Fateh Benmahmoud, Pierre Lemaitre-Auger, Smail Tedjini
©SHUTTERSTOCK.COM/ GROUND PICTURE
A folded planar dipole antenna was designed and customized to operate close to the human body. Backward radiation was reduced to protect the human body from unwanted electromagnetic (EM) emissions and increase the antenna’s operating range. The antenna design was based on the concept of a folded planar dipole, which presents favorable impedance bandwidth (BW) characteristics. For integrability and robustness during usage in real environments, the antenna was completely covered by resin, taking into account the tradeoff between the antenna’s radiation performance, protection, and isolation. Measurements demonstrated a 10-dB impedance BW of 470 MHz (18.76%) covering the 2.27–2.74-GHz band. The maximum gain at 2.45 GHz was 1.4 dB isotropic (dBi). The antenna has a nearly omnidirectional pattern, occupies 77 × 35 × 11.15 mm, and weighs 120 g, making it easy to integrate with clothes. The resin packaging increased the mechanical robustness and improved the design’s isolation from external interference and perturbations; however, it reduced the radiation efficiency to 48.35%. The presented antenna is an excellent candidate for many wireless applications, in particular, applications where withstanding exposure to external mechanical strains and EM perturbations is required.
In wireless communications and networking, on-clothing antennas are becoming essential for many applications, including military and security. The integration of wireless communication devices with clothes has become useful and necessary and requires consideration of both performance and ergonomics. At this time, on-clothing antennas are not everyday products that can be found on the shelf. Also, the way of integrating the antenna on clothes is usually challenging. Indeed, many specific problems must be solved concerning such systems. User safety is an essential factor to consider. The specific absorption rate (SAR) and maximum radio frequency (RF) power thresholds must be carefully maintained below the permitted threshold. Moreover, the presence of the human body could alter radiation from the often-desired omnidirectional pattern. It could also provoke a detuning effect, causing a decline in the range of communication. Using a ground plane or a high-impedance surface (HIS) is a popular method to limit such a negative effect. Proximity to the human body reduces antenna efficiency; therefore, a minimum distance separating the antenna and the body must be guaranteed without affecting ergonomics. In addition, feeding cables must be firmly attached to the antenna to withstand frequent body movements, which is a crucial task. Finally, the clothes must be washable, and the antenna and its feed cables must resist corrosion, fatigue, and deformations.
Many applications that require on-clothing communication systems are faced with extreme operative conditions. These challenges are generally related to climates in which temperatures are extremely low (e.g., polar exploration) or high (e.g., deserts and fire-affected areas). Humidity rates can also reach a climax in some climates, such as tropical forests and on the decks of ships on the high seas. Accounting for these factors is essential when designing items with on-clothing antennas that can be worked seamlessly into the jackets for firefighters, sailors, military personnel, and explorers. Other factors are the operator’s posture and sudden body movements that can make nearby materials and items, such as the operator’s body and clothing, tangent with the antenna structure. This considerably affects the impedance matching and can make the antenna inoperative if it is insufficiently isolated and reinforced. In extreme cases, the operator’s movements and actions can lead to permanent malfunctioning and deformations.
Usually, antenna design efforts focus more on achieving high radiation characteristics than improving protection and isolation, which produces antennas capable of operating in regular and comfortable environments. However, the radiation efficiency drops when such antennas are deployed under harsh conditions, which is the case in most real-life situations. A harsh environment includes phenomena that affect the antenna physically, such as rust and corrosion, and mechanically, such as fatigue and deformation caused by continuous usage and external strain. These factors change the antenna behavior, usually for the worse, and sometimes lead to a permanent interruption of the communication link. Another real-life problem is EM interference caused by nearby electronic devices that could obstruct the antenna and, thus, the communication chain. All these problems should be addressed if the antenna is intended to operate in real environments.
An efficient design approach for obtaining robust yet flexible antennas is to use the silicon-based elastomer polydimethylsiloxane (PDMS) [1]. This material offers mechanical and electrical properties that are tunable. It is possible to control rigidity and permittivity by controlling the inclusion of this material. Several studies have shown promising results for PDMS-based antennas. In [2], mesh conducting sheets embedded in a substrate were used to design an antenna operating at 365 MHz. The antenna was entirely encapsulated inside the PDMS, with copper mesh sandwiched between the layers of the substrate. The antenna showed waterproof and mechanical robustness, without any noticeable difference before and after the washing operation. In [3], a planar inverted-F antenna (PIFA)-like structure for long-distance voice communication was designed and tested. The antenna’s immersed conductive part was produced through a perforation process with hexagonal holes. The antenna occupies 130 × 80 × 11 mm and resonates at 383.5 MHz. It has been found that PDMS reduces antenna impedance when deployed near the operator’s chest. Based on a tripod kettle structure, the copper antenna part in [4] was encapsulated inside the PDMS. Its small circular shape made it easily embeddable.
The work in [5] proposed robust and compact PDMS antennas for on-clothing applications. In addition to reducing the antenna size, it was found that the PDMS made the design resilient in different environments. The antenna was embedded in a ballistic jacket, demonstrating a roughly 600-MHz 10-dB return loss BW within the Industry, Science, Medicine (ISM) radio band. The work in [6] presents a flexible antenna that combines conductive fabric and PDMS material that also serves as a substrate and an encapsulation material. The antenna operates around 2.46 GHz, with a 10-dB return loss BW ranging from 3.3% to 5.7%. Many other antennas, as in [7], [8], [9], and [10], and antenna arrays, as in [11], with PDMS encapsulation have been reported to obtain robust and durable features against environmental changes.
