Design of wearable antenna sensors and Path loss Modeling for Internet of Wearable Things applications

The Internet of Wearable Things (IoWT) is a revolutionary technology of Wireless Body Area Networks (WBAN). IoWT applications are widely employed in diverse sectors including medical, civil, and military domains. In these emerging applications, human bodies are equipped with wearable and interconnected sensors, which can be used on, inside, or around the human body. However, the deployment of these devices in close proximity of the human body faces several electromagnetic factors such as channel fading due to several reasons like energy absorption, reﬂection, multipath, and shadowing by the body. These electromagnetic phenomena can degrade the quality of the communication link budget between wearable sensor systems in WBAN. Therefore, investigation in channel modeling between wearable sensors has become a major challenge in Wireless Body Area Networks (WBAN), especially when using these devices for real-time monitoring and remote control of vital signs and physiological data of patients in medical applications. This paper presents a novel approach of a rigorous and eﬀective mathematical voxel-based channel model for channel modeling between wearable wireless antenna sensors in free space and human body environments. Also, this work presents a design ﬂow and performance analysis of the proposed 3D-voxel human body model and the designed wearable antenna at 2.4 GHz. Moreover, a performance evaluation studies between IEEE 802.15.6 CM3 channel model based on the proposed analytical voxel based-human body at 2.4 GHz for various distances between the coordinator and the on-body wearable antenna sensors are presented. These studies reveal a good agreement between IEEE 802.15.6 CM3A and our model.


INTRODUCTION
Nowadays, the development of wireless wearable technology has led to a remarkable evolution in the healthcare field.In this context, a wide range of application domains have been covered by WBAN to perform continuous and non-invasive monitoring of vital signs of persons such as blood pressure, body temperature, Oxygen saturation, electrocardiogram (ECG) to track heart activities, and electromyography (EMG) to evaluate the functioning of nerves and muscles, etc 1 .In particular, WBAN networks as an integral component of the Internet of Wearable Things paradigm, are extensively employed in various domains such as biomedical domain 2 , military domain 3 4 5 , and sports domain 6 7 .Currently, the advancement of intelligent healthcare systems plays a crucial role in facilitating remote medical care for patients, reducing the need for hospital visits, and enabling early detection of life-threatening illnesses as a preventive measure.For example, as indicated in the study conducted in 8 , rapid intervention is particularly necessary during pandemics such as the recent pandemic Coronavirus Disease 2019 (COVID- 19).Therefore, the IoMT can be used to detect the Coronavirus symptoms.However, the deployment of wearable sensors faces several electromagnetic factors that can degrade the performance of WBAN and the stability of Wireless channels such as fading and interference effects caused by signal attenuation, human body shadowing, multipath, and energy absorption by human body tissues 9 10 .According to the study 10 , there is no doubt that the human body consists of natural absorbent layers.As a result, these tissue layers act as a low-loss dielectric medium which causes attenuation of propagation signals along the human body.As per the IEEE 802.15.6 standard, a Body Area Network (BAN) is composed of sensor nodes and a central node known as a hub, sink, or coordinator node.This standard adopts a star topology, which supports two types of communication: simple one-hop and extended two-hop star topology 11 .In simple one-hop topology data are collected by sensors and handled by the coordinator node.Therefore, in our electromagnetic channel characterization, the receiver antenna serves as a coordinator sensor positioned on the waist of the human body, while the transmitter antennas act as sensor nodes located in various positions on the human body in accordance with the requirements of IEEE 802.15.6.
To ensure dependable body surface-to-body surface communication between the coordinator and the sensor nodes within the human body propagation environment, it becomes imperative to mitigate various electromagnetic factors.These factors include the avoidance of diffraction, reflection, and absorption of radio signals by human body tissues.Additionally, due to the complexity of the human body environment, it becomes crucial to characterize and analyze the behavior of the wireless communication channel between devices placed on the human body.The purpose of this task is to characterize propagation characteristics, signal attenuation, interference, and other body-centric impairments.Furthermore, accurate channel modeling plays a crucial role in the design and optimization of communication protocols, antenna designs, power control mechanisms, and resource allocation strategies in Body Area Networks (BANs).Moreover, this latter enables researchers and designers to evaluate and improve the performance of wireless links, ensuring reliable and efficient communication between wearable or implantable devices and the external network infrastructure.
A WBAN is governed by IEEE 802.15.6 standard.This standard provides a medium access control (MAC) and physical (PHY) layers to WBAN systems.Therefore, these layers are based on rigorous wireless propagation channel models and antenna designs for different frequency bands 10 .Introduced in 2007, the IEEE 802.15.6 standard was specifically developed to meet the needs of diverse medical applications reliant on WBAN.These applications include remote patient monitoring within healthcare facilities as well as the monitoring of elderly individuals in their own homes, as stated in a study conducted by Al Barazanchi et al. (2022) in 12 .This standard focuses on providing wireless communication that is suited for applications inside, on, and around the human body and that is also effective in terms of energy efficiency and suitable for short distances.Moreover, communications in BAN networks are categorized into three types : (I) In-Body communication, (II) On-Body communication, and (III) Off-Body communication 13 9 .For that reason, a variety of channel models have been classified according to the type of body communication link (in-body, on-body, and off-body) 14 15 .
The WBAN standard proposes three potential deployment scenarios: (I) Implant node injected under the skin in the human body, which could be placed in the deep tissues or simply under the skin, (II) Surface node located on the skin surface, and (III) External node, known as the Gateway Node.This latter is an off-body node positioned a few centimeters from the skin 9 .Furthermore, the BAN standard has defined four channel models : (I) CM1 Implant to Implant; (II) CM2 Implant to Body surface model; (III) CM3 Body surface to Body surface model, and (IV) CM4 Body surface to the external model. 9, as illustrated in figure 1. 9 In this way, to investigate the propagation of radio waves and subsequently analyze the performance of channel models within or on the body surface, it becomes necessary to study rigorously the radio channel analytically and experimentally.In other words, it's essential to understand the impact of electromagnetic phenomena and the effect of biological human body dielectric tissue parameters on the stability of path loss models.Moreover, the necessity to study the impact of absorbed power by human body tissues layers on the effectiveness and radiation patterns of wearable devices on human bodies.

