Design and Implementation of a 28-GHz Four-Phase Beam-Steering Antenna Based on a Butler Matrix Network

Abstract

This paper presents a four-phase beam-steering antenna for 28 GHz wireless communication, targeting the demand for high-efficiency and low-complexity beam-steering solutions in millimeter-wave systems. The proposed design employs a Butler matrix network to achieve multi-directional beam switching while reducing implementation complexity. The antenna is realized using microstrip technology on a printed circuit board (PCB), and the overall architecture consists of a 1 × 4 microstrip antenna array and a 4 × 4 Butler matrix network. Each component is carefully designed and analyzed to ensure optimized performance and proper system balance. The proposed antenna exhibits excellent performance in terms of bandwidth and compact size, while also providing advantages that include low cost, ease of fabrication, and structural simplicity. The beam-steering capability is experimentally verified through far-field measurements. The measurement results indicate that the four beam directions are −38°, −13°, +19°, and +41°, with corresponding gains of 8.79, 9.66, 10.8, and 9.21 dBi, respectively. In addition, a good agreement between the measurement and simulation results is observed, which validates the effectiveness and feasibility of the proposed design.

1. Introduction

Millimeter-wave antennas with multiple switchable beams have been widely studied for 5G communication systems. Various multibeam antenna designs and beamforming techniques have been proposed and applied [1,2]. In particular, beamforming networks (BFNs) and their integration with antenna arrays play a critical role in achieving efficient beam-steering performance. Due to the increase in the number of wireless communications and sensors in each wireless communication device, electromagnetic interference in each device and in different propagation paths must be adequately suppressed to mitigate multipath degradation and channel interference in complex environments [3,4]. Millimeter-wave multibeam antenna arrays have the advantage of high directivity radiation capability and wide coverage area. Compared to conventional phase-controlled antenna arrays, multibeam antenna arrays are gaining interest in modern wireless communication applications such as cellular phones, radar, and satellite communications. There are already multi-beam control using switching circuits [5,6,7].

With the beam switching network, the desired amplitude and phase characteristics of the antenna can be achieved. Blass matrices [8,9], Nolen matrices [10,11,12,13], Rotman lenses [14,15,16,17], and Butler matrices [18,19,20,21,22,23,24,25] are commonly mentioned in beamforming networks to enable beam switching antenna arrays to adjust various amplitudes and phases. Among these types, Butler matrix is one of the most promising techniques. The reason is its easy implementation with microstrip line technology and its advantages of simple architecture, compact size, cost-effectiveness, and wide beam coverage [26,27,28,29].

There have been many studies on beam control using Butler matrix with various techniques, such as single-layer substrate design [28,29,30,31], multi-layer substrate design [32,33,34,35], substrate integrated waveguide (SIW) [36,37,38,39,40,41,42,43,44], waveguide (WG) [45,46,47] techniques, etc. In [31], three novel designs are proposed for 27 GHz wireless communication, namely, a branch-line coupler with defective ground structure, a crossoverless crossover, and an LC phase shifter with high quality factor and quick response. The phase control of LC is faster and therefore applicable to 5G, but produces a low antenna gain. In [36], a Butler matrix multibeam antenna array with dual polarization was proposed. The two polarized radiation beams can provide a wide angular range between ±48° with a peak gain of 9.38/9.51 dBi. Although the realized gain is relatively high, the substrate synthesized waveguide structure is a complex structure, making the design difficult. In [45], the authors present an end-fire dipole antenna fed by a waveguide, where a broadband Butler matrix is implemented to provide power to a 1 × 4 antenna array. The multibeam antenna achieves wide bandwidth and high radiation efficiency; however, it remains challenging to fabricate in practical communication hardware.

