A novel approach to design compact, wideband, and high-gain circularly polarized MIMO antenna using dual-polarized elements and hybrid couplers Authors: H Tran, Phuong Kim-Thi, Nguyen Tran, Thao Hoang-Thi-Phuong, Yem Vu-Van Published: 2025 ABSTRACT AbstractThis paper presents a novel design approach for a compact, circularly polarized (CP) multiple-input multiple-output (MIMO) antenna that achieves wide bandwidth, high gain, and multi-port operation. Unlike conventional MIMO configurations, where each port feeds a single-polarized radiating element and gain enhancement relies on frequency-selective surfaces or T-dividers, the proposed design integrates dual-polarized CP radiating elements with hybrid couplers. Through this configuration, a four-port MIMO array is realized using only two radiating elements, significantly reducing the overall antenna size. The hybrid coupler enables simultaneous excitation of both elements, resulting in enhanced gain and improved radiation characteristics. A prototype is fabricated and experimentally verified. The measured results show that the antenna, with compact dimensions of 1.41 1.06 0.07, exhibits an operating bandwidth of 9.5% (5.0–5.5 GHz), port isolation better than 10 dB, and gain ranging from 7.8 to 8.8 dBi. Owing to its compact, wideband, and high-gain CP performance, the proposed antenna is well suited for modern wireless communication systems, including WLAN, Wi-Fi 6, C-band radar, and satellite IoT applications.Subject terms: Energy science and technology, Engineering Introduction Multiple-input multiple-output (MIMO) antennas have been widely used in modern communication systems due to their capability to enhance channel capacity without requiring additional spectrum1. Compared to linearly polarized (LP) antennas, circularly polarized (CP) antennas are more effective in mitigating multi-path effects. This work focuses on the design of compact CP MIMO antennas featuring multiple ports and high-gain performance. Various MIMO CP antennas utilizing microstrip patch structure to achieve low-profile configuration have been reported in2–8. In such designs, one MIMO port is connected to one radiating element. Besides, additional decoupling networks are also necessary to suppress the mutual coupling between the MIMO elements. Consequently, this type of MIMO antenna generally possesses extremely large overall dimensions as a number of MIMO ports is similar to a number of radiators. In addition, low gain performance is also another drawback of such designs. For gain enhancement, the conventional approach is a combination of single-polarized radiator and a frequency selective surface (FSS) layer9–14. Accordingly, high gain radiation can be achieved. However, this approach significantly enhances the antenna profile due to the FSS layers are positioned at a distance of at least quarter-wavelength from the radiators. Another method combines single-polarized radiator and T-junction power divider15–21. Here, one MIMO port is connected to at least two radiators by a T-divider. When the array elements are fed with equal magnitude and phase, high gain radiation can be achieved. Nonetheless, this method requires many radiators, which significantly enhance the antenna dimensions. Besides, the element spacing must keep around 0.5 to 0.7, which also increases the antenna size dramatically. To conclude, designing MIMO antennas with multiple ports and high gain with compact size is very challenging. The conventional methods commonly used single-polarized radiators as MIMO elements and FSS layers/T-dividers for gain enhancement. This paper proposed a novel approach to overcome the limitations of current designs. Instead of using a single-polarized radiator, the proposed approach employs two dual-polarized radiators and two hybrid couplers. Accordingly, a 4-port MIMO antenna system with high gain and compact size can be realized. The measured data demonstrates that the antenna with compact overall dimensions of 1.41 1.06 0.07 can achieve good performance over an operating bandwidth from 5.0 to 5.5 GHz, equivalent to about 9.5%. Across this band, the gain varies in the range from 7.8 to 8.8 dBi. Proposed solution This paper focuses on the solution to design multi-port MIMO antennas with high gain radiation and compact size features. To improve antenna gain, the possible solution is to increase the radiating elements. For this purpose, a power divider is necessary as it can be used to