Main Contributions and Paper organization
The current research work contributes to the field of channel modeling and the optimal design and deployment of wearable antenna sensors.Therefore, the main objective is to propose a rigorous analytical and simulation-based approach to analyze the path loss characteristic between on-body wearable antenna sensors at 2.4 GHz, and also to propose an optimized design for 3D voxel human body using wearable microstrip patch antennas.
The main contributions of this article are : • A rigorous analytical voxel-based path loss model for on-body channel modeling in free space and human body environments.
The proposed model can be applied for channel modeling in WBAN field.• A design flow and performance analysis of the proposed antenna in different parts of the proposed voxel human body model • A path loss performance evaluation and comparison of IEEE 802.15.6 CM3A channel model with our proposed mathematical on-body voxel-based channel model for on-body communications.• Study of the behavior of the deployed antenna sensors in free space and in the presence of a voxel human body considering the realistic biological properties of the human tissues.
Our numerical simulations are performed using the CST microwave simulator.In our simulations, we have examined the channel attenuation between the coordinator antenna placed on the body waist and antenna sensors placed on different positions using S 21 parameter as a radio channel parameter for path loss modeling.By numerical simulations, we aim to prove the effectiveness of our mathematical voxel-based channel model in both free space and human body environments and to evaluate on-body channels around the body.
The remainder of this article is organized as follows: related work is presented in Section II.Section III exhibits the proposed mathematical voxel based-channel model and CM3A channel model.Section IV presents the 3D voxel human model and deployment scenario.In Section V we present the wearable antenna design flow and performance analysis in terms of return loss, radiation pattern, and Specific Absorption Rate (SAR) analysis.Section VI presents a path loss performance evaluation and comparison of the CM3A path loss model with the proposed on-body mathematical voxel-based channel model.Section VII concludes the paper and provides some future work directions.

