A Mechanically Rollable Reflectarray With Beam-Scanning Capabilities

A novel 1-D beam-steerable reflectarray antenna (RA) is proposed for the <inline-formula> <tex-math notation="LaTeX">${\text{K}}_{u}$ </tex-math></inline-formula>-band. The RA aperture consists of <inline-formula> <tex-math notation="LaTeX">$24\times48$ </tex-math></inline-formula> variable size square patch elements printed on a flexible plastic substrate. The aperture is wrapped around two cylinders to create a <inline-formula> <tex-math notation="LaTeX">$12\lambda \times 12\lambda $ </tex-math></inline-formula> illumination window at the operating frequency of 14 GHz. Beam-steering is achieved by mechanically rolling the aperture. To mitigate antenna pattern degradation, an aperture phase distribution optimization technique is presented. The performance of this RA system is studied analytically using array theory. The effects of aperture size and unit-cell size are also discussed. The performance of the proposed RA is validated using simulations and measurements. The results illustrate that our design achieves 1-D beam-steering from −21° to +21° in the elevation plane while maintaining towards the broadside direction a maximum realized gain of 25.1 dBi. In this beam-scanning range, the gain variation is less than 1.5 dB, beamwidth variation is less than 1.6°, and the side lobe level is maintained below −15 dB. In summary, the main advantage of our design is its ability to steer its high gain beam without using complicated feed networks or complex mechanical systems. Such antennas are especially needed in SmallSat and space applications.


I. INTRODUCTION
T HE REDUCTION in volume and mass of small satellites (SmallSats) has enabled more efficient and robust missions in space [1], [2], which has led to exponential growth in the number of SmallSats launched each year. The communications networks proposed by OneWeb, SpaceX, and Telesat alone are expected to deploy at least 8,000 SmallSats in lower earth orbit by the year 2024 [3]. However, for most satellite applications, it is challenging to meet the communication system requirements due to the small size of SmallSats. Specifically, in such platforms, the large attenuation losses of transmitted signals must be overcome using systems that meet the space constraints of SmallSat buses. Therefore, high gain antennas (HGA) have been proposed to achieve longer transmission distances with higher efficiency thereby minimizing the volume and mass of the needed power systems. Furthermore, a key requirement in advanced SmallSat applications is beamscanning [4]. However, traditional beam-scanning HGAs are not practical for SmallSat applications given the power and space constraints. Therefore, current research is focusing on developing compact/deployable beam-scanning HGAs for space applications [4], [5], [6]. The proposed solutions are generally based on traditional HGA technologies, such as, parabolic reflector antennas [7], [8], [9], phased array antennas [10], [11], [12], [13], and reflectarray antennas (RAs) [14], [15], [16], [17], [18]. Compared to the parabolic reflector and phased array antennas, RAs provide highly directive and versatile beams while maintaining low mass, low fabrication cost, and low complexity [19], [20]. Also, RAs have recently demonstrated the ability to achieve pattern reconfigurability [21]. However, current reconfigurable RA designs suffer from various limitations including lossy and complex DC biasing networks that drive reconfigurable This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ elements [22], or inefficient mechanical systems [23]. In this paper, we present a novel 1-D beam-steering RA system that uses a simple and efficient scrolling mechanism to steer its beam along a single cut plane. Our approach uses a flexible RA aperture that is wrapped around two columns, which can rotate, causing the aperture to roll, as shown in Fig. 1. The aperture is first synthesized to steer its main beam towards the broadside direction, as shown in Fig. 1(b). Then, the aperture is rolled as shown in Fig. 1(a) and Fig. 1(c), thereby defocusing the RA feed antenna causing the main beam to scan towards the direction of rotation.
Our novel rollable RA was first studied in [24], but only a preliminary study to validate the phase optimization procedure was presented. In this work, we present the theoretical background of our optimization procedure, and analytically investigate how the aperture layout and unit-cell size affect the beam-scanning performance of our design. Then, a rollable RA is synthesized to operate in the K u -band, and it is simulated using ANSYS HFSS R full-wave simulation software. A prototype is then fabricated and measured to validate our design. An accurate characterization of the flexible substrate is performed by conducting parametric simulations and unit-cell measurements. Finally, the effectiveness of our proposed optimization procedure is validated. The rollable RA prototype achieves 42 • of lateral beam-scanning while maintaining a maximum realized gain of 25.1 dBi with a 1.5 dB variation. Our rollable RA uses a simple and low profile mechanism, thereby eliminating the need for complicated, bulky, and inefficient feed networks or mechanical systems. This makes our design an ideal candidate for reconfigurable HGA applications in compact spaces, such as, SmallSats or tactical equipment.
The paper is organized as follows: Section II provides a brief background on SmallSat RAs and reconfigurable RA technologies. Section III details our optimization procedure for realizing the element phase distribution across the aperture. Section IV presents the synthesis of the RA system, the fabrication details of our prototype, and our simulation and measurement results. Finally, Section V provides our conclusions and highlights the key findings of our work.

