Paper-Based Flexible Antenna for Wearable Telemedicine Applications at 2.4 GHz ISM Band

This paper demonstrates the performance of a potential design of a paper substrate-based flexible antenna for intrabody telemedicine systems in the 2.4 GHz industrial, scientific, and medical radio (ISM) bands. The antenna was fabricated using 0.54 mm thick flexible photo paper and 0.03 mm copper strips as radiating elements. Design and performance analyses of the antenna were performed using Computer Simulation Technology (CST) Microwave Studio software. The antenna performances were investigated based on the reflection coefficient in normal and bent conditions. The total dimensions of the proposed antenna are 40 × 35 × 0.6 mm3. The antenna operates at 2.33–2.53 GHz in the normal condition. More than an 8% fractional bandwidth is expressed by the antenna. Computational analysis was performed at different flexible curvatures by bending the antenna. The minimum fractional bandwidth deviation is 5.04% and the maximum is 24.97%. Moreover, it was mounted on a homogeneous phantom muscle and a four-layer human tissue phantom. Up to a 70% radiation efficiency with a 2 dB gain was achieved by the antenna. Finally, the performance of the antenna with a homogeneous phantom muscle was measured and found reliable for wearable telemedicine applications.


Introduction
Wearable communication technologies can offer promising solutions in biomedical, consumer electronics, military, and smart home applications. Zimmerman was the very first to discuss body-centric wireless communication systems [1]. The operating frequency of his prototype device was 330 KHz. Since then, body-centric communication systems' operating frequencies have undergone various changes. In addition to the ultrahigh frequency (UHF) bands that are used for intrabody telemedicine applications, 2.4 GHz is used extensively as an industrial, scientific, and medical (ISM) band [2,3]. Low-profile and lightweight antennas are usually preferred for such applications. However, the height or thickness of the antenna can occasionally be an issue for on-body applications. Moreover, thick antennas lack flexibility; even though they might have flexible features, they can become fragile. To address this problem, flexible paper substrate-based antennas can be utilized since paper is light in weight but has a suitable flexibility limit, depending on the thickness of the substrate. According to the literature, paper-based [4], inkjet-printed antennas have been used for wearable applications [5,6].

Design and Methodology
The design and computational analysis of the antenna were performed using Computer Simulation Technology (CST) Microwave Studio software. The primary design process of the antenna started with a basic microstrip-fed antenna, and modification of the radiating patch led to the proposed design. Three designs were simulated to obtain the final design, as depicted in Figure 1. In the proposed design, the length of the microstrip feed is 15 mm and the length and width of the substrate paper material are 40 mm and 35 mm, respectively, providing the antenna with a low profile. Further, 0.54 mm thick photo paper was used as the substrate, and 0.03 mm thick copper strips were used as radiating elements. A partial grounding technique was utilized to obtain a satisfactory operating bandwidth. The corresponding reflection coefficients of the computed paper-based antennas are shown in Figure 2, and the final design of the proposed antenna is illustrated in Figure 3. heat. In addition, textile antennas can be somewhat flexible [7], but they can also be quite costly, depending on the type of textile. In addition to these types of antennas, other antennas can be found in the literature on wearable communication: for instance, an artificial magnetic conductor (AMC)backed wearable antenna was described in [8,9]. The antenna used two layers of AMC and Styrofoam between the radiating element and the ground. These layers gave the antenna a total thickness of 9.5 mm, which is quite high. This paper proposes a new design concept of a paper-based antenna to overcome the shortcomings of other paper-based antennas. The proposed design exploits the lightweight and flexible properties of paper for its application in telemedicine and wearable smart electronic devices for health monitoring. The proposed antenna operates at 2.33-2.53 GHz and resonates at 2.43 GHz. Photo paper was used to construct the 0.54 mm thick substrate, which provides a balance between flexibility and stability in the antenna. A slotted microstrip feed line was used to excite the antenna. The total electrical dimensions of the antenna are 0.31λ × 0.27λ × 0.004λ. First, the antenna was simulated and measured in the normal straight condition. Then, the antenna was measured in the bent condition. In both conditions, the performance is stable in the desired 2.4 GHz ISM band. Then, to validate the flexibility performance of the antenna, computation analysis was performed while bending the antenna into side-by-side and upside-down orientations at different curvature radii. Moreover, a homogeneous phantom muscle and a four-layered human tissue model were mounted with the antenna for computational performance analysis. Finally, the antenna was mounted on a homogeneous phantom muscle, and the reflection coefficient and radiation patterns were studied.