Different beamforming architectures can be implemented in RF, IF, or LO paths depending on system requirements and implementation strategies. Among these approaches, RF-path beam switching based on passive Butler matrix networks provides advantages in terms of structural simplicity, low implementation complexity, and straightforward integration with antenna arrays. In addition, passive RF beamforming architectures avoid additional active control circuits while maintaining stable beam-switching characteristics through passive phase distribution. Alternative beamforming approaches, including IF-path and LO-path beam steering methods, have also been investigated in communication systems. In particular, adaptive DDS-PLL beamsteering architectures based on real-time direction-of-arrival (DoA) estimation provide another implementation strategy for adaptive beam control. However, such approaches generally involve additional active circuit complexity and control mechanisms.

Butler matrix antennas provide commendable performance; nonetheless, they frequently depend on intricate implementations, including multilayer substrates, substrate integrated waveguides (SIWs), and waveguide structures, which pose obstacles for system integration. This study offers a switchable-beam array antenna system functioning at 28 GHz, constructed with a single-layer microstrip design to implement the Butler matrix network. The employment of a low-loss, single-layer substrate facilitates diminished fabrication complexity and expense while preserving satisfactory antenna performance. The proposed antenna is evaluated using ANSYS HFSS 2022 R2 high-frequency simulation software and corroborated by construction and measurement.

2. Design Approach

The Butler matrix is a beam-controlled feeding network that can generate different radiation directions. It can be manufactured with microstrip lines and does not require any expensive active circuit equipment. The Butler matrix architecture is shown in Figure 1. The four beam directions for input ports 1–4 are denoted 1L, 2R, 2L, and 1R, respectively. This section integrates Butler matrix antennas on the same substrate. A Rogers RT5880 substrate with a thickness of 0.254 mm and a dielectric constant (${\epsilon }_r)$ of 2.2 is used to implement the Butler matrix antenna.

2.1. Analysis of the 4 × 4 Butler Matrix

The realization of progressive phase difference (PPD) at the output ports of the Butler matrix is analyzed and derived based on Figure 2. It consists of four 90° couplers and two crossovers forming four inputs and four outputs. When the input port 1 is stimulated, the signal is transmitted to the output ports 5–8, and the important positions are marked with capital letters, and the transmission paths are: A → B → C → D → E, A → B → C → F → G, A → H → I → J → K and A → H → I → L → M.

The PPD modulation factor that changes the direction of the beam is modified by a phase shift value. When input port 1 is activated, signals of −φ, −90°, −φ−90° and −180° phases are output from ports 5 to 8. The output port must have the same phase difference to satisfy (1):

$$\mathrm{\angle }{S}_{61}-\mathrm{\angle }{S}_{51}=\mathrm{\angle }{S}_{71}-\mathrm{\angle }{S}_{61}$$

(1)

The value derived from φ from (1) is 45°, and the PPD of the input port 1 is equal to −45°.

$$-90°+\phi =-\phi -90°+90°\to \phi =45°$$

(2)

Similarly, when the input ports 2 to 4 are excited, the output phase will satisfy (1) as shown in (3)–(5), respectively.

$$0°+90°+\phi =-90°-\phi -90°+0°$$

(3)

$$-90°-\phi -90°+90°=-0°+90°+\phi +90°$$

(4)

$$-\phi -90°+90°+90°=-90°+\phi +90°$$

(5)

Similarly, substituting φ = 45° in (3) through (5) produces a PPD of 135° for input port 2, a PPD of −135° for port 3, and a PPD of 45° for port 4. Therefore, it can be concluded that this structure produces PPD values of ±45° and ±135°.

Table 1. Output phase of Butler matrix with different input ports.

Port#	P1	P2	P3	P4
P5	$\phi$	$90°+\phi$	$90°$	$180°$
P6	$90°$	$0°$	$180°+\phi$	$\phi +90°$
P7	$90°+\phi$	$180°+\phi$	$0°$	$90°$
P8	$90°+90°$	$90°$	$90°+\phi$	$\phi$
PPD	−45	135	−135	45

shows the theoretical phase values of the Butler matrix output.