excite multiple radiating elements simultaneously. Generally, the conventional designs15–21 utilize T-junction power divider and single-polarized radiator, designated as Solution-1 in Fig. 1. This configuration has the drawback of employing many radiating elements, resulting in large antenna dimensions. The other solution is to combine the T-divider with a dual-polarized radiator, as Solution-2. It can be seen obviously that this combination has the advantage of reducing the number of radiating elements. For example, to design a 2-port MIMO antenna, Solution-1 requires 4 radiating elements, while Solution-2 only needs 2 radiating elements. Now, let’s consider when the hybrid coupler is used as a replacement for the T-junction divider, as a proposed solution in Fig. 1. Obviously, a 4-port MIMO antenna can be realized with only 2 radiators, which has a larger number of ports than Solution-2 despite using similar radiating elements. This contributes to significantly reducing the overall antenna size when designing multiple ports high-gain MIMO antenna arrays. Possible solutions to design multi-port high gain MIMO antennas. There are two critical tasks that need to be considered in this paper. Firstly, a dual-CP antenna with high inter-port isolation and wideband operation as well has to be designed. Secondly, using the proposed dual-CP antennas to design MIMO array configuration with the aid of hybrid couplers. Wideband dual-CP antenna Principles for isolation enhancement The most challenging task is to design dual-CP antenna with high isolation. Theoretically, the coupling of a right-hand CP (RHCP) and left-hand CP (LHCP) antenna can be shown to be minimum when their phase center is collocated (due to the orthogonality of RHCP and LHCP). Consider a typical asymmetrical dual-CP antenna structure, which uses a dual-fed patch with truncated corners. When both RHCP and LHCP are excited in a patch that has an asymmetrical geometry, the isolation becomes very poor. The main reason is the strong mutual coupling between modes that share the same polarization (horizontal-to-horizontal and vertical-to-vertical), much like what occurs when two linearly polarized antennas are placed close together. This issue is consistently seen in many other dual-CP antenna designs with asymmetric configurations, where achieving high isolation is notably difficult22–24. On the other hand, a symmetrical configuration inherently supports two orthogonal modes, which can be excited to generate RHCP and LHCP. When the phase centers coincide and the CP modes are excited correctly−meaning equal amplitude with a 90 phase difference−high isolation is achieved. As a result, the isolation in antennas with symmetric primary radiators depends heavily on polarization purity. Any deviation from ideal RHCP/LHCP purity (i.e., a non-ideal axial ratio) leads to leakage between the two senses. This behavior has been reported in numerous studies25–28, where different slot geometries are used to excite symmetrical patches or metasurfaces. Antenna design Figure 2 shows the geometrical configuration of the wideband dual-CP patch antenna. The patch and four parasitic elements are printed on Taconic RF-35 substrates with dielectric constant of . The dimensions of the patch and the parasitic elements are chosen about half-guided wavelength in the desired frequency ranges. For dual CP realization, two coaxial cables are utilized to excite the patch at different feeding positions, designated as P-1 and P-2. When P-1 is excited, the antenna produces LHCP waves and vice versa, RHCP is the dominant mode with P-2 excitation. The optimal dimensions of the proposed dual-CP antenna are as follows: , , , , , , , , (unit: mm). Geometry of the proposed dual-CP antenna with parasitic elements. Figure 3 shows the simulated S-parameter (reflection coefficient , transmission coefficient ) and AR results of the dual-CP antennas with and without parasitic elements. It can be seen obviously that the antenna without parasitic elements can achieve good matching and AR performance with the overlapped bandwidth of 6.4% (5.25–5.6 GHz). However, exciting both RHCP and LHCP in the same asymmetrical geometry patch results in extremely poor isolation 3.6 dB. This is due to the strong coupling between the modes with similar polarization (horizontal with horizontal polarizations, vertical with vertical polarizations), which is similar when placing two linear polarized antennas in