Wide CPW-Fed Multiband Wearable Monopole Antenna with Extended Grounds for GSM/WLAN/WiMAX Applications

A novel wide coplanar waveguide(CPW-) fed multiband wearable monopole antenna is presented. The multiband operation is achieved by generating slanted monopoles of different lengths from an isosceles triangular patch. The different operating frequencies of the proposed antenna are associated with the lengths of the slanted monopoles, which are determined under quarter wavelength resonance condition. The CPW line is used as a multiband impedance-matching structure. The two grounds are slightly extended for better impedance matching. The proposed antenna is designed to cover the 1800MHz GSM, 2.4 GHz/5.2GHz WLAN, and 3.5GHz WiMAX bands. The measured peak gains and impedance bandwidths are about 4.18/3.83/2.6/2.94 dBi and 410/260/170/520MHz for the 1550-1960MHz/2.3-2.56GHz/3.4-3.57GHz/5.0-5.52GHz bands, respectively. The calculated averaged specific absorption rate (SAR) values at all the resonant frequencies are well below the standard limit of 2W/kg, which ensures its feasibility for wearable applications. The antenna performance under different bending configurations is investigated and the results are presented. The reflection coefficient characteristics of the proposed antenna is also measured for different on-arm conditions and the results are compared. A good agreement between experimental and simulation results validates the proposed design approach.


Introduction
In today's world, for applications which involve wearable antennas, like health monitoring of patients and firefighting and military personal or body-centric communications, it becomes absolutely necessary that a single antenna operates on multiple bands, with SAR values well below the standard limits and its performance should not degrade under bent conditions.In order to satisfy these specific wireless communication requirements, multiband wearable antennas which can operate at 880-960 MHz/1710-1880 MHz for Global System for Mobile (GSM) communications, 2.4-2.484GHz/5.15-5.825GHz for wireless local area network (WLAN), and 3.4-3.69GHz/5.25-5.85GHz for Worldwide Interoperability for Microwave Access (WiMAX), are desired.
The specific absorption rate of a wearable antenna should be calculated to ensure the safety of the user.Either single-layer or multilayer human head and human tissue models are used to calculate the specific absorption rate (SAR) of an antenna [1,2,7,8,10,13,[20][21][22][23].On the body performance of a wearable antenna, the investigation of its characteristics under bent conditions is very important.It is a well-established fact that the resonant frequency of a wearable antenna shifts under bending conditions.The on-arm performance and bending effect of a wearable antenna have been analyzed in [1-4, 7, 9, 10, 13, 24].
In this paper, we propose a four-band wearable monopole antenna fed by a wide CPW line.The small values of the substrate's dielectric constant increase the width of the CPW line.This limitation is exploited in this design.The proposed design has used the wide CPW line as a multiband impedance-matching structure.An isosceles triangular patch is slotted to generate four slanted monopoles for multiple wide-band operations.The antenna is tested under bent conditions and the results are presented.Antenna performance is also investigated for different on-arm conditions, with and without a cloth, and the results are compared.The specific absorption rate of the proposed antenna is found to be within standard limits.The geometrical structure and design procedure of the proposed antenna are presented in Section 2. Section 3 describes the simulation and measurement methodology.The results of the simulated and handmade proposed antenna are presented, compared, and analyzed in Section 4. The estimation of SAR values, effect of bending, and on-arm performance of the proposed antenna are also explained in this section.The paper is concluded in Section 5.

Antenna Geometry and Design Procedure
The geometry of the proposed multiband wearable antenna is shown in Figure 1.The antenna is formed by slanted monopoles using the approach applied in [17].Generally, the planar quarter-wavelength monopoles have small bandwidths.To increase the bandwidth, taper-shaped elements are often used [17].Hence, slanted monopoles are generated by slotting an isosceles triangular-shaped patch.The length of each monopole is set approximately equal to a quarter-wavelength of the desired resonant frequency.The length of each monopole is defined as where L mn is the length and λ n is the respective guided wavelength of the nth monopole.The antenna is fabricated on a polyester cloth substrate with 1 mm thickness, dielectric constant ε r = 1 39, and loss tangent tanδ = 0 01 [3].
The thickness of the substrate is achieved by stitching multiple layers of the polyester cloth.The presence of an air gap between substrate layers can reduce the effective permittivity of the substrate.To eliminate the impact of the air gap, multiple layers of the polyester cloth are stitched in such a manner that they tightly adhered to each other leaving a negligible air gap.The width of the CPW strip (W 1 ), gap (g), thickness (h), and dielectric constant (ε r ) of the substrate determine the effective dielectric constant (ε eff ) and characteristic impedance (Z 0 ) of the CPW line.The effective dielectric constant and characteristic impedance (Z 0 ) of a coplanar waveguide on a dielectric substrate of finite thickness [25] is given by where the modulus of the complete elliptic integrals of the first kind K k 0 , K k 0 ′ , K k 1 , and K k 1 ′ are as follows: The size of the proposed antenna considering the substrate size is 45 5 × 85 × 1 mm 3 .The design procedure of the antenna, which consists of five steps, is elaborated and shown in Figure 2.
Step 2. The first slanted monopole m 1 at a resonant frequency of 1800 MHz is generated by slotting the triangular patch.The theoretical length of the slanted monopole m 1 under quarter-wavelength resonance condition is 39.19 mm.The length L T and width W T of the triangular patch is taken to match the length L m1 of the first slanted monopole m 1 , which is given by For the selected length L T and width W T of the isosceles triangular patch, the value of θ comes out to be approximately 85.5 °, which is estimated as Due to the mutual coupling between the slanted monopole m 1 and the rest of the triangular patch which affects the resonant frequency, the final length L m1 of the slanted monopole m 1 is modified to be 38.62 mm.The width W m1 of the first slanted monopole m 1 is 1.0 mm.This slanted monopole with the remaining triangular patch is denoted Antenna II.It is shown in Figure 2(b).The width W s1 of the first slot is 0.4 mm.It is considered in such a way to match the length of the second monopole m 2 .
Step 3. The second slanted monopole m 2 at a resonant frequency of 2.4 GHz is generated by making another slot in the remaining triangular patch.Theoretically, the quarter wavelength at 2.4 GHz is 29.39 mm, whereas practically, the length L m2 of the slanted monopole m 2 is kept at 30.45 mm with a width W m2 of 1.2 mm.Two slanted monopoles m 1 and m 2 along with the remaining triangular patch as shown Step 4. The third slot with a width of W s3 = 0 45 mm is cut in a triangular patch to generate the third slanted monopole m 3 at a resonant frequency of 3.5 GHz.The quarter wavelength at 3.5 GHz is 20.15 mm.The final length L m3 of the third slanted monopole m 3 is kept at 20.53 mm with a width W m3 of 1.5 mm.This antenna with three monopoles and the remaining triangular patch is designated Antenna IV as shown in Figure 2(d).The remaining triangular patch is the fourth monopole which resonates at 5.38 GHz.The length L m4 and width W m4 of this fourth monopole is 8.578 mm and 1.4 mm, respectively.
Step 5.In the case of Antenna IV, the reflection coefficient and bandwidth of the four bands at 1800 MHz, 2.4 GHz, 3.5 GHz, and 5.38 GHz are not adequate and need to be improved.To improve the reflection coefficient and bandwidth of the four bands, both ground patches are extended with length L ag and width W ag .The required four-band operation is obtained when L ag = 9 mm and W ag = 12 mm.These extended grounds provided voltage standing wave ratio (VSWR) values close to 1 at four resonating frequencies, which proves improved impedance matching.Improved impedance matching further decreases the reflection coefficient and increases the bandwidth of the four respective bands.The antenna with four monopoles along with extended grounds is the final required design denoted Antenna V. Antenna V is shown in Figure 2(e).The photograph of the handmade Antenna V is shown in Figure 2(f).

Simulation and Measurement Methodology
The antenna design is simulated using a high-frequency structure simulator (HFSS).The simulation is performed for the reflection coefficients (S 11 ) of the proposed antenna with varying L ag and W ag , for 3-dimensional and 2-dimensional radiation patterns, gain, radiation efficiency, surface current density distribution, and SAR values at all resonant frequencies.The reflection coefficient (S 11 ) of the handmade Antenna V is measured using the ANRITSU MS46322A vector network analyzer.The expression used to measure the gain of the Antenna V is given as [13] follows: where G is the gain of the reference pyramidal horn antenna known over the anisotropic antenna, P T and P R are the power received by the test antenna and the power received by the reference antenna, respectively.A three-layer human tissue model is used for the estimation of SAR values at the measured resonant frequencies of the proposed antenna.The permittivity (ε r ) and conductivity σ (S/m) of the skin (dry and wet), fat, and muscle layers at 1780 MHz, 2.40 GHz, 3.46 GHz, and 5.26 GHz are listed in Tables 1 and 2 [7, 13, 23, 26].The thickness, mass density, and loss tangent [7] of the skin, fat, and muscle layers are summarized in Table 3.The surface size of the human tissue model used for SAR calculations is 200 × 200 mm 2 .The IEEE C95.3 standard, in which SAR is averaged over 10 g of biological tissue, is used to calculate averaged SAR [2,13].
In order to investigate the effects of slight and severe bending on antenna performance, two diameters have been taken for bending analysis.The 160 mm diameter offers slight bending, while the 110 mm diameter offers severe bending.The reflection coefficients (S 11 ) of the proposed antenna have been measured for light and severe bending configurations along x and y planes, and are compared with simulation results.Two on-arm conditions, with and without cloth, have been used to measure the performance of the proposed antenna on human body.

Results and Discussion
The simulated reflection coefficients (S 11 ) of Antenna I to Antenna V are summarized in Figure 3.The effect of L ag and W ag on the overall reflection coefficient (S 11 ) is illustrated in Figures 4 and 5  From the simulated and measured results, a slight shift in the resonant frequencies accompanied by a change in reflection coefficient values has been found.This can be attributed to the inconsistency in the fabrication process and soldering tolerance.
Refereeing to these results, the proposed antenna can satisfy the 1800 MHz GSM, 2.4/5.2GHz WLAN, and 3.5 GHz WiMAX bands, resulting in a four-band operation.
Figure 7 shows the surface current distributions of Antenna V at 1800 MHz, 2.44 GHz, 3.5 GHz, and 5.38 GHz.The surface current densities are at the maximum in the corresponding monopoles at the respective simulated resonant frequencies of 1800 MHz, 2.44 GHz, 3.5 GHz and 5.38 GHz.In monopole elements that are sufficiently thin electrically and not too long, the element current distribution is approximately sinusoidal [27].The first quarter-wavelength monopole m 1 corresponding to 1800 MHz is the thinnest among all four monopoles, and its current distribution is nearly sinusoidal.
Since the measured resonant frequencies of Antenna V are 1780 MHz, 2.40 GHz, 3.46 GHz, and 5.26 GHz, Figure 8 shows the 3-dimensional simulated radiation patterns of Antenna V at measured resonant frequencies.Figures 9-12    International Journal of Antennas and Propagation reason for the shapes of their radiation patterns.However, at 3.46 GHz and 5.26 GHz, the radiation patterns are distinct.
Figure 13 shows the measured and simulated peak gains versus the frequency of Antenna V.The measured and simulated peak gains in the 1550-1960 MHz band vary from 3.3 to 4.18 dBi and 8.51 to 9.33 dB, respectively.The measured and simulated results in the 2.3-2.56GHz band described that the gain variations are from 3.24 to 3.83 dBi and 5.63 to 7.15 dB.For the third and fourth operating bands of 3.4-3.57GHz and 5.0-5.52GHz, the measured peak gains are about 1.86-2.6dBi and 2.01-2.94dBi.The simulated peak gains for the 3.4-3.57GHz and 5.0-5.52GHz bands lie within 1.74 to 2.41 dB and 1.59 dB to 2.75 dB, respectively.The simulated radiation efficiency of the proposed Antenna V is shown in Figure 14.For the 1550 MHz-1960 MHz band, the radiation efficiency is about 91.92 to 94.94%.In the second band of 2.3-2.56GHz, the radiation efficiency varies from 82.16 to 95.69%.
The results in the 3.4-3.57GHz and 5.0-5.52GHz bands described that the radiation efficiency variations are from 80.89 to 94.10% and 80.20 to 87.82%, respectively.4.1.Specific Absorption Rate (SAR) Estimation. Figure 15 shows the three-layer human tissue model [13] used for SAR calculations at the measured resonant frequencies of Antenna V.The estimated SAR values for dry and wet skin are investigated and the results are compared in Table 4.
Referring to these results, SAR values tend to increase with an increase in the simulating frequency of the human tissue model except at 5.26 GHz.It is due to the pointed The estimated SAR value for the human tissue model with wet skin is higher than that of the model with dry skin for 1780 MHz, whereas for 2.4 GHz, 3.46 GHz, and 5.26 GHz, the SAR values for the model with dry skin are higher than that of the model with wet skin.

Antenna Performance under Bending Configurations.
Bending is investigated in the x and y planes by keeping the handmade Antenna V around the PVC (polyvinyl chloride) pipes [3,13] with diameters of 160 mm and 110 mm as shown in Figure 16. Figure 17 shows the four cases of bending during HFSS simulation.Figures 18-21 compare the measured reflection coefficient of the flat Antenna V in free space with the simulated and measured reflection coefficients of Antenna V, bent along the x and y planes on PVC pipes with radii of 55 mm and 80 mm.Frequency detuning, which is observed when simulated and measured results are compared for all four cases of the bending of Antenna V, is presented in Table 5.
For both simulated and measured results, the 1780 MHz band shifts to the right side of the spectrum for all four bending cases of Antenna V. Similarly, in the case of the simulated results, the 2.4 GHz band shifts to the right side of the spectrum for all four cases of bending of Antenna V along the x and y planes.The measured results described that the 2.4 GHz band shifts to the right side of the spectrum for Cases 2 and 3, but it does not shift for Cases 1 and 4.
The simulated results of the four bending cases illustrated that the 3.46 GHz band shifts to the left side for Cases 1, 3, and 4, whereas it shifts to the right side of the spectrum for Case 2. However, the measured results show that the 3.46 GHz band shifts to the left side for Cases 1 and 3, does not shift for Case 2, and shifts to the right side of the spectrum for Case 4.
The simulated as well as measured results show that the last 5.26 GHz band shifts to the left side of the spectrum for all four bending cases of Antenna V.In our proposed antenna, it could be easily seen that the antenna has strong currents on its ground planes, especially around the CPW feed line.The ground planes are actually a part of the proposed CPW-fed antenna.Taking this into account, it could be deduced that the distribution of the currents in the ground plane will change with the size and or/shape of the ground plane.As a result, the impedance and radiation performance    9 International Journal of Antennas and Propagation of the antenna will change, similar to that in [6].The bending of Antenna V along the x plane has more effects on ground planes, giving rise to an impedance mismatch, which is more significant for the 3.46 GHz band.This band is also severely degraded during bending along the x plane, for both the 55 mm and 80 mm bending radius cases.Similarly, the  For bending along the x plane, the measured and simulated frequency detuning values of Case 2 are either equal to or higher than that of Case 1, except for the 3.46 GHz band, for the measured value.On the contrary, for bending along the y plane, the measured and simulated frequency detuning values of Case 4 are either equal to or less than that of Case 3.
Maximum frequency detuning is observed for Case 3 at the 5.26 GHz band.In the measured results of the 2.4 GHz and 3.46 GHz bands, no frequency shift is observed for Cases 1, 4, and 2, respectively.Minimum frequency detuning has been found at the 3.46 GHz band for Case 4.
4.3.On-Arm Performance of the Proposed Antenna. Figure 22 illustrates the setup for two on-arm conditions of Antenna V during reflection coefficient measurement.The first condition is Antenna V on arm, without cloth, and the second condition is Antenna V on arm, with cloth [13,24].
The measured reflection coefficients for the two on-arm conditions are compared with the measured reflection coefficient of the flat Antenna V in free space, and the results are presented in Figure 23.For both on-arm conditions, the 1780 MHz, 2.4 GHz, and 3.46 GHz bands shifted to the left side, whereas the 5.26 GHz band shifted to the right side of the spectrum.The reflection coefficient of all bands of Antenna V for the on-arm, without-cloth condition, decreases.
For the on-arm, with-cloth condition, the reflection coefficient of the 1780 MHz and 5.26 GHz bands decreases, while that of 2.4 GHz band increases and that of 3.46 GHz band remains unchanged.
The frequency detuning of Antenna V for the two on-arm conditions is given in Table 6.The frequency detuning values of Antenna V for the on-arm, without-cloth condition are higher than that for the on-arm, with-cloth condition.
Maximum frequency detuning is calculated at the 1780 MHz band for the without-cloth condition during on-arm performance measurements.Minimum frequency  Finally, we compare the proposed antenna with other wearable antennas having single-band [1,2,4,5], dual-band [6][7][8][9][10], tri-band [11,12], and quad-band [13] operations, and the results are listed in Table 7.The single-band wearable antennas in [1,2] have comparatively lower SAR values at 2.4 GHz, but their sizes are larger than those of our proposed antenna.It can be seen that the sizes of the CPW-fed wearable antennas in [4-6, 8, 9] are smaller than those of our proposed four-band wearable antenna, but they are exhibiting only single-band and dual-band operations, respectively.
The gains of the dual-band wearable antennas in [7, 10] are better; however, their sizes are larger and their radiation efficiencies and bandwidths are lower than that of our proposed antenna.The gain, radiation efficiency, and SAR values of our proposed antenna are better than the tri-band wearable antennas proposed in [11,12].The quad-band wearable antenna in [13] has a larger size and lower bandwidths than that of the proposed antenna; however, its gain is higher for first, second, and fourth band, its radiation efficiencies are lower for the first three bands, and its SAR values are lower for the first two bands in comparison to the proposed antenna.The CPW line width of the proposed antenna is wider than the other works [4-6, 8, 9] illustrated in Table 7.

Conclusion
A novel wide CPW-fed multiband wearable monopole antenna and its design procedure have been proposed and successfully implemented.The antenna has a simple geometry and is easy to design on a polyester substrate.It has been shown that the proposed antenna with a wide CPW line is sufficient to cover the 1800 MHz GSM, 2.4/5.2GHz WLAN, and 3.5 GHz WiMAX bands, resulting in a four-band operation.The antenna provides good radiation   12 International Journal of Antennas and Propagation pattern characteristics and appreciable gain over each of the operating bands.The calculated SAR values at all the resonant frequencies of the wearable antenna are well below the acceptable limit of 2 W/kg, ensuring the safety of the user and viability of the proposed antenna for wearable applications.Further, the effect of bending in the x and y planes on antenna performance has been investigated and frequency detuning is discussed.Finally, the reflection coefficient characteristics has been measured for two on-arm conditions, with and without cloth, and the results are presented.

Figure 1 :
Figure 1: A wide CPW-fed multiband wearable monopole antenna with extended grounds.

Figure 2 :
Figure 2: Geometries of (a) Antenna I, (b) Antenna II, (c) Antenna III, (d) Antenna IV, and (e) Antenna V and (f) photograph of the proposed handmade Antenna V.

Figure 3 :Figure 4 :
Figure 3: Simulated reflection coefficient (S 11 ) of Antenna I to Antenna V.

Figure 15 :
Figure15: Modelling of (a) Antenna V on a three-layer human tissue model for SAR calculation and (b) Antenna V over a three-layer human tissue model in HFSS[13] (reproduced courtesy of the Electromagnetics Academy).

Figure 14 :
Figure 14: Radiation efficiency of the proposed Antenna V.

Figure 13 :
Figure 13: Measured and simulated peak gain of the proposed Antenna V.

Figure 17 :
Figure 17: HFSS simulation photographs of Antenna V under four bending conditions on a PVC pipe (a) bent along the x plane on an 80 mm radius, (b) bent along the x plane on a 55 mm radius, (c) bent along the y plane on an 80 mm radius, and (d) bent along the y plane on a 55 mm radius.

Figure 16 :
Figure 16: Photographs of Antenna V under four bending conditions on a PVC pipe (a) bent along the x plane on an 80 mm radius, (b) bent along the x plane on a 55 mm radius, (c) bent along the y plane on an 80 mm radius, and (d) bent along the y plane on a 55 mm radius.
space (measured) Bent along y plane (simulated) Bent along y plane (measured)

Figure 20 :Figure 21 :Figure 19 :Figure 18 :
Figure 20: Simulated and measured reflection coefficients of Antenna V bent along the y plane on a PVC pipe with an 80 mm radius.

Figure 22 :Figure 23 :
Figure 22: Setup for performance measurement of Antenna V for on-arm (a) without-cloth and (b) with-cloth conditions.

Table 4 :
Estimated SAR (W/kg) values for the three-layer human tissue model with dry and wet skin.

Table 5 :
Frequency detuning of Antenna V for bending in the x and y planes.

Table 6 :
Frequency detuning of Antenna V for on-arm conditions.

Table 7 :
Comparison of the proposed antenna with existing literature.