Beamforming MIMO Array Antenna for 5G-Millimeter-Wave Application

The high gain, high speed/data rate, high capacity, and beamforming antennas are required for the present generation of mobile and wireless applications to satisfy the exponentially growing demands of the users. This paper presents low mutual coupling multiple input-multiple output (MIMO) array antenna for millimetre-wave (mmw) application. The MIMO-array beamforming antenna with 2:1 VSWR band is proposed for 28.0 GHz and covers 27.04–28.35 GHz frequency band, which is suitable for mm-wave n261 5G-band which cover the frequency range from 27.5–28.35 GHz. It consists of 2×12\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$2\times 12$$\end{document} antenna array elements and the prototype is designed on low loss Rogers Duroid 5880 substrate of size 51.45×36.87mm2.\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$51.45 \times 36.87\,{\text{mm}}^2.$$\end{document} The beamforming MIMO antenna covers ±200\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\pm \,20^{0}$$\end{document} main lobe directions. The mutual coupling at the MIMO-array ports is less than 28.0 dB. The radiation efficiency and the gain in the presented band are more than 93.0% and more than 13.99 dBi. The ECC in the presented frequency band is ≤10-4\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\le 10^{-4}$$\end{document}, which is one of the advantages of the proposed design. The design covers indoor and outdoor Gaussian applications, and has 1.31 GHz TARC active bandwidth. It has 4.65% simulated and 4.73% measured fractional bandwidths.


Introduction
Due to the tremendous increase in the high speed data rate, wireless communication may requires 1000 times more capacity in the future. In order to meet this exponentially demand, the fifth generation of wireless technologies are currently being developed for the millimetre-wave (mmw) communications, waveforms, multiple access, massive MIMO with beam-steering and dense networks. Antenna is the most primary constituents for the successful deployment of the wireless communication. MIMO uses spatial multiplexing where each data stream can be beam-formed to increase throughput. It is based on the principle that when the received signal quality is high, it is better to receive multiple streams of data with reduced power per stream than one stream with full power, as is done by single-input-single-output (SISO). Using non-line-of-sight (NLOS) communication, MIMO solves many drawbacks of SISO. Fifth generation of mobile networks are expected to provide very high connectivity speeds for uplink and downlink (upto 10 Gbps and more), reduction in per-bit-cost, low power consumption, larger data distribution, and intended/ wide coverage for existing networking standards and allows integration of artificially intelligent (AI) devices, human-to-human communications, indoor hotspot (IH), dense urban (DU), and rural connectivity [1,2].
5G technology is expected to deliver the high data rate with low latency. The frequency band n261 27.5-28. 35 GHz mmw have numerous advantage as it offers high resolution, data transfer at high speed, cost effectives and increased security making the mm-wave band an ideal candidature for 5G Technology. It is also depicted that most of the countries are considering the 27/28 GHz band for mmw 5G communication. Beamforming is the technique to guide the main lobe radiation beam in the desired directions at the transmitters and receivers with the help of MIMO/array elements, nullifying the undesired for spatial selectivity. Beamforming in present scenario is the most deserving candidate in RADAR, seismology, biomedical, SONAR etc. The arrays, MIMO and Beamforming technologies are considered to be the key enablers for 5G mm-wave communication [3][4][5].
A T-shaped MIMO antenna resonating in 26.83-33.13 GHz and 34.17-38.13 GHz with split-ring resonator (SRR) and defected ground structure (DGS) was designed with Rogers Duroid 5880 for vehicular communication applications [15]. Quasi-Yagi antenna with Rogers RT/Duroid 5880 substrate was used to obtain broadband as well as multi-band operations in 27.0-29.0 GHz and 36.0-40.0 GHz frequency bands [16]. Plexiglass substrate can be used to cover 23.92-43.80 GHz frequency band, having more than 87.45% efficiency in band [17].
Dielectric resonator antenna (DRA) fabricated with Rogers 5880 substrate, resonating between 27.25-28.59 GHz and produced 24.0 dB isolation [26]. A synthetic aperture radar (SAR) having 8 rotated slot antennas covered 27.20-28.20 GHz frequency band was used for 5G communication [27]. 5G with coplanar waveguide-fed (CPW) and T-shaped radiating patch element designed with Rogers RT Duroid 5880 substrate were used to achieve better frequency response between 25.1-37.5 GHz frequency band [28]. An omnidirectional antenna having radius and height equal to 0.19 mm and 0.15 mm respectively were used to achieve minimum 2.2 dBi gain using Rogers 6006 substrate in 26.5-28.7 GHz frequency band [29]. Techniques such as SIW corrugated printed with Rogers 5880 substrate in 26.8-28.4 GHz frequency range [30], and enhanced central force optimization-Nelder-Mead (ECFO-NM) fabricated with Rogers Duroid RT 5880 substrate covering 28.0-38.0 GHz [31] were considered for better performance characteristics for 5G. For mobile phone circuit board, CP antenna with N9000 PTFE, covering frequency range of 27.0-40.0 GHz [32], and 28.0 GHz beam steering array on Rogers 5880 covering frequency range 27.5-30.0 GHz and having 16 cavity-backed slot antennas were designed [33] for 5G applications.
Above discussed antennas have limited gain in comparatively high sizes and very high values of envelope correlation coefficients (ECCs). Also, these considered references have not discussed about the mean effective gain (MEG) for indoor and outdoor applications, and total active reflection coefficient (TARC) for excitation angles at ports and active bandwidth. At very high frequencies, effects of these parameters play a very important role in designing of high isolation, high gain, and high efficiency antennas.
In this paper, two port 50 Ω MIMO-array antenna with 2:1 voltage standing wave ratio (VSWR) impedance band is proposed to resonant at 28.0 GHz frequency. The designed antenna covers 27.04-28.35 GHz frequency band (1.31 GHz bandwidth) and more than 28.0 dB isolation between ports. Radiators at ports are uniquely arranged in parallel fashion, having 45 0 chopping at the corner patches to have very low return loss. The designed MIMO-array antenna provides greater than 13.99 dBi gain in presented frequency band.

MIMO Beamforming Antenna Design
The computer simulation tool microwave studio (CST MWS) is used for optimizing and designing the proposed structure. The initial design parameters are decided using particle swarm organization (PSO). A compact 2-port MIMO-array antenna on Rogers Duroid 5880 substrate with dielectric constant (r) = 2.2 , loss tangent (tan( )) = 0.0009 is designed with full copper ground for the 5G application. The dielectric thickness of the substrate is 0.79 mm, while copper thickness is 0.035 mm. The 2 × 12 MIMO-array antenna is made on substrate dimension of 51.45 × 36.87 mm 2 . The connectors with the proper frequency range are used at the ports. The used connectors have proper calibration before the measurements. The center to center separation between two ports is 26.22 mm. This separation is required for the proper isolation between the radiating ports and for the maximization of the considered antenna performance parameters i.e. gain, efficiency, ECC, MEG, and TARC. This is done with the PSO optimization tool. For the proper tuning of the proposed MIMO-array antenna at the resonant frequency, electrical length is required. The electrical length ( ) for the designed MIMO-array antenna is 7.18 mm (this is corresponding to the length of the antenna equals 47.10 mm). Due to this the length of the antenna can be 6.57 (using mm unit).
The widths of rectangular patches are according with 50 Ω feed lines at each port. To decrease the return loss at ports, patches situated at the edges are cut at the corners. Elements at each patch are connected with 50 Ω feed with the help of 70.7 Ω quarter wave line for better impedance matching and desired performance parameters. The CST optimized values are given in Table 1 In sixth step, antenna is connected to 12 patches to single port and is the proposed structure to resonate at 28.0 GHz frequency. The considered single port antenna covers − 10.0 dB return loss band of 27.10-28.33 GHz, and provides 1.23 GHz bandwidth. Finally, seventh step is the    GHz is obtained using VNA. The measured bandwidth is 1.48 GHz. The difference in simulated and measured bandwidths is due to the port coupling losses and fabrication errors. The simulated and measured results are compared in Fig. 5. As gain is inversely proportional to bandwidth, therefore simulated gain of SISO is 16.07 dBi while that of MIMO is 15.46 dBi at 28.0 GHz. Due to simplicity, only S 1 1 and S 1 2 parameters are considered in this paper (because S 1 1 = S 2 2 and S 1 2 = S 2 1 ). The effect of mutual coupling or isolation between the radiating ports can be studied in terms of surface current distribution. Surface current distribution at each port can be obtained by exciting any one port at a time and vice versa of proposed antenna. For this, port 1 is excited and other port is terminated by 50 Ω impedance to observe the effect of current distribution. Each patch antenna gets the power from the lower patch antenna via feed arm. The red arrow is the representation of the maximum current concentrated on the feed arm and the different colored arrows represent the varying amount of current on upper sides of the antenna patches and feed arms. As shown in Fig. 6, maximum   The variation in length of patch (l5) from 1.67 to 5.67 mm is observed in Fig. 9. The width w5 is fixed at 4.23 mm. The length of patch is fixed at 3.67 mm. Single or multiple bands are obtained at other than 28.0 GHz frequency with high return losses as compared to considered length of patch. No major changes are observed in isolation parameter S 1 2 . The diversity parameters ECC, MEG, and TARC are analyzed here for the effectiveness of the proposed MIMO-array structure. The ECC includes all ports and their return-losses and isolation parameters of the designed MIMO-array antenna and can also be obtained in terms of the far field radiation patterns. Formula of ECC ( e ) is given in Eq. 1 using far field patterns [39]. Simulated value of ECC lies in the range ⩽ 10 −4 , which is very much lesser than the ECC value specified by ITU. The measured values of ECC using both the S-parameters and radiation patterns lie also in the range ⩽ 10 −4 . Due to ⩽ 10 −4 ECC values, it looks close to zero values on x-axis. ECC also shows the proper correlation/isolation/coupling between radiating ports. A comparison between the simulated and measured ECC is shown in Fig. 11. Only minute changes are observed in simulated and measured where F 1 and F 2 are far-field patterns of antenna 1 and antenna 2 respectively and Ω is the solid angle.
Another diversity parameter MEG is analyzed here for the applicability of the proposed MIMO-array design for indoor and outdoor environments. The proposed design is cross checked for isotropic environment with cross-polarization ratio (XPR) = 0 dB and 6.0 dB, and for Gaussian environment with XPR = 0 dB and 6.0 dB respectively [39,40]. The value of XPR equals 0 dB for outdoor environment, and XPR equals 6.0 dB for indoor environment respectively are considered. A comparison among all the MEG values with different environments with different XPR values is shown in Fig. 12.
Formulas of the MEGs are given by Eqs. 2 and 3 respectively: where G j and G j are the gain parameters in azimuthal and elevation planes, P j and P j are the probability distribution functions in azimuthal and elevation planes respectively. It is observed that, for isotropic environment with XPR = 6.0 dB, values of MEG are slightly greater than − 3.0 dB, although for XPR = 0 dB its value is around − 3.0 dB. Therefore, proposed MIMO-array antenna shows its suitability for outdoor environment only, and is not suitable for indoor environment. Similarly, for Gaussian environment with XPR = 0 dB and 6.0 dB, MEG values are lesser than − 5.0 dB. Therefore, proposed antenna shows the suitability for indoor as well as for outdoor environments. Hence, proposed MIMO-array antenna shows its strong candidature for Gaussian environment conditions. Now, a variation in XPR is also carried out to show the extension in the values of MEG. Figures 13 and 14 , show the variation in MEG for both the isotropic and Gaussian mediums with respect to the XPR values varying in the range − 50.0 dB to 50.0 dB. For isotropic medium, when XPR < 0 dB, MEG is slightly greater than − 3.0 dB, while the rest values are lower than − 3.0 dB in the considered 27.04-27.3 GHz frequency band. At XPR = 0 dB, its value is constant at − 3.0 dB for the whole frequency band. Similarly when XPR > 0 dB, MEG is slightly lower than − 3.0 dB in 27.04-27.17 GHz band, while the rest values are greater than − 3.0 dB for the remaining band. Similarly, for Gaussian medium, when XPR < 0 dB, MEG values are lesser than − 7.0 dB for the whole band. At XPR = 0 dB, MEG values are lower than − 6.0 dB, while for XPR > 0 dB, its value is lower than − 6.0 dB. A favorable figure of merit is obtained in terms of the MEG diversity parameter for validation of indoor and outdoor environmental conditions. Another diversity parameter TARC ( Γ t a ) is used to relate total incident power to total outgoing power for MIMO-array antennas. It is obtained using Eq. 4 [40]. The analysis of TARC is carried out in two steps. In first step, different angles of excitations at different ports are selected to find the variation in TARC. In second step, same excitation angles are selected at  where S i i and S j j are the return loss and isolation parameters at ports 1 and 2, and q is the input excitation angle. The gain of the proposed MIMO-array antenna is measured in an anechoic chamber using E-field, in presence of standard horn antennas and microwave generator suitable to cover the presented band. The simulated and measured gains in the presented band are achieved between 13.99-15.46 dBi for the two ports, and are 15.03-16.07 dBi for single port. Slightly higher value of gain is obvious, as bandwidth is slightly lower in single element. This is the main reason that the gain is slightly lower in MIMO antennas as compared to single element. The FRIIS equation is used to find the gain after the measurement of the received power in anechoic chamber, and is given by Eq. 5 [39]. where P r is the received power by proposed array antenna, P t is the transmitted power by the standard horn antenna, G t is the gain of the transmitting horn antenna, G r is gain of the receiving antenna, is wavelength, and R is distance between the antennas. The simulated radiation efficiency in the presented band lies in the range of 93.3%-94.56%, and the simulated total efficiency lies in the range 82.69%-93.3%. At 28.0 GHz resonant frequency, simulated and measured gains are 15.46 dBi and 15.21 dBi, and simulated radiation efficiency is 93.43%. The overall efficiency is calculated as per given in Eq. 6. The reason of high radiation efficiency is due to the lower losses and high impedance matching.
where e 0 is total efficiency, e r is reflection (mismatch) efficiency, e c is conduction efficiency, and e d is dielectric efficiency.
The product of e c and e d is known as the radiation efficiency. All the gains and efficiencies are compared in Fig. 17. The directivity (D) of the proposed array antenna can be obtained using the formula given in Eq. 7 [39].
where E is the half power beamwidth in E-plane, and is the half power beamwidth H in H-plane of the proposed array antenna.
The proposed structure can be tested for the real time environmental conditions for multipath propagation in CST simulation software as well as in an anechoic chamber. But the separation between transmitting horn (reference antenna) and device under test (proposed MIMO-array) must be minimum (around 10 cm or less) due to mmWave frequencies. Also the availability of the instruments for the outdoor environment is limited or not available in developing countries. The cost of investments on the research on mmWave frequencies and instruments can be in future at the full swing.

Conclusion
A 2-port MIMO-array antenna has been proposed for 5G applications to work for millimeter-wave frequency band. With 2:1 VSWR, proposed antenna resonated at 28.0 GHz and covered 27.04-28.35 GHz frequency band. The design exhibited more than 28.0 dB isolation between ports, more than 93.0% radiation efficiency, and 13.99 dBi minimum   (Table 2). At 28.0 GHz resonant frequency, obtained gain is 15.46 dBi. The ECC in the presented band is ≤ 10 −4 , which is one of the most important aspects of the designed prototype. The design showed its applicability in indoor and outdoor Gaussian applications completely, and has 1.31 GHz TARC active bandwidth. The future of the beamforming can be visible in high security, intended coverage, and in high speed communications with the spectrum transformation with lowest interference with living/nonliving bodies (Figs. 18 and 19).