II. BACKGROUND
Recently, the viability of deployable RAs on SmallSat systems has been demonstrated with NASA missions for deep space [25], and lower earth orbit [26] applications. However, these missions relied on either a reaction wheel system or an Attitude Determination and Control System (ADCS) to meet the pointing accuracy requirements of the antennas. This method of beam-scanning adds significant cost, mass, and complexity to the overall mission, and reduces the lifespan of SmallSats. Furthermore, some missions involve instrumentation that precludes the adjustment of spacecrafts' pitch, roll, and yaw [27]. Alternatively, RAs with beam-scanning capability provide unique advantages to distributed swarm antenna arrays [28], satellite clusters [29], and SmallSat constellations [30], [31]. Beam-scanning RAs are typically realized using either phase-tuning, or feed-tuning. In each of these approaches, the phase distribution, φ RP (x i , y i ), across the RA aperture is tuned to steer the main beam towards the desired direction, as shown in (1): where φ PP and −k 0 R i are the progressive phase and spatial phase delay of the i th unit-cell of the aperture, respectively [21]. The phase-tuning technique achieves pattern reconfigurability by tuning the φ PP of each unitcell. Proposed approaches include the use of electromechanical components [32], [33], [34], electronic components [35], [36], [37], [38], optical components [39], [40], [41], and smart materials [42], [43], [44]. Phasetuning approaches for RAs can achieve high-speed beamscanning but require complicated DC biasing networks, and materials, which in turn increase the cost of RAs and make them less competitive compared to phased array technologies [27], [45], [46], [47]. Alternatively, the feed-tuning technique tunes the spatial phase delay of the unit-cells. This is accomplished by changing the distance, R i , between the unit-cells and the feed antenna phase center. Proposed approaches include single feed systems that mechanically tune R i [48], [49], [50], or multiple feed systems that create discrete values of R i [51], [52], [53]. Multiple feed systems are not well suited for SmallSat applications since they add significant mass to the spacecraft and require individual feed networks and complicated deployment mechanisms. Also, current single feed systems are not ideal for SmallSats since they require inefficient mechanical systems to physically move either the feed antenna or the aperture. Furthermore, the large sweeping movements of feed tuning approaches cover a large volume of space, which could obstruct other instrumentation. Recently, alternative approaches have been proposed [54], [55], [56], which tune R i by folding the RA aperture instead. However, the drawback of this approach is its limited beam-scanning angle. In [57], Legay et al., proposed a reconfigurable reflectarray that scrolls through several radiating apertures printed on a single flexible substrate. However, their aperture design is bulky and does not provide continuous beam-scanning capability.
The benefits of our rollable RA design are that we can achieve high gain and continuous beam-scanning with a lowprofile, simple, and efficient aperture rolling mechanism. The trade-offs to such a simple design are that the direction of beam-scanning is limited to the axis of aperture rotation and the scanning speed is limited to the speed of rotation. However, 2D beam-scanning could be achieved by tilting the RA system along an orthogonal axis with a simple tilting mechanism, like a motor and hinge. It should be noted that tilting the aperture on an orthogonal axis changes the incidence angle on the array which could lead to degradation of the radiation pattern [20], but these effects could be mitigated by characterizing the elements under different angles of incidence and using a lookup table during aperture synthesis. Also, there are several SmallSat applications that do not require high-speed beam-scanning. Such applications can benefit from the beam-steering capabilities of our design. For example, our proposed design can use its beam-steering capabilities to: (1) satisfy the pointing accuracy requirements of satellite relay stations, (2) adjust the scanning area of remote sensing satellites, or (3) tune the focal point of a distributed swarm antenna array.