2.2. Hybrid Coupler

The hybrid coupler [48] is composed of one input, two outputs and one isolation port. The function of the hybrid coupler is to equally distribute the power across the output ports, and the two outputs have a phase difference of 90°.

Figure 3a shows the design of the 28 GHz hybrid coupler. When the impedances are matched, port 1 power outputs to port 2 and port 3 with the same amplitude and 90° phase difference. The hybrid coupler is mainly coupled between two sets of quarter wavelength microstrip lines and the auxiliary line, one with two 50 Ω and the other with two 35 Ω transmission lines, thus forming a square with a perimeter approximately equal to one wavelength. The width of the impedance line can be calculated by bringing the result of (7) into (6).

$$W={\displaystyle \frac{8h{e}^B}{e^{2B}-2}}$$

(6)

$$B={\displaystyle \frac{Z_0}{60}}\times \sqrt{{\displaystyle \frac{{\epsilon }_r+1}{2}}}+{\displaystyle \frac{{\epsilon }_r-1}{{\epsilon }_r+1}}(0.23+{\displaystyle \frac{0.11}{{\epsilon }_r}})$$

(7)

where W is the line width, the height (h) is 0.254 mm, and the dielectric constant (${\epsilon }_r)$ is 2.2. The matrix describing the parameters of the hybrid coupler S is shown in (8).

$${\left[S\right]}_{Hybrid}={\displaystyle \frac{-1}{\sqrt{2}}}\left[\begin{array}{cc}\begin{array}{cc}0& j\\ j& 0\end{array}& \begin{array}{cc}1& 0\\ 0& 1\end{array}\\ \begin{array}{cc}1& 0\\ 0& 1\end{array}& \begin{array}{cc}0& j\\ j& 0\end{array}\end{array}\right]$$

(8)

Figure 3b is the result of the simulated hybrid coupler S parameters, and it resonates between 27.5 and 29.5 GHz. At 28 GHz, $S_{21}$ and $S_{31}$ are about −3 dB, which justifies the hybrid coupler that divides the transmitted power equally between the output ports. In Figure 3c, the phase difference between the outputs is |−90.43°|, which fits the design requirements.

2.3. Crossover

The Butler matrix has the main problem of signal overlap. Therefore, a high isolation crossover can be overcome in adjacent ports [48]. Figure 4a shows the crossover architecture with a total of four ports, including two inputs and two outputs, and each neighboring port is isolated to complete the crossover design. In addition, the crossover is designed using a 50 Ω microstrip transmission line. The S matrix of the crossover is shown in (9).

$${\left[S\right]}_{Crossover}=\left[\begin{array}{cc}\begin{array}{cc}0& 0\\ 0& 0\end{array}& \begin{array}{cc}j& 0\\ 0& j\end{array}\\ \begin{array}{cc}j& 0\\ 0& j\end{array}& \begin{array}{cc}0& 0\\ 0& 0\end{array}\end{array}\right]$$

(9)

Figure 4b shows the simulated S-parameters when Port 1 is excited at 25.53–31.8 GHz. $S_{11}$, $S_{21}$ and $S_{41}$ are below −10 dB. $S_{31}$ is −0.82 dB at 28 GHz. These results satisfy reflection and isolation parameters and are suitable for use in the Butler matrix.

2.4. Phase Shifter

The purpose of using a phase shifter [48] is to change the phase direction of the wave. It is implemented in a microstrip design. Each line of the reference line ΔL introduces the phase shift angle given by (10).

$$\begin{array}{c}{\lambda }_g={\displaystyle \frac{{\lambda }_0}{\sqrt{{\epsilon }_{reff}}}}\end{array}$$

(10)

where ${\lambda }_g$ is the waveguide wavelength in the substrate, ${\lambda }_0$ is the wavelength in free space, and ${\epsilon }_{reff}$ is the effective dielectric constant of the microstrip line.