In this study, a robust antenna that can endure harsh weather conditions and external strains was designed. It also had to be sufficiently flexible to minimize the risk of cracking and increase its operational durability. This approach ensured a tradeoff between flexibility and robustness to overcome the limitations of highly flexible and rigid designs. The structure was based on a planar folded dipole antenna covered with a resin protection layer. The antenna was designed to operate on clothing in the ISM band, covering the spectrum from 2.3 to 2.7 GHz. The fabricated prototype was passed through different mechanical and environmental proofs and showed high robustness and stable mechanical, electric, and EM performance.
This article is organized as follows. The “Antenna Configuration” section describes the designed antenna structure, including the resin packaging and the steps of the fabrication process. The “Simulations and Experimental Validation” section discusses the evaluation of the performance of the fabricated prototype, including detailed studies of the SAR, isolation from harsh weather conditions, and immunity against mechanical constraints. Concluding remarks are presented in the “Conclusions” section.
The antenna design was based on the architecture of a half-wave folded dipole, which can produce higher gain and a larger impedance BW than an ordinary half-wave dipole [12]. The topology of the studied folded dipole antenna is shown in Figure 1. It consists of three main parts: the radiating element (the folded dipole), an excitation structure (the antenna feed and matching stub slot), and a reflective ground plane [Figure 1(a)]. Slots of the same length were introduced to the folded dipole to adjust the antenna input impedance. This promoted proper matching over a wide frequency band and increased the antenna’s radiation resistance. The number of slots was limited to three, taking into account a balanced compromise between impedance matching and the size of the antenna. A stub slot was introduced and optimized to adjust the impedance matching between the folded dipole and a 50-Ω feed line. All the ground copper parts of the antennas were connected to the bottom face of the Kapton substrate by using ground vias. The purpose of these connections was to prevent parasitic ground inductance caused by ground current return paths, thus reducing parasitic signals.
Figure 1. The antenna design: the (a) printed design (on Kapton), (b) antenna dimensions, and (c) antenna layers.
As illustrated in Figure 1(c), a separate ground plane was placed 6.35 mm below the bottom face of the printed antenna. This plane served as a reflector for the connected radiating part. It directed the radiation pattern toward the desired zone (i.e., forward) to protect the operator. It was placed at an optimal distance to boost the antenna’s proper matching. Maintaining a short distance between the reflector and the radiator was essential since doing so reduces the amount of loaded resin, thus lightening the system’s weight and volume. All the geometric parameters in Figure 1 were optimized using the CST Microwave Studio (MWS) computer simulation tool (Table 1).
Table 1. The optimal antenna parameters (in millimeters).
The input signal propagates from the antenna feed point through the coplanar waveguide (CPW) structure to be converted to the even mode of the coupled CPW. This feeding architecture can be influenced by the presence of an even mode, which is excited by numerous elements, especially the following:
At each slot, the CPW line curvature provokes a difference in the distance traveled by the wave, creating a phase difference between the waves. A technique that helps to eliminate the effect of the even mode is the establishment of cross connections known as air bridges [13]. These connections, positioned near the extremities of the line curvature, oblige both ground planes to the same potential, which inhibits the even mode. Nevertheless, using air bridges has drawbacks. It results in capacitive coupling with the central feed line, which locally degrades the characteristic impedance ${Z}_{c} = {\sqrt{{L} / {C}}}$ [14], where C and L are the capacitance and the inductance of the line. A slight narrowing of the line (step compensation) locally increases the inductance, which counterbalances the capacitance effect [13], [15], [16] [Figure 1(a)]. In our case, the air bridges are connected to the ground plane at the bottom face of the substrate through vias.
The antenna was first printed on a 0.15-mm-thick Kapton substrate with a dielectric constant of 3.4 and a low loss tangent of 0.005. The dielectric constant of 3.4 sufficiently minimized the antenna’s dimensions in anticipation of its integration with clothes. The 0.15-mm Kapton sheet thickness offered high flexibility, allowing the antenna to bend if needed. Flexible thin Kapton was used as a substrate upon which to print the whole structure, including the reflective ground plane on the same sheet face and in one operation. The ground plane was placed at the appropriate distance below the antenna structure by simply bending the Kapton sheet [Figure 1(c)]. This approach accelerates the fabrication process and reduces its complexity and cost, especially in the case of mass production. Figure 2 displays the simulated antenna’s current distribution at 2.45 GHz, using CST MWS. The surface current is concentrated mainly on the feed line and at the edge of the slot that forms the folded dipole. There is also a high current intensity on the edges of the matching stub slot that assures impedance matching between the feed line and the folded dipole.
Figure 2. The simulated current distribution on the designed antenna at f = 2.45 GHz.