III. DESIGN METHODOLOGY AND THEORETICAL STUDY
In this section, our proposed synthesis process for our rollable beam-scanning aperture is described. Also, analytical studies are presented here to characterize performance of our proposed RA and demonstrate the effectiveness of our optimized synthesis process.

A. BEAM-SCANNING BY DEFOCUSING THE FEED ANTENNA
It has been shown that reflector antennas can achieve beamscanning by defocusing the feed antenna [58], [59], [60]. The effects of defocusing an RA are presented in [50], where it was found that the main beam can be scanned by laterally shifting the position of the feed. Alternatively, the RA can be defocused by axially shifting the aperture while maintaining the feed position fixed. Fig. 2(a) shows an RA, which is synthesized to point its beam towards the broadside direction. This RA is centered on the coordinate system with origin O 0 , which is shown in red. Fig. 2(b) shows the geometric configuration of the RA after the aperture is shifted along the −x-axis a distance d i , so that it is centered on the relative coordinate system with origin O i , which is shown in black. This lateral shift creates a similar defocusing effect as presented in [50], only in this case there is no secondary mechanism that re-aligns the feed antenna pattern with the center of the aperture. The defocusing angle, θ f i , caused by the shift creates a proportional beam-scanning angle, θ b i , that directs the main beam along b i , as shown in Fig. 2(b). However, laterally shifting the aperture causes the same problems as laterally shifting the feed antenna: (1) reduction in effective aperture size by a factor of cos θ f i , (2) reduction of aperture efficiency from increased spillover and non-uniform aperture illumination, and (3) introduction of phase errors due to the change in spatial phase delay of the elements [50].

B. APERTURE PHASE DISTRIBUTION OPTIMIZATION
To minimize all these effects, we propose a rollable RA design, as shown in Fig. 2(c). As the aperture rolls, the elements rolled out of the illumination window are out of the sight of the feed antenna. Simultaneously, optimized elements, shown in green, are rolled in the illumination window. The optimized elements serve three purposes: (1) maintain the effective aperture size constant, (2) maintain high aperture efficiency by minimizing spill over and maintaining uniform amplitude tapering, and (3) minimizing total phase errors of the illumination window. To minimize the total phase errors, the reflection phase of the elements rolled in the illumination window are optimized to radiate in the direction of the scanned beam, b op = b i , as shown in Fig. 2(c).
The flowchart for our proposed phase distribution optimization process is shown in Fig. 3. In Step 1, the phase distribution of a N×M element RA is synthesized to radiate a main beam towards the broadside direction (φ = 0 • , θ = 0 • ) using the ray-tracing method [20]. Then, in Step 2, the RA elements are rolled laterally a distance d, which is equal to the width of one element column, to the first rolling position, i = 1. Next, in Step 3, the column of elements shifted out of the illumination window is eliminated to simulate it being rolled behind the ground plane and out of sight of the feed antenna. In Step 4, the radiation pattern of the remaining N × (M − 1) element array is calculated using array theory [20]. In Step 5, the direction of the main beam, (φ b i , θ b i ), from the radiation pattern calculated in Step 4 is found. In Step 6, a column vector of N elements, positioned in the empty column created by the roll in Step 2, is synthesized to radiate towards direction (φ b i , θ b i ). In Step 7, the column of elements obtained from Step 6 are concatenated with the N × (M − 1) elements generated in Step 3 to form a new N × M phase distribution in the illumination window. Then, Steps 2 through 7 are repeated on the new N × M phase array to generate an optimized column of elements for the second rolling position, i = 2. This process is repeated for the number of column rotations required to achieve the desired beam-scanning angle. The result of the procedure is a N × (M + 2i) phase distribution matrix.
To illustrate how our optimization process works, we examine an example aperture with dimensions 10λ × 10λ with 0.5λ inter-element spacing. Fig. 4 shows the 20 × 20 element matrix for this aperture at three different rolling positions i = 0, 5, and 10. In Fig. 4(a), the RA aperture is shown in the reference position, i = 0, where the elements shown in black are synthesized in Step 1 to scan the main beam towards the broadside. After 5 iterations of the optimization procedure in the −x-direction, the RA is at the fifth rolling position, i = 5, resulting in 5 columns of optimized elements in the illumination window, shown in green in Fig. 4(b). After 10 iterations, the RA is at the tenth rolling position, i = 10, and there are 10 optimized columns on the illumination window, as shown in Fig. 4(c). If the optimization process is stopped after 10 column rotations, i = 10, in both directions, the resulting aperture phase distribution would consist of a 20 × 40 phase matrix, where the first 10 columns of the matrix are optimized for rolling in the +x-direction, the next 20 columns contain the phase distribution for the reference position, i = 0, and the last 10 columns are optimized for rolling in the −x-direction.