RELATED WORKS
In the literature, various approaches have been proposed for channel modeling and path loss performance evaluation of a BAN network by applying different analytical channel models and several experiments and simulations.Several approaches have been introduced in the literature for channel modeling between on-body sensors.In WBANs, electromagnetic signals are transmitted along the human body.For this reason, the human body is considered a propagation medium for propagating waves.For instance, in the study conducted in 16 , authors have talked about the importance of channel modeling in Body Area Networks to estimate the path gain and link loss with an Electromagnetic propagation wave technique.In addition, Several approaches have been introduced in the literature to prove the importance of BAN channel models between on-body sensors for the deployment of these sensors and their importance in the performance evaluation of Body area networks based on IEEE 802.15.6 requirements.
For example, in the studies conducted in 17,18 , the authors have proposed a simulation and a channel modeling of a BAN network composed of one coordinator and eleven sensor nodes to characterize the channel based on lognormal shadowing path loss model for the evaluation of the performance of the proposed BAN in terms of consumed energy, packet loss, and radio modulation type.Therefore, in some studies 19 and 20 , the authors have indicated that the lognormal model is more accurate than the Nakagami and Rayleigh channel models for indoor environments.Furthermore, in other studies 21,22,23,24 the authors have employed the CM3A and CM3B on-body path loss models proposed by the IEEE 802.15.6 standard for channel modeling between on-body sensors.In these studies, the authors have applied the two path loss models to the frequency ranges of 2.4-2.45GHz.The effectiveness of CM3 has been tested in both hospital rooms and anechoic chambers.In addition, in the study conducted in 23 , authors have used the CM3B path loss model to study the impact of power and modulation schemes variation on the performance of a BAN in terms of packets received at the coordinator, Packet Loss Rate and latency at 2.45 GHz frequency.Moreover, the authors in 24 have proposed the use of the CM3B path loss model.This work aims to study and evaluate the impact of inter-BAN interference on the performance of a BAN in terms of energy consumption.
On the other hand, numerous research studies have been proposed to characterize the on-body channel between wearable antennas at different frequency bands based on voxel-human body models and human-body phantoms.
For example, in studies conducted in 25,26,27 , the authors have proved the relevance of S 21 parameter in channel modeling between wearable antennas and a capsule endoscope implanted in a small intestine inside the human body.In these, studies, the different biological tissues have been considered.Furthermore, in the study performed in 28 , the authors have proposed a patch antenna design at 2.4 Ghz and an analytical channel characterization between two wearable antenna sensors placed on a simple 3D cylindrical phantom using CST microwave studio.However, in this study, only the muscle tissue was considered in the human body phantom.
In the same context, more recently, in studies conducted in 29,10 , the authors put forward an analytical and experimental approach for on-body path loss and path gain channel modeling between two wearable antenna sensors for Wireless Body Area Networks.In general, these studies 28,29,10 have some common points, because the authors have developed an on-body channel model based on a previous voxel-based channel model proposed in Hall et al. 30 and Alves et al. 31 .Also, they proposed similar analytical and experimental contributions.Although they are innovative approaches, in these studies 28,29,10 , the authors did not take into consideration the SAR study to examine the behavior of the different antenna sensors in the presence of a human body and especially the impact of power absorption on the body safety, also they did not employ the skin, fat, and muscle biological tissue for exact and realistic channel modeling.Therefore, the authors consider only the muscle tissue.In addition, in the studies conducted by Hamdi et al. 29 and Hamdi et al . 10, the authors have changed the conductivity and permittivity of the human body muscle randomly without taking into account the realistic biological properties of the conductivities and permittivities of each human body tissue.As a consequence, simulation results will produce approximate return loss and path loss results that are not representative of the realistic conditions in the channel modeling of the human body environment.
In this paper, we aim to contribute to the existing literature by introducing new findings for on-body channel modeling between wearable sensors, by proposing a rigorous on-body path loss modeling theoretically and experimentally considering the real dielectric biological tissues parameters of a human body and based on our voxel-based channel model, also considering the power absorption effects on the behavior of wearable antennas in free space and in presence of a human body.For this reason, according to IEEE 802.15.6 draft, attenuation caused by the human body should also be considered, in addition to frequency and distance between Tx and Rx antennas.

ANALYTICAL CHANNEL MODELS FOR WBAN SYSTEMS
This section presents CM3A channel model and the proposed voxel-based analytical model.