The aims of antenna optimization are impedance matching, minimizing size, and reducing undesired EM emissions during usage. The reflection coefficient and the SAR, in Watts per kilogram, were analyzed. The SAR describes the absorbed power per human tissue mass when the human body is exposed to the EM field. The SAR for EM energy is related to the electric field within the tissue, as follows: \[{\text{SAR}} = \frac{1}{V} \mathop{\int}\nolimits_{\text{Sample}}{\frac{{\sigma}{(}{r}{)}\,{\vert}\,{E}{(}{r}{)}\,{\vert}\,^{2}}{{\rho}{(}{r}{)}}}{dr} \tag{1} \] where V represents the volume of the tissue sample, v is the conductivity of the sample, t is the sample density, and r represents the calculation points position. Every part of the human body has a SAR limit that must be respected. Complying with SAR limit standards is necessary for the safe use of the antenna. In particular, in the case of a trunk and with an injected power of 1 W, the allowed limit values are 1.6 W/kg for 1 g of tissue [17] and 2 W/kg for 10 g of tissue [18].
The structure in Figure 1 is covered in resin. The upper layer of this resin (the superstrate) ensures the high isolation of the radiating element. In addition, as a dielectric, it affects impedance and radiation efficiency. Whereas impedance matching was manageable by readjusting and optimizing the antenna geometry, the radiation efficiency drop was an inevitable sacrifice since it was related to the resin loss tangent. The other motivation for adding protection resin was the reinforcement of the EM isolation of the antenna against the hosting environment. The superstrate thickness was studied to determine the minimum value required to achieve good isolation. The central resonance frequency shifting is represented with respect to the superstrate thickness in mm, as demonstrated in Figure 3(b). A polyethylene-based resin that offers a compromise between the mechanical and dielectric characteristics (see the “Parametric Study” section) was selected.
Figure 3. The impact of the variation of the resin’s top layer thickness. The simulation was obtained using the CST MWS EM simulation tool. The (a) simulated reflection coefficient and (b) frequency response shifting.
To see the impact of the superstrate on the antenna response ${(}{\left\vert{S}_{11}\right\vert}{)}$, we incremented its thickness from 0.5 to 4.8 mm (Figure 3). The space separating the Kapton sheet and the reflective plane was filled with the resin for all the presented cases. As shown in Figure 3(a), the antenna response shifted gradually from around 2.62 to 2.45 GHz. As the superstrate thickness increased, its variation impacted the antenna reflection coefficient less. Shifting toward lower frequencies was expected, as the resin used had a higher permittivity ${\left({\epsilon}_{r} = {3.8}\right)}$ than air. The antenna was better matched to 50 Ω at 4.8 mm compared to the lower thicknesses. Figure 3(b) conveys that the resonance frequency shift ${\Delta}{F}$ was very sensitive to the superstrate thickness. The high sensitivity was translated by a fluctuating and unstable derivative ${d}{\left\vert{\Delta}{F}\right\vert} / {dH}_{\text{Superstrate}}$ for lower thicknesses (the highlighted area). Beyond a superstrate thickness of 5 mm, the antenna response sensitivity dropped considerably, which meant more stability or minor frequency shifting. The frequency shift toward a lower frequency was expected owing to the high dielectric permittivity of the resin.
Given that a stable response also means better antenna isolation, we decided to fix the thickness of the superstrate to 4.8 mm. This was also done to compromise between isolation and lightness. The simulation indicated a 20-MHz/mm sensitivity of the frequency response for this thickness. This value is acceptable considering that the accuracy of the fabrication technique is as small as 0.05 mm. In this case, the fabrication uncertainty is roughly a ±1-MHz frequency shift, which is acceptable.
The proposed design, presented in Figure 1, includes numerous parameters (Table 1). Variations of each of these parameters inevitably influence the design’s characteristics; however, only a few deeply impact the antenna response. The major parameters are generally located around the areas that show high current density, as detailed in Figure 2. These parameters are related to the CPW feed line (W1, W2, W3, G1, and G2), the inserted slot stub (L4), and the folded dipole structure (L5, L6, L7, and G6). A logarithmic scale on the x-axis in Figure 4 was used to clearly represent the parametric study. The y-axis presents the relative shift in the resonant frequency. Here, ${\Delta}{f}_{r}$ is defined as follows: \[{\Delta}{f}_{r} = \frac{{f}{-}{f}_{\text{optimal}}}{f}{(}{\%}{)} \tag{2} \]
Figure 4. The impact of the major parameters on the antenna response: the (a) radiating part parameters and (b) CPW feeding parameters.
where f represents the resonance frequency and foptimal is the resonance frequency at 2.45 GHz. As illustrated in Figure 4(a), we can see that parameter L7, which represents the length of the folded dipole, had the most significant impact on the resonant frequency. This was expected due to its relation to the emitted wavelength. The impacts of L5, L6, and G6 remained below 5%. As evident in Figure 4(b), the antenna response was more sensitive to the parameters of the CPW feeding structure (W1, W2, W3, G1, and G2). Parameter W3 had a higher impact than the rest, with a relative frequency shift between –33% and 38%. Regarding the high sensitivity of the antenna response to the CPW feeding structure variation, the realization of this part of the design had to be done carefully (i.e., with high accuracy).