Design and Methodology
The design and computational analysis of the antenna were performed using Computer Simulation Technology (CST) Microwave Studio software. The primary design process of the antenna started with a basic microstrip-fed antenna, and modification of the radiating patch led to the proposed design. Three designs were simulated to obtain the final design, as depicted in Figure 1. In the proposed design, the length of the microstrip feed is 15 mm and the length and width of the substrate paper material are 40 mm and 35 mm, respectively, providing the antenna with a low profile. Further, 0.54 mm thick photo paper was used as the substrate, and 0.03 mm thick copper strips were used as radiating elements. A partial grounding technique was utilized to obtain a satisfactory operating bandwidth. The corresponding reflection coefficients of the computed paperbased antennas are shown in Figure 2, and the final design of the proposed antenna is illustrated in Figure 3.   The length of the upper copper strip has a great impact on the resonant frequency. A parametric study was conducted to determine the optimum length of the upper radiating element. The results are shown in Figure 4. Increasing the length of 'a' results in a lower-frequency resonance. Since the length of 'a' has a great impact on the resonant frequency, varying the length from 20 to 36 mm shifted the resonant frequency from 2.2 to 2.42 GHz, approximately. The other parameters were obtained through a parametric study, and the optimized design parameters are listed in Table 1.    The length of the upper copper strip has a great impact on the resonant frequency. A parametric study was conducted to determine the optimum length of the upper radiating element. The results are shown in Figure 4. Increasing the length of 'a' results in a lower-frequency resonance. Since the length of 'a' has a great impact on the resonant frequency, varying the length from 20 to 36 mm shifted the resonant frequency from 2.2 to 2.42 GHz, approximately. The other parameters were obtained through a parametric study, and the optimized design parameters are listed in Table 1.  The length of the upper copper strip has a great impact on the resonant frequency. A parametric study was conducted to determine the optimum length of the upper radiating element. The results are shown in Figure 4. Increasing the length of 'a' results in a lower-frequency resonance. Since the length of 'a' has a great impact on the resonant frequency, varying the length from 20 to 36 mm shifted the resonant frequency from 2.2 to 2.42 GHz, approximately. The other parameters were obtained through a parametric study, and the optimized design parameters are listed in Table 1.  The length of the upper copper strip has a great impact on the resonant frequency. A parametric study was conducted to determine the optimum length of the upper radiating element. The results are shown in Figure 4. Increasing the length of 'a' results in a lower-frequency resonance. Since the length of 'a' has a great impact on the resonant frequency, varying the length from 20 to 36 mm shifted the resonant frequency from 2.2 to 2.42 GHz, approximately. The other parameters were obtained through a parametric study, and the optimized design parameters are listed in Table 1.   The prototype was fabricated with commercially available photo paper and a metallic copper strip. The reflection coefficient of the proposed antenna was measured using a performance network analyzer (N5227A) that was calibrated using the Electronic Calibration Module (N4694-60001). Figure 5 represents the fabricated prototype and reflection coefficient measurement of the proposed antenna. The antenna far-field characteristics were investigated using a Satimo near-field measurement system.
Moreover, a homogeneous phantom muscle was also made to verify the simulated reflection coefficient of the proposed antenna with the phantom muscle, as shown in Figure 6. The homogeneous phantom muscle was fabricated using the methodology described in [10]. The dielectric constant of the fabricated muscle phantom was measured in the Centre for Communication Engineering Studies Lab, Universiti Teknologi MARA (UiTM), Malaysia. The dielectric characteristics of the phantom can be seen in Figure 6b,c.  The prototype was fabricated with commercially available photo paper and a metallic copper strip. The reflection coefficient of the proposed antenna was measured using a performance network analyzer (N5227A) that was calibrated using the Electronic Calibration Module (N4694-60001). Figure 5 represents the fabricated prototype and reflection coefficient measurement of the proposed antenna. The antenna far-field characteristics were investigated using a Satimo near-field measurement system.
Moreover, a homogeneous phantom muscle was also made to verify the simulated reflection coefficient of the proposed antenna with the phantom muscle, as shown in Figure 6. The homogeneous phantom muscle was fabricated using the methodology described in [10]. The dielectric constant of the fabricated muscle phantom was measured in the Centre for Communication Engineering Studies Lab, Universiti Teknologi MARA (UiTM), Malaysia. The dielectric characteristics of the phantom can be seen in Figure 6b,c.   The prototype was fabricated with commercially available photo paper and a metallic copper strip. The reflection coefficient of the proposed antenna was measured using a performance network analyzer (N5227A) that was calibrated using the Electronic Calibration Module (N4694-60001). Figure 5 represents the fabricated prototype and reflection coefficient measurement of the proposed antenna. The antenna far-field characteristics were investigated using a Satimo near-field measurement system.
Moreover, a homogeneous phantom muscle was also made to verify the simulated reflection coefficient of the proposed antenna with the phantom muscle, as shown in Figure 6. The homogeneous phantom muscle was fabricated using the methodology described in [10]. The dielectric constant of the fabricated muscle phantom was measured in the Centre for Communication Engineering Studies Lab, Universiti Teknologi MARA (UiTM), Malaysia. The dielectric characteristics of the phantom can be seen in Figure 6b,c.