proximity. Simulated performance of dual-CP antennas without and with parasitic elements. With the presence of parasitic elements, wider bandwidth and better isolation can be achieved. There is strong power coupled from the excited patch to the parasitic elements, which act as the main radiator of the antenna while the asymmetrical patch only acts as the feeding element. The operating bandwidth with isolation of better than 10 dB is about 14.8%, ranging from 5.0 to 5.8GHz. Here, the parasitic elements contribute to producing additional and AR bands in higher frequency regions, leading to wider operation bandwidth. Besides, the parasitic elements also act as the main radiating aperture of the antenna. Its symmetrical geometry make the phase center of the RHCP and LHCP antennas close to each other. Therefore, high isolation can be obtained. It is noted that the isolation of the antenna with four parasitic elements is strongly affected by the polarization purity. Imperfect RHCP/LHCP purity (axial ratio not ideal) produces leakage between senses. This can be demonstrated in Fig. 4, which shows the AR and for different corner truncation, s. Obviously, the isolation behavior strongly follows the AR characteristics. In Fig. 4, when the polarization purity is degraded, the dual-CP antenna exhibits extremely poor isolation between RHCP and LHCP ports. It is also worth noting that the feeding position (d)strongly affects the matching performance. Meanwhile, the AR performance is determined by the corner truncation of the patch (s). Simulated AR and of Design-2 for different sizes of truncated corner, s. Further demonstration can be observed in Fig. 5a, which shows the current distributions on the patch at 5.3 GHz. Here, the electromagnetic energy around the non-excited port of the antenna with parasitic elements is lower than that of the antenna without parasitic elements. Accordingly, better isolation can be obtained. Figure 5b shows the simulated vector current distribution on the patch (5.2 GHz) and the parasitic elements (5.7 GHz) with Port-1 excitation. In can be seen that two orthogonal fields are produced at phase 0 and 90. The rotation of vector current is clockwise direction, demonstrating the LHCP realization of the proposed dual-CP antenna. It is noted that the vector rotation will be inverse when the antenna is excited at Port-2. Simulated (a) current distribution with and without parasitic elements, and (b) vector current distribution on the patch (5.2 GHz) and the parasitic elements (5.7 GHz). Four-port MIMO array Antenna geometry Figure 6 depicts the geometry of the proposed 4-port MIMO antennas. Here, two dual-CP radiators are employed as the MIMO elements. To excite these elements, two hybrid couples (Coupler-1 and Coupler-2) are utilized, and they are positioned underneath the ground plane to minimize their effects on the radiation characteristic of the antenna. These couplers are printed on a low-cost FR-4 substrate. Coupler-1 has two input ports namely Port-1 and Port-2 and its output ports are connected to P-1 and P-2 of the radiators. Meanwhile, Coupler-2 has Port-3 and Port-4 as input ports, and it excited the antenna at P-3 and P-4. Accordingly, LHCP waves can be radiated with either Port-1 or Port-2 excitation. On the other hand, RHCP radiation can be produced when either Port-3 or Port-4 of Coupler-2 is excited. The radiating element and the coupler are separated by a continuous ground plane, which acts as an effective isolation layer. As a result, the feeding network does not disturb the radiation characteristics or isolation performance of the antenna. The optimized design parameters are as follows: , , , , , , , , , , , , , , , , , (unit: mm). (a) 2-D and (b) 3-D geometry of the proposed MIMO antenna. For example, RF signal comes to Port-1. This signal will not come to Port-2 due to the hybrid coupler is designed so that the isolation between their input ports is very high. The signal will come to radiators (truncated corner patches) at P-1 and P-2. Since the dual-CP radiators exhibit high isolation between the input ports, P-1 and P-3, P-2 and P-4 (as demonstrated in section “Wideband dual-CP antenna”). Consequently, there will be strong isolation between Port-1 and Port-3, -4. Similar phenomena can be obtained when the RF signal comes to Port-2, -3, as well as -4. It is also noted that Port-1 and Port-2 both provide RHCP, but different beam directions (due to different phases excitation). Thus, pattern