Since the antenna was designed to be deployed in harsh environments, we protected it with a shielding material, which protected the radiating element from all external disrupting factors that could alter its performance. We chose a polyethylene-based resin, as it offered an excellent tradeoff between rigidity (a kilogram-force of 33.6 dN/cm3) and the dielectric loss tangent. The resin was initially characterized using several techniques to determine the resin’s dielectric constant and loss tangent. Two main techniques were employed to achieve this purpose. The first, an open-ended coaxial probe technique, was carried out using a N1501A dielectric probe kit from Keysight [19]. The kit is based on an impedance analyzer that measures the response to a signal transmitted into the material being tested. The second technique utilized cavity resonance, which made the test results available only for a finite number of discrete frequencies. Several sheets of the polyethylene-based resin material, with a fixed thickness of 2, 4, and 6 mm, were tested using a resonant cavity from MuEpsln [20]. The use of different techniques was necessary to reduce the uncertainty of the measurements and avoid any unexpected behavior of the employed resin. Analyses have shown some fluctuation between one sample and another and even in different locations for the same sample. Such fluctuations are attributed to impurities caused by air bubbles within the fabricated resin, which forced us to be rigorous about the purity of the resin samples. The mostly pure resin samples had a dielectric constant of 3.8 and a loss tangent of 0.03. These values changed slightly for less-pure samples, which generally had lower dielectric constant values.
The conductive copper areas were printed on a thin Kapton sheet. The Kapton sheet was then bent and fixed inside a mold according to the optimized dimensions and distances between different layers. The mold dimension gave the antenna its final cuboid shape, with dimensions of 77 × 35.98 × 11.15 mm. The corners of the cuboid were rounded for esthetic and ergonomic reasons. The resin also covered the SMA RF connectors and part of the 50-Ω RF cable to reinforce the SMA connector soldering and avoid any parasitic signal from the connector–antenna liaison. Tests have revealed that this antenna can always be operational, with good radiation performance after immersion. This makes the immersion task beneficial despite the process adding extra weight to the antenna.
The immersion of the design in the resin was elaborated using a casting technique in which the resin fills the gap between the antenna and the reflective ground plane, which allows the optimal spacing to be maintained. As previously discussed, the thickness of the superstrate was studied to obtain a compromise among size/weight, isolation, and frequency response stability. A thickness of 4.8 mm was maintained, resulting in a whole design thickness of 11.15 mm. Figure 5(a) shows the antenna after immersion in the resin (black). During the immersion step, reducing and controlling the randomly distributed air bubble rate in the resin was one of the biggest challenges. These bubbles give the resin fluctuating electrical properties, which results in unpredictable behavior. This issue was critical since the presence of undesired air volume within the resin means lowering the dielectric constant, thus shifting the reflection coefficient response of the whole design. The final prototype had high rigidity and cohesion and low heat conductivity. The antenna was then equipped with a metallic hook on the bottom to facilitate its attachment to the operator’s clothing [Figure 5(b)]. The weight of the entire design was 120 g, making it easily portable and embeddable into clothes and, ultimately, on the human body. The prototype also showed good solidity and robustness against mechanical strains and shocks.
Figure 5. The realized prototype: the (a) final antenna structure and (b) antenna attached to the human body.
The fabricated antenna was intended to be deployed in harsh environments under different EM and mechanical conditions. Therefore, we conducted several experimental tests to evaluate and quantify the impact of the surrounding conditions on the designed antenna’s performance. The tests mainly concerned radiation and emissions near the operator’s body, the influence of weather conditions, and the impact of geometrical deformations on the antenna’s response.
The antenna was simulated using CST MWS, with the final optimal dimensions listed in Table 1. The simulation included the antenna reflection coefficient, the radiation efficiency, the radiation pattern, and the localized SAR calculated for 10 g of tissue. The simulations used the Gustav model, an approximation of the human body in the simulation tool library. It allows us to predict the impact of the body on the antenna RF performance and the calculation of the SAR. Figure 6 gives the simulated SAR distribution when positioning the antenna on the central part of the chest. Hotter body areas indicate a higher SAR rate. A 3-mm spacing was kept between the antenna packaging and the model to simulate realistic conditions in the case of on-clothes integration. As shown in Figure 6, the calculated localized SAR remained well below the allowed threshold of 2 W/kg. The nearest tissue areas to the antenna packaging bottom (the lowest gap, of 3 mm) had a maximum SAR value of 0.291 and 0.162 W/kg for 1 and 10 g of tissue, respectively, for an injected power of 1 W. These values were well below the fixed European standards (1.6 and 2 W/kg). Body tissues far from the antenna location had much lower SAR values. These results show that the antenna is ultimately safe for deployment on personnel. Owing to the isolation offered by the resin, the antenna maintained its response when the lowest gap (the spacing between the skin and the bottom of the design packaging) was reduced. It maintained an almost steady response, even when it came into contact with the operator’s skin. A similar result was obtained for the maximum SAR values, where a minor increase occurred.
Figure 6. The calculated SAR on a Gustav voxel model in CST MWS (1 W of input power).
Figure 7 describes the simulated and measured reflection coefficient ${\left\vert{S}_{11}\right\vert}$ (in decibels). The measurements were done inside an anechoic chamber in the air and on the human body. We can see that the antenna had low sensitivity to external disturbances. The measured BW was relatively broader than the simulated BW, with several ripples of the measured S11. These ripples are an indication of multimodal operation. However, this is usually caused by the measuring coaxial cable attached directly to dipoles and pseudomonopoles without some sort of balun. This factor was not taken into consideration during the simulation process. Nevertheless, the simulation results remained close enough to the measured response. In contrast to the in-air case, no considerable impact on the antenna response was measured for the on-body case. The S11 BW was slightly wider at 470 MHz (2.27–2.74 GHz). It included the ISM band (2.4–2.5 GHz) with a reflection coefficient lower than –10 dB, offering a comfortable margin for fabrication process accuracy.