Results and Discussion
The simulated and measured reflection coefficients of the proposed antenna were analyzed, and the results are illustrated in Figure 7. There is a slight frequency shift in the measured results, which might be due to fabrication and measurement tolerances. In the simulated result, the resonance frequency is 2.43 GHz, and, in the measured result, the resonant frequency is found to be 2.47, which seems to be in good harmony with the computational result.

Results and Discussion
The simulated and measured reflection coefficients of the proposed antenna were analyzed, and the results are illustrated in Figure 7. There is a slight frequency shift in the measured results, which might be due to fabrication and measurement tolerances. In the simulated result, the resonance frequency is 2.43 GHz, and, in the measured result, the resonant frequency is found to be 2.47, which seems to be in good harmony with the computational result. The reflection coefficient performance of the proposed antenna in the normal and bent conditions was also investigated. It turns out that the antenna shows a stable performance in the bent condition, which authenticates its flexible operation, depicted in Figure 8. There is a slight deviation in the resonant frequency in the bent condition. Bending the antenna yields a resonant frequency that is lower than that in the normal condition by approximately 20 MHz. The antenna's far-field properties were investigated using a Satimo near-field measurement system. In Figure 9, the normalized radiation patterns are depicted for Phi 0 and Phi 90 degrees. One can observe that the antenna shows bidirectional radiation patterns for both planes. The reflection coefficient performance of the proposed antenna in the normal and bent conditions was also investigated. It turns out that the antenna shows a stable performance in the bent condition, which authenticates its flexible operation, depicted in Figure 8. There is a slight deviation in the resonant frequency in the bent condition. Bending the antenna yields a resonant frequency that is lower than that in the normal condition by approximately 20 MHz.

