide frequency range. To utilize this principle, two units of CAMS are arranged in a chessboard configuration. These two units need to reflect the $x$ - and $y$ -polarized incident waves with a reflection phase difference of ideally 180°. A typical satisfactory range for this phase difference is $180^{\circ }\pm 30^{\circ}$ [11], [12], [13], [14], [15], [16], [17].

B. Unit Cell Design

The proposed unit cell is illustrated in Fig. 2(a). It consists of two layers of metallic patterns placed on both sides of a single-layered substrate (F4BM sheet, $\varepsilon _{r}$ = 2.65, tan$\delta$ = 0.002). The bottom side is a metallic coat with a symmetrical cross slot, while the top side is a metallic strip. Fig. 2(b) shows its simulation model in the ANSYS Electronics Desktop. The unit cell is characterized using periodic boundary conditions along the $x$ - and $y$ -directions, and two Floquet ports along $\pm z$ directions. This CAMS-integrated PRS is designed to operate around 6.0 GHz and its optimized parameters are given in the caption of Fig. 2.

FIGURE 2.

(a) Structure of PRS integrated CAMS: $P$ = 13, $W_{d}$ = 2, $L_{d}$ = 12, $H_{sub}$ = 4, $W_{s}$ = 3.6, $L_{s}$ = 12 (unit: mm); (b) simulation model for the unit-cell.

In order to realize both PRS and CAMS functionality, the following conditions need to satisfy

1. For both $x$ -polarized and $y$ -polarized wave, the reflection magnitude of PRS, i.e., $|\Gamma _{PRS}|$ (or $|S_{11}^{x}|$ and $|S_{11}^{y}|$ ) has to be sufficiently high, according to (3).

2. The reflection phases of the CAMs for the incident $x$ -polarized and $y$ -polarized incoming waves exhibit about 180° phase difference, i.e., $\angle S_{22}^{x} - \angle S_{22}^{y} \approx 180^{\circ }$ , according to the phase cancellation principle using CAMS [10].

Fig. 3 shows the three steps to realize the proposed unit cell structure and their corresponding scattering properties. The initial design (Step 1) consists of a metallic strip printed on a grounded substrate. A slot is etched on the bottom side of the initial design to create Step 2. Finally, a cross slot is used at the bottom side to establish the proposed structure. For demonstration, the parameters in all steps are the same as those of the final design (given in the caption of Fig. 2).

FIGURE 3.

(a) Three steps to establish the unit-cell of PRS integrated CAMS and (b-f) their simulated scattering properties.

As shown in Fig. 3(b), at 6 GHz, Step-1 design acts as an AMC for the $x$ -polarized incident wave ($\angle S_{22}^{x} \approx 0^{\circ}$ ), but as a nearly PEC for the $y$ -polarized incident wave ($\angle S_{22}^{y} \approx 150^{\circ}$ ). This results show the potential of using metallic strips in a chessboard arrangement to reduce RCS. Nevertheless, the design in Step 1 totally blocks the incident wave from the feed since the bottom side of substrate is fully metal, i.e., $|S_{22}^{x,y}| = 1$ , and $|S_{11}^{x,y}| = 1$ .

To allow the structure to work as a PRS, a slot along $x$ -direction at etched on the GND in Step-2. This configuration allows the $y$ -polarized waves passing through itself, but reflects the $x$ -polarized waves. Accordingly, the reflection property for the $y$ -polarization is impacted significantly, whereas it hardly changes for the $x$ -polarization (see the all red curves in Fig. 3(b)-(f)). It is noted that the Step-2 design yields a reflection phase difference of about 90° between the two polarizations, which should be re-adjusted if one wants the design to work as a CAMS for RCS reduction.

Thanks to the cross slot in the GND, the proposed design (Step 3) achieves all desired scattering properties (see the all blue curves in Fig. 3(b)-(f)): For the illumination from Floquet port 2, the reflection phase difference is $180^{\circ} \pm 30^{\circ}$ from 5 to 7 GHz. For the illumination from the Floquet port 1, the structure yields a resonant frequency of $f_{r} = 6$ GHz for the broadside radiation of FPA and a high reflectivity ($|S_{11}^{x,y}| \geq 0.8$ ) at around 5.8 to 6 GHz. Following equation (3), the proposed PRS can theoretically provide an enhanced directivity of about 9.5 dB for the FPA.