Voxel-based human body analytical channel model
In our work, we propose a mathematical voxel-based channel model.The proposed WBAN channel model is based on the principle of simple wireless communication between multiple wearables transceiver (T x ) and receiver (R x ) antennas for on-body communications.The proposed antenna operates at a frequency of 2.4 GHz in the Industrial, Scientific, Medical (ISM) band.The suggested voxel-based human body model illustrated in figure 2, takes into account all dimension parameters and all electromagnetic properties of propagation mediums (1) and ( 2).We propose mathematical rigorous formulas to predict the channel model based on the interaction between the electromagnetic waves of the free space (medium 1) and the human body (medium 2).The proposed voxel-model used in medium 2 consists of three tissue layers (Skin, fat, and muscle) which introduces an additional impact on the propagation of electromagnetic waves due to the variability of the physiological and dielectric characteristics and parameters of a human body.

F I G U R E 2 Radio link communication between two antennas
The radio link between the two antennas is illustrated by the Friis formula in equation (1).By using Friis formula, the transmitted power (P Tx ) can be extracted with transmitter antenna gain (G Tx ), and the received power P Rx can be extracted with the receiver antenna gain (G Rx ), knowing the distance (d) between T x and R x antennas.Where P Tx and P Rx are expressed in dBm, however G Tx and G Rx in dBi.
The Friis formula is presented in equation ( 1) as follows : Using the equation 1, we neglect the heights of T x and R x antennas, the Friis equation is simplified as follows indicated in equation 2. In this equation A t parameter expressed in dB represents an essential and critical parameter for the optimization of Tx/Rx communication link.This latter depends on three factors : (I) the distance between Tx and Rx antennas; (II) the operating frequency; and (III) the attenuation of the propagation medium.
Therefore, to evaluate the path loss at 2.4 GHz for the two proposed propagation environments, we focus on calculating the approximative path loss in free space and on various human body positions using on-body antennas polarized normally with the body surface.

Free space channel model
Based on the Friis formula illustrated in equation ( 1), the path loss can be written as follows : Thus, PL d0 also is considered as a path loss at a reference distance d 0 .This parameter is calculated using the equation (4).
According to the study conducted in 32 , for the typical radio applications, when the frequency fr is expressed in MHz and the distance d is expressed in kilometers, PL d0 becomes equal to 32.44 dB.However, in the proposed model the operating frequency is in the range of 2.4 GHz, and the T x /R x distance is significantly small, usually less than 1 meter when considering real human bodies.
Thus, When the operating frequency fr is expressed in GHz, we add 60 dB to PL d0 .Therefore, PL d0 becomes equal to 92.44 dB and the path loss in the free space environment is presented in equation ( 5) as follows : If the frequency f exceeds 1 GHz, PL FreeSpace becomes expressed as the product of two path losses PL FreeSpace (f ) and PL FreeSapce (d), which depend on the frequency f and the Tx/Rx distance d, respectively.In this case, the expression of PL Freespace(dB) is given by equation ( 6) : PL Freepace (fr, d) = PL FreeSpace (fr).PL FreeSpace (d) (6)