Results and Discussion
The simulated and measured reflection coefficients of the proposed antenna were analyzed, and the results are illustrated in Figure 7. There is a slight frequency shift in the measured results, which might be due to fabrication and measurement tolerances. In the simulated result, the resonance frequency is 2.43 GHz, and, in the measured result, the resonant frequency is found to be 2.47, which seems to be in good harmony with the computational result. The reflection coefficient performance of the proposed antenna in the normal and bent conditions was also investigated. It turns out that the antenna shows a stable performance in the bent condition, which authenticates its flexible operation, depicted in Figure 8. There is a slight deviation in the resonant frequency in the bent condition. Bending the antenna yields a resonant frequency that is lower than that in the normal condition by approximately 20 MHz. The antenna's far-field properties were investigated using a Satimo near-field measurement system. In Figure 9, the normalized radiation patterns are depicted for Phi 0 and Phi 90 degrees. One can observe that the antenna shows bidirectional radiation patterns for both planes. The antenna's far-field properties were investigated using a Satimo near-field measurement system. In Figure 9, the normalized radiation patterns are depicted for Phi 0 and Phi 90 degrees. One can observe that the antenna shows bidirectional radiation patterns for both planes.

Bending Deformation Study
The performance of the antenna in the side-by-side bending condition at different bending radii was also investigated. Figure 10 depicts the perspective view of the antenna at different bending radii, and Figure 11 shows the reflection coefficients at the corresponding values of the radii. It is evident that the reflection coefficient at the desired band can be found at all the radii when the prototype is bent in a side-by-side orientation. During the side-by-side bending condition, 10   Then, the performance of the antenna was studied during the upside-down condition at different radii, as depicted in Figure 12. The resulting reflection coefficients are depicted in Figure 13.

Bending Deformation Study
The performance of the antenna in the side-by-side bending condition at different bending radii was also investigated. Figure 10 depicts the perspective view of the antenna at different bending radii, and Figure 11 shows the reflection coefficients at the corresponding values of the radii. It is evident that the reflection coefficient at the desired band can be found at all the radii when the prototype is bent in a side-by-side orientation. During the side-by-side bending condition, 10

Bending Deformation Study
The performance of the antenna in the side-by-side bending condition at different bending radii was also investigated. Figure 10 depicts the perspective view of the antenna at different bending radii, and Figure 11 shows the reflection coefficients at the corresponding values of the radii. It is evident that the reflection coefficient at the desired band can be found at all the radii when the prototype is bent in a side-by-side orientation. During the side-by-side bending condition, 10   Then, the performance of the antenna was studied during the upside-down condition at different radii, as depicted in Figure 12. The resulting reflection coefficients are depicted in Figure 13.