diversity can be obtained. Meanwhile, Port-1 and Port-3 have similar beam directions, but different polarizations (LHCP for Port-1 and RHCP for Port-3). Thus, polarization diversity can be achieved. Antenna performance The simulated S-parameter results of the proposed MIMO antenna are shown in Fig. 7. Since the S-parameter results are quite similar for all ports, only the results for Port-1 and Port-4 are shown here. As seen in all cases, the antenna shows good impedance matching and inter-port isolation. The common operating bandwidth with isolation values of better than 10 dB is from 4.8 to 5.8 GHz. The simulated radiation patterns at 5.2 GHz in Fig. 8 indicate that the main beam is slightly tilted off in the broadside direction of about . This is attributed to the different excitation phases of these radiating elements due to the use of hybrid coupler, which has two outputs with equal magnitude and phase difference. The maximum gain values are quite high of around 9.2 dBi. The simulated AR results at the highest radiated direction are depicted in Fig. 9. Due to the asymmetric configuration, the AR results are quite different for each port. The data demonstrates that the antenna can work effectively in the frequency range from 5.0 to 5.6 GHz, in which the AR values for all ports are less than 3 dB. Across this operating bandwidth, the simulated radiation efficiency is better than 80%. Simulated S-parameter results of Port-1 and Port-4. Simulated gain radiation patterns at 5.2 GHz for different ports. Simulated AR for different ports. The most important parameter in this design is the element spacing, which not only affects the operating performance but also determines the compactness of the proposed antenna. Here, the matching and AR performance can be optimized through the patch, MS, and feeding position. The isolation will be better when the spacing between the radiating element increases. Figure 10 shows the antenna radiation patterns at 5.2 GHz for different values of element spacing (). As seen, increasing the spacing between the elements results in higher grating lobe level. The optimal value of is 30 mm. Simulated gain radiation patterns at 5.2 GHz for different values of . Finally, the gain enhancement and CP realization of the proposed antenna is confirmed by observing the surface current distribution at 5.2 GHz. As seen in Fig. 11, the power is distributed from Port-1 of the coupler to its output ports. Accordingly, both MIMO elements are excited simultaneously with equal magnitude and 90 phase difference, leading to high gain radiation. In terms of CP realization, the vector currents on the parasitic elements rotate in the clockwise direction when the phase changes from 0 to 90. This demonstrates the LHCP radiation of the proposed MIMO antenna when Port-1 is excited. Simulated surface current distributions at 5.2 GHz of the proposed antenna. Diversity performance One of the most important parameters to evaluate the efficiency of the MIMO system is the envelop correlation coefficient (ECC). The ECC is a key metric used to evaluate how independently multiple antennas in a MIMO system operate. Low ECC means the antennas provide diverse, uncorrelated signals, which improves system performance. Figure 12 shows the ECC calculation based on S-parameter results for Port-1 and Port-4 operation. It can be seen obviously that over the band from 5.0 to 5.6 GHz, the ECC values are all lower than 0.05. This demonstrates that signals are mostly uncorrelated, and then good isolation and diversity performance. Calculated ECC of the proposed antenna. Figure 13 shows the calculated Total Active Reflection Coefficient (TARC) and MIMO efficiency of the proposed antenna. The TARC quantifies the overall reflected power of a multi-port antenna under simultaneous excitation, accounting for both mismatch and mutual coupling effects. Meanwhile, MIMO efficiency measures the real transmission effectiveness of a MIMO antenna under simultaneous multi-port excitation. According to the data shown in Fig. 13, it is obvious that in the operating frequency band from 5.0 to 5.5 GHz, the TARC is below − 10 dB, which meets the excellent criteria for MIMO performance. Besides, the efficiency is in the range from 70 to 75%, demonstrating good overall multi-port radiation capability. Calculated TARC and total efficiency of the proposed 4-port MIMO antenna. Design guideline The steps to design the proposed 4-port MIMO antenna