The total length of the theoretical transmission line of the 45° phase shifter is 8.547 mm at 28 GHz, calculated by (10). Figure 5a depicts the 45° phase shifter layout. $S_{21}$ is −0.5 dB and $S_{11}$ is −24 dB at 28 GHz in Figure 5b. For the phase shift, the phase of port 1 to port 2 is 45° at 28 GHz as shown in Figure 5c.

2.5. Design of Butler Matrix

The Butler matrix layout is composed of 4 inputs and 4 outputs via a hybrid coupler, divider, and phase shifter. When a signal is input to any of the input ports, the power is divided equally among the four outputs, and these four outputs have a phase shift and equal amplitude. Figure 6 shows the Butler matrix configuration designed at 28 GHz and simulated for this design architecture.

The simulation of Butler matrix S-parameters is shown in Figure 7 and Figure 8. The reflection coefficients of input ports 1 to 4 are below −15 dB at 28 GHz, presenting a good matching result and a large bandwidth between 26.5 GHz and 30 GHz. The Butler matrix in Figure 7 is a symmetric structure, so the reflection coefficient of port 1 is quite similar to Port 4, as well as for port 2 and 3. Simulation of the transmission coefficient results in the input power of each input port being equally distributed among the four output ports. When one of the inputs provides 1 W of power, each output port will receive 0.25 W of power, which can also be expressed as 10 log (0.25) = −6.02 dB. Figure 8 shows the transmission coefficients corresponding to the four input ports, the simulated S5i, S6i, S7i and S8i at 28 GHz in the range of −6.85 dB to −8.17 dB, where i = 1, 2, 3, 4.

According to Figure 9, the phases of output ports 5 to 8 at 28 GHz obtained when input port 1 is activated are −46.7°, −89.1°, −134.2° and −178.6° respectively, which are close to the ideal results of −45°, −90°, −135° and −180°. The average phase difference of the adjacent output ports turned out to be −43.96°, which is a tolerable error of 1.04° from the ideal −45°. The error of 1.04° is tolerable. The simulated phases of output ports 5 to 8 obtained from input port 2 excitation at 28 GHz are −133.6°, +9.1°, +135.3°, and −83.1°, as shown in Figure 10 and the results are close to the ideal −135°, 0°, +135°, and −90°. The average phase difference between adjacent output ports was 136.83°, which is approximately 1.83° off from the ideal phase difference of 135°. There is an error of about 1.83° from the ideal phase difference of 135°. Since the designed Butler matrix is symmetrical, the phase difference for the input port 3 and port 4 excitation is close to the value of the result for port 2 and port 1, but there is a negative sign difference. The phase simulation data for input port 3 and port 4 at 28 GHz are listed in

Table 2. Simulated phase difference.

No.	Adjacent Output Port Phase			Average Phase Difference	Theoretical Goals	Phase Deviation
	Port 5–6	Port 6–7	Port 7–8
Port 1	−42.4	−45.1	−44.4	−43.96	−45	1.04
Port 2	142.7	126.2	141.6	136.83	135	1.83
Port 3	−150	−116.9	−141.4	−136.1	−135	1.1
Port 4	45.7	38.1	42.8	42.2	45	2.8

.

Ideally, the output phase difference is −45°, +135°, −135°, and +45° for exciting input ports 1 to 4, respectively. The average phase differences obtained for the four simulated input ports are −43.96°, +136.83°, −136.1°, and +42.2°, which are within the acceptable range of phase errors for the output ports. These differences are due to the phase errors generated by the couplers, which indirectly cause the directional shifts in the beam generated by the matrix.

