Proposed antenna configuration and antenna evolution (single radiating patch)

This research discusses the evolution of the proposed single-element antenna which is modified to work with a four-port configuration and mitigates all the demerits possessed by a single radiating patch. Figure 1 shows the complete configuration of the proposed unit element antenna with a radiating patch resembling Calendula flower and rectangular ground etched with the rectangular slot. Figure 1(a) shows the perspective view of the antenna is fabricated on an FR4 substrate for the design of the proposed antenna. Figure 1(b) shows the detailed dimensions which are optimized by using an EM simulator. Also, it can be observed that the radiating patch is fed by an asymmetric 50 Ω microstrip line. The partial rectangular ground which is printed on the opposite plane of the substrate is etched with a rectangular slot behind the microstrip line. Figure 1(c) represents the detailing of the Calendula flower leaves and these leaves are aligned at different angles concerning origin (E1 = 165°, E2 = 015°, E3 = 90°, E4 = 105°, and E5 = 70° (anticlockwise direction)). All the optimized dimensions shown in Fig. 1 are recorded in Table 1 given below.

Fig. 1. Proposed 5G-NR antenna design with single radiation patch; (a) perspective view, (b) front view with a magnified patch.

Table 1. Proposed single and four-element 5G radiators parameter with values

The next step is to study the evolution of the proposed antenna which is shown in Fig. 2. Figure 2(a) depicts a circular patch antenna with the ground on the opposite plane containing the patch. This antenna #1 exhibits very poor matching of the impedance as observed in Fig. 2(e). The improvement of the impedance is noted by modifying antenna #1 to antenna #2 where the circular patch is etched by a circular slot and the partial ground is etched with a rectangular slot which is shown in Fig. 2(b). This modification improves the matching of impedance and the partial 5G NR bands are achieved. Further antenna #3 shows the addition of three fractal leaves with the patch connected to asymmetric feed. The lower bandwidth is improved when compared with the previous version. The final version of the antenna shown in Fig. 2(d) which is antenna #4 achieves all the 5G NR bands including n77, n78, and n79 which is the objective of the design. The modification includes the addition of two more cornered leaves and a rectangular stub attached to the ground providing an impedance bandwidth of 3.18–5.53 GHz.

Fig. 2. Evolution of the antenna (a) Antenna #1 (b) Antenna #2 (c) Antenna #3 (d) Antenna #4 (e) S11 result of Antenna#1–Antenna#4.

Study of key parameters affecting the bandwidth and surface current distribution

The proposed single-element antenna discussed in section “Proposed antenna configuration and antenna evolution (single radiating patch)” was the optimized version shown in Fig. 1. However, the key parameters which are optimized including Wf, axial ratio, L2, and Lg offer a vital role in the matching of the impedance and designing of the antenna, thus, change in their respective physical length does affect the operating bandwidth to a larger extent. These effects are seen by studying the parametric study shown in Fig. 3. The width of the microstrip, Wf is calculated by following equations

(1)$$Wf = \left({\displaystyle{{7.48 \times h} \over {e^{Zo{{\sqrt {\varepsilon r + 1.41} } \over {87}}}}}} \right)-1.25 \times t, \;$$

(2)$$Zo = \displaystyle{{120\pi } \over {\sqrt {\varepsilon eff\;\left[{\displaystyle{{Wf} \over h} + 1.393 + \displaystyle{2 \over 3}\ln \left({\displaystyle{{wf} \over h} + 1.444} \right)} \right]} }}, \;$$

(3)$$\varepsilon eff = \displaystyle{{\varepsilon r + 1} \over 2} + \displaystyle{{\varepsilon r-1} \over 2}\;\left[{1 + 12\;\displaystyle{h \over {wf}}} \right]^{{{-1} \over 2}}.$$

Fig. 3. Parametric analysis on; (a) feed width (Wf), (b) axial ratio (A.R), (c) truncated ground slot length (L2), (d) truncated ground length (Lg).