are summarized as follows: Step 1: Design dual-CP patch antenna The initial patch size () is chosen about half-effective wavelength at the desired frequency.Tune the corner truncation (s) to achieve good AR value.Tune the feeding position (d) to achieve good matching performance. Step 1: Design dual-CP patch antenna The initial patch size () is chosen about half-effective wavelength at the desired frequency.Tune the corner truncation (s) to achieve good AR value.Tune the feeding position (d) to achieve good matching performance. The initial patch size () is chosen about half-effective wavelength at the desired frequency. The initial patch size () is chosen about half-effective wavelength at the desired frequency. Tune the corner truncation (s) to achieve good AR value. Tune the corner truncation (s) to achieve good AR value. Tune the feeding position (d) to achieve good matching performance. Tune the feeding position (d) to achieve good matching performance. Step 2: Design dual-CP patch antenna with four parasitic elements The size of the parasitic element () is chosen about half-effective wavelength at the desired frequency, which is slightly higher than the frequency produced by the patch.Tune the corner truncation (s) to achieve good CP purity. Accordingly, high isolation can be obtained.Final tune all design parameters for the best performance in terms of matching, isolation, as well as AR characteristics. Step 2: Design dual-CP patch antenna with four parasitic elements The size of the parasitic element () is chosen about half-effective wavelength at the desired frequency, which is slightly higher than the frequency produced by the patch.Tune the corner truncation (s) to achieve good CP purity. Accordingly, high isolation can be obtained.Final tune all design parameters for the best performance in terms of matching, isolation, as well as AR characteristics. The size of the parasitic element () is chosen about half-effective wavelength at the desired frequency, which is slightly higher than the frequency produced by the patch. The size of the parasitic element () is chosen about half-effective wavelength at the desired frequency, which is slightly higher than the frequency produced by the patch. Tune the corner truncation (s) to achieve good CP purity. Accordingly, high isolation can be obtained. Tune the corner truncation (s) to achieve good CP purity. Accordingly, high isolation can be obtained. Final tune all design parameters for the best performance in terms of matching, isolation, as well as AR characteristics. Final tune all design parameters for the best performance in terms of matching, isolation, as well as AR characteristics. Step 3: Design a hybrid coupler working at the desired frequency The hybrid coupler with impedance values of and for its branches.The length of coupler is quarter-effective wavelength at the desired frequency. Step 3: Design a hybrid coupler working at the desired frequency The hybrid coupler with impedance values of and for its branches.The length of coupler is quarter-effective wavelength at the desired frequency. The hybrid coupler with impedance values of and for its branches. The hybrid coupler with impedance values of and for its branches. The length of coupler is quarter-effective wavelength at the desired frequency. The length of coupler is quarter-effective wavelength at the desired frequency. Step 4: Combine 2 hybrid couplers and 2 dual-CP radiators in Step 2 Final tune all design parameters for final optimization. Step 4: Combine 2 hybrid couplers and 2 dual-CP radiators in Step 2 Final tune all design parameters for final optimization. Final tune all design parameters for final optimization. Final tune all design parameters for final optimization. Measurement results To validate the design concept, a 4-port MIMO antenna prototype shown in Fig. 14 is fabricated and measured. The S-parameter is conducted with the aid of the Vector Network Analyzer type N5242A. Meanwhile, the far-field performance is tested in the anechoic chamber. Overall, the well-matched between simulations and measurements are obtained. The difference could be caused by the tolerance of fabrication and the imperfection in measurement setup. Photographs of the fabricated antenna. The simulated and measured S-parameter results with all-port excitations are presented in Fig. 15. It can be seen obviously that the common operating bandwidth for these ports is from 4.8 to 5.8 GHz, in which