C. CHARACTERIZING THE ROLLABLE RA PERFORMANCE ANALYTICALLY
To study the performance of our proposed RA, we consider a square aperture with length A discretized by a square lattice with unit-cell length d, as shown in Fig. 4(a). In this setup, the feed is assumed to be a point source positioned at the broadside direction at a distance f /A = 0.8, where f is the focal distance. The feed antenna pattern is defined by cos 2q (θ ) [20], with a q-factor of 4.25. Four different aperture sizes are considered in our analysis: A = 10λ, 20λ, 30λ, and 40λ. For each of the these apertures, four unit-cell sizes, d, are analyzed: d = 0.3λ, 0.4λ, 0.5λ, and 0.6λ. Specifically, for each combination of the aperture and unit-cell size, a rollable aperture is synthesized using the procedure described in Section III-B. In total, 16 rollable RA apertures were synthesized. Then, the performance of each aperture is characterized by calculating the radiation pattern for each rolling position, i, analytically using array theory [20]. The results of our analytical study are shown in Fig. 5. Specifically, the directivity and beam direction are plotted versus the rolling position, i, in the −x-direction until the grating lobes of the rollable RA overtake the main beam (i.e., SLL = 0). The results for rolling in the +x-direction are symmetric; therefore, they are omitted for brevity. From the results, it  can be seen that for all the cases beam direction is a relatively linear function with respect to the rolling position, i. In other words, for any given case of aperture and unitcell size, the beam direction at any i can be approximated by a line with slope equal to the scanning precision (SP), which is the degrees of beam-scanning achieved per wavelength distance of translation ( • /λ). The maximum deviation of beam direction from the linear approximation was < 2 • for all cases.

1) EFFECT OF APERTURE SIZE ON BEAM-SCANNING PERFORMANCE
The theoretical maximum directivity of a reflecting aperture with area, A ap , is given by D 0 = 4π A ap /λ 2 , where λ is the wavelength at the frequency of operation. When the side length of the aperture, A, increases, the maximum directivity (D 0 ) of the aperture increases. However, increasing A reduces the scanning range due to a faster gain degradation. For example, in Fig. 4(c) it can be seen that for a maximum allowable directivity variance of 1.5 dB, an aperture with A = 10λ can achieve a scan angle of 33.4 • , while a A = 40λ aperture can only achieve 1.2 • . The maximum scanning angle is inversely proportional to A because larger apertures have more reflecting elements that introduce larger cumulative phase errors when the columns of the aperture are rolled. In turn, these larger cumulative phase errors generate side lobes and degrade the main beam. Therefore, increasing the aperture size to achieve higher directivity comes at the expense of smaller beam-scanning angle.
Furthermore, our results show that SP is inversely proportional to the aperture size, A. Specifically, as A increases, the focal distance, f , should increase to maintain high aperture efficiency [56]. Therefore, for larger apertures, a smaller defocusing angle θ f = arctan(d/f ) is created after each column roll. For example, Fig. 4(c) shows that an aperture with size A = 10λ or A = 40λ can steer the beam with a SP of 6.21 • /λ or 1.71 • /λ, respectively. Therefore, a larger aperture requires a higher number of rolled columns to reach a specific beam-scanning angle. It should be mentioned that SP is a function of the feed position and can be increased by using a less directive feed antenna which would reduce the f /A ratio.

2) EFFECT OF UNIT-CELL SIZE ON BEAM-SCANNING PERFORMANCE
In our design, the unit-cell size, d, determines the rolling step of the aperture. Therefore, the effects of d on the performance of the rollable RA were also investigated. Our results illustrate that d has negligible effect on the performance of our rollable RA. Specifically, Fig. 4 shows that for a given aperture size, A, the beam-scanning angle and SP remain relatively constant for variable d at the −1.5, −3, and −4 dB directivity levels. This demonstrates that our optimization process can be applied to any design regardless of unit-cell size. Notably, for a fixed size aperture, the interelement spacing of an RA is determined by the unit-cell size, and the amount of mutual coupling between the elements of an RA depends on its inter-element spacing [20], [61], [62], [63], [64]. Our study was conducted analytically using array theory; therefore, the effects of mutual coupling were not included, but should be considered in the design process.
This analysis shows that our novel phase synthesis technique achieves good beam-scanning performance for different combinations of A and d. However, it should be mentioned that the rollable RA aperture is a platform that can support any phase synthesis technique used for translating apertures. For example, a phase synthesis approach for a translating aperture is presented in [65]. Similar designs using a bi-focal approach were presented in [66] and [67]. These techniques, among others, could be implemented on the rollable RA aperture to achieve the desired beam-scanning performance.

IV. EXAMPLE ROLLABLE RA DESIGN
Our optimization procedure is used here to design a rollable beam-scanning RA. Specifically, in this section, we present the unit-cell characterization for our RA as well as our simulation and measurement results.

A. UNIT-CELL CHARACTERIZATION
Our proposed rollable aperture must use a flexible substrate that can wrap around two rotating columns. Apart from being flexible, the substrate must possess appropriate dielectric material properties to achieve adequate phase range and acceptable losses. Therefore, by carefully reviewing different materials, a sheet of 0.762mm thick low-density polyethylene (LDPE) was chosen for our application. Even though the properties of this material were reported by the manufacturer (i.e., r = 2.3 and tanδ = 0.001 at microwave frequencies), it was not clear what the material properties are at our design's operational frequency of 14 GHz. To accurately design our proposed RA and minimize any phase errors that would occur from not accurately knowing the material properties of our substrate, we chose to perform a material characterization. A waveguide measurement setup was established using a VNA and a WR-75 waveguide to replicate an infinite array scenario, as shown in Fig. 6. A custom flange was fabricated out of brass to fit the WR-75 waveguide. Reflecting unitcells with different sizes of square patches were fabricated and measured. Specifically, three batches of unit-cells with different sizes were made to characterize how the dielectric properties of our substrate vary across the material sample that we purchased. Each batch consisted of 13 different sizes of unit-cells, ranging from 5 to 8mm at 0.25mm increments. The unit-cells were made by applying copper tape with conductive adhesive to the top and bottom surfaces of the LDPE material. The patches were created on the top surface with a CNC cutting machine and the excess tape was removed. Then, the reflection phase and magnitude of each unit-cell were measured and recorded. The average curves of the reflection phase and magnitude of the three batches were calculated and fitted using a polynomial function.
Then, the measured unit-cells were simulated on ANSYS HFSS by modeling a waveguide with PEC boundaries fed by a wave port, as shown in Fig. 7(a). The dielectric properties of the material were found numerically by running parametric simulations until the results of the simulation matched our measurements. Through this process we arrived at these properties for our substrate: r = 2.1 and tanδ = 0.01. Finally, the material properties were simulated with a single unit-cell using master/slave boundaries, as shown in Fig. 7(b). The simulated results for the waveguide (WG Sim) and master/slave boundary (MS Sim) models are compared to the measured data and the polynomial fitting function of the average data of the three measured batches (AVG) in Figs. 7(c) and 7(d). The excellent comparison of the simulated and measured data validates the material properties that we calculated following our process. Also, Figs. 7(c) and 7(d) show that our unit-cell achieves a phase range of 300.4 • and a maximum reflection loss of 0.9 dB, which are both sufficient for our rollable RA design.

