MIMO Antenna with High Isolation for WBAN Applications

Given the fast progress of wireless communication technologies, wireless body area network (WBAN) systems have received considerable attention for various applications [1, 2]. Because WBAN systems can be placed on or in the human body with high permittivity and conductivity at microwave frequencies, the gain and efficiency of an antenna for WBAN systems can be degraded [3–6]. Furthermore, to operate in WBAN communication environment, the antenna should have compact size, low height, insensitiveness to human body effects, and low specific absorption rate (SAR) [7]. Many researchers have conducted various studies on the performance analysis of on-body WBAN communication systems in the industrial, scientific, and medical (ISM) bands [8–11]. Because of reflections/scatterings that occur in a neighborhood environment and/or on the human body, severe multipath fading can arise in on-body communication links [9]. Multipath fading not only decreases the communication reliability of multisignals but also worsens the efficiency of a WBAN system [12]. To improve communication performance under the influence of multipath fading, a diversity technique such as multiple-input and multipleoutput (MIMO) is necessary. Since the independence of the multisignals can be improved by the high isolation of a MIMO antenna, a MIMO technique is frequently used to overcome the deterioration of communication performance due to multipath fading [13]. To achieve high isolation between MIMO antenna elements, an isolator has been used by many researchers [14– 17]. A MIMO antenna with a high isolation characteristic is proposed by using a shorted strip and two slits in the ground plane [17]. The antenna has isolation lower than −32 dB for on-body WBAN applications in the 2.4GHz ISM band. The MIMO antenna utilizes two planar inverted-F antenna (PIFA) elements with lengths of λ/4 at a resonant frequency to achieve a compact size [18]. The performance of the proposed antenna on the human muscle-equivalent flat phantom is analyzed by examining data such as the S-parameter characteristic, current distribution, radiation pattern, SAR, and envelope correlation coefficient (ECC).


Introduction
Given the fast progress of wireless communication technologies, wireless body area network (WBAN) systems have received considerable attention for various applications [1,2].Because WBAN systems can be placed on or in the human body with high permittivity and conductivity at microwave frequencies, the gain and efficiency of an antenna for WBAN systems can be degraded [3][4][5][6].Furthermore, to operate in WBAN communication environment, the antenna should have compact size, low height, insensitiveness to human body effects, and low specific absorption rate (SAR) [7].
Many researchers have conducted various studies on the performance analysis of on-body WBAN communication systems in the industrial, scientific, and medical (ISM) bands [8][9][10][11].Because of reflections/scatterings that occur in a neighborhood environment and/or on the human body, severe multipath fading can arise in on-body communication links [9].Multipath fading not only decreases the communication reliability of multisignals but also worsens the efficiency of a WBAN system [12].To improve communication performance under the influence of multipath fading, a diversity technique such as multiple-input and multipleoutput (MIMO) is necessary.Since the independence of the multisignals can be improved by the high isolation of a MIMO antenna, a MIMO technique is frequently used to overcome the deterioration of communication performance due to multipath fading [13].
To achieve high isolation between MIMO antenna elements, an isolator has been used by many researchers [14][15][16][17].A MIMO antenna with a high isolation characteristic is proposed by using a shorted strip and two slits in the ground plane [17].The antenna has isolation lower than −32 dB for on-body WBAN applications in the 2.4 GHz ISM band.The MIMO antenna utilizes two planar inverted-F antenna (PIFA) elements with lengths of /4 at a resonant frequency to achieve a compact size [18].The performance of the proposed antenna on the human muscle-equivalent flat phantom is analyzed by examining data such as the -parameter characteristic, current distribution, radiation pattern, SAR, and envelope correlation coefficient (ECC).

Antenna Design and Analysis
2.1.Basic Geometry.The proposed MIMO antenna consists of two PIFAs, a shorted strip, and two slits in a ground plane, as shown in Figure 1.The shorted strip, along with the two slits, acts as an isolator.The PIFAs, which have dimensions of 12 × 10.5 × 2 mm 3 , are located on a FR4 substrate (  = 4.4) with a 1 mm thickness and an area of 40 × 40 mm   PIFAs are symmetrically placed with a separation distance of 8 mm (0.06  ∘ ) in the -axis direction.The isolator and the ground plane are printed on the upper side of the substrate.
To consider the effects on the human body, the simulation setup of an antenna on a human muscle-equivalent flat phantom is illustrated in Figure 2. The phantom, with a volume of 200 × 270 × 60 mm 3 , has the relative dielectric constant (  = 52.7)and the conductivity ( = 1.95 S/m) of human tissue [19].Considering practical applications such as wearable Bluetooth services, the antenna is placed on the phantom with a separation distance of 10 mm to satisfy the required clearance to assemble the cover [3].The antenna geometry was designed and analyzed by utilizing the high frequency structure simulator (HFSS 14) [20].
The simulated -parameter characteristics of the proposed antenna with and without the isolator are compared in Figure 3.The optimized design parameters are  1 = 27.5 mm and  2 = 15.5 mm.The proposed antenna has a −10 dB reflection coefficient bandwidth ranging from 2.27 GHz to 2.6 GHz, which fully covers the desired 2.4 GHz ISM band.When the isolator is added between the two PIFAs, the isolation is improved significantly with a slight shift of the resonant frequency.The antenna has isolation below −32 dB in the 2.4 GHz ISM band.
To investigate the effect of the isolator, the simulated current distributions with and without the isolator at 2.4 GHz  are shown in Figure 4.By exciting port 1, substantial current is induced at PIFA 2 in the absence of the isolator.After the isolator is added, the induced current on PIFA 2 becomes weak.This is because the impedance of the /4 isolator becomes large at 2.4 GHz so that the current flowing from port 1 to PIFA 2 is blocked by the isolator [21,22].

Key Parameter Analysis.
The -parameters of the proposed antenna with respect to a variation in the length ( 1 ) of the PIFA radiator are shown in Figure 5.As  1 decreases, the resonance frequency shifts toward a higher frequency, while the isolation is improved.However, as  1 decreases beyond 27.5 mm, the isolation deteriorates.When  1 = 27.5 mm, the antenna satisfies the 2.4 GHz ISM band with optimum isolation.In Figure 6, the -parameters of the proposed antenna with respect to the various lengths ( 2 ) of the shorted strip are illustrated.A variation in  2 also changes the slit's length because the tip of the shorted strip is fixed.As  2 increases, the isolation performance is improved and good impedance matching is obtained.However, after  2 = 15.5 mm, the isolation degrades.To achieve optimum isolation performance,  2 = 15.5 mm is chosen.

Antenna Performance
Photographs of the fabricated antenna and the human muscle-equivalent flat phantom for the measurement setup are shown in Figure 7 [23].By utilizing the fabricated phantom (  = 52.1 and  = 0.94 S/m), the performance of the antenna on the human body can be investigated.
The simulated and measured -parameter characteristics of the proposed antenna are compared in Figure 8.The measured and simulated results are in reasonably good agreement.The measured −10 dB reflection coefficient bandwidth of the fabricated antenna is 490 MHz from 2.11 GHz to 2.6 GHz, which satisfies the entire 2.4 GHz ISM band.The fabricated antenna has a measured isolation below −38 dB in the 2.4 GHz ISM band.
To analyze the effects on the human body when the proposed antenna operates, the SAR value of the antenna is calculated at 2.4 GHz, as shown in Figure 9.The Federal Communications Commission (FCC) of the United States requires that the SAR be lower than 1.6 W/kg over a volume of 1 g of tissue [24].The antenna has a SAR of 1.52 W/kg when an input power of 100 mW is applied.Thus, the proposed antenna can be used for low power Bluetooth device applications [25].
The simulated and measured far-field radiation patterns of the proposed antenna on the phantom are compared at    2.4 GHz, as shown in Figure 10.The simulated and measured radiation patterns agree very well, except for those in the back radiation direction (150 ∘ ≤  ≤ 210 ∘ ) of the  plane.The difference between the simulated and measured radiation patterns of the  plane occurs because the phantom has different conductivity values (conductivity for simulation: 1.95 S/m, conductivity for measurement: 0.94 S/m).Because the radiated field reflected by the phantom with high conductivity used in the simulation is stronger than that of the measurement, the backward radiation decreases.When the conductivity for the simulation becomes 0.94 S/m, the simulated radiation patterns are almost the same as the measured ones, as shown in Figure 11.The radiation patterns of PIFA 1 are somewhat analogous to those of PIFA 2. The antenna has quasi-omnidirectional radiation patterns in the  plane and has its peak radiation in the outward normal to the phantom surface in the  plane.
Measured radiation characteristics of the two PIFAs at 2.4 GHz are compared in Table 1.The efficiencies of PIFA 1 and PIFA 2 are 21.34% and 21.37%, respectively.
The envelope correlation coefficient (ECC) is commonly utilized to evaluate the diversity capability of a MIMO   antenna system.The ECC must be computed by using threedimensional radiation patterns [26].ECCs computed by simulated radiation patterns and measured radiation patterns are shown in Figure 12.The measured ECC agrees very well with the simulated one.The proposed antenna has an ECC value that is lower than 0.5 in the 2.4 GHz ISM band.When the computed ECCs with and without the isolator are compared, ECC performance improves owing to the improved isolation between PIFA 1 and PIFA 2.

Conclusion
A high isolation MIMO antenna is designed to operate in the 2.4 GHz ISM band for WBAN applications.When installed on a human muscle-equivalent flat phantom, the antenna satisfies the −10 dB reflection coefficient bandwidth of the 2.4 GHz ISM band.An isolator, consisting of a shorted strip and two slits, is added between the two PIFAs to improve the isolation.Although the two PIFAs are placed close to

Figure 2 :
Figure 2: Proposed antenna on the phantom for simulation setup.

Figure 3 :
Figure 3: Simulated -parameter characteristics for the proposed antenna with and without isolator.

Figure 8 :
Figure 8: Simulated and measured -parameter results for the proposed antenna on the human muscle-equivalent flat phantom.

Figure 9 :
Figure 9: Simulated SAR distribution of the proposed antenna on the phantom at 2.4 GHz (input power: 100 mW).

Figure 10 :
Figure 10: Simulated and measured radiation patterns of the proposed antenna at 2.4 GHz.(a)  plane; (b)  plane ( for simulation: 1.95 S/m,  for measurement: 0.94 S/m).

Figure 11 :
Figure 11: Simulated and measured radiation patterns ( plane) of the proposed antenna at 2.4 GHz ( for simulation and measurement: 0.94 S/m).

Figure 12 :
Figure 12: Simulated and measured envelope correlation coefficients of the proposed antenna with and without isolator.

Table 1 :
Measured peak gain and radiation efficiency.