the matching performance is less than – 10 dB and the inter-port isolation values are always higher than 10 dB. Measured S-parameter of the proposed antenna. As the symmetrical configuration and the identical in the measured S-parameter results. The far-field performance with Port-1 and -2 excitations is chosen to present. The simulated and measured peak gain results are shown in Fig. 16. As observed, the maximum gain of 8.8 dBi can be achieved at 5.3 GHz, and the gain within the operating bandwidth is higher than 7.8 dBi. The gain radiation patterns in two principal planes at 5.2 GHz are plotted in Fig. 17. The dominant mode is LHCP, and the main beam is tilt-off the broadside direction in the plane. Meanwhile, the radiation patterns in the plane are almost symmetric. The back radiation is quite low of about 12 dBi. The simulated and measured AR at the direction of and are also illustrated in Fig. 18. It observed an overlapped 3-dB AR bandwidth for all ports from 5.0 to 5.5 GHz, equivalent to about 9.5%. This band is also covered by the impedance matching bandwidth. Simulated and measured peak gain values of the proposed antenna. Simulated and measured gain radiation patterns at 5.2 GHz. Simulated and measured AR of the proposed antenna. Performance comparison The comparison among high gain MIMO antennas is summarized and shown in Table 1. As seen, the novel features of the proposed antenna are multi-port operation and high gain with compact dimensions and the lowest number of radiators. Unlike the conventional approaches of using FSS layer, single-polarized radiator, and T-divider, the proposed approach utilized hybrid couplers and dual-polarized radiators to achieve high gain 4-port MIMO system with only two radiating elements. Using FSS layers can achieve high gain radiation, but trade-off with antenna’s profile10,13,14. Meanwhile, despite having low-profile structure, using the conventional method of using T-junction power divider and single-polarized radiating element requires large numbers of radiating elements, leading to bulky antenna structures18,19,21. Table 1Performance comparison among high-gain MIMO antennas.Ref.Overall size ()No. of radiatingelementsNo. ofportsHigh gainapproach10-dB isolationBW (%)15-dB isolationBW (%)Peak gain(dBi) 10 1.82 1.82 0.2522FSS layer2.02.09.8 13 2.06 2.06 0.1022FSS layer3.51.09.1 14 2.09 2.09 0.1644FSS layer43.443.410.5 18 2.24 2.24 0.0784T-divider16.516.510.3 19 6.00 2.30 0.0842T-divider8.78.79.5 21 1.87 0.54 0.0342T-divider1.01.05.4Prop.1.41 1.06 0.0724Coupler9.57.28.8 Performance comparison among high-gain MIMO antennas. Conclusion A MIMO CP antenna with multi-port, compact size, and high gain radiation characteristics has been presented and investigated in this paper. These novel features can be achieved by combining the dual-polarized CP radiating elements and the hybrid couplers. As a result, 4-port MIMO antenna can be realized with only two radiators. This is different from the conventional approach, which employs multiple radiators and thus significantly increases the antenna size. Meanwhile, the use of hybrid coupler can excite both elements simultaneously, leading to high gain radiation. Measurement on the fabricated antenna prototype demonstrates that the antenna can achieve operating bandwidth of 9.5% with isolation of better than 10 dB. Additionally, the gain within this band is always higher than 8.4 dBi and maximum value of 9.2 dBi can be obtained at 5.3 GHz. With its compact, wideband, and high-gain CP performance, the design is highly suitable for WLAN (5.15–5.35 GHz), Wi-Fi 6/6E (5.17–5.85 GHz), C-band radar (5.0–5.4 GHz), and satellite IoT communication systems (5.25–5.35 GHz). Author contributions Y.V.-V. and T.H.-T.-P. developed the conceptualization and supervised the project. H.T. performed the simulations, optimized the design, and wrote the original draft. N.T. fabricated the antenna prototype and conducted the experimental measurements. P.K.-T. curated the data and assisted with formal analysis. H.T. and N.T. contributed to reviewing and editing the manuscript. All authors have read and agreed to the published version of the manuscript. Funding This work was supported by the Vietnam National Foundation for Science and Technology Development (NAFOSTED) under Grant number 102.04-2023.28. Data availability Data is provided within the manuscript. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement Data is provided within the manuscript.