Figure 7. The fabricated prototype: the simulated and measured reflection coefficient.
Illustrating the antenna’s radiation pattern during deployment is more accurate and realistic than a pattern measured in free space. The radiation pattern considers the environmental variants that characterize a real scenario: the presence of the human body, the slight inclination introduced by the operator’s body, and the clothes. For far-field measurements, we used the anechoic chamber at the Systems Design and Integration Laboratory (LCIS), located in Valence, France. It is equipped with a remote rotative platform that can handle weights as high as 100 kg. This platform allowed us to perform the measurements for an adult, as depicted in Figure 8(e).
Figure 8. The simulated and measured realized gain at f = 2.45 GHz: the (a) horizontal plane, (b) vertical plane, (c) on-body (chest) normalized radiation pattern measurement, (d) simulated and measured gain versus the frequency, and (e) measurements inside the LCIS anechoic chamber. CP: co-polarization; XP: cross polarization; VNA: vector network analyzer.
Figure 8(a) and (b) graphs the antenna simulated and measured radiation patterns in both the H- and E-planes at 2.45 GHz. These measurements represent the radiation pattern for the in-air case. The simulated and measured patterns show good agreement, with slight differences related to the measurement setup. The radiation pattern’s main direction was toward the front half space facing the operator, covering almost the half space with a beamwidth of 279 and 60º on the horizontal and vertical planes, respectively. An additional measurement with the antenna positioned at the center of the operator’s chest was also elaborated. Figure 8(c) indicates a normalized on-body (chest) radiation pattern measurement at the horizontal plane, where we can see a difference between the forward and the backward radiation of almost 40 dB.
The maximum measured gain and simulated radiation efficiency were 1.4 dBi and 48.35%. The efficiency decrease was expected due to the resin’s thick superstrate. Deposing a 0.03-loss-tangent resin will inevitably have this negative impact, especially when deposed directly on the radiating copper part of the antenna. The Kapton substrate’s contribution to the decreasing radiation efficiency was less relevant thanks to its low loss tangent (only 0.005). One significant supplementary contributor to the low efficiency was the short distance separating the folded dipole and the backing reflector. This distance was only 6 mm, corresponding to ${\lambda} / {20}$ in the air at 2.45 GHz and almost ${\lambda} / {10}$ within the resin. A high-efficiency value requires a good ${\lambda} / {4}$, which was not reasonable for our case because it would have made the antenna bulkier and heavier and thus less practical for wearable applications. Nevertheless, including the resin in the antenna design was beneficial. This increased the forward radiation and the antenna directivity by 44.9% (from 3.62 to 5.23 dBi).
The purpose of the experimental setup shown in Figure 9(a) was to evaluate every degradation of the antenna reflection coefficient caused by variations in the surrounding temperature and humidity. By enabling and disabling the heat source, we could control the flow of hot steamy air being passed through the channel and thus the temperature and humidity inside the closed container hosting the antenna. A vector network analyzer (VNA) and a temperature–humidity sensor were used to monitor variations in the reflection coefficient, temperature, and humidity values, respectively. We measured temperature and humidity variations at intervals between 22 and 51 ºC and 61% and 100%, respectively. Variation of a single parameter was unfeasible since the parameters were interdependent. The tested intervals represented temperature and humidity variations in different geographical regions (climate zones). These intervals are generally present in most realistic scenarios, from ambient conditions (22 ºC/61%) to extreme conditions (51 ºC/100%) [21]. Figure 9(b) reveals that the antenna had a stable measured frequency response with minor variations in the reflection coefficient. This remained true even for the extreme case of 51 ºC and 100% humidity. The antenna’s external structure was wet at this stage, but this did not affect its performance at the band of interest.
Figure 9. The environmental conditions test: the (a) measurement setup and (b) measured reflection coefficient magnitude.
Figure 10(a) and (b) presents a setup that has been mounted to measure the impact of mechanical strains and, thus, the antenna’s geometrical deformation on its matching performance. Deformations caused by regular daily use are usually very slow and occur at negligible rates. However, in some cases, applied forces could lead to considerable and fast deformations, causing significant variations in antenna response. We tested deformations by applying a separate force along the x- and z-axes. Deformations along the y-axis were not tested since the axis is parallel to the structure’s most significant dimension (L) and, thus, highly resistant to deformations [Figure 10(a) and (b)].
Figure 10. The physical deformation test: the (a) measurement setup, (b) measured 10-dB impedance BW variation (Δ BW−10dB), (c) measured resonance frequency shifting (Δ fr), and (d) measured reflection coefficient magnitude degradation (Δ |S11|) AUT: antenna under test.