B. ROLLABLE RA SYNTHESIS, PROTOTYPING, AND MEASUREMENT
After characterizing the material properties of our flexible substrate, we proceeded to synthesize our rollable RA at 14 GHz. Specifically, our design has an illumination window of 12λ × 12λ and uses 0.5λ square unit-cells (i.e., A = 12λ, d = 0.5λ). The aperture phase distribution of our RA was synthesized using our optimization process and was designed to accommodate a maximum of 12-column rolls in either direction. Therefore, the total RA aperture size is 24 × 48 elements. A linearly polarized MVG SH2000 horn antenna was used as the feed. Its radiation pattern was measured and its maximum gain was 12.7 dBi at 14 GHz (i.e., q-factor = 4.25). The optimal feed position occurred at f /A = 0.75. This was calculated considering a feed offset of 20 • , to reduce feed blockage, and it corresponds to a maximum aperture efficiency of 70%.
Our RA design was then modeled and simulated using ANSYS HFSS, as shown in Fig. 8. The RA is excited by a linked far-field source whose radiation pattern is the imported measurements of the SH2000 horn antenna. A total of 25 simulations were performed, one for each rolling position. The simulation setup considered ideal conditions to reduce the computational load. These ideal conditions reduced the required computing time and memory by 96% and 86%, respectively, saving 915 hours of total computing time. To estimate the gain penalty imposed by deviations from the ideal simulation setup, an extensive tolerance analysis was performed. The effects of feed blockage, conductor losses, machining tolerances, surface roughness, aperture bowing, and alignment errors were accounted for by conducting parametric simulations. Table 1 lists the sources of loss and shows that for the worst-case scenario of fabrication and misalignment errors, the estimated gain is 24.9 dBi and the estimated aperture efficiency is 17.1%.
A prototype of our RA design was fabricated to validate its performance. Specifically, the aperture was fabricated using a CNC cutting technique. First, a layer of copper tape (which has conductive adhesive) was applied to the top and bottom surfaces of a 700 × 300 × 0.762 mm 3 sheet of LDPE with a press. Then, the 24 × 48 elements array of variable size patches was etched into the top layer with a precision CNC cutting machine, the Silhouette Cameo 4, which was found to have a machining tolerance of 250μm. Then, the ground plane was etched into the bottom layer using the same CNC cutting technique. Figs. 9(a) and 9(b) show our manufactured RA aperture. To create our rollable design, we sewed the ends of the aperture and re-enforced this connection with epoxy. Due to space limitations of our antenna chamber, a structure was designed to hold the aperture upright and roll vertically, as shown in Fig. 9(c). The flexible aperture was then wrapped around two pipes and tension was applied to create a flatter surface. However, approximately 1-2mm of quadratic bowing on the aperture was exhibited due to the low elasticity of the material. This bowing accounted for the largest source of loss (1.5 dB) which was determined through full wave simulation and measurement of an equivalent square aperture held flat by adhering it to a rigid plastic sheet. The performance of the proposed rollable RA could be significantly improved by using materials or manufacturing techniques that minimize aperture bowing. Notably, our manufacturing method was chosen due to our available equipment, low-cost material, simplicity, and adequate fabrication tolerances to demonstrate the rollable beam-scanning concept. However, this prototype is not optimized for space applications. A space bound aperture should consider the effects of several environmental and material factors, such as, temperature gradients, solar radiation, material stresses, and expected life cycle.  The measurements of our RA were conducted in an MVG StarLab near-field measurement system. The different rolling positions were realized by rotating the aperture manually. However, faster and more precise rotation can be achieved by using motors. Figs. 10(a) and 10(b) show the simulated and measured xz-plane co-pol normalized gain patterns for the RA rolling towards the −x and +x directions, respectively. These results show good agreement between the simulations and the measurements. Specifically, the maximum predicted and measured realized gain are 24.9 and 25.1 dBi, respectively. It is also seen that the measured gain is higher than the predicted gain which indicates that the errors in the measurement setup and fabrication process were lower than the most conservative estimate. This resulted in maximum predicted and measured aperture efficiency (η ap ) of 17.1% and 17.8%, respectively. The maximum deviation between the predicted and measured gain for all beam directions is 0.2 dB, and the average deviation is only 0.09 dB. The maximum deviation between the simulated and measured beam direction is only 1.8 • . This deviation is due to phase errors from the machining tolerance and alignment errors due to the manual rolling of the aperture. Notably, our prototyped design achieved 42 • of lateral beam-scanning while maintaining a maximum gain deviation of only 1.5 dB over the entire scanning range. A summary of the measured and simulated performance of our RA design is shown in Table 2.

V. CONCLUSION
In this work, a novel rollable RA with beam-scanning capabilities was presented. Specifically, beam-scanning is achieved by defocusing the reflecting elements of the aperture as it rolls. Also, a phase distribution optimization process was proposed to reduce the phase errors and achieve optimal performance. Simulations and measurements were used to characterize our RA design and they exhibited good agreement. Our prototyped design achieved realized gain of 25.1 dBi and 42 • of continuous lateral beam-scanning with a gain deviation of only 1.5 dB, while the SLL was maintained below −15 dB. Therefore, our proposed RA provides high directivity and beam-scanning capabilities using a simple, low mass, low profile, and efficient design, which makes it ideal for SmallSat applications.