Here Zo is the impedance (Ω), Wf is the width of microstrip (mm), t is the trace width (mm), h is the height of the dielectric material used in this design, ɛr is the permittivity of the microwave substrate used.

As per the observations from Fig. 3(a), the width of the microstrip line changes the impedance of the feed which is governed by three equations, equations (1)–(3). The impedance of the feed line is designed for 50 Ω and the variation of width Wf also changes the impedance which in turn affects the overall impedance of the antenna. The values for Wf are changed from 2.10 to 2.50 mm, achieving the required bandwidth, but more matched impedance is observed for Wf = 2.30 mm with S11 = −44.86 dB at 4.00 GHz. Similarly, the partial ellipse used in the design with axial ratio = 2.0 shown in Fig. 3(b) also provides the required impedance bandwidth of 3.18–5.51 GHz with corresponding S11 = −44.86 dB at 4.00 GHz. The variation of L2 which is the height of the etched rectangular slot on the ground observes a larger deviation of the S11 parameter recorded in Fig. 3(c). For L2 = 3.00 mm which is an optimized value that achieves the 5G NR bands. Another important parameter Lg, which is the length of the ground, observes very poor matching of impedance for Lg = 11.50 mm and deviation of center resonance frequency to 4.48 GHz for Lg = 13.50 mm. After optimization, for Lg = 12.50 mm, the impedance bandwidth covers n77, n78, and n79 bands with S11 = −44.86 dB at 4.00 GHz.

The illustration of the surface current density at 3.30, 3.50, 4.00, 4.50, 4.80, and 5.00 GHz is shown in Fig. 4. Figures 4(a) and 4(b) show the distribution of surface current for 3.30, 3.50, and 4.00 GHz which covers the n77 and n78 bands. From the observations, the electric current is strongly distributed on the portion of microstrip and rectangular stub attached to the ground and confirms that these three-antenna areas form part of the radiating structure. It can also be depicted from Fig. 3(d) that the length of the ground does affect the matching of the impedance. Figures 4(d)–4(f) show the simulated current density distribution for 4.50, 4.80, and 5.00 GHz respectively for the n79 band. The same observations are noted that the surface current density is maximum within the feedline. For the remaining area of the antenna structure, the distribution of surface current is maximum, thereby ensuring all the input signals are radiated.

Fig. 4. Surface current distribution of 5G single radiator at various frequencies; (a) at 3.30 GHz, (b) at 3.50 GHz, (c) at 4.00 GHz, (d) at 4.50 GHz, (e) at 4.80 GHz, (f) at 5.00 GHz.

Modified 4 × 4 MIMO antenna (without and with stub)

The high throughput in the limited bandwidth environment is solved in multiple antenna configurations or MIMO systems. The demand for the faster transmission of higher data concerning time lead to the evolution of the MIMO antenna maintaining enhanced bandwidth/throughput of the data under different conditions including interference, fading of the signal, and multipath followed by signal at the receiver side. It is known from the Shannon–Hartley theorem [Reference Shannon36] that the capacity of the channel is given by

(4)$$C = BW\;{\rm lo}{\rm g}_2( {1 + SNR} ) , \;$$

where C is the channel capacity, BW is the bandwidth, and SNR is a signal-to-noise ratio.

From above equation (4), it can be concluded that the channel capacity can be increased by increasing more number of radiating antennas rather than a single one. A MIMO channel capacity is calculated by

(5)$$C_{MIMO} = N \times BW\;{\rm lo}{\rm g}_2( {1 + SNR} ) .$$

Equation (5) suggests that increasing the number of radiating elements rather than increasing channel SNR which ends up in marginal gains increases the channel bandwidth and hence higher throughput is achieved.

Based on the above concept, a 4 × 4 MIMO antenna configuration is developed which is the modified version of the single radiating antenna discussed in previous sections. All the key results corresponding to MIMO configuration including bandwidth, diversity performance, and far-field results are discussed below.

Figures 5(a)–5(d) show the design of the 4 × 4 MIMO antenna which is the modified version of the antenna shown in Fig. 1.

Fig. 5. Parametric analysis on substrate of isolated ground; (a) S11 values, (b) S21 or S41 values, (c) S31 values, (d) S-parameters of 60 × 60 mm² size 5G radiator with isolated ground.

Figure 5(a) shows the reflection coefficient (S11 in dB) curves with a 4 × 4 MIMO antenna configuration. The unit cell or single radiating antenna is placed in orthogonal sequence with their respective ground to achieve spatial diversity. Figure 5(a) shows the simulated S11 curves for different sizes of the antenna (54 × 54, 56 × 56, 58 × 58, and 60 × 60 mm²). The optimized size of the antenna with 60 × 60 mm² achieves the required 5G bands and also the center frequency. It is worth noting that there is no decoupling structure used, but Fig. 5(b) which shows the transmission coefficient curve between port 1-port 2 and port 1-port 4 observes poor isolation between them due to the reason that they are placed adjacently. Figure 5(c) which shows the isolation graph (S31 in dB) between port 1-port 3 records better values of isolation in the operating band suggesting that these two radiating patches are not adjacent but are diagonally placed in an orthogonal fashion. Also, the proposed antenna as observed in Fig. 5(c) shows better S13 results. Figure 5(d) compares the reflection and transmission coefficients of port 1 concerning port 2, port3, and port 4. The required bandwidth for 5G bands is achieved but the matching of impedance is poor and can be further improved when a decoupling structure is used. Also, the S31 shows good isolation, but S21 and S41 offer poor isolation between them and need to be improved.

Figure 6 shows the distribution of surface current density for a four-element 5G radiator with the isolated ground for frequencies 3.30, 3.50, 4.00, 4.50, 4.80, and 5.00 GHz. This simulation is obtained by exciting port 1 and matching all the remaining ports (port 2, port 3, and port 4) with the matched impedance of 50 Ω. As per the observations from all the six figures shown in Fig. 6 for different frequencies, it can be observed that the interference between the adjacent radiating antennas is more while the antenna placed diagonally has less interference which is already shown in Figs 5(c) and 5(d). The above-discussed results for MIMO antenna will not result in very good diversity performance, and hence further modification or addition of decoupling structure becomes necessary.

Fig. 6. Surface current distribution of four-element 5G radiator isolated ground at various frequencies when port 1 is excited; (a) at 3.30 GHz, (b) at 3.50 GHz, (c) at 4.00 GHz, (d) at 4.50 GHz, (e) at 4.80 GHz, (f) at 5.00 GHz.

The 4 × 4 MIMO antenna discussed in Fig. 5 inherits demerits such as poor isolation and hence poor diversity performance. This problem is overcome in the proposed 4 × 4 MIMO antenna configuration shown in Fig. 7. Figure 7(a) shows the prototype of the proposed 4 × 4 MIMO antenna configuration with a decoupling structure connected to all the grounds. The modification of the MIMO antenna ensures better isolation between all the ports which was not available in the earlier proposed MIMO design. The rectangular strip which is interconnected with the ground as shown in Fig. 7(b) ensures good impedance matching for the desired n77, n78, and n79 bands as shown in Fig. 7(d). It can be observed that the simulated S11 parameter offers the maximum value of −44.68 dB at 4.18 GHz, while the measured values are −50.28 dB at 4.52 GHz. In both the simulated and measured results comparison, the MIMO antenna covers all the 5G bands. Similarly, the proposed MIMO antenna also offers better isolation of more than −15 dB for both simulated and measured results.

Fig. 7. (a) Prototype of 4 × 4 MIMO antenna with decoupling structure; (b) simulation snapshot; (c) simulated and measured S-parameters; (d)–(i) surface current density distribution at 3.30, 3.50, 4.00, 4.50, 4.80, and 5.00 GHz.