Applying a force ${\vec{F}}$ perpendicular to the surface of the antenna under test made the antenna structure flex due to the presence of two fixed supports (green). The force was applied using a rigid cylindrical tube (yellow) parallel to the antenna surface to guarantee a uniform distribution of the strain. By increasing ${\vec{F}}$, the distance ${\Delta}{H}$ was monitored using a dial indicator from Mitutoyo, which has a sensitivity higher than micrometers. During the deformation of the antenna structure, we monitored variations in three characteristics: the BW ${(}{\Delta}{BW}_{{-}{10}{\text{dB}}}{)}$, resonance frequency shifting ${(}{\Delta}{f}_{r}{)}$, and reflection coefficient magnitude at the resonance frequency ${(}{\Delta}{\left\vert{S}_{11}\right\vert}{)}$. The maximal deformation along the x- and z-axes was limited to 1.3 and 4 mm, respectively. These limits were fixed with respect to the geometrical shape of the antenna, which made the bending operation along the z-axis easier than that along the x-axis (since ${H}_{x}\,{\gg}\,{H}_{z}$, then ${\Delta}{H}_{{X}{\max}}\,{\ll}\,{\Delta}{H}_{{Z}{\max}}{)}$. This guaranteed that the antenna bending remained reversible, thus preventing destructive deformations.
As illustrated in Figure 10(c), the BW decreased gradually due to the deformation; however, this deterioration remained very small (<3 MHz), representing 0.64% of the measured BW. The resonance frequency was more sensitive to deformations along both axes. Its maximum variation was 13 MHz (–8.7 and +4.3 MHz) for deformations along the x-axis and 5.6 MHz (–0.8 to +4.8 MHz) for deformations along the z-axis [Figure 10(d)]. In both cases, the frequency response shifting remained low compared to the operational BW. No significant deterioration in the reflection coefficient at the resonance frequency ${\Delta}{\left\vert{S}_{11}\right\vert}$ was observed. As indicated in Figure 10(e), the measured deterioration was less than 1 dB for both deformations.
The antenna was designed to be integrated on cloths near the chest location. However, it is possible to deploy it elsewhere on the body. In this brief study, we quantified the impact of the antenna’s on-body location on its resonance frequency. Figure 11(a) provides the resonance frequency variation for different locations:
Figure 11. The impact of the on-body location on the antenna response: the (a) tested on-body locations and (b) frequency shifting ${\left\vert{\Delta}{f}_{r}\right\vert}$.
Figure 12. The path loss (S21) measurements for on-body communication (on chest).
Figure 11(b) shows that the resonance frequency remains reasonably steady for the tested on-body locations ${(}{\left\vert{\Delta}{f}_{r}\right\vert}\,{<}\,{5}{\text{ MHz}}{)}$. This steadiness can be explained by the high isolation of the antenna design, which included a sufficiently thick superstrate. The reflector also made the antenna’s on-body location less relevant in determining the antenna’s electrical properties. The antenna performance, mainly the radiation pattern, was sensitive only to the device’s orientation. This factor must be taken into consideration to guarantee reliable communication with the counterpart receiving antenna. The reflection coefficient was almost the same (a variation of <1 dB) for all the tested on-body locations. An on-body measurement that corresponds to the on-chest position is available in Figure 12. This measurement was performed inside the anechoic chamber at a distance of 3 m, using an SA-571 double-ridged horn antenna.
As described in Table 2, our design’s performance was comparable to several other recent works from the literature with comparable sizes [3], [6], [7]. It exhibited a wide BW covering the ISM band, with a relatively low gain view of its quasi-omnidirectional radiation pattern. However, when considering the robustness and isolation of every item in tangency/proximity, we can see the clear superiority of the proposed design, due to its high immunity to harsh environmental conditions. The proposed antenna’s high isolation qualifies it as a good candidate for many applications, such as in military and firefighting jackets.
Table 2. The advantages of the proposed design in comparison with other antennas reported in the literature.
As illustrated in Figure 13, we can see that the proposed antenna possesses a high-enough quality factor with a moderate gain value. The gain drop could be explained as follows:
Figure 13. The gain and quality (Q) factor characteristics of the designed antenna compared to other designs from the literature.
Another factor that affects the gain is the antenna size. Increasing the antenna size would be a way to increase its gain. However, this would drastically affect the ergonomics, as in [22].
Many other smaller designs do not always perform better. It is not helpful to compare our design to others in which only the radiation performances (the BW, reflection coefficient, gain, and others) were considered. Usually, antenna designs are proofs of concept and do not fulfill the constraints of durable operational usage. Compared to other similar studies, the distinct feature of this study was the importance given to the shielding factor and the robustness of the design. This resulted in an antenna designed to be operational in extremely harsh environments, without drastic degradation in performance. Clearly, including extra parameters would entail subjecting the antenna to additional tests and experimental checkpoints in addition to the classical radiation and matching performance. These extra tests and design steps will guarantee that the antenna fulfills its purpose while operating in harsh environments.
The development of wireless communications using specialized jackets and clothing is a topic of great importance, as it can allow personnel (e.g., soldiers and firefighters) to operate effectively in very harsh conditions while remaining safe and connected with headquarters. On-clothing antennas are vital elements in wireless communication systems. In the real world, designing an antenna that can be integrated into clothes is a challenge, as several properties and constraints must be satisfied, and some of these are contradictory (e.g., the antenna’s compactness and gain). In addition to the necessary ergonomic character of the antenna, it should exhibit a reduced SAR and high radiation efficiency.
