Various Textiles-Based Comparative Analysis of a Millimeter Wave Miniaturized Novel Antenna Design for Body-Centric Communications

A 60GHz compact and novel shaped microstrip-fed antenna based on a textile substrate for body-centric communications has been proposed in this paper. The antenna has a partial ground, and the textile substrate is made up of 1.5 mm thick 100% polyester. Two rectangular sections from the patch antenna’s radiator were removed to give the antenna a swan-shaped appearance. The antenna was designed and simulated using computer simulation technology (CST) microwave studio software. Simulated results show that, in free-space, the antenna achieved a high bandwidth of 11.6GHz with a center frequency of 60.01 GHz. With 89.4% radiation eﬃciency, the maximum gain of the antenna was 8.535dBi. For the on-body scenario, the antenna was simulated over ﬁve diﬀerent distances from a human torso phantom. At the closest distance from the phantom, the antenna’s gain was 5.27dBi, while the radiation dropped signiﬁcantly to 63%. The highest bandwidth of 12.27GHz was attained at 8mm, while the lowest bandwidth of 5.012GHz was attained at 4mm away from the phantom. Gain and radiation eﬃciency were comparable to free-space results at the furthest distance. The antenna was also simulated with ten diﬀerent textile substrates for both free-space and on-body scenarios. Among these ten substrates, denim, tween, and Quartzel fabric had similar performance results as polyester. This design achieved similar performance compared to other 60GHz textile antennas while being a bit more compact. This antenna will be a promising choice for body-centric communications because of its compact size, textile-based substrate, and excellent on-body performance.


