UWB MIMO antenna configurations

Based on the previous discussion, the 4-element antenna is arranged as depicted in Fig. 6. We combined the two orientations of previous configuration of antenna 1 and 2. The antenna element at port 2 is oriented orthogonally on antenna at port 1 and 3 and oriented in the opposite direction with the antenna at port 4. The spaces between the edges of the antenna at port 1 and antenna at port 2, antenna at port 3 and antenna at port 4 equal 15 mm ((0.135λ ₀ at 2.7 GHz)), 29 mm (0.26λ ₀ at 2.7 GHz) and 11 mm (0.099λ ₀ at 2.7 GHz), respectively, as shown in Fig. 6(a).

Fig. 6. The configurations of suggested 4-element MIMO antenna (a) 3-D configuration with dimensions (b) fabricated photograph.

Scattering parameters and current distribution results

The suggested 4-element MIMO antenna with band notched behavior is fabricated as shown in Fig. 6(b) using R&S ZVB 20 four ports VNA to validate the antenna design by measuring its S-parameters performance as shown in Figs 7 and 8. Figure 7 displays the simulated and measured return loss results (S ₁₁, S ₂₂, S ₃₃, S ₄₄) using 4-ports VNA as shown in Fig. 7. It is seen from measuring results that, first the antenna is worked at frequency bands from 2.7 up to 11 GHz (UWB range) with good reflection coefficients performance lower than −10 dB at the four ports excluding the WiMAX band from 3 up to 4 GHz (the band notched region). Also, Fig. 8 presents the measured and simulated coupling coefficients (S ₂₁, S ₃₁, S ₄₁, and S ₄₂). The measured results have mutual coupling lower than −20, −24, −28, −27 dB between ports 1, 2, ports 1, 4, ports 1, 3 and ports 2, 4. The reason for the high isolation is the best choice of the antenna orientations as discussed before.

Fig. 7. The (S ₁₁, S ₂₂, S ₃₃, S ₄₄) performance (measured and simulated) of the suggested 4-element MIMO antenna.

Fig. 8. The coupling coefficients results (measured and simulated) of the proposed 4-element MIMO antenna (a) S ₂₁, S ₃₁, (b) S ₄₁, S ₄₂.

The measured results have good trend with the simulated ones. Nevertheless, the measured results have a little shift with the simulated results and this is due to the fabrication tolerance and soldering process.

Figure 9 illustrates the surface current behavior of the 4-element antenna at 3.5 GHz (the center band of the notch) and at 8 GHz. The current distribution is studied to show the behavior of the MIMO antenna at different frequency bands. First at 3.5 GHz as shown in Fig. 9(a), it is seen that the current is collected around the stub and prevented the antenna from radiation at this frequency band. Second at 8 GHz, the current is distributed along the radiator (ordinary radiation), also there is very weak current which is transferred to other antenna elements which means the high isolation between the four elements and this is actually based on the antenna orientations without using any decoupling structures.

Fig. 9. The surface current distributions of the suggested 4-element MIMO antenna (a) at 3.5 GHz, (b) at 8 GHz.

Radiation pattern results

In order to show the pattern diversity of the suggested MIMO antenna, the simulated 3-D pattern is investigated at different frequency bands as shown in Fig. 10. The desired port is excited and the other ports are terminated with matched loads as depicted in Fig. 10. It is noticed there is a progressive phase difference of 90° between each two successive ports that confirms the suitability of our proposed configuration to achieve the desired polarization diversity which in turn will improve the MIMO performance.

Fig. 10. The 3-D radiation patterns of the suggested 4-element MIMO antenna (a) at 3.5 GHz, (b) at 8 GHz.

The simulated x−z plane and y−z plane co-pol and X-pol radiation patterns at 6.5 and 9.5 GHz of port 1 are shown in Fig. 11. It is clear that the radiation pattern of the x−z plane is semi-omnidirectional with cross polarization lower than −15 dB. However, the radiation pattern of the y−z plane is bidirectional with cross polarization lower than −12 dB. The co-polarization has a higher and reasonable value than the cross polarization in both planes at 6.5 and 9.5 GHz.