In this study, a folded planar dipole antenna was designed and customized to operate close to the human body. Backward radiation was reduced to protect the human body from unwanted EM emissions and increase the antenna’s operating range. The proposed design constitutes a qualified candidate for this type of application in terms of its balanced radiation and shielding qualities. Although the shielding provoked a decrease in the radiation efficiency, this was inevitable due to the resin’s dielectric losses. The presence of a reflector reduced the SAR considerably and promoted near-hemispherical coverage; however, this also caused the efficiency to drop. To overcome this limitation, the use of low-loss resins is recommended.
In our figure of merit, in addition to radiation properties, we also considered enhancing the ergonomics and the robustness of the design. To do so, the antenna was designed to be entirely covered by resin, which resulted in an acceptable tradeoff between rigidity and radiation properties. The electrical and mechanical properties of the shielding polyethylene-based resin were carefully chosen to ensure a balanced tradeoff between rigidity and robustness. Tests demonstrated that the designed antenna has high immunity to mechanical strains and deformations. We also demonstrated that the antenna has high robustness to weather conditions, particularly rain and fog, making it capable of enduring extreme weather conditions.
While the proposed design allows a balanced tradeoff among performance, rigidity, and robustness, it can be evolved to use, instead of employing a simple reflector, the design of HISs, which is a suitable way to limit radiation losses. The antenna size also remains a factor that needs improvement since lighter and smaller wearable antennas are usually more ergonomic. These potential evolutions will constitute the objectives of future studies.
This work is part of the project Gilet Intégrant ANTennE, a partnership between Safran and the LCIS. We thank the company Éolane for its contribution to the prototyping part of this work. The authors would like to thank Tsitoha Andriamiharivolamena, who initially contributed to this work during his Ph.D. preparation at Grenoble Institute of Engineering and Management/LCIS.
Fateh Benmahmoud (benmahmoud.fateh@gmail.com) is a researcher at Université de Grenoble Alpes, Laboratoire d’intégration et conception des systèmes (LCIS), 20902 Valence, France. He works on designing and integrating antennas in harsh environments, namely, fully metallic 3D antennas, robust wearable antennas, and properties of artificial surfaces. He is a Member of IEEE.
Pierre Lemaitre-Auger (pierre.lemaitre-auger@lcis.grenoble-inp.fr) is an associate professor at Université de Grenoble Alpes, Grenoble-INP, Laboratoire d’intégration et conception des systèmes (LCIS), 20902 Valence, France. His research activities include electromagnetic wave generation and propagation: antennas, array antennas, frequency-selective surfaces, and localized waves.
Smail Tedjini (smail.tedjini@lcis.grenoble-inp.fr) is a full professor at Université de Grenoble Alpes, Institue of Engineering and Managment, Laboratoire d’intégration et conception des systèmes (LCIS), 20902 Valence, France. He founded the LCIS lab and served as its director. He has more than 35 year experience in education, research and management of university affairs. His research interests include the design and integration of RFID systems, antennas, and RF power harvesting applications. He is a Senior Member of IEEE.
[1] M. Adler, R. Bieringer, T. Schauber, and J. Gnther, Materials for Stretchable Electronics Compliant With Printed Circuit Board Fabrication. Hoboken, NJ, USA: Wiley, ch. 7, pp. 161–185. [Online] . Available: https://onlinelibrary.wiley.com/doi/abs/10.1002/9783527646982.ch7
[2] J. Trajkovikj, J.-F. Zrcher, and A. Skrivervik, “Soft and flexible UHF antennas for W-BAN applications,” in Proc. 2014 IEEE Antennas Propag. Soc. Int. Symp. (APSURSI), pp. 305–306, doi: 10.1109/APS.2014.6904484.
[3] J. Trajkovikj, J.-F. Zrcher, and A. K. Skrivervik, “PDMS, a robust casing for flexible W-BAN antennas [EurAAP corner] ,” IEEE Antennas Propag. Mag., vol. 55, no. 5, pp. 287–297, Oct. 2013, doi: 10.1109/MAP.2013.6735538.
[4] D. D. Cara, J. Trajkovikj, R. Torres-Snchez, J.-F. Zrcher, and A. K. Skrivervik, “A low profile UWB antenna for wearable applications: The tripod kettle antenna (TKA),” in Proc. 2013 7th Eur. Conf. Antennas Propag. (EuCAP), pp. 3257–3260.
[5] R. Joshi, C. Constantinides, S. Podilchak, M. Ramli, H. Lago, and P. Soh, “Robust and compact PDMS antennas for search and rescue operations and emergency communications,” in Proc. 12th Eur. Conf. Antennas Propag. (EuCAP), pp. 1–5, doi: 10.1049/cp.2018.0975.
[6] R. B. B. Simorangkir, S. Feng, A. Sadat Sayem, K. P. Esselle, and Y. Yang, “PDMS-embedded conductive fabric: A simple solution for fabricating PDMS-based wearable antennas with robust performance,” in Proc. 2018 12th Int. Symp. Med. Inf. Commun. Technol. (ISMICT), pp. 1–3, doi: 10.1109/ISMICT.2018.8573690.
