An endoscope is a device placed in a body cavity or more generally in the digestive tract and allows the transmission of medical data in image/video format between the body and the outside world. Wireless endoscopes are increasingly used in comparison to wired endoscopes because they offer a certain comfort to the patient during the clinical examination. Since wireless endoscopes are usually ingested, they can explore several cavities of the digestive tract during their passage. This allows them to explore certain cavities such as the small intestine that are inaccessible to wire endoscopes. According to [3] , an endoscopic capsule includes an optical dome, a camera, one or more LEDs, a radio frequency transmitter, an antenna and a battery. The antennas of the endoscopes must be efficient (broadband, good adaptation) to allow the visualization of High Definition (HD) images or videos and the early diagnosis of certain diseases (Crohn’s disease, tumors, cancers...).

They are used in measurement and data transfer applications of gadgets worn on or near the body (arm, neck, clothing...). Wearable antennas can be used as an interface between implantable or endoscopic antennas and external receivers. They can also directly transmit data collected by the equipment which is an integral part of other sensors.

Due to their proximity to the body, the performance of the antenna is impaired, so the use of suitable antennas is essential. Artificial Magnetic Conductors (AMC) antennas can be used in the one proposed by Agarwal K, Guo Y and Salam B in [4] , where the authors used a double layer of AMC material to increase the radiation efficiency of a Yagi antenna to 78.97% and reduce its SAR to 0.714 W/Kg.

To minimize the effect of the body on the performance of wearable antennas, metamaterial can also be used. In [5] authors used 2 × 2 metamaterial cells to improve the performance of a U-shaped patch antenna. After placement of the metamaterial cells the gain of the antenna reached 7.5 dBi and the relative bandwidth was 14.5%. The weakness of this antenna is the complexity of the structure for mass production.

An implant is a device placed in a body tissue other than a cavity for the purpose of collecting and monitoring key physiological data from a patient and sending it to a remote node on or off the body. The most well-known implants today are intraocular implants, cardiac defibrillators (ICDs), neural brain monitoring implants, and Electromyogram (EMG) implants [6] . But increasingly, they are also being used to measure vital parameters in people exposed to major pathologies. Implants must be distinguished from endoscopes that are necessarily placed in cavities as described in the subsection dealing with endoscopy antennas.

Antennas used in biomedical telemetry must comply with many limitations. According to [7] , for an implantable antenna, these constrains concern size, resonance frequency, good biocompatibility, good impedance matching. According to [8] , in-body antennas must also have omnidirectional radiation to be able to communicate with sensors outside the body. Other criteria are also important, such as the miniaturization criteria of the antennas defined by Harold W [9] , the dimensions of the implants or medical capsules, the flexibility of the substrate and the radiating element, the optimization of the power consumption, the adaptation of the bandwidth to the video transmission and above all the relatively low frequency band defined for the Radio medical communications (Med Radio).

Numerous studies have been conducted on the propagation inside the body and the antennas used. It has been found that satisfying one or more of the constraints often degrades the performance of the other specifications. For example, in paper [10] , Ala Alemaryeen proposed a miniature of 18 × 10 mm² folded shape antenna for implants of various tissue characteristics. Simulated in an 80 × 80 × 80 mm³ Phantom, for a reflection coefficient of -10 dB, the antenna exhibits good impedance matching (−30 dB) and a bandwidth of 532 MHz. In addition, the maximum gain of the antenna is −28 dBi, which seems very low for efficient long-range transmission. Furthermore, the author does not specify the size of the implant around which the antenna is wrapped, nor the adaptation of the antenna in the presence of electronic components inside the implant. Different frequency bands have been regulated for biomedical transmission. In [2] [10] these bands have been listed according to the following classification:

Different biocompatible materials are used as substrate or superstrate in the design of antennas radiating into the human body. Table 1 hereafter illustrates the main substances used. As most of the tissues are dielectric supers, their permeability is very close to that of air. This leads us not to specify this characteristic in this table. Complex permittivity’s are measured at frequencies ranging from 100 MHz to 1 GHz, so there is little variation in these values for frequencies near this range.

Most of the materials used are polymers. The radiation of an antenna inside a tissue depends on the characteristics of the tissue. The characteristics vary according to the type (stomach, intestine, muscle...), the age, the gender of the individual and especially the location of the tissue (skull, abdomen, trunk, legs...). The tissues have high permittivity and low conductivity. In the following Table 2 extracted from the survey [2] we can find the average characteristics of the main tissues.