Fig. 11. The co and cross polarization radiation pattern performance results of the suggested 4-element MIMO antenna at port 1 (a) x−z plane, (b) y−z plane.

The peak gain and efficiency of the proposed MIMO antenna at port 1 is depicted in Fig. 12. The antenna has an average gain and efficiency around 3.5 dBi and 85% within the entire frequency bands from 4.1 up to 11 GHz. However, the antenna has lower gain and efficiency at the notched band around −5 dBi and 18%, respectively. The normalized radiation patterns of the suggested 4-elements MIMO antenna is measured inside anechoic chamber at 3.5, 6.5 and 9.5 GHz in three planes (x−z, y−z and x−y). In the measuring setup; we excite port 1 and port 2 independently. When port 1 is excited, the other ports are terminated with matched load and vice versa.

Fig. 12. The peak gain and efficiency performance results of the proposed 4-element MIMO antenna at port 1.

It is obvious from Figs 13 and 14 that, first, the pattern has semi omnidirectional at both x−z and x−y planes at both two ports with 90^o rotation between them. Second, the radiation patterns at y−z plane are the same with the bidirectional shape when any of the two ports is excited. Third, the simulated and the measured results performance have the same trend with small shift between them because of the ACS feeding structure and measurements setup. Finally, it can be concluded that the proposed MIMO antenna produces the desired polarization diversity.

Fig. 13. The radiation pattern performance results of the suggested 4-element MIMO antenna at port 1 (a) at 3.5 GHz, (b) at 6.5 GHz, (c) at 9.5 GHz.

Fig. 14. The radiation pattern performance results of the suggested 4-element MIMO antenna at port 2 (a) at 3.5 GHz, (b) at 6.5 GHz,, (c) at 9.5 GHz.

The ECC

The first important parameter to evaluate the MIMO system is ECC. The ECC is defined as the correlation between the elements in MIMO system. Higher MIMO performance means lower ECC between elements. There are two equations that are used to extract the ECC parameter. The first one is extracted from the S-parameters assuming the environment with uniform multipath as [Reference Blanch, Romeu and Corbella24, Reference Abdalla and Ibrahim25]

(2)$${\rm ECC} = \rho _e = \vert {\rho_{ij}} \vert = \displaystyle{{{\vert {S_{ii}^\ast S_{ij} + S_{\,ji}^\ast S_{\,jj}} \vert }^2} \over {( {1-( {{\vert {S_{ii}} \vert }^2 + {\vert {S_{\,ji}} \vert }^2} ) } ) \mathop {}\limits^{} ( {1-( {{\vert {S_{\,jj}} \vert }^2 + {\vert {S_{ij}} \vert }^2} ) } ) }}$$

The second one is extracted from the radiation patterns results as [Reference Blanch, Romeu and Corbella24]

(3)$${\rm ECC} = \rho _e = \displaystyle{{{\left\vert {\mathop \int \nolimits \mathop \!\!\int \nolimits 4\Pi [ {F_1( {\theta , \;\varphi } ) \bullet F_2( {\theta , \;\varphi } ) {\rm d}\Omega } ] } \right\vert }^2} \over {\mathop \int \nolimits \mathop \!\!\int \nolimits 4\Pi {\vert {F_1( {\theta , \;\varphi } ) } \vert }^2{\rm d}\Omega \mathop \int \nolimits \mathop \!\!\int \nolimits 4\Pi {\vert {F_2( {\theta , \;\varphi } ) } \vert }^2{\rm d}\Omega }}$$

where F_i(θ, Ф) is the radiation pattern field with the excitation at port 1 and ● is the Hermitian product.

Based on Ref. [Reference Mao and Chu26], the agreeable limits of ECC is lower than 0.5. The ECC results (simulated and measured) between (ports (1,2), (1,3), (1,4), and (2,4)) are extracted from the previous two equations and displayed in Fig. 15. The value of the ECC at the frequency band (4–11 GHz) is lower than 0.0002, 0.0004 and 0.008 extracted from S- parameters, the measured results, and the radiation patterns parameters, respectively. However, the obtained values at the notched frequency band (3–4 GHz) are 0.15, 0.03, and 0.18, respectively. The lower values of the ECC between the antenna elements mean good MIMO features.

Fig. 15. The ECC performance of the suggested 4-element MIMO antenna (a) from S-parameter simulation, (b) from S-parameters measured, (c) from radiation patterns simulation.

Diversity gain (DG)

The second parameter to judge the behavior of MIMO systems is DG which is linked to ECC by the following equation [Reference Rosengren and Kildal27]:

(4)$${\rm DG} = 10 \times \sqrt {1-\vert {{\rm ECC}} \vert } $$

The DG results (measured and simulated) between (ports (1,2), (1,3), (1,4), and (2,4)) of suggested MIMO antenna are extracted and shown in Fig. 16. The DG equals around 9.97, 9.96 dB extracted from S-parameters (Fig. 16(a)) and the measured results (Fig. 16(b)), respectively within the frequency band (4–11 GHz), while the achieved values in the notched frequency band (3–4 GHz) are 9.1 and 9.85 dB for both simulation and measurement cases, respectively.

Fig. 16. The DG performance of the suggested 4-element MIMO antenna (a) from S-parameter simulation, (b) from S-parameters measured.

The CCL

The third parameter to estimate the MIMO performance is CCL (bit/s/Hz). The channel capacity is considered the rate of data which can be transferred over the communication channel [Reference Choukiker, Sharma and Behera28]. The CCL can be determined using correlation coefficient matrix and can be calculated using equations (5), (6) with agreeable value lower than 0.4 bit/s/Hz [Reference Choukiker, Sharma and Behera28, Reference Shin and Lee29].

(5)$$C( {{\rm Loss}} ) = {-}\log _2\det ( \psi ^R) $$

$$\psi ^R = \left[{\matrix{ {\rho 11} & {\rho 12} \cr {\rho 21} & {\rho 22} \cr } } \right], \;\rho _{ii} = 1-( {{\vert {S_{ii}} \vert }^2 + {\vert {S_{ij}} \vert }^2} ) $$

and

(6)$$\rho _{ij} = {-}( {S_{ii}^\ast S_{ij} + S_{\,ji}^\ast S_{ij}} ) , \;\,{\rm for}\,i, \;j = 1\,{\rm or}\,2$$

The CCL results (simulated and measured) between (ports (1,2), (1,3), (1,4), and (2,4)) are calculated using equations (5), (6) and illustrated in Fig. 17. The value of the CCL at frequency band from 4 GHz up to 10.5 GHz is lower than 0.4 bit/s/Hz except the notched frequency band.

Fig. 17. The CCL performance of the suggested 4-element MIMO antenna (a) from S-parameter simulation, (b) from S-parameters measured.

Multiplexing efficiency (η _mux)

The last parameter used for MIMO evaluation is η _mux. The η _mux is the ratio between the power of the real antenna to the power of ideal antenna [Reference Tian, Lau and Ying30]. The η _mux can be extracted assuming uniform environment and high signal to noise ratio (SNR) as [Reference Tian, Lau and Ying30]

(7)$$\eta _{{\rm mux}} = \sqrt {\eta _1\eta _2( {1-{\vert r \vert }^2} ) } $$

The η ₁, η ₂ are the total efficiency of the two real antenna elements and r is the magnitude of complex correlation coefficients between the antenna elements. The simulated η _mux between (ports (1,2), (1,3), (1,4), and (2,4)) are calculated using equations (7) and shown in Fig. 18. The η _mux equals around −1 dB within the band from 4 to 11 GHz, whereas the minimum value equals −8 dB within the notched frequency band.

Fig. 18. The simulated multiplexing efficiency performance of the suggested 4-element MIMO antenna.

Finally, to compare our study with recently published studies, Table 1 is constructed as illustrated. Table 1 shows that our antenna has compact size and provides high isolation with significant MIMO performance which permits our antenna to be utilized in UWB communications.

Table 1. Comparison between the suggested antenna and recently reported antennas