Bending Deformation Study
The performance of the antenna in the side-by-side bending condition at different bending radii was also investigated. Figure 10 depicts the perspective view of the antenna at different bending radii, and Figure 11 shows the reflection coefficients at the corresponding values of the radii. It is evident that the reflection coefficient at the desired band can be found at all the radii when the prototype is bent in a side-by-side orientation. During the side-by-side bending condition, 10   Then, the performance of the antenna was studied during the upside-down condition at different radii, as depicted in Figure 12. The resulting reflection coefficients are depicted in Figure 13. Then, the performance of the antenna was studied during the upside-down condition at different radii, as depicted in Figure 12. The resulting reflection coefficients are depicted in Figure 13. In this case, like in the side-by-side bending condition, the reflection coefficients remain consistent during the upside-down bending condition. Most of the resonant frequencies remain in good harmony at different radii. At a curvature radius of 10 mm, the reflection coefficient is −17.20 dB at 2.38 GHz, which is in the desired 2.30-2.48 GHz ISM band. When the radius is 35 mm, the operating band is found at 2.30-2.53 GHz. The minimum bandwidth of 180 MHz is found at a 10 mm curvature radius, and the maximum bandwidth of 230 MHz is found at a 35 mm curvature radius. ors 2018, 18, x 7 of his case, like in the side-by-side bending condition, the reflection coefficients remain consiste ing the upside-down bending condition. Most of the resonant frequencies remain in go mony at different radii. At a curvature radius of 10 mm, the reflection coefficient is −17.20 dB GHz, which is in the desired 2.30-2.48 GHz ISM band. When the radius is 35 mm, the operati d is found at 2.30-2.53 GHz. The minimum bandwidth of 180 MHz is found at a 10 mm curvatu ius, and the maximum bandwidth of 230 MHz is found at a 35 mm curvature radius.  From Figures 11 and 13, it can be seen that the antenna reflection coefficient remains adequate sistent with various bending radii, and the reflection coefficient can still be found within t ired band. However, a slight change in frequency can be observed for different bending radii. nge in resonant length occurs due to variation in the diameter of bending radii, which impacts t rating frequency [11]. Table 2 represents all the data related to the antenna's bendi formance. The minimum operating band is found to be 170 MHz with a fractional bandwidth In this case, like in the side-by-side bending condition, the reflection coefficients remain consistent during the upside-down bending condition. Most of the resonant frequencies remain in good harmony at different radii. At a curvature radius of 10 mm, the reflection coefficient is −17.20 dB at 2.38 GHz, which is in the desired 2.30-2.48 GHz ISM band. When the radius is 35 mm, the operating band is found at 2.30-2.53 GHz. The minimum bandwidth of 180 MHz is found at a 10 mm curvature radius, and the maximum bandwidth of 230 MHz is found at a 35 mm curvature radius.  From Figures 11 and 13, it can be seen that the antenna reflection coefficient remains adequately consistent with various bending radii, and the reflection coefficient can still be found within the desired band. However, a slight change in frequency can be observed for different bending radii. A change in resonant length occurs due to variation in the diameter of bending radii, which impacts the operating frequency [11]. Table 2 represents all the data related to the antenna's bending performance. The minimum operating band is found to be 170 MHz with a fractional bandwidth of 7.08%. The maximum operating bandwidth is 230 MHz with a fractional bandwidth of 9.58%.  From Figures 11 and 13, it can be seen that the antenna reflection coefficient remains adequately consistent with various bending radii, and the reflection coefficient can still be found within the desired band. However, a slight change in frequency can be observed for different bending radii. A change in resonant length occurs due to variation in the diameter of bending radii, which impacts the operating frequency [11]. Table 2 represents all the data related to the antenna's bending performance. The minimum operating band is found to be 170 MHz with a fractional bandwidth of 7.08%. The maximum operating bandwidth is 230 MHz with a fractional bandwidth of 9.58%.