Introduction
Advances in wireless technologies have meant an exponential increase in demand for smart wireless devices for the last few decades. ese devices include emerging wearables for health and fitness. In today's world, the expectation of having every device connected to a network for the transfer of a huge amount of data is quite the norm. Due to this ever-increasing demand for connectivity, it is apparent that current wireless bands are getting congested, and future communication will have to be in the millimeter-wave (mmwave) spectrum [1,2]. e mm-wave spectrum consists of frequencies between 30 to 300 GHz, which means the spectrum has a wavelength range from 10 mm to 1 mm. In free space, mm-wave has slightly higher attenuation than 4G frequencies due to high oxygen absorption. At 60, 180, and 380 GHz frequency bands, the attenuation of mm-wave is significantly higher. Short wavelength and higher attenuation mean communication in the mm-waves spectrum is suitable for short-range indoor obstacle-free applications [1].
e Federal Communications Commission (FCC) allocated the 60 GHz band for unlicensed use in 2001 [3,4]. Currently, the 60 GHz band is used for Wireless Gigabit Alliance (WiGig) wireless local area network (WLAN) devices. e 60 GHz band (57-64 GHz) offers a very high data rate in the gigabit per second range [3].
Wearable devices work in the vicinity of body-centric wireless communications (BCWC). BCWC includes wireless body area network (WBAN), a network of sensors around the human body. Wearable devices under WBAN have three modes of operation: on-body, off-body, and in-body. Communication between wearable devices placed in close proximity to the human body is known as on-body communication. Off-body communication is between a device placed on the human body and another device placed a few meters away from the human body. When an implant sends data to a device placed outside of the human body, it is known as in-body communication [2]. Body-centric devices may consist of medical equipment for monitoring health conditions such as heart rate and blood pressure [5]. Such devices designed for working in the 60 GHz band will have very directional and small antennas in physical size [3].
Textile antennas for WBAN are an attractive design prospect for their potential wearable uses. In general, these antennas are made from flexible conductive materials known as electro-textiles and substrate materials. Electrotextiles are made by weaving or knitting conductive threads like copper with nonconductive fabric materials. e substrate is a nonconductive material usually made of regular fabrics such as denim, polyester, and cotton. Textile antennas are practical due to their ease of integration into different fabrics, making them suitable for hospital patients' clothing, rescue workers' clothing, etc. e construction of textile antennas makes them low profile, lightweight, and low maintenance [6,7].
Antennas designed for body-centric communication need to be evaluated appropriately under comparable human body conditions. e human body acts as a lossy medium, and it can greatly degrade the antennas' performance [8]. In [9][10][11][12], narrowband and ultra wideband antenna designs have been discussed by various authors. For these cases, none of the antennas are designed on a textile substrate. However, in [13][14][15], the authors presented 60 GHz antennas that are not designed on textile substrates. In addition, there are very few mm-wave antenna designs presented in [16][17][18][19]. In [15], a 60 GHz body-centric antenna is designed for on-body operation. In free space, the antenna achieved a high gain of 9.37 dB with a bandwidth of 5 GHz. For on-body performance evaluation, a phantom of the human torso was created. Even though mm-waves have a very low penetration depth on the human body, the phantom was created, taking skin, fat, and muscle layers. As a result, the antenna produced a slightly lower bandwidth but a higher maximum gain. Puskley et al. designed a disc-like low-profile antenna that is well matched between 59.3 and 63.4 GHz [16]. e antenna lost about 25% efficiency when kept 1 mm away from a 5 mm thick skin phantom. Similar to the previous paper, depending on the antenna's distance from the phantom, the on-body maximum gain increased compared to free space. A low-profile slotted patch 60 GHz antenna was designed in [17]. e antenna has a wide bandwidth of 9.8 GHz with 74% efficiency in free space. For the on-body scenario, the antenna's efficiency decreased to 63%, but again, like in previous cases, the gain increased from 10.6 dBi to 12.1 dBi. e abovementioned mm-wave antennas are not textile based.
A 60 GHz textile antenna based on a microstrip patch array is proposed in [18]. Due to the fabrication process's complexity, instead of electro-textiles, copper foils of 0.07 mm were used for all radiating materials. For textile material, only a cotton-woven fabric of 0.2 mm was used. e antenna was fed by a 15.2 mm-long microstrip line and a V connector for performance measurements. e reflection coefficients were lower than −10 dB in the 57-64 GHz range with 40.7% efficiency in free space. For human body performance, the antenna was kept at 0 and 5 mm away from a skin phantom of size 2 × 100 × 100 mm 3 . At the closest distance from the phantom, both the efficiency and gain of the antenna decreased. At 5 mm away from the phantom, these values increased closer to free-space values. e antenna was also subject to bending and crumpling, both of which had little effect on the antenna's performance. In [19], a similar sort of analysis has been made on a microstrip-fed Yagi-Uda antenna. e antenna consists of a driven dipole and 10 directors printed on a 0.2 mm thick fabric made of cotton. For the previously mentioned reason, the radiators are made from 0.07 mm thick copper foil. e antenna produced an end-fire radiation pattern over the 55-64 GHz range. Computed and measured maximum gain in free space was around 9.2 dBi. e antenna was evaluated on a homogeneous skin-equivalent phantom measuring 10 × 100 × 100 mm 3 at two different distances like in [18]. Compared to free space, the antenna's efficiency dropped to 48% from 78%, while the gain increased to 11.9 dBi. ese two antennas are designed for only cotton substrates [18,19]. One was designed for 60 GHz array antennas and another one was designed to get end-fire radiation.
Most of the studies conducted for the wearable textile antenna are based on designs working in much lower frequency bands [20][21][22][23][24][25][26][27]. Although the antennas presented in [18,19] are textile based, they are designed for only one textile material, cotton. e antenna presented in [18] was an array antenna. e objectives and research investigation of the work [18,19] are different from the research objectives and investigation of this paper. is paper's main objective is to propose a design of a 60 GHz antenna based on a 100% polyester substrate for wearable devices. Various textilesbased comparative analysis of a millimeter wave miniaturized novel antenna design for body-centric communications is presented here. e performance of the antenna has been compared with ten different textile substrates. Parametric studies are performed by changing various antenna parameters. e on-body performance was also investigated. In addition, the performance also investigated placing the antenna at different distances from the human body model. Finally, the on-body performance of the antenna for various substrates has been investigated and compared.
is paper consists of six sections. e antenna's structure and its material selection are given in Section 2. Free-space performance is described in Section 3. Parametric changes were made where lengths of different parts of the antenna are altered to observe performance changes. e polyester textile substrate was replaced with ten other substrates and evaluated in free space. e simulation results are included in Section 3. For the wearable scenario, the antenna was placed over a torso phantom at 5 different distances.
e on-body simulation results are detailed in Section 4. On-body simulation results for other textile substrates closest to the phantom are also included in this section. In Section 5, a brief comparison of the proposed antenna has been made with other available 60 GHz textiles. Lastly, in Section 6, conclusions are made.