3. Antenna Element Design

3.1. Rectangular Patch Antenna

The four outputs of the Butler matrix are integrated with four antenna elements, and these antennas radiate beams in four different directions. A rectangular patch antenna is used as the radiating element [49], and its dimensions can be determined by (11)–(14).

$$W={\displaystyle \frac{c}{2f_0}}\sqrt{{\displaystyle \frac{2}{{\epsilon }_r+1}}}$$

(11)

$${\epsilon }_{eff}={\displaystyle \frac{{\epsilon }_r+1}{2}}+{\displaystyle \frac{{\epsilon }_r-1}{2}}{\left[1+12{\displaystyle \frac{h}{W}}\right]}^{-0.5}$$

(12)

$${\displaystyle \frac{∆L}{h}}=0.412{\displaystyle \frac{\left({\epsilon }_{eff}+0.3\right)\left(\frac{W}{h}+0.264\right)}{\left({\epsilon }_{eff}-0.258\right)\left(\frac{W}{h}+0.8\right)}}$$

(13)

$$L={\displaystyle \frac{c}{f_0\sqrt{{\epsilon }_{reff}}}}-2∆L$$

(14)

where $f_0$ is the design frequency, ${\epsilon }_r$ the substrate dielectric constant, ${\epsilon }_{eff}$ is the effective dielectric constant, W is the patch width, L is the patch length, and h is the substrate height. Since the edge effect, the patch length is extended by ΔL.

The 28 GHz rectangular patch antenna designed by electromagnetic simulation and optimization is shown in Figure 11a. The reflection coefficient and radiation direction are given in Figure 11b,c. The reflectance coefficient is below −10 dB in 27.7–28.44 GHz. The peak antenna gain at 28 GHz is 8.2 dBi and the front-to-back ratio is 18 dB.

3.2. Integration of Butler Matrix and Antenna Arrays

A Butler feeding network feeds the signal into a 1 × 4 antenna array operating at 28 GHz in Figure 12a. The rectangular patch antenna is connected to the output port of the Butler matrix, and the final beam formation can be guided by the Butler matrix and the selected beam direction. The dimensions of the Butler antenna are 40 × 43 × 0.254 ${\mathrm{m}\mathrm{m}}^3$ (3.733λ₀ × 4.013λ₀ × 0.0237λ₀). For practical measurement considerations, the Butler matrix antenna is designed with ground and feed lines at the front of the input port for connection and measurement verification using a 1.85 mm End Launch connector. Finally, a detailed dimensional drawing of the overall Butler matrix antenna, hybrid coupler, crossover, phase shifter and antenna is shown in Figure 12. The dimensions are listed in

Table 3. Detailed dimensions of Butler matrix antenna (Unit: mm).

Parameter	Value	Parameter	Value
WS	40	L1	4.7
LS	43	L2	0.725
W1	7	L3	1.2
W2	1.6375	L4	1.975
W3	2.725	L5	0.725
W4	1.725	L6	1.975
W5	8	L7	0.725
W6	0.2	L8	1.625
W7	0.725	L9	4.42
W8	1.5375	L10	0.35
W9	1.9	L11	0.725
W10	3.0125	R1	2
W11	3.425	h	0.254
W12	3.42
W13	1.0

.

The proposed Butler matrix antenna is simulated for each input port as shown in Figure 13. The calculated impedance bandwidth is 35.7% in 27.5–28.5 GHz with −10 dB as the reference. At 28 GHz, the reflection coefficient is lower than −25 dB, so the design simulation shows good reflection coefficient and bandwidth.

Figure 14 illustrates the xz-plane radiation patterns, where four distinct beams are generated by exciting input ports 1 to 4, respectively. Due to the symmetrical structure of the Butler matrix, the beam corresponding to port 1 is approximately symmetric to that of port 4, while the beam corresponding to port 2 is symmetric to that of port 3, with opposite beam directions.