One interesting feature of this design concept is that it can enhance the filtering (frequency selectivity) characteristics of the FPA. The results in Fig. 3(e) indicate that the FPA only works with high gain enhancement in the vicinity of $5.5- 6$ GHz. Out of this range, especially at high frequency, the magnitude of $|S_{11}^{y}|$ is low, which significantly reduces the antenna gain at high frequency. This effect will be demonstrated in the later section.

C. Antenna Geometry

After the scattering properties of the PRS unit-cell are fully characterized, the CAMS is formed and applied to establish the FPA (see Fig. 4). The PRS is suspended above a conducting GND to form a Fabry-Perot resonator cavity with height of $H_{c}$ . A T-shaped feed patch adapted from [35] is employed as the source element of the FPA, as shown in Fig. 4(c). The source element is built on a $40\times 40$ -mm² Roger Duroid RT5880 substrate ($\varepsilon _{r} = 2.2$ , tan$\delta = 0.0009$ , and thickness of $h_{s1}$ ) and mounted at the center of GND. The T-shaped feed consists of a microstrip-line, a pair of $\lambda /4$ resonators, and a shorting via, which provides a broad bandwidth and harmonic suppression. The antenna parameters optimized via ANSYS HFSS 2022 are given in the caption of Fig. 4. The final design has an overall size of $160\times 160\times28.5$ mm³.

FIGURE 4.

Geometry of the CAMS-integrated FPA: (a) top view, (b) side view, and (c) the source element of T-shaped feed patch. ($H_{c} = 24.5$ , $H_{sub} = 4$ , $h_{s1} = 1.5748$ , $W_{p} = 15$ , $d_{v} = 1.5$ , $L_{r} = 9.5$ , $S_{r} = 2$ , $W_{f} = 4.85$ , $L_{f} = 7$ ; unit: mm.)

D. Antenna Performances

To illustrate the roles of the CAMS-integrated PRS, the antenna in different configurations, including the feed only, FPA without CAMS, and the proposed FPA design shown in Fig. 4, are investigated and compared in Fig. 5.

FIGURE 5.

Performance comparison of the different antennas: (a) reflection coefficient, (b) broadside gain, and (c) monostatic RCS with Y-polarized incident wave.

Thank to the T-shaped feed, the source element achieves a −10-dB impedance bandwidth (BW) of 5.7 - 6.2 GHz with two resonances at 5.75 and 6.10 GHz [Fig. 5(a)]. Also, the source yields a moderate filtering characteristic and an averaged gain of about 6.0 dB within its impedance matching BW [see Fig. 5(b)]. With the large conducting GND, the source yields a monostatic RCS value similar to the same-size copper plate [Fig. 5(c)].

By adding the PRS without CAMS, the resonances slightly shifted toward the lower frequency, whereas the broadside gain is enhanced significantly. The FPA configuration without CAMS achieves a BW of $5.6-6.1$ GHz for −10-dB reflection coefficient [Fig. 5(a)] and a 3-dB gain BW of $5.55-6.0$ GHz with the peak value of 15.7 dBi [Fig. 5(b)]. From Fig. 5(c), the antenna without CAMS yields a RCS reduction within a small frequency range of $5.2-6.0$ GHz.

With the CAMS-integrated PRS, the antenna achieves a broadband RCS reduction. At the same time, the filtering characteristic is significantly enhanced with a much higher roll-off rate at the upper frequency range. These results agree well with the theoretical prediction about filtering performance mentioned in Section II-B, and demonstate the effectiveness of the proposed design approach. This is the main contribution of the paper. Finally, Fig. 5(c) indicates an avarage of about 5-dB RCS reduction in a wide bandwidth from 4.5 GHz to 8.5 GHz.

