Proposed antenna configurations

The geometrical description and the design process of quad-port UWB-MIMO antenna are discussed in this section. Figure 1 reveals the geometry of the proposed single band-notched UWB antenna. First, a rectangular-shaped patch antenna (Ant-1) considering W = 8 and L = 7 mm with partial ground plane is designed over an FR4 Epoxy substrate material using equations (1)–(5) [Reference Balanis2], as depicted in Fig. 2(a). However, it failed to satisfy the criteria of the specified UWB frequency range. Now to attain wide impedance bandwidth, Ant-1 is modified by vertically truncated rectangular slots on the bottom edge of the patch referred to as Ant-2 (see Fig. 2(b)). Unfortunately, Ant-2 dims to produce the desired operational bandwidth. Furthermore, to achieve better impedance matching within the desired UWB frequency range, Ant-2 is modified by introducing a bent-shaped conductor on the top of patch (see Fig. 2(c)). Moreover, incorporation of bent-shaped conductor provides uniqueness in the antenna geometry:

(1)$$W = \displaystyle{c \over {2f_0}}\sqrt {\displaystyle{2 \over {\varepsilon _r + 1}}} {\rm \;\;}\,$$

(2)$${\cal E}_{reff}{\rm \;} = \displaystyle{{{\cal E}_r + 1} \over 2} + \displaystyle{{{\cal E}_r-1} \over 2}{\rm \;}\left[{1 + 12\displaystyle{h \over W}} \right]{\rm \;\;\;\;}$$

(3)$$\Delta L = 0.412h\;\displaystyle{{{\rm \;}{\cal E}_{reff} + 0.3{\rm \;}[ {W/h + 0.264} ] } \over {{\cal E}_{reff}-0.258[ {W/h + 0.8} ] }}$$

(4)$$\;\;L_{eff}{\rm \;} = \displaystyle{c \over {2f_0\sqrt {{\cal E}_{reff}} }}{\rm \;}$$

(5)$$L = L_{eff}-2\Delta L\;$$

Now, to predict the behavior of radiating element, lower cut-off frequency for Ant-3 has been roughly obtained using the standard equation (6) given for the printed rectangular monopole antenna [Reference Ray29]; here, T1 and R1 are the length and width of the patch, “P” is the gap between ground and patch and K = $\surd \varepsilon _{reff}$:

(6)$$f_{L\;} = \displaystyle{c \over \lambda } = \displaystyle{{7.2} \over {\{ {( {T1 + r + P} ) \times K} \} }}\;\,\,{\rm GHz\;}$$

where r  =  R1/2π.

Fig. 1.

Suggested antenna geometry: (a) top view and (b) bottom view.

Fig. 2.

Step-wise design: (a) Ant-1, (b) Ant-2, (c) Ant-3, (d) Ant-4, and (e) Ant-5.

Now, to reject the unwanted interferences within the UWB band, the Ant-3 is modified by placing a pair of spiral EBG unit cells over the ground plane which is referred to as Ant-4 (see Fig. 2(d)), which possesses a very compact dimensions of 12 × 19 × 0.8 mm³. The other optimized dimensions (mm) are as follows: L_S = 19, W_S = 12, h = 0.8, W_p = 2.25, C ₁ = 2.5, L_p = 4.5, W_f = 0.6, L_g = 5, g = 0.3, L ₁ = 2.3, L ₂ = 1.92, W ₂ = 0.19, C_L = 21.4, T1 = 9, R1 = 11, and P = 2.5.

Once the desired requirements are achieved, the reported Ant-4 is further modified into a cuboidal four-port UWB-MIMO antenna by accumulating four antenna elements vertically wrapped around the polystyrene foam substrate material (ɛ_r = 1.2) to achieve overall compactness of the system. Generally, in planar MIMO antenna systems there is a restriction of antenna size, due to the placement of multiple antenna elements over the same plane. Simultaneously, the isolation among them also reduces. Thus, non-planar MIMO antenna systems have the benefits of miniaturize antenna structures over the planar counterparts. In addition to that, a 3D MIMO antenna also provides good spatial diversity by making the communication system robust and secured [Reference Srivastava, Kanaujia, Dwari, Kumar and Khan28]. The proposed MIMO design consists of microstrip line feeds which are represented as Port-1, Port-2, Port-3, and Port-4, followed by four identical F-shaped UWB antennas, which are arranged vertically in a cuboidal pattern referred to as Ant-5 (see Fig. 2(e)). The performance behaviors observed within all ports are almost identical and provides an acceptable mutual coupling reduction. In general, MIMO antenna systems should have a common ground plane rather than separated ground plane, for proper interpretation of signal levels based on the reference level. Although separated ground plane provides good isolation, in a real system it is not practical [Reference Sharawi30]. However, in open literature several MIMO antenna designs with separated ground plane are readily available [Reference Singhal8, Reference Bilal, Saleem, Abbasi, Shafique and Brown26, Reference Srivastava, Kanaujia, Dwari, Kumar and Khan28, Reference Li, Hei, Subbaraman, Shi and Chen31, Reference Tang, Wu, Zhan, Hu, Xi and Liu32]. Hence, due to the vertical orientation and separation of each antenna elements in our suggested design, practically it is not feasible to have a common ground plane.