Materials and Antenna Design
To minimize losses in wearable scenarios, an appropriate design should be considered for textile antennas. e antenna should be of low-profile and flexible enough so that the performance does not significantly degrade under stress [7]. is paper proposes the design of a microstrip patch antenna with the partial ground (Figures 1(a)-1(d)). Instead of electro-textile, the radiating patch and ground of the antenna are made of perfect electric conductor (PEC) with 0.035 mm thickness. e textile substrate is made of 1.5 mm thick 100% polyester, which has a dielectric constant of 1.9. e dielectric constant is the real part of relative permittivity and is an important parameter as it affects the impedance bandwidth of the antenna. Low dielectric constant reduces surface wave losses within a substrate and hence increases impedance bandwidth [7]. e overall dimension of the textile substrate is 12.2 × 12 × 1.57 mm 3 . e electrical size of the antenna for the overall width and length of the substrate is 0.42λ and 0.41λ, respectively. e antenna is fed by a 4.9 mm-long and 2.15 mm-wide microstrip line. e list of antenna materials and their thickness and dielectric constant are given in Table 1. e antenna is designed and simulated using computer simulation technology (CST). e antenna's radiating patch has a 7.3 mm × 12 mm dimension and is connected to a 2.15 mm-wide 4.9 mm-long feedline. e electrical size of the antenna for the overall width and length of the radiating element is 0.68λ and 0.42λ, respectively. e ground measures 4.5 mm × 12 mm. Two rectangular patches have been cut from the middle and right-hand sides of the radiator. e removed middle section from the patch has a 5.38 mm × 2.9 mm dimension, while the removed section from the side is 6.3 mm × 3.1 mm. e removed radiator sections also give the antenna a swanshaped design. e total dimension of every section of the antenna is marked in Figure 1(c), and the values are given in Table 2. Figure 2 shows the return loss curves of the proposed textile antenna. e antenna produced a very wide −10 dB impedance bandwidth of 11.632 GHz with a center frequency of 60.01 GHz in free-space conditions. e antenna shows a very good return loss value of close to −19 dB at the operating frequency. With a radiation efficiency of 89.4%, the antenna achieved a maximum gain of 8.535 dB, as shown in Table 3. Figure 3(a) shows the 3D radiation pattern at 60 GHz, while the E-plane and H-plane radiation patterns are shown in Figures 3(b) and 3(c), respectively. e radiation pattern of this antenna shows nearly omnidirectional radiation patterns with slight distortion at various angles.

Free-Space Simulation Results
is could be due to the lower ground plane of the backside of this antenna. e antenna's E-and H-plane radiation patterns at 60 GHz are shown to have good characteristics, except for slight null and few angles. e voltage standing wave ratio (VSWR) is shown in Figure 4. e figure shows that the antenna has a VSWR value at 60 GHz, close to 1.3, which is excellent. For a surface, the current distribution is shown in Figure 5.

Parametric Studies.
For the parametric study, the antenna's feedline width (fw), length (l), and ground length (lg) were changed to observe the changes in the return loss curve. e rectangular patch size removed from the middle section was also varied by changing the values of the variables "ls" and "ws."

Feedline Width (fw).
e center frequency shifted slightly towards the right to 60.164 GHz when the feedline width was made smaller from 2.15 mm to 2 mm. Bandwidth decreased by 300 MHz while both gain and radiation efficiency remained virtually unchanged. When the feedline width was increased to 2.5 mm, the center frequency shifted to the left to 59.66 GHz. e bandwidth significantly increased by more than 1 GHz when compared to free space. Gain remained unchanged for both smaller and larger widths. While the radiation efficiency slightly increased for feedline width. Figure 6 shows a return loss comparison curve for different feedline widths with the free-space result. Table 3 contains the detailed values of the parameters for feedline width changes.

Total Length of Antenna (l)
. When the antenna "l" length was decreased from 12.2 mm to 12 mm, a shift of the center frequency to the right was observed (Figure 7). Bandwidth remained unchanged, but both gain and radiation efficiency marginally increased. For a larger antenna length of 12.5 mm, the center frequency shifted significantly to 58.834 GHz. All three parameters, bandwidth, gain, and radiation efficiency, decreased slightly. e details are given in Table 4, and the return loss comparison curves are given in Figure 7.
International Journal of Antennas and Propagation

Ground Plane Length (lg).
Decreasing the size of the ground length, lg, by 0.5 mm, made the center frequency shift to the right while increasing the impedance bandwidth ( Figure 8). Gain remained unchanged, but the radiation efficiency decreased by more than 1%. Opposite observations were made when the size of the ground length was increased. e center frequency shifted to the left, bandwidth decreased by 0.6 GHz (Figure 8), radiation efficiency increased, and the gain remained almost the same (Table 5). Figures 9(a) and 9(b) show the return loss for change in radiator dimension. A huge shift of the center frequency to the left was observed when the length "ls" was decreased from 5.38 mm to 5 mm (Figure 9(a)). Even though both bandwidth and gain increased, the radiation efficiency decreased slightly. Very few changes were observed when the length was increased to 5.5 mm. Altering the width "ws" from 2.9 mm to 2.6 mm and 3.2 mm had an insignificant effect on all four parameters. Table 6 shows all the detailed values from these changes.

Simulation Results Using Different Textiles.
e antenna's polyester substrate was replaced by 10 different fabrics to analyze the antenna design in free-space further. ese fabrics are as follows: Jeans, Denim, Silk, Tween, Panama, Felt, Moleskin, Cotton, Quartzel Fabric, and Cordura/Lycra. e dimensions of these substrates were kept the same as 100% polyester, but the thickness was altered depending on the fabric used. Results indicate that denim, tween, and quartzel fabric performed closest to the initial design in terms of center frequency and bandwidth (Figure 10(a)). Except for cordura/lycra, the gain was above 8 dBi for all the substrates. Due to the similar thickness and relative permittivity of jeans and quartzel fabric, the results were very similar to polyester fabric. A summary of this fabric variation free-space analysis is given in Table 7. Figures 10(b) and 10(c) show the antenna's E and H plane radiation patterns for different textile substrates. Due to different textile substrates, the E and H plane radiation patterns do

On-Body Simulation Results
For body-centric simulations, a human torso-equivalent phantom is created.
is phantom contains three layers representing skin, fat, and muscle. mm-waves have very low penetration capability, and at 60 GHz, they can penetrate only 0.5 mm of human skin [15]. For this reason, most of the research works for 60 GHz are based on skin-equivalent phantom [18,19]. We have chosen three layers for our work to represent a more accurate model of the human body. ese three layers can be modeled accurately by taking their relative permittivity and conductivity. In Table 8, detailed values of the dimensions, relative permittivity, and conductivity of the torso phantom are given [28]. Figure 11(a) shows the different layers of the human torso phantom. e antenna was positioned at 5 different distances from the phantom to analyze the antenna's performance in wearable scenarios. Starting from 2 mm, the antenna was

100% Polyester
Results. e return loss curves for different antenna placement distances from the three layers of the human body model are shown in Figure 12(a). At 2 mm and 6 mm away from the phantom, the center frequency of the antenna shifted by around 0.5 GHz. e shift was not as significant for 4, 8, and 10 mm, while at 10 mm, the center frequency was closest to the free-space center frequency. From distances of 6 mm and above, the impedance bandwidth was over 11 GHz, with 8 mm having the highest bandwidth. e bandwidth was lowest at 4 mm, and at 2 mm, it was just above 10 GHz. e maximum gain was around 8.7 dBi for distances of 6, 8 and 10 mm. At 2 mm, the gain dropped to a minimum value of 5.27 dBi. Radiation efficiency increased as the antenna's distance from the phantom increased but was significantly lower compared to free-space. In this study, it is noted that the antenna's onbody radiation efficiency is slightly lower than free space.
is happens due to the absorption of different lossy tissues of the human body. From Table 9, it is noted that the reduction of radiation efficiency is lower when the distance between the antenna and the body increases. is happens because when the antenna is placed at a higher distance from the human body, it has less effect on the radiation efficiency from lossy human tissues. Table 9 summarizes the results of on-body antenna performance with distance variation. E-Plane and H-plane radiation pattern with distance variation are shown in Figures 12(b) and 12(c). e radiation patterns for E and H planes do not change much for changing the distance between the antenna and the human body model, whereas the variation of power loss is observed.

Different Textile
Results. All ten substrates were also simulated for an on-body scenario by keeping the antenna 2 mm away from the phantom. Antenna's performance    International Journal of Antennas and Propagation     International Journal of Antennas and Propagation 7    International Journal of Antennas and Propagation       International Journal of Antennas and Propagation using denim and quartzel fabric was closest to on-body polyester performance. Figure 13(a) shows an on-body return loss comparison for all 10 different substrates along with the original polyester substrate. For denim, even though the maximum gain decreased to 5.239 dBi, the center frequency was exactly at 60 GHz and bandwidth close to 10 GHz. With a low gain of 4.376 dBi, 43.53% radiation efficiency, and center frequency of 62. 19 GHz, jeans achieved the highest bandwidth of 23.51 GHz. Overall, every substrate achieved a bandwidth higher than 8.6 GHz, while the radiation efficiency never exceeded 64.02%. Detailed values from the on-body analysis are given in Table 10. Figures 13(b) and 13(c) show the antenna's E and H plane radiation patterns for different textiles when the antenna is placed 2 mm away from the human body model. For changing different textile substrates, the on-body radiation patterns of the antenna change slightly; however power loss is noticed.

Comparison with Previous Works
As mentioned previously, there aren't many 60 GHz textile antennas studied for body-centric communications. Compared to works in [18,19], the antenna proposed in this work is smaller in length but a bit wider. Overall, the dimension is a little bit more compact than other works. e bandwidth of our proposed antenna was higher in both free-space and onbody. e proposed design in this study achieved a similar result in free-space in terms of maximum gain, but for onbody, gain decreased significantly. Table 11 provides a comparative overview of our work and other known designs.

Conclusions
is paper's main purpose was to design a compact 60 GHz antenna based on a textile substrate for body-centric communication. Designed on a 100% polyester substrate, the microstrip patch antenna showed high impedance bandwidth in the 60 GHz band with good radiation efficiency. From parametric studies, it is shown that by decreasing the ground length, the bandwidth increases. Both gain and radiation efficiency are not affected much when changes are made in radiator and ground size. For on-body arrangement, the antenna performed best when kept at 8 mm away from a phantom. We have seen both bandwidth and gain increase at this distance compared to free-space with a slight degradation in radiation efficiency.
e antenna's lowest gain was at the closest distance from the phantom. In general, the antenna's performance kept on improving as the distance between the phantom and the antenna kept on increasing.
eir relative permittivity value characterizes textile substrates.
is was evident as the antenna performed equally well when the polyester substrate was replaced by denim, tween, or quartzel fabric. Compared to previous works on 60 GHz textile antennas, our design achieved higher impedance bandwidth. Even though the free-space gain was similar, the on-body gain was much lower. ere are not many textile antennas designed for mm-wave wearable devices. Our work will be a good addition to this class of antennas because of its compact and relatively simple design.

Data Availability
e data used to support the findings of this study are freely available at http://niremf.ifac.cnr.it/tissprop/

Conflicts of Interest
e authors declare that they have no conflicts of interest to report regarding the present study.