E. CAMS Arrangements

With the aim of achieving a better wideband RCS reduction, in this sub-section, we will investigate how the performance varies with different CAMS arrangements, as shown in Fig. 6. The results are illustrated in Fig. 7. Since the PRS integrated CAMS is designed to realize both PRS and CAMS functionality simultaneously, the size of effective CAMS cell strongly impacts the RCS reduction, but hardly impacts the radiation performance of the proposed antenna. As shown in Figs. 7(a, b), the reflection coefficient and broadside gain slightly change for different CAMS arrangements. From Fig. 7(c), the effective cell of CAMS needs at least two adjacent cells contributing together to obtain the wideband RCS reduction. Moreover, there is a trade-off between the gain roll-off rate and the RCS reduction performance. Filtering performance is the best with $\#2$ configuration while the RCS reduction performance is maximized with a wider bandwidth for $\#4$ configuration. As our primary aim is to achieve an optimal performance in RCS reduction while filtering performance is partly provided by the feeding structure, we select $\#4$ design for fabrication and validation. Nevertheless, it should be indicated that $\#2$ should be used if a high roll-off rate in realized gain is required.

FIGURE 6.

Different CAMS configurations of the proposed FPA.

FIGURE 7.

Performances of the proposed FPA for different CAMS arrangements: (a) reflection coefficient, (b) broadside gain, and (c) monostatic RCS with Y-polarized illumination.

F. Arbitrary Polarization Incidences

To further confirm the feature of RCS reduction, the proposed FPA (with configuration $\#4$ in Fig. 6) is illuminated by arbitrary polarization incidences, including $X$ - ($\phi = 0^{\circ}$ ), $Y$ - ($\phi = 90^{\circ}$ ), and $D$ - ($\phi = 45^{\circ}$ ) polarizations and its monostatic RCS values are given in Fig. 8. As compared to a reference metallic plate with the same aperture of $160 \times 160$ mm², the proposed FPA using CAMS-integrated PRS achieves a wideband RCS reduction with different polarizations of the incident waves. Although the RCS-reduction bandwidths for the $X-$ and $D-$ polarized illuminations are slightly narrower than the $Y$ -polarization, they are still relatively large, i.e, from 4.5 to 9.0 GHz. Also, for all polarizations, the RCS reduction is slightly degraded at the in-band. This has been a common problem in most of the RCS-reduction antennas using CMAS structures, e.g., [11], [12], [13], [14], [15], [16], [17], [18], and [19]. Nevertheless, the proposed design achieves the minimum RCS reduction of about 3 dB at 6 GHz, as shown in Fig. 8.

FIGURE 8.

Simulated monostatic RCS under normal incidences with arbitrary polarization.

SECTION III. Fabrication and Measurement

A. Fabricated Prototype

As mentioned above, the proposed FPA with configuration $\#4$ of CAMS (in Fig. 6) is fabricated and measured. Fig. 9 shows a fabricated prototype with an overall size of $160\times 160\times 28.5\,\,\text {mm}^{3}$ ($3.2\lambda _{{0}} \times 3.2\lambda _{{0}} \times 0.57\lambda _{{0}}$ at 6 GHz). The components, including the source and PRS-integrated CAMS, are fabricated via a standard printed circuit board (PCB) technology, meanwhile a copper plate with 1-mm thickness is used as the GND. Plastic hexagonal posts and screws ($\varepsilon _{r} = 2.0$ , tan$\delta = 0.02$ , and diameter of 6 mm) are employed to construct the antenna prototype.

FIGURE 9.

Fabricated prototype of the proposed antenna: (a) perspective view, (b) side view, (c) the source element of T-shaped feed patch, and (d) back side of PRS integrated CAMS.

B. Radiation Performances

Fig. 10(a) shows the simulated and measured reflection coefficients of the antenna prototype. There is a good agreement between the simulation and measurement and both indicate that the antenna has very low realized gain in the out-of-band region (i.e., there is not any higher-order radiating mode in the high frequency region up to 15 GHz). The measurement results in a −10-dB impedance bandwidth of 5.73 – 6.23 GHz, whereas the simulated bandwidth is 5.70 – 6.07 GHz. A slight discrepancy between the simulation and measurement could be attributed to the different permittivities of the substrates in full-wave simulator and realization.

FIGURE 10.

(a) S-parameters and (b) realized broadside gain of the prototype.