Parametric analysis and discussion

The simulated performance characteristics for Ant-1 to Ant-4 are compared in Fig. 3. Thus, Ant-1 with its arrangement is resonating at 9.0 GHz (6.7–10.8 GHz), meanwhile Ant-2 is operating over a wide range of 3.7–9.1 GHz. Finally, a wide operational bandwidth of 7.6 GHz (3.1–10.7 GHz) is realized in Ant-3 by modifying Ant-2 with a bent-shaped conductor. The calculated lower cut-off frequency for Ant-3 is also obtained at 3.05 GHz using equation (6). A parametric analysis concerning L_g (length of the ground) is performed to evaluate the performance characteristics of Ant-3. It can be seen from Fig. 3(b) that on increasing L_g from 3 to 5 mm, the antenna contributes to improving the impedance matching. However, it is also observed that at L_g = 6 and L_g = 7 mm, the antenna suffers from impedance mismatching resulting in the degradation of performance. Hence, L_g = 5 mm is considered as the optimized value.

Fig. 3.

Performance parameters: (a) comparison of Ant-1 to Ant-4, (b) parametric analysis of L_g, and (c) parametric analysis of g.

Now, Ant-3 is equipped with a pair of spiral EBG structures (see Fig. 2(d)) to obtain notched band characteristics at 5.6 GHz (WLAN band) as shown in Fig. 3(a). The EBG structures are placed over the ground plane aligning with the feed line. A parametric analysis with respect to “g” (gap between EBG structures) is performed to understand the occurrences of notched frequency. From Fig. 3(c), it can be seen that on increasing gap the coupling between EBG structure decreases resulting in a decreasing notch peak level.

Experimental validation

Finally, a cuboidal four-port UWB-MIMO antenna is designed using Ant-4 to attain diversity scheme and its fabricated prototype is shown in Fig. 4. Thus, to observe the effects on performances, suggested antenna (Ant-5) is simulated considering air as a dielectric medium between the antenna elements and it is noticed that the reflection coefficient curves for both the cases are similar within all ports. The simulated and measured |S ₁₁| curve for the proposed quad-port UWB-MIMO antenna and its corresponding |S ₂₂|, |S ₃₃|, and |S ₄₄| curves are illustrated in Figs 5(a) and 5(b) for comparison. It is obvious that on inserting multiple antenna elements the notched center frequency is shifted to 5.4 GHz, having an identical performance within all ports. Figure 6 displays the simulated and measured isolation curve for the reported four-port UWB-MIMO antenna. It is seen from Figs 6(a) and 6(b) that mutual coupling between Port-1 and Port-3, Port-2 and Port-4 is around −25 dB within the working frequency band. This is due to the opposite placement of antenna elements. However, isolation between Port-1 and Port-2, Port-1 and Port-4, Port-2 and Port-3, and Port-3 and Port- 4 is around −15 dB within the working frequency band of 3–5.5 GHz, because of adjacent placement of antenna elements. Moreover, a close agreement is observed between simulated and measured results, considering slight variation at upper frequency due to lossy substrate materials.

Fig. 4.

Fabricated prototype for the proposed four-port UWB-MIMO antenna.

Fig. 5.

Simulated and measured reflection coefficient curves for the suggested antenna: (a) |S ₁₁| and |S ₂₂| and (b) |S ₃₃| and |S ₄₄|.

Fig. 6.

Simulated and measured isolation curves for the suggested antenna: (a) |S ₁₂|, |S ₁₃|, and |S ₁₄| and (b) |S ₂₃|, |S ₂₄|, and |S ₃₄|.

To validate the proposed antenna to be a MIMO antenna, certain parameters need to be calculated such as the envelope correlation coefficient (ECC) and diversity gain (DG). Moreover, these are the key parameters to ensure the performance of the MIMO antenna. The ECC can, therefore, be determined from S-parameters using the generalized formula presented in equation (7) [Reference Naji33]. Thus, in our proposed design, the ECC value is <0.2 within the working band except for 5–6.5 GHz as plotted in Fig. 7(a). The maximum DG of almost 10 dB is accomplished, except for the 4.8–6.5 GHz band (see Fig. 7(b)) which can be calculated using equation (8) [Reference Kumar, Urooj and Alrowais34]. Now, to understand the radiating mechanism of the antenna, the current concentration along the surface of the antenna is studied at notched frequency. Figure 8(a) shows the surface current distribution at 5.4 GHz when Port-1 is excited terminating other ports with 50 Ω matched loads. It can be seen that a strong current concentration is observed around the EBG structures; meanwhile, weak current concentration is observed on other ports. Similar behavior has also been observed when Port-2, Port-3, and Port-4 have been excited as shown in Figs 8(b)–8(d). The simulated and measured radiation patterns for the proposed quad-port UWB-MIMO antenna at two different resonating frequencies (3.4 and 7.4 GHz) are shown in Fig. 9(a). It is obvious that the antenna produces a fairly omni-directional radiation pattern in both planes. However, some fluctuation is observed in the measured result. Besides this, the antenna attains a maximum gain above 5.5 dBi at 10 GHz in both simulated and measured parameters as shown in Fig. 9(b). Finally, a comparison has been made between the existing literature studies in Table 1:

(7)$${\rm ECC\;}( {i, \;j, \;N} ) = \displaystyle{{\mathop \sum \nolimits_{n = 1}^N \vert S_{i, n}^{\rm \ast } S_{n, j}\vert ^2} \over {\mathop \prod \nolimits_{k = i, \;j} [ 1-\mathop \sum \nolimits_{n = 1}^N S_{k, n}^{\rm \ast } S_{n, k}] }}$$

(8)$${\rm Diversity\;}\,\,{\rm Gain}\,{\rm \;}( {{\rm DG}} ) = \sqrt { 1-{\rm EC}{\rm C}^ 2} {\rm \;}$$

Fig. 7.

Simulated performance parameters: (a) ECC and (b) DG.

Fig. 8.

Simulated surface current density by exciting (a) Port-1, (b) Port-2, (c) Port-3, and (d) Port-4.

Fig. 9.

Simulated and measured (a) radiation pattern and (b) gain curve.

Table 1.

Comparison with the existing literature