Figure 15 depicts the radiation pattern for theta from −90° to +90° at 27.5, 28 and 28.5 GHz. The simulated beam directions at 27.5 GHz are +13°, −37°, +37°, and −13° with gains between 8.44 and 9.71 dB. The beam directions at 28 GHz are +12°, −35°, +35°, and −12°, with gains between 9.02 and 10.5 dB, respectively. The beam directions at 28.5 GHz are +11°, −37°, +37° and −11° with gains between 9.21 and 10.1 dB. Figure 16 shows the normalized radiation direction at different frequencies, from which it can be observed that the side channel of port 1 and port 4 starts from −6 dB, while the side channel of port 2 and port 3 has a tendency to be larger, probably because the input ports 2 and 3 of the Butler matrix are designed to be closer to each other, causing them to influence each other, so that the side channel has a tendency to be larger.

4. Fabrication, Measurement and Discussion

The phase-controlled millimeter wave Butler matrix antenna for the proposed beamforming system is depicted in Figure 17, consisting of a 90° hybrid coupler of a Butler matrix, a frequency divider, a 45° phase shifter, a phase-modulated microstrip line and a 28 GHz rectangular antenna. The antenna is printed on the top layer of 0.254 mm thick Rogers RT5880 substrate material, as well as reserved for 1.85 mm End Launch connector ground and feed transmission line, and finally the bottom layer is a perfect ground. The total dimensions of the manufactured antenna system including the reserved clamp space are 40 mm × 50 mm (3.733λ × 4.666λ) and 40 mm × 43 mm (3.733λ × 4.013λ) without clamps.

The reflectance coefficient was observed by an Agilent PNA N5227A network analyzer (Agilent Technologies, Santa Clara, CA, USA), and the instrument was calibrated using the SOL calibration method with a coaxial cable. The input signal is fed from the 1.85 mm End Launch connector to one of the input ports, while the three unused input ports are connected by 50 Ω resistors in Figure 18. The measured reflection coefficients of the four input ports are in Figure 19. Since the Butler antenna is symmetrical, the consistency of the simulation reflection coefficients for port 1 and 4 can be seen in Figure 19a, and the measurement results of port 1 and 4 both show a low point of resonance at 27.6 GHz and 28.6 GHz. However, the reason for the difference between the simulation and measurement results may be the effect of the bending of the substrate produced by the manufacturing. Then, port 2 and port 3 are in symmetric mode, and the reflection coefficients in Figure 19b show good consistency between the simulations and measurements, although there is a slight frequency bias between the two. The reflection coefficients for the four input port simulations were all below −20 dB at 28 GHz, and all measurements showed slight attenuation. However, all showed no significant effect on the antenna below −10 dB. This degradation may be caused by manufacturing, such as solder surface adhesion of resistive components with non-flat tin surface and bent substrate.

The radiation direction was measured in a reflectionless chamber using a manufactured antenna as the receiving antenna, and the measurement was carried out using the millimeter wave antenna measurement system (BWant-MW6) equipment. Theta angles of the non-reflective chamber equipment range from −90° to +90°, mainly in the northern hemisphere of the spherical coordinate system, as shown in Figure 20. Figure 21 illustrates the radiation directions corresponding to each of the four input ports by simulations and measurements at 28 GHz. The simulated peak gains for 2L, 1L, 1R and 2R directions are +9.02, +10.5, +10.2, +9.25 dBi, and the beam steering angles are −41°, −10°, +11° and +41°. The measured peak gains in the 2L, 1L, 1R and 2R directions are +8.79, +9.66, +10.8, +9.21 dBi, and the measured beam steering angles are −38°, −13°, +19° and +41°. Figure 22 illustrates the gain simulation and measurement when port 1 to port 4 are excited. The simulated gain at 27.5 to 28.5 GHz is 8.4 to 10.4 dBi, and the actual measured gain is between 6.1 and 10.9 dBi. The simulated and measured gains were in the range of 6.1 to 10.9 dBi. Except for a slight difference in gain and a significant difference in the 1L beam steering angle, the other input ports match well with the simulated results, and the errors between simulation and measurement were mainly caused by the substrate bending effect, the connector effect, and inaccuracies in the antenna set-up. The normalized beam direction at 28 GHz is shown in Figure 23. For 2L, 1L, 1R and 2R, the simulated side lobes are −1.59 dB, −5.71 dB, −5.47 dB, −3.73 dB; and the measured results are −2.9 dB, −3.74 dB, −4.25 dB, −4.74 dB. This may be due to the effect of the substrate fabrication causing bending.

5. Literature Comparison and Discussion

The performance of the proposed design is evaluated through comparison with recently reported millimeter-wave Butler matrix beam-switching antennas, as summarized in

Table 4. Comparison of recent literature.

Ref.	Proposed	[2]	[32]	[37]	[38]	[45]	[51]	[50]
fc (GHz)	28	28	28	28	29	30	28	27
Dielectric Layer	1	2	2	5	2	2	2	4
Substrates	RT5880	RF-30	Polytetrafluoroethylene	RO4003C	RT5880	Metal Waveguide	Merck GT7- 29001	TLY-5 RO4450B
Dielectric Constant	2.2	3	2.45–3.53	3.55	2.2	-	2.17/6.15	2.2/3.5
Degree	±14/ ±39	−43, +34	±60	±48	±135°	±14/ ±42°	−39 ~+36	±61°/ ±75°
Lobe (dB)	−3	−5.3	−6	−7	−15	−10	−6	-
Gain	8.8–10.8	5.8–6.7	5.6	9.51	11.6–14.3	10.6–13.9	8.5–9.9	6.3–8.9
Size (λ0 × λ0)	3.73 × 4.01	3.4 × 4.6	-	11.6 × 98.0	8.6 × 4.0	0.51 × 2.0	1.7 × 2.1	-
Design method	MS	SP4T	LC	SIW	SIW	WG	MS	Multi-layer

. The proposed antenna achieves competitive gain performance compared with [2,32,38]. In comparison with designs employing more complex architectures, such as those in [37,45,50,51], the proposed microstrip-based structure provides comparable beam angle coverage and radiation performance while maintaining a simpler configuration. One of the main advantages of the proposed design is its single-layer implementation, which reduces fabrication complexity and facilitates system integration. In addition, the compact structure and reduced antenna thickness make it suitable for practical 5G millimeter-wave applications. At the same time, the design reflects the typical trade-offs of planar passive beamforming networks. For instance, the microstrip implementation may introduce higher insertion loss compared with waveguide-based approaches, and the radiation characteristics can be influenced by layout and fabrication conditions. Overall, the proposed design provides a balanced trade-off between structural simplicity, compactness, and beam-steering performance.

6. Conclusions

In this paper, a 4 × 4 Butler matrix beam-switching antenna operating at 28 GHz has been designed, fabricated, and experimentally validated. The Butler matrix network and antenna array were thoroughly analyzed through full-wave simulations and measurements. The measured results demonstrate that the proposed antenna is capable of steering the radiation beam toward −38°, −13°, +19°, and +41°, with side-lobe levels ranging from −2.9 dB to −4.7 dB. The overall size of the integrated antenna system, including the Butler matrix and antenna array, is only 40 mm × 43 mm. Compared with recently reported Butler matrix beam-switching antennas, the proposed design achieves a favorable balance among antenna performance, structural simplicity, fabrication cost, and system integration. The single-layer microstrip implementation simplifies the fabrication process while maintaining stable beam-switching performance and favorable radiation characteristics. The measured results confirm the feasibility of the proposed architecture for practical 28 GHz 5G millimeter-wave communication systems. Although adaptive beamforming architectures may provide additional beam-control flexibility, they generally require more sophisticated active circuits and control mechanisms. In contrast, the proposed passive Butler matrix architecture emphasizes implementation simplicity, low-cost fabrication, compact size, and straightforward antenna integration. These features make the proposed design attractive for practical millimeter-wave beam-switching applications and provide a useful reference for future beamforming antenna developments.