On body channel model
Defining an accurate channel model for on-body communications between wearable antenna sensors is a critical and challenging task.This propagation environment is characterized by high attenuation due to signal absorption by different body tissues or some electromagnetic phenomena such as fading, reflection, and shadowing.The signal attenuation is affected by transceiver and receiver distances as well as the variability of the dielectric properties of biological tissues along the propagation path.The dielectric properties of a human body used in our on-body channel model are tissue body permittivity (ε r ) and tissue body conductivity (σ).Therefore, by inspiration from the study of Alves and al. 31 , disruptive effects are modeled as an attenuation called α Body in equation (7) and expressed in Np/cm or Np/m.
As a result, we have derived our path loss model from equation (7) and with inspiration from the study of 10 .The proposed on-body path loss model is expressed in function of three variables parameters: Tx/Rx distance, human body conductivity, and human body permittivity.This model is expressed in dB and presented in equation (8).
Thus, realistic channel modeling requires at least the consideration of the skin, fat, and muscle tissues of the human body.In 28 , the authors considered only the muscle tissue.In our model, at 2.4 GHz, we have determined the permittivities (ε r ) and the conductivities (σ) of these three human body biological tissues according to the U.S. Air Force Report AFOSR-TR-96 defined by the Federal Communications Commission (FCC).These parameters were defined using the body tissue dielectric parameters tool.
We present in Table 9 the permittivities and conductivities the skin, fat, and muscle layers.These parameters were used in our analytical model and also in our simulations.The total path loss for free space and human body environments can be calculated by the summation of equations ( 5) and ( 8).In our analytical on-body model, we have considered that we have a muscled body where ε r ) = 50 and σ = 1.705 s/m.Also, after mathematical reasoning, we found that for 1 meter, the attenuation α Body is equal to 400 dB, thus for 25 cm distance between the transceiver and the receiver antennas, the attenuation α Body is approximately equal to 100 dB.The expression of total path loss is illustrated in equation ( 9) as follows : We present in figure 3 the path loss in the two environments in terms of the distance between Tx/Rx sensors in free space and on-body environments.From these results illustrated in figure 3, it is clear that the path loss in the two propagation environments is affected by Tx/Rx distance.
F I G U R E 3 Path-loss characteristic of proposed model According to Hall et al. 30 , Alves et al. 31 , and Hamdi et al. 10 , the proposed model is accurate, generic, and effective for on-body channel modeling in WBAN systems and it can be used for channel characterization for on or off-body applications.Moreover, this model can be used for path loss modeling between wearable sensors in wireless body area networks.

CM3 channel model for wearable WBAN
The CM3A (Body surface to Body surface channel model) is a specific channel model defined by IEEE 802.15.6 standard for On-body communications at 2.4 GHz.This model is designed to simulate the wireless communication channel between devices located on the surface of the human body.It takes into account factors such as signal attenuation, interference, and path loss.This model has been subjected to testing in various environments, including hospital rooms and anechoic chambers as mentioned in table 2. The path loss model at a calculated distance is expressed by the following equation : Where : a and b are coefficients of linear fitting, d represents the Tx-Rx distance and N represents the normally distributed variable with standard deviation σN .The CM3A model considers the body shadowing effects and the reflection of the signal from the body surface.This model allows the prediction of path loss of radio signals traveling between two nodes on a human body surface.This model takes into account critical factors such as distance, frequency, and the specific characteristics of the human body.

General WBAN architecture
The proposed architecture allows hospital staff to manage data collected from BAN nodes to monitor sick patients.This architecture is composed of two categories of users, the patients and the hospital staff.The collected data are sent to medical databases.Medical applications installed on computers or smartphones allow health professionals to access patient data in real-time.
F I G U R E 6 3D human Body model for WBAN system

ANTENNA DESIGN FLOW AND PERFORMANCE ANALYSIS
In this work, we have proposed a rectangular microstrip patch antenna operating at 2.4 GHz using CST microwave studio.To analyze the impact of the human body on antenna performances we have investigated a variety of design flows and numerical simulations.The proposed antenna has been considered as a highly directional antenna for human bodies applications.Our wearable antenna is printed on an FR-4 dielectric substrate with a Perfect Electrical Conductor (PEC) material.This substrate material is flexible and it can be implanted into the body or cloth.Another challenging aspect consists of designing and placing our antenna on the human body in a way that reduces the power absorption effects by the human body tissues.To reduce the impact of propagation waves on the human body and to have an effective antenna in terms of radiation patterns (Gain and directivity, etc.), we have chosen a substrate with a low dielectric constant ε r = 4.3 and low dielectric height i.e., SH = 1.5mm.The feed line and ground plane are made of copper annealed material, which is useful for on-bodies applications because it reduces power absorbed by the human body tissues caused by the propagating creeping waves of 2.4 GHz frequency.To design an effective antenna with stable performance we considered critical factors such as antenna impedance matching, radiation efficiency, and also direction of the printed circuit with their angle of radiation.
To optimize the performance of the proposed antenna in term of ad impedance adaptation between the printed patch and the feed line, we have adjusted the width (PW) and the length (PL) of the patch as well as the widths of the two inset gaps INW1 and INW2 .

5.1
Wearable Antenna performance in free space and human body environments

Return Loss Analysis
We start first with the performance analysis of the return loss for different wearable antenna sensors.In telecommunications measurements, the return loss term represents the loss of power, which have been returned or reflected from an antenna to a transmