Decoupled and Descattered Monopole MIMO Antenna Array with Orthogonal Radiation Patterns

Abstract

This chapter introduces a novel design concept to reduce mutual coupling among closely-spaced antenna elements of a MIMO array. This design concept significantly reduces the complexity of traditional/existing design approaches such as metamaterials, defected ground plane structures, soft electromagnetic surfaces, parasitic elements, matching and decoupling networks using a simple, yet a novel design alternative. The approach is based on a planar single decoupling element, consisting of a rectangular metallic ring resonator printed on one face of an ungrounded substrate. The decoupling structure surrounds a two-element vertical monopole antenna array fed by a coplanar waveguide structure. The design is shown both by simulations and measurements to reduce the mutual coupling by at least 20 dB, maintain the impedance bandwidth over which S11, is less than −10 dB, and reduce the envelope correlation coefficient to below 0.001. The boresight of the far-field radiation patterns of the two vertical monopole wire antennas operating at 2.4 GHz and separated by 8 mm (λo/16), where λo is the free-space wavelength at 2.45 GHz, is shown to be orthogonal and inclined by 45° with respect to the horizontal (azimuthal) plane while maintaining the shape of the isolated single antenna element.

1. Introduction

Contemporary wireless systems including, but not limited to, 4G/5G LTE/LTE-A, radar, RFID, Wi-Fi, WiMAX, GPS, geolocation, biomedical imaging, and remote sensing dictate the use of miniaturized MIMO antenna arrays on mobile terminals. They can also be permanently installed on fixed structures for increased gain, which will improve link reliability and quality of service, increase communication range, and increase battery life through a variety of diversity schemes [1, 2], and/or increase data rate/throughput through MIMO spatial multiplexing schemes [3]. Moreover, these structures can provide esthetic, miniaturized wireless consumer devices. Recent trends toward miniaturized, esthetically appealing, battery efficient handheld wireless devices, and green wireless systems require antenna arrays that should be implemented within a restricted physical space.

In this chapter, we report a new approach of how to reduce the complexity of prior existing designs for mutual coupling reduction such as periodic metamaterial/metasurface constructs [4, 5, 6, 7, 8], defected ground plane structures [9, 10], soft/hard electromagnetic surfaces [11, 12], parasitic elements [13], matching and decoupling networks [14, 15, 16], and neutralization lines [17]. Our proposed approach can still maintain, and in certain circumstances exceed the performance metrics of traditional approaches well-known in previous art, [18, 19, 20] using much simpler, cost-effective, and novel design alternatives.

Numerous techniques were proposed for the mutual coupling reduction among the elements of MIMO antenna arrays, the most notable and highly embraced is the use of metamaterials, which subsequently turned into an unique topic of its own used in various disciplines including acoustics, RF, optics, laser, and nanotechnology [21]. A defected ground plane consisting of a slitted pattern is disclosed in [9] to suppress mutual coupling between two monopole antennas separated by 0.093 λ₀, where λ₀ is the free-space wavelength at 2.53 GHz. The two monopoles share a 40 × 25 mm ground plane and are separated by 11 mm. Each slit is 11 × 1 mm. A periodic array of three slits was used to achieve a mutual coupling better than −20 dB. However, no information on the behavior of far-field radiation patterns of the array was provided. A simple yet elegant decoupling approach was presented in [22] using a parasitic element to decouple two monopole antennas operating at 900 MHz. The parasitic element is situated at the center between the two driven elements on a 330 × 250 mm FR4 substrate, 1.5 mm in thickness. The substrate has a permittivity of 4.4 and a loss tangent of 0.02. The radiation behavior shows an end-fire pattern. The measured pattern correlation is close to 0.02. This method achieved coupling coefficient, S₁₂ close to −40 dB over the bandwidth in which S₁₁ is less than −10 dB. The method requires careful attention to the choice of the distance between the parasitic element and the driven elements for a given separation between the driven elements. A drawback of this technique is the undesired radiation from the parasitic element, which may degrade the decoupled radiation pattern.

A complicated approach for mutual coupling suppression between two monopole antennas operating at 2.49 GHz and separated by 0.09 λ₀, where λ₀ is the free-space wavelength at 2.49 GHz, with a length of 32.5 mm over a ground plane containing four slits, 1 × 11 mm in size and two defected wall structures situated at the edge of the outermost slits are proposed in [20]. The defected wall structure consists of a metallic plane with C-shaped cuts mimicking split ring resonators. The spacing between the slots in the defected wall is 0.09 λ₀. The structure achieved a measured isolation of about 44 dB. No information has been provided on the shape of the far-field radiation patterns. One notable approach for the isolation of two high-profile monopole antennas laid on a common copper ground plane (1.25 λ₀ × 1.25 λ₀), involves the use of highly negative-permeability, broadside coupled split-ring resonators etched on both sides of a dielectric substrate. In this approach, the two monopoles are separated by λ₀/8 (30.25 mm), where λ₀ is the free-space wavelength at 1.24 GHz [18]. A 3-D stack employing 10 arrays is aligned vertically using plastic support in the space between the two monopoles, where each array consists of four broadside coupled split-ring resonators.

It should be noted, however, that these artificial magnetic inclusions obviously suffer from extremely narrow resonance bandwidth. The design methodology is based on creating artificial negative permeability, which consequently presents extremely high attenuation to the near-field that exists in the region between the two monopoles for both propagating and evanescent fields. The authors claim a reduction of at least 20 dB in S₁₂ but over an extremely small bandwidth. The structure is neither convenient nor rigid for miniaturized handheld devices since the decoupling structure occupies a significant volume in the space between the two monopoles. Also, the fabrication costs as well as the dielectric and conductor losses introduced by the decoupling structure even at frequencies in the decoupled regime create an inconvenience. We should note that a similar approach was reported in [19] in which a metamaterial spacer is employed to reduce the mutual coupling between two monopoles, each 27.8 mm in height, installed on a common (2 λ₀ × 2 λ₀) aluminum ground plane and separated by 0.156 λ₀ (18 mm) at 2.6 GHz. The principle of operation is explained by considering the metamaterial acting as an artificial magnetic conductor (AMC). The decoupling structure consists of four existing design of a bidirectional AMC printed on a dielectric substrate with width-height-thickness dimensions of 64 × 33 × 10.5 mm.

Several defected ground planes [9, 10, 11, 12], soft artificial electromagnetic surfaces, matching and decoupling networks [14, 15, 16], and neutralization lines [17] have received considerable recent research interest in both academia and industry. However, they were not applied specifically to vertical monopole antennas, the subject of this chapter. It should be noted that these traditional approaches suffer from:

1. Extremely narrow bandwidths since metamaterials, defected ground structures, and soft electromagnetic surfaces embrace inherently high-resonant structures;

2. Repeated periodic structures are employed periodically in 1-D, 2-D, or in 3-D using vias, which further complicates the design and increases the fabrication cost;

3. Insertion of many 2D cells (arranged in multiple rows and/or columns) and sometimes 3D unit cells [7, 8] between the antenna elements limits the empty space between the antenna elements and hence complicates the task of miniaturizing the MIMO antenna array; and

4. Distortion of the far-field radiation patterns and/or reduction the operating bandwidth [18, 19, 20] from these artificial structures.

To conclude, most of the related research utilized meta-surfaces inclusions inserted in the space between the antenna elements for which the mutual coupling is to be reduced. This is another major obstacle for applications involving extremely miniaturized antenna arrays, such as MIMO systems on handheld wireless devices. In this chapter, however, we report a novel approach in which the mutual coupling is reduced by inserting two vertical monopole antenna elements inside a rectangular conducting ring loaded with two conducting strips. This technique allows much smaller separations (λ_o/16), where λ_o is the free-space wavelength. Unlike the traditional techniques that are based on inserting artificial resonant structures between the radiating elements, the methodology reported in this chapter is based on enclosing the radiating elements by simple and versatile planar conducting structures, the performance of which exceeds the performance of traditional/existing design approaches.

2. MIMO array structure

We will elaborate on one embodiment involving a miniaturized two-element antenna array system composed of two vertical monopole antennas separated by 8 mm, corresponding to λ_o/16, where λ_o is the wavelength in free space at 2.4 GHz. The unique design scheme of the coupled array produces orthogonal radiation patterns even without the inclusion between the monopoles. Additionally, the decoupling structure is designed to reduce the mutual coupling and enhance isolation between the signals at the array terminals. A very low envelope correlation coefficient between the received signals is realized to accommodate operating with the different MIMO modes. To summarize, the design of the antenna array disclosed in this chapter seeks to achieve three functions: decoupling, descattering, and orthogonalization of the far-field radiation patterns in the boresight directions. Descattering in this context implies that the radiation pattern of each antenna element in the array is the same as the radiation pattern when in isolation.

To obtain a boresight oriented by 45° relative to that of a conventional vertical monopole, the design uses a partial reference ground plane instead of a traditional full ground as shown in Figure 1. The monopole is mounted on a dielectric substrate and is fed by a coplanar waveguide (CPW) circuit. The ground plane of the coplanar feeding circuit serves as a partial ground for the antenna as well. The height of the monopole h and the feeding circuit geometry are tuned such that the antenna operates at center frequency of 2.4 GHz serving applications in the ISM band. The coplanar waveguide feeding circuit is selected to have a characteristic impedance of 50 ohm at its input. The geometry, reflection coefficient, and radiation pattern of the single element are shown in Figures 1–3, respectively.

Figure 1.

Geometry of a single monopole antenna fed by CPW.

Figure 2.

Reflection coefficient of the monopole antenna fed by a CPW.

Figure 3.

Realized gain of the monopole antenna fed by a CPW.

This behavior is exploited to construct an array consisting of two vertical monopoles, the radiation patterns of which are identical in shape but orthogonally oriented in space. The antenna array is formed by placing monopole 2 and its feeding circuit as a mirror image of monopole 1 around the y-z plane as depicted in Figure 4. Both antennas are assembled on a common substrate of a relative permittivity, ε_r = 10.2, and height, h_s = 1.27 mm.

Figure 4.

The coupled monopole antenna array.

When the two antennas are positioned in a close proximity, λ_o/16, in this design, the far-field radiation patterns will be distorted while the correlation coefficient increases due to mutual coupling. To investigate the mutual coupling effects, the array is simulated using the commercial software package, high-frequency structure simulator (HFSS) of ANSYS. The geometry of the design is shown in Figure 5 and the dimensions are presented in Table 1.

Figure 5.

Geometry of the coupled antenna array. (a) Prospective view, (b) Side view, (c) Bottom view and, (d) Top view.

Monopoles separation (s)	8
Monopole height (h)	25
Monopole diameter (d)	1.6
Feeding probe length(l2)	11
Feeding probe width (w2)	2
Ground width (w3)	10.85
Ground length (l3)	5
Center strip width (w4)	3
CPW gap width (g1)	0.5
Substrate (l1 × w1 × hs)	(40 × 40 × 1.27)
Dielectric constant (εr) of the substrate	10.2

Table 1.

Dimensions of the coupled antenna Array.

Despite the reduction in transmission level between the feeding ports, the distribution of surface currents depicted in Figure 6, resulted from exciting port 1 while terminating port 2 with 50 ohms, shows the existence of significant interfering currents distributed on the CPW feeding of monopole 2. The current induced on monopole 2 and its feeding circuitry is due to the mutual coupling between the two antennas.

Figure 6.

Magnitude of surface current distribution for the coupled monopole antenna array, dB(1 A/m).

In Figure 7, we show the reflection coefficient, S₁₁ and transmission coefficient, S₂₁ plotted against frequency. The coupled array provides a bandwidth of 230 MHz over which the reflection coefficient is below −10 dB. At 2.4 GHz, the partial ground plane reduced S₂₁ to −10 dB compared to −4 dB if the antennas share a common full ground plane and are separated by 8 mm. The radiation patterns in terms of the realized gain are plotted in Figure 8 for two cases; in the first case, an excitation signal is injected into port 1, while the second port is terminated with its internal impedance, 50 ohms. In the second case, the excitation signal is injected into port 2, and port 1 is terminated by 50 ohms. Having the antennas placed extremely close to each other (8 mm at 2.4 GHz) resulted in disturbing the near field and increasing the coupling level. This proximity causes the radiation patterns to be directed into the lower part below the plane containing the substrate.

Figure 7.

Frequency response of the coupled monopole antenna.

Figure 8.

Mutual coupling effects on the far-field radiation patterns. (a) Exciting port 1 and, (b) Exciting port 2.

The proposed MIMO antenna array consists of two monopole antennas positioned on a printed circuit board. The antennas are surrounded by a decoupling circuit printed on the top side of the substrate and a feeding circuit on the bottom side. For short, the design will be called the decoupled coplanar waveguide fed monopole antenna array (CPW-MAA). The substrate has a length and width of 40 × 40 mm, and a thickness of 1.27 mm.

In order to further enhance the performance of pattern diversity for the two-element antenna array and minimize the contribution of mutual coupling among the elements of the antenna array to the channel correlation, the mutual coupling effect needs to be reduced with minimal distortion to the radiation patterns of the individual elements, preferably such that the 3-D far-field patterns of the individual elements being identical in shape and orthogonal in space. This applies to the signals at the output of the antennas when transmitting and at the input of the feeding circuit when operating in the receive mode.

The decoupling structure, shown in Figure 9, consists of a simple 2D planar rectangular metallic ring and two tuning metallic strips printed on the surface of the substrate.

Figure 9.

Decoupling circuit.

First, the performance of the CPW-MAA is compared against the coupled monopole antenna array shown in Figure 5 using the same substrate. Furthermore, as a design alternative and to further validate the disclosed design concept, another substrate has been used with a dielectric constant ε_r of 6.15 and height, 1.27 mm. Each monopole is fed by a copper trace printed on the bottom side of the substrate. Each monopole makes a contact with the copper trace by extending the monopole body inside the substrate to reach the feeding copper trace at the bottom layer. A CPW at the bottom side of the substrate is used to deliver the signal to the antennas in case of transmitting and from the antennas in case of receiving.

The CPW consists of a central strip having a length of l₃ and a width of w₄ with the finite ground planes positioned on both sides of the central strip with a width of w₃. An SMA is utilized to launch the wave into the array. The impedance of the SMA is 50 ohm and is used to connect the antenna to a vector network analyzer used in the measurements. The array system dimensions are tuned to operate at a 2.4 GHz with a bandwidth of 260 MHz over which the reflection coefficients, S₁₁ and S_22, are less than −10 dB and transmission coefficients, S₁₂ and S₂₁ less than −20 dB. The two-element MIMO antenna array is shown in Figure 10, while dimensions are depicted in Table 2 for the two substrates used where all dimensions are in millimeters.

Figure 10.

The decoupled CPW-MAA. (a) Prospective view, (b) Side view, (c) Top view, (d) Bottom view.

Dimension defined name	Substrate RO3210 (εr = 10.2)	Substrate RO3006 (εr = 6.15)
Monopoles separation (s)	8	8
Monopole height (h)	28	27
Monopole diameter (d)	1.6	1.6
Feeding probe length (l2)	11	11
Feeding probe width (w2)	2	2
Ground width (w3)	10.5	10.8
Ground length (l3)	5.0	5.0
CPW trace width (w4)	3.0	3.0
CPW gap width (g1)	0.5	0.2
Rectangular ring outer length (l5)	22	22
Rectangular ring outer width (w5)	24	30
Rectangular ring trace width (w7)	2.5	2.5
Tuning strip length (l4)	16	16
Tuning strip width (w6)	1.5	4
Tuning strips separation (s1)	6.5	8
Substrate (l1 × w1 × hs)	(40 × 40 × 1.27)	(40 × 40 × 1.27)

Table 2.

Dimensions of the CPW-MAA for two substrates.

Prototypes are fabricated and measured at the University of Arkansas at Little Rock’s Antennas and Wireless Systems Research Laboratory (AWSRL). The prototype which is shown in Figure 11 used Roger RO3210 substrate. The frequency response has been measured using E5071B and N5242A VNAs. The radiation patterns are measured inside an anechoic chamber. Using Roger RO3210 (dielectric constant ε_r of 10.2) as a substrate, and in order to demonstrate that adding this decoupling circuit isolates the input and output ports, the magnitude of the surface current dB (1 A/m) has been plotted for different excitation phase delays as depicted in Figure 12. A signal of 1 W is assigned through port 1 to excite the first monopole, the monopole in the right side of the array, while port 2 is terminated by its internal impedance of 50 ohm. It seems quite evident that the amount of current coupled to the feeding circuit of monopole 2 at different phase delays is very low, minimum at −8 dB (1 A/m) (blue color), compared to other excited monopole, where a high current is observed with a maximum value at 32 dB (1 A/m) (red color). This contrast of 40 dB (1 A/m) reveals the high isolation level achieves due to the decoupling network compared to the coupled current before adding the decoupling network shown in Figure 6.

Figure 11.

The decoupled CPW-MAA and decoupled MS-MAA prototypes.

Figure 12.

Two monopole antenna array with decoupling network; surface current distribution (magnitude) dB(1 A/m).

The simulated and measured results of the scattering parameter, S₁₁ are presented in Figure 13. The amount of coupling expressed in terms of scattering parameter, S₂₁ is depicted in Figure 14. These figures reveal good agreement between the simulated models and the measurements using the fabricated prototype.

Figure 13.

Reflection coefficient S₁₁, simulation versus measurements.

Figure 14.

Coupling coefficients S₂₁, simulation versus measurements.

Introducing the decoupling network resulted in a reduction of −19 dB in mutual coupling level, from −4 to −23 dB. The antenna array system offers a bandwidth of 10%, in which the coupling coefficient, S₂₁ is less than −20 dB and 14% over which the reflection coefficient, S₁₁ is less than −10 dB. The slight shift of 2% in the frequency at which the measured coupling coefficient S₂₁ is at its minimum in the bandwidth of interest compared to simulated results is due to the variations associated with the fabrication of the prototype and the limited precision of simulations and measurements. The fabricated prototype also provides a wider bandwidth. This is due to the dielectric and metallic losses inherent in the substrate and dipoles, which lead to reduced Q factor.

Table 3 summarizes the performance metrics of the simulated models and the measured prototype. In this table, f_L and f_H denote the lower and upper frequency limits of the operating bandwidth, respectively. The array meets and exceeds the bandwidth requirement of 802.11 Wi-Fi systems operating in the ISM band at 2.4 GHz.

Performance metric	Conventional array	HFSS	CST	Measured
S11 (min) dB	−8	−24	−40	−20
S22 (min) dB	−8	−23	−40	−33
S21 (min) dB	−4	−29	−45	−23
fL (S11, S22 < −10 dB) GHz	—	2.25	2.22	2.12
fH (S11, S22 < −10 dB) GHz	—	2.58	2.56	2.74
fL (S12, S21 < −20 dB) GHz	—	2.28	2.26	2.32
fH (S12, S21 < −20 dB) GHz	—	2.51	2.51	2.58
BW (S11, S22 < −10 dB) GHz	—	0.33	0.34	0.62
BW (S21, S12 < −20 dB) GHz	—	0.23	0.25	0.26
% BW (S11, S22 < −10 dB)	—	13.6%	14.2%	25.5%
% BW (S21, S12 < −20 dB)	—	9.6%	10.4%	10.6%

Table 3.

Decoupled array performance at 2.4 GHz in terms of scattering parameters (dB).

The maximum value of the total realized gain is plotted versus frequency in Figure 15. Another advantage of the proposed antenna array is the enhancement achieved in the realized gain and hence the efficiency. The total realized gain of a single element varies over the range from 1.6 to 1.69 dB over the frequency range 2.25–2.52 GHz over which S₁₁ and S₂₁ are less than −10 and −20 dB, respectively. This gain is obtained when one port is excited while the other is terminated with its characteristic impedance. An improvement of 2.1 dB is observed in the realized gain with the proposed antenna array as compared to a conventional array antenna consisting of two monopoles separated by λ₀/16 over a λ₀ × λ₀ full ground plane. It is noteworthy that the maximum value of the total realized gain of the array system is close to that of a stand-alone monopole antenna.

Figure 15.

Maximum realized gain.

The 3-D radiation patterns in terms of the realized gain are shown in Figure 16.

Figure 16.

(a) The realized gain scale. (b) Realized gain for a stand-alone monopole, while (c) and (d) are for the coupled monopole antenna array, then (e) and (f) are for the decoupled CPW-MAA. Examining the upper right section of the 3-D pattern of model 2, (π ≤ ϕ ≤ 2π, and θ ≤ π/2) shows red spot, which indicates high gain which is reduced for the decoupled CPW-MAA, where more power is directed in the lower left side (0 ≤ ϕ ≤ π and θ ≥ π/2). This distortion has been corrected in the disclosed antenna array due to the low mutual coupling caused by adding the decoupling network. Furthermore, a descattered pattern is realized. Figure 17 shows a comparison between simulated and measured co/cross radiation patterns of the decoupled CPW-MAA (Figure 18).

Figure 17.

Realized gain; scale (a), stand-alone monopole (b), two-monopole array without decoupling structure (c) and (d), decoupled CPW-MAA (e) and (f).

Figure 18.

Realized gain; simulation versus measurements.

3. Conclusions

A compact MIMO antenna array with identical yet orthogonal radiation patterns and high isolation level is reported in this chapter. The shape and size of the partial ground plane and its distance from the monopole allowed the boresight of the radiation pattern of the isolated monopole to be oriented at an elevation angle of 45^o. The 45^o elevation achieved for the single monopole made it possible to place two vertical monopoles back to back to achieve an array with orthogonal radiation patterns. However, the mutual coupling between the two elements when placed in close proximity resulted in a distortion of the radiation patterns. To mitigate the effects of mutual coupling and increase the isolation between the two ports of the array, a decoupling element that surrounds the array elements which provides more flexibility on how close the elements need to be from each other. The dimensions of the rectangular ring and the tuning strips have been tuned to achieve the desired isolation level and restore the radiation pattern of the isolated element. The CPW-MAA with orthogonal far-field radiation patterns opens new opportunities for modern handheld devices and/or fixed wireless access points implementing MIMO techniques. Unlike existing design approaches that are based on inserting artificial resonant structures between the radiating elements, the design presented in this chapter encloses the radiating elements by simple and versatile planar conducting structures. The proposed design is demonstrated by simulations and measurements to significantly reduce the mutual coupling when the monopoles are spaced by 8 mm (λ_o/15.3). The performance achieved in terms of impedance matching, isolation, gain, bandwidth, and radiation pattern is comparable to network decoupling and parasitic approaches and better in terms of bandwidth, size, and complexity in comparison to existing designs which employ periodic AMC. A matching bandwidth of 0.634 GHz over which the reflection coefficient, S₁₁ is less than −10 dB, and a 0.246 GHz bandwidth over which the transmission coefficient, S₂₁ is less than −20 dB have been achieved. A total realized gain of 1.69, 1.95, and 1.68 dB are achieved at 2.26, 2.4, and 2.52 GHz, respectively.