Antenna Performance with Phantom
A homogeneous phantom muscle was mounted with the antenna to investigate the antenna's performance. The simulation setup is shown in Figure 14. The cylindrical phantom shown in Figure 14a was created in CST Microwave Studio using the muscle material available in the material library. This phantom muscle has a dielectric constant of 52. Then, the performance of the antenna attached to a human tissue model with four layers was investigated. The model consists of skin, fat, muscle, and bone, as shown in Figure 14b. The dielectric constants of the modeled skin, fat, muscle, and bone are approximately 37, 5.3, 52, and 11.4, respectively. The thicknesses of the skin, fat, muscle, and bone layers are 3, 7, 15, and 30 mm, respectively. Even though a partial ground plane is used in the antenna, the reflection coefficient remains in good agreement with the reflection coefficient of the antenna in free space.
The reflection coefficient performance of the antenna with the homogeneous phantom muscle and four-layered human tissue is illustrated in Figure 15. To analyze the directivity of the proposed antenna while mounted on the human tissue, Figure 16 represents the 3D directivity of the antenna. The main direction of propagation is outward from the antenna, which is good for using with the human body.
nd bone are approximately 37, 5.3, 52, and 11.4, respectively. The thicknesses of the skin, fat, muscle nd bone layers are 3, 7, 15, and 30 mm, respectively. Even though a partial ground plane is used in he antenna, the reflection coefficient remains in good agreement with the reflection coefficient of the ntenna in free space. The reflection coefficient performance of the antenna with the homogeneous phantom muscle nd four-layered human tissue is illustrated in Figure 15. To analyze the directivity of the proposed ntenna while mounted on the human tissue, Figure 16 represents the 3D directivity of the antenna   There were noticeable variations in the reflection coefficient between the measured and simulated results with the homogeneous phantom muscle, as shown in Figure 17. It is clear, however, that the bandwidths are in good agreement. One reason for the slight mismatch might be the difference in the dielectric properties of the computational and realistic phantoms. Some other factors could increase the losses of the coupling medium, and the measurement tolerance could impact the results. The radiation patterns of the antenna with the muscle are shown in Figure 18b. The radiation patterns with the phantom muscle remain stable to a satisfactory degree. The overall performance of the antenna seems to be in good harmony when in free space and with the phantom muscle.  There were noticeable variations in the reflection coefficient between the measured and simulated results with the homogeneous phantom muscle, as shown in Figure 17. It is clear, however, that the bandwidths are in good agreement. One reason for the slight mismatch might be the difference in the dielectric properties of the computational and realistic phantoms. Some other factors could increase the losses of the coupling medium, and the measurement tolerance could impact the results. The radiation patterns of the antenna with the muscle are shown in Figure 18b. The radiation patterns with the phantom muscle remain stable to a satisfactory degree. The overall performance of the antenna seems to be in good harmony when in free space and with the phantom muscle. . (a) Directivity and (b) front-to-back ratio of the proposed antenna with huma re noticeable variations in the reflection coefficient between the m lts with the homogeneous phantom muscle, as shown in Figure 17. It is c idths are in good agreement. One reason for the slight mismatch e dielectric properties of the computational and realistic phantoms. Som the losses of the coupling medium, and the measurement tolerance cou iation patterns of the antenna with the muscle are shown in Figure 18b. he phantom muscle remain stable to a satisfactory degree. The overall p ms to be in good harmony when in free space and with the phantom m ure 17. Reflection coefficients of the proposed antenna with the muscle phantom   Figure 19 presents the total efficiency and gain of the proposed paper-based flexible antenna. It achieves up to an 83% efficiency in free space and up to a 70% efficiency when mounted on the phantom at 2.45 GHz; it also obtains more than a 2 dB gain in the operating band. Achieving a good performance from a compact antenna is a challenge when the application is wearable communication. From Table 3, it can be seen that the proposed antenna is sufficiently compact with a minimum thickness of only 0.6 mm, which makes the antenna a perfect candidate for wearable telemedicine applications.  Figure 19 presents the total efficiency and gain of the proposed paper-based flexible antenna. It achieves up to an 83% efficiency in free space and up to a 70% efficiency when mounted on the phantom at 2.45 GHz; it also obtains more than a 2 dB gain in the operating band.  Figure 19 presents the total efficiency and gain of the proposed paper-based flexible antenna. It achieves up to an 83% efficiency in free space and up to a 70% efficiency when mounted on the phantom at 2.45 GHz; it also obtains more than a 2 dB gain in the operating band. Achieving a good performance from a compact antenna is a challenge when the application is wearable communication. From Table 3, it can be seen that the proposed antenna is sufficiently compact with a minimum thickness of only 0.6 mm, which makes the antenna a perfect candidate for wearable telemedicine applications. Achieving a good performance from a compact antenna is a challenge when the application is wearable communication. From Table 3, it can be seen that the proposed antenna is sufficiently compact with a minimum thickness of only 0.6 mm, which makes the antenna a perfect candidate for wearable telemedicine applications.

Conclusions
This paper illustrates a new design and performance analysis of a paper substrate-based flexible antenna. The prototype was fabricated for wearable telemedicine applications in the 2.4 GHz band. This paper presents the reflection coefficients for both normal and flexible conditions. The antenna is only 0.6 mm thick and can serve as a wearable system with a satisfactory flexibility range. In addition, the performance of the antenna with human tissue was investigated. The results for the reflection coefficient and radiation patterns are found to be satisfactory for wearable telemedicine applications.