[7] A. S. Za’aba, S. N. Ibrahim, N. F. Abdul Malek, and A. M. Ramly, “Development of wearable patch antenna for medical application,” in Proc. 2017 IEEE Regional Symp. Micro Nanoelectron. (RSM), pp. 260–263, doi: 10.1109/RSM.2017.8069166.
[8] A. Y. Ashyap et al., “Flexible antenna with HIS based on PDMS substrate for WBAN applications,” in Proc. 2018 IEEE Int. RF Microw. Conf. (RFM), pp. 69–72, doi: 10.1109/RFM.2018.8846555.
[9] R. B. V. B. Simorangkir, Y. Yang, L. Matekovits, and K. P. Esselle, “Dual-band dual-mode textile antenna on PDMS substrate for body-centric communications,” IEEE Antennas Wireless Propag. Lett., vol. 16, pp. 677–680, 2017, doi: 10.1109/LAWP.2016.2598729.
[10] R. B. V. B. Simorangkir, D. Le, T. Bjrninen, A. S. M. Sayem, M. Zhadobov, and R. Sauleau, “Washing durability of PDMS-conductive fabric composite: Realizing washable UHF RFID tags,” IEEE Antennas Wireless Propag. Lett., vol. 18, no. 12, pp. 2572–2576, Dec. 2019, doi: 10.1109/LAWP.2019.2943535.
[11] A. S. M. Alqadami, M. F. Jamlos, P. J. Soh, and G. A. E. Vandenbosch, “Assessment of PDMS technology in a MIMO antenna array,” IEEE Antennas Wireless Propag. Lett., vol. 15, pp. 1939–1942, 2016, doi: 10.1109/LAWP.2015.2513960.
[12] W. L. Stutzman and G. A. Thiele, Antenna Theory and Design. New York, NY, USA: Wiley, 1998.
[13] A. Omar, Y. Chow, L. Roy, and M. Stubbs, “Effects of air-bridges and mitering on coplanar waveguide 90 degrees bends: Theory and experiment,” in Proc. 1993 IEEE MTT-S Int. Microw. Symp. Dig., pp. 823–826, doi: 10.1109/MWSYM.1993.276748.
[14] A. Bessemoulin, C. Algani, G. Alquié, and V. F. Hanna, “A simple airbridge analytical model in coplanar waveguides for MMIC applications,” Microw. Opt. Technol. Lett., vol. 17, no. 4, pp. 265–267, Dec. 1998, doi: 10.1002/(SICI)1098-2760(199803)17:4<265::AID-MOP14>3.0.CO;2-4.
[15] T. M. Weller, R. M. Henderson, K. J. Herrick, S. Robertson, R. Kihm, and L. P. Katehi, “Three-dimensional high-frequency distribution networks. I. Optimization of CPW discontinuities,” IEEE Trans. Microw. Theory Techn., vol. 48, no. 10, pp. 1635–1642, Oct. 2000, doi: 10.1109/22.873890.
[16] P. M. Watson and K. C. Gupta, “Design and optimization of CPW circuits using EM-ANN models for CPW components,” IEEE Trans. Microw. Theory Techn., vol. 45, no. 12, pp. 2515–2523, Dec. 1997, doi: 10.1109/22.643868.
[17] R. E. Fields, “Evaluating compliance with FCC guidelines for human exposure to radiofrequency electromagnetic fields,” OET Bull., vol. 65, no. 10, pp. 1–57, Nov. 1997.
[18] The International Commission on Non-Ionizing Radiation Protection, “ICNIRP statement on the guidelines for limiting exposure to time-varying electric, magnetic, and electromagnetic fields (up to 300 GHz),” Health Phys., vol. 97, no. 3, pp. 257–258, Sep. 2009, doi: 10.1097/HP.0b013e3181aff9db.
[19] “N1501A dielectric probe kit 10 MHz to 50 GHz: Technical overview,” Keysight Technologies, Santa Rosa, CA, USA, 2015. [Online] . Available: https://www.keysight.com/us/en/assets/7018-04631/technical-overviews/5992-0264.pdf
[20] MuEpsln, LLC. Accessed: Oct. 12, 2022. [Online] . Available: https://muepsln.com/
[21] “U.S. winter outlook: Warmer, drier South with ongoing La Niña.” Climate.gov. Accessed: Oct. 12, 2022. [Online] . Available: https://www.climate.gov/
[22] L. Zhou, S. Fang, and X. Jia, “Dual-band and dual-polarised circular patch textile antenna for on-/off-body WBAN applications,” IET Microw., Antennas Propag., vol. 14, no. 7, pp. 643–648, Jun. 2020, doi: 10.1049/iet-map.2019.1073.
[23] R. B. V. B. Simorangkir, Y. Yang, R. M. Hashmi, T. Bjrninen, K. P. Esselle, and L. Ukkonen, “Polydimethylsiloxane-embedded conductive fabric: Characterization and application for realization of robust passive and active flexible wearable antennas,” IEEE Access, vol. 6, pp. 48,102–48,112, Aug. 2018, doi: 10.1109/ACCESS.2018.2867696.
[24] R. A. Sheikh et al., “Design of a compact meander-based antenna for search and rescue using PDMS,” in Proc. 2019 IEEE Conf. Antenna Meas. Appl. (CAMA), pp. 258–260, doi: 10.1109/CAMA47423.2019.8959691.
Digital Object Identifier 10.1109/MAP.2022.3216783