Figures 7(d)–7(i) show the distribution of surface current for 3.30, 3.50, 4.00, 4.50, 4.80, and 5.00 GHz. Port 1 corresponding to antenna #1 is excited while all the remaining ports (port 2, port 3, and port 4) are terminated with a matched impedance of 50 Ω. The transmission coefficients shown in Fig. 7(c) offer isolation of more than −15 dB and this is due to the reason that the additional decoupling structure used in the ground provides an additional path for the flow of current and hence prevents the interaction between all the remaining radiating antennas, thereby reducing the interference which was observed for MIMO antenna without decoupling structure.

Diversity performance

There are different diversity schemes such as space diversity, frequency diversity, angle diversity, time and multiple path diversity, and polarization diversity. The fading of the two signals will differ when two or more radiating antennas are separated by a minimum distance of 0.5λo (where λo is the operating wavelength corresponding to the frequency of 4.00 GHz in the proposed 4 × 4 MIMO antenna configuration (Table 2).

Table 2. Comparison of S11, S21 (or) S41, and S31 of 5G radiators

Envelope correlation coefficient (ECC) which is a very important parameter of diversity performance is calculated by using either a far-field radiation pattern or using scattering-parameter. Calculation of ECC from S-parameter makes assumptions that all the signals which are fed to the antenna are uniform spread, as well as all the radiation elements are well matched and possess no loss. The following equations show the calculation of ECC using radiation patterns concerning fields radiated and S-parameter (2 × 2 and 4 × 4 MIMO) [Reference Chandel, Gautam and Rambabu3, Reference Karimian, Soleimani and Hashemi27]

(6)$$ECC_{m \times n} = \displaystyle{{{\left\vert {{\rm \iint }\overrightarrow {F_m} ( {\theta , \;\phi } ) .{\rm \iint }\overrightarrow {F_n} ( {\theta , \;\phi } ) \;d\Omega } \right\vert }^2} \over {{\rm \iint }{\vert {\overrightarrow {F_m} ( {\theta , \;\phi } ) } \vert }^2d\Omega {\rm \iint }{\vert {\overrightarrow {F_n} ( {\theta , \;\phi } ) } \vert }^2d\Omega }}, \;$$

(7)$$ECC_{m \times n} = \displaystyle{{{\vert {S_{11}^\ast \;S_{12} + S_{21}^\ast \;S_{22}} \vert }^2} \over {( {1-{\vert {S_{11}} \vert }^2-{\vert {S_{21}} \vert }^2} ) \;( {1-{\vert {S_{22}} \vert }^2-{\vert {S_{12}} \vert }^2} ) }}.$$

For the four-port MIMO antenna configuration, ECC is evaluated by

(8)$$ECC_{m \times n} = \displaystyle{{{\vert {S_{11}^\ast \;S_{12} + S_{12}^\ast \;S_{22} + S_{13}^\ast \;S_{32} + S_{14}^\ast \;S_{42}} \vert }^2} \over { {( {1-( {{\vert {S_{11}} \vert }^2 + {\vert {S_{21}} \vert }^2 + {\vert {S_{31}} \vert }^2 + {\vert {S_{41}} \vert }^2} } ) } ) {( {1-( {{\vert {S_{12}} \vert }^2 + {\vert {S_{22}} \vert }^2 + {\vert {S_{32}} \vert }^2 + {\vert {S_{42}} \vert }^2} } ) } ) }}.$$

In equation (6), Fm and Fn have radiated fields of the mth and nth antennas. For the ideal MIMO array, the ECC is zero, but for practical cases, these values are expected to be <0.50.

The effective diversity in the communication channel is given by the diversity gain which shows the dissimilarity between the time-average SNR signals. This compares the diversity of several radiating antennae when compared with the single antenna system. The DG which is evaluated by equation (9) is related to ECC as given by

(9)$$DG_{m \times n} = \sqrt {1-ECC_{m \times n}^2 } .$$

The values of directive gain (DG) should be ideally 10 dB, and in the proposed antenna, these values are approximately 10 dB as noted in Fig. 8(b).

Fig. 8. Diversity performance; (a) ECC_4×4 (simulated and measured), (b) DG_4×4 (simulated and measured), (c) CCL_4×4 (simulated and measured), (d) MEG_4×4 (simulated and measured), (e) TARC_4×4 (simulated and measured).

The capacity of the channel signifies the efficient transmission of the signal with no distortion or loss of data bits in the communication environment. However, the ideal condition is not achieved, and hence channel capacity loss (CCL) has to be evaluated which is given by the following equations

(10)$$Channel\;capacity\;loss_{m \times n} = -{\rm lo}{\rm g}_2( {\Omega^{MIMO}} ) , \;$$

where

(11)$$\Omega ^{MIMO} = \left[{\matrix{ {\Omega_{11}} & {\Omega_{12}} \cr {\Omega_{21}} & {\Omega_{22}} \cr } } \right], \;$$

(12)$$\Omega _{11} = 1-[ {{\vert {S_{11}} \vert }^2 + {\vert {S_{12}} \vert }^2} ] , \;$$

(13)$$\Omega _{22} = 1-[ {{\vert {S_{22}} \vert }^2 + {\vert {S_{21}} \vert }^2} ] $$

(14)$$\Omega _{12} = {-}[ {S_{11}^\ast S_{12} + S_{21}^\ast S_{12}} ] , \;$$

(15)$$\Omega _{21} = {-}[ {S_{22}^\ast S_{21} + S_{12}^\ast S_{21}} ] .$$

The maximum allowable CCL values or ideally CCL < 0.40 b/s/Hz and in the proposed 4 × 4 MIMO antenna configuration, the values of CCL are 0.04 b/s/Hz which are better than 10 times the ideal values as observed in Fig. 8(c).

The antenna receiving averaged signal is computed by the diversity parameter known as mean effective gain (MEG) and is defined as the ratio of the power received by the receiving antenna to the total power which is incident on it. The MEGs are calculated between any two antennas by generalized formula given for the mth and nth antenna

(16)$$MEG_m = 1-\vert {S_{mm}} \vert ^2-\vert {S_{mn}} \vert ^2, \;$$

(17)$$MEG_n = 1-\vert {S_{nn}} \vert ^2-\vert {S_{nm}} \vert ^2.$$

The ratio calculates the MEG which is given by

(18)$$\displaystyle{{MEG_m} \over {MEG_n}} = \displaystyle{{1-{\vert {S_{mm}} \vert }^2-{\vert {S_{mn}} \vert }^2} \over {1-{\vert {S_{nn}} \vert }^2-{\vert {S_{nm}} \vert }^2}}.$$

For the proposed antenna, MEGs calculated for antenna #1-antenna #2, antenna #1-antenna #3, and antenna #1-antenna #4 were simulated and measured values are plotted in Fig. 8(d). As per the observations, the ratio of MEG values is approximately −3.0 dB for both simulated and measured values.

In MIMO antenna configuration, the radiating elements are closely placed to each other and are operated at the same time, but this arrangement of radiating antennas also affects the performance in terms of interference. This suggests that the S-parameters obtained for the MIMO antenna only do not ensure the merit of interference, and hence a new parameter called total-active-reflection-coefficient (TARC) needs to be evaluated. This parameter is defined as the “square root of the ratio of total reflected power to the total incident power and overall apparent loss”. The TARC between any two ports is calculated by

(19)$$TARC_{m \times n} = \sqrt {\displaystyle{{{( {S_{mm} + S_{mn}} ) }^2 + {( {S_{nm} + S_{nn}} ) }^2} \over 2}} , \;$$

(20)$$TARC_{14} = \sqrt {\displaystyle{{{( {S_{11} + S_{14}} ) }^2 + {( {S_{41} + S_{44}} ) }^2} \over 2}} .$$

Figure 8(e) shows the simulated and measured TARC values which are below −10 dB in all the three bands of 5G and the expected values are <0 dB indicating more power is incident rather than reflected and this is due to the highly achieved matched impedance in the proposed design.