The realized broadside gain of the FPA prototype is given in Fig. 10(b). The measurements result in the peak gain of 14.6 dBi and a 3-dB gain bandwidth of 5.76 – 6.17 GHz, while the simulations result in the corresponding values of 15.4 dBi and 5.76 – 6.12 GHz, respectively. Moreover, the antenna yields a very good filtering characteristic; i.e., both simulation and measurement yield an out-of-band suppression of ≥ 24 dB in the low-frequency region ($3 - 5.5$ GHz) and ≥ 20 dB in the high frequency region (6.5 – 15.0 GHz).

Fig. 11 illustrates the 6-GHz normalized radiation patterns of the fabricated prototype. It is observed that the proposed FPA achieves an excellent broadside radiation with quite symmetrical pattern, low cross-polarization level, and high front-to-back ratio. At 6.0 GHz, the measurements result in a cross-polarization level of $\le -20$ dB, front-to-back ratio of ≥ 15 dB, and half-power beamwidths of 26° and 24° in the $xz$ and $yz$ -planes, respectively.

FIGURE 11.

Normalized radiation pattern of the FPA prototype at 6.0 GHz.

C. Scattering Performances

The RCS of the proposed FPA is measured in a laboratory environment. We used two single-ended wideband dual-polarized Vivaldi antennas in [36] as transmitter (Tx) and receiver (Rx) which were connected to two ports of a Keysight N5244A PNA-X microwave network analyzer. The proposed FPA was illuminated at a distance of 1 m away from the Tx antenna. For normal incidence, the separation angle between Tx and Rx was set 10° because of the finite size of the prototype. Fig. 12 shows the simulated and measured RCS reduction of the FPA prototype for the $x$ - and $y$ -polarized illuminations. The RCS reduction values are obtained by comparing the scattering fields of the FPA prototype and a reference metallic plate with the same aperture of $160 \times 160$ mm². Both simulation and measurement indicate that the proposed antenna achieves an RCS reduction within a wide frequency range. The measurements result in the peak RCS reduction of 20 dB, whereas the simulations result in the corresponding values of 28.4 dB.

FIGURE 12.

Simulation and measurement RCS reduction of the FPA prototype.

D. Comparison and Discussion

Table 1 shows a comparison of the proposed design and the related works, including non-FPA antennas with high gain and filtering performances [11], [12], [13], [14], [15], [16], [17], [28], [29], [30], [31], [32]. It should be noted that this study targets a design with filtering performance, which is often required for narrow band operation so the radiation bandwidth is not a figure of merit in this particular case of comparison.

TABLE 1 A Comparison of the Proposed FPAs With the Related Works

For the radiation characteristic, as compared to the priors [11], [12], [13], [14], [15], [16], [17], the proposed FPA yields comparable performances in terms of aperture size and broadside gain. Thanks to the utilization of CAMS, all configurations achieve a wideband RCS reduction. Nevertheless, it should be emphasized that compared to all other FPAs, our design introduces the filtering characteristics using a very simple CAMS-integrated PRS (one substrate layer with only one type of PRS unit-cell), which is unique among others. As compared to the existing designs with high-gain and filtering features [28], [29], [30], [31], [32], only our FPA achieves RCS reduction.

SECTION IV. Conclusion

A high-gain FPA with wideband RCS reduction has been demonstrated. It employs a single-layered PRS-integrated CAMS which is deliberately designed to achieve three functions simultaneously: (i) ensuring a high reflectivity for the high-gain of FPA; (ii) reducing RCS across a wide frequency range; (iii) enhanced filtering performance. The final prototype with overall size of $3.2\lambda _{{0}}\times 3.2\lambda _{{0}}\times 0.57\lambda _{{0}}$ at 6.0 GHz yields a −10-dB reflection coefficient BW of $5.73-6.23$ GHz and a 3-dB gain BW of $5.76-6.17$ GHz with the peak gain of 14.6 dBi. Meanwhile, the prototype achieves an average of roughly 5-dB RCS reduction across a wide bandwidth from 4.0 to 9.0 GHz. A very good filtering performance has been realized with more than 24 dB and 20 dB suppression level in the lower and upper frequency range, respectively. Thanks to these advantages, the proposed design can be a good candidate for used in radar systems, satellites, aircraft, as well as unmanned aerial vehicles.