¹Liquid Crystalline Polymer, ²Poly Di Methyl Siloxane; ${\epsilon }_r$ and tan $\delta$ represent the relative permittivity and loss tangent, respectively.

To design an antenna that will radiate with minimal loss and acceptable gain, it is necessary to know the characteristics of the tissue in which the antenna will be inserted.

There are different methods to miniaturize and adapt biomedical telemetry antennas. The most well-known techniques are among others the winding of the radiating strand, the creation of slots, the insertion of electronic components and the use of metamaterials. In [15] , authors proposed a broadband capsular antenna for biotelemetry applications. The antenna with dimensions of 12 × 6 × 0.035 mm3 is an E-shaped meander etched on a capsule of 11 mm diameter and 24 mm length. The biocompatible Ultem substrate used has a relative permittivity of 3.15 and a conductivity of 0.0013 S/m. Implanted in a body, for a return loss limited to −10 dB, the authors obtained a bandwidth of 7.31 GHz and gains of −25.23 dBi in the human model of three layers (80 mm of muscle, 8 mm of fat, 2 mm of skin). The weakness of this antenna lies in the drop of the antenna gain in the presence of electronic components, particularly the battery.

The antenna developed by R. Alrawashdeh et al. in [18] inspires the shape of the suggested antenna but its shape has been revised to optimize its operation.

In the 1st step of the design, we unified the width of the radiating strands at 1 mm and replaced the vertical elements with circular arc elements with a diameter of 2 mm. By integrating the circular arcs, we aim to achieve harmonious radiation and lengthen the electron path. In fact, the gradual 180˚ turn of the electrical charges when changing direction eliminates unwanted radiation and lowers the resonant frequency.

In the second phase of antenna design, 4 arcs were symmetrically integrated into the upper and lower horizontal strands. The arcs allow the resonant frequency to drop to a much lower frequency and balance the current distribution.

Figure 1 shows the geometry of the antenna when extended and when wrapped around an implant of 3.5 mm radius and 10 mm height [18] .

The length (L) of the lower strand is 20 mm and that of the upper strands (L₁) and (L₂) is 3.37 mm. The diameter (D₁) and (D₂) of the upper arcs is 2.26 mm. The gap (E) between the central arcs has been dimensioned at 1.6 mm and the horizontal sections (S₁), (S₂), (S₃) have common lengths of 2.66 mm. The height (H) of the implant is 10 mm.

Electromagnetic simulation software was used to perform the simulations. Figure 2 and Figure 3 show the simulation results of the new antenna structure in the muscle characterized by the paper [18] . The muscle has a permittivity and dielectric constant equal to ε_rmuscle = 57 and σ = 0.79 s/m respectively. The dimensions of the human phantom were also kept at 180 × 100 × 50 mm³. We have chosen the elliptical cylinder shape of the Phantom to better approximate the real shape of most human muscles.

As shown in Figure 2 , the antenna has two modes of resonance which correspond respectively to bandwidths of 203.7 MHz and 673.7 MHz. The first resonance occurring at the frequency of 304.6 MHz extends from 224.3 MHz to 428 MHz and the second resonance that occurs at the frequency of 1010.8 MHz covers the frequency range from 576.3 MHz to 1200 MHz. The first resonance mode has therefore moved to low frequencies. The integration of the circles has also improved the adaptation. At the first resonance, the simulated reflection coefficient is −22.56 dB and at the second resonance, it is −25.20 dB. Figure 3 also shows that the maximum gain achieved has increased from −28.4 dBi to −26 dBi. Furthermore, the radiation is mainly in the plane (xoz) orthogonal to the antenna axis. However, it is obvious that neither bandwidth covers the 400 MHz bands (401 – 457 MHz) dedicated to medical communication services.

Table 3 . summarizes the comparative study of the performance of the initial antenna with that of the antenna integrating the arcs.

The drop of the first resonance mode to 304.6 MHz after arcs integrating gives the possibility to reduce the dimensions of the antenna.

We wanted to miniaturize the antenna so that once implanted in the body the bandwidth covers all biomedical frequencies up to 1 GHz. The targeted bands constitute 60% of the frequency bands allocated to the transmission of data for biomedical applications. These bands are the most used because of their low propagation losses in the body and the low risk of ionization of human tissue. To achieve this objective, the first resonance mode should be at 500 MHz. By application of the three rules, this theoretically gives a reduction of 40%. At this reduction ratio, two modes of resonance appear respectively at 554.4 MHz and at 1337 MHz but most of the medical bands are not covered. Also, the antenna is almost broadband. To therefore unify the two passbands, we sought by simulation an intermediate resonance frequency between the two resonance frequencies. The simulation proved that at a size of 9.4 × 4.7 mm², the antenna resonates at 562.8 MHz with coverage of all biomedical frequency bands up to 1 GHz. Applying the same rule of three, this corresponds to a theoretical reduction of 45.88%. But if we consider the simulated size of the antenna, it appears that the antenna has been reduced to 47% of its initial size. The area of the reduced antenna is thus reduced from 20 × 10 m² to 9.4 × 4.7 m².

The miniaturized antenna was simulated in the same conditions of the human Phantom as in the paper [18] . Figure 4 and Figure 5 show the simulation results of the S11 and the reduced antenna radiation pattern. The antenna has been reduced by 53% compared to its initial size. The occupied volume of the proposed antenna is 1.122 mm³ (9.4 × 4.7 × 0.0254 mm³).

Through Figure 4 and Figure 5 , with the reduction, the antenna remains dual band but each of these bands has been widened. The first band, which covers the range from 10 Hz to 270 MHz has a bandwidth of 270 MHz. This band can be used in the downlink transmission of biomedical communication systems [19] for the administration of products or the activation of a patient treatment process. The second bandwidth extending from 397.2 MHZ to 1159.2 MHZ is 762 MHz wide. This mode covers the MIC (402 - 405 MHz), MediRadio (401 - 457 MHz), ISM (902 - 928 MHz) and WMTS (608 - 614 MHz) bands which are regulated for the transmission of uplink data in biomedical communication systems. The antenna can be used in implantable connected objects to communicate with long-range IoT gateways (LoRa, Sigfox) that are currently transmitted in the 868 MHz band (ISM868).

The matching and maximum realized gain of the antenna have also improved. Table 4 shows the simulated performance of the reduced antenna.

Inside the implants are assembled small electronic equipment providing useful functions. That equipment, which is mostly made of conductive materials, can affect the performance of the antenna. The most unpleasant of these devices is the battery. It can act as an unidimensional ground plane whose effect on the matching will be unpredictable. It can also generate a static electric field, which can alter the radiation pattern of the antenna. To study the effect of internal components, we modeled the battery by a perfect cylindrical conductor (PEC) of 5 mm diameter and 10 mm height. We varied the distance (d) between the antenna and the battery by varying the diameter of the battery.

Figure 6 shows a modelled battery inside an implant and Figure 7 shows the variations of the simulated reflection coefficient for different distances.

The analysis of this curve shows (see Figure 7 ) that when the battery is moved away from the antenna, the antenna matching improves reaching −50.57 dB for a distance d of 2 mm. However, as the battery gets closer to the antenna, its matching degrades, and the resonance shifts to higher frequencies. However, the antenna retains its performance in the presence of internal electronic components with a distance (d) more than 1mm.

Since the characteristics (permittivity, conductivity) of human body can vary according to organ, gender, and age, we wanted to analyze the effect of any variation of these quantities on antenna performance. We therefore varied relative permittivity from 54 to 60 and conductivity from 0.59 S/m to 1.09 S/m. These values are those that a muscle can assume, depending on age and part of the body, at frequencies above 1 G Hz.

Figures 8(a) and Figures 8(b) illustrate the variation curves of the reflection coefficient for these different permittivity and conductivity values.

For relative permittivity below 56, the S11 parameters of the first resonance mode and the resonance frequency of the second resonance mode drop, respectively.

But for relative permittivity values above 56, the reflection coefficient remains unchanged. Moreover, despite the change in permittivity, the reflection coefficient remains below −10 dB overall. Analysis of the curve in Figures 8(a) shows that, although the conductivity of the body varies, the reflection coefficient remains below −10 dB, and the two resonance modes retain their same resonance frequencies. These two simulations prove that the antenna can be used in different parts of the body while maintaining the same performance.

In Table 5 , we compared the antenna performances with those of antennas listed in the literature. The proposed antenna dimensions (44.18 mm²) appear to be the smallest but it performs better than the other antennas. It covers 5 bands (402, 433, 614, 868, 915 MHz) including the 868MHz band which is currently successful in IoT connectivity. This band could be used to popularize the IoMT (Internet of Medical things). Moreover, the size of the antenna allows it to be implanted in small muscles such as those of the fingers. The antenna also radiates in the 160 MHz band used for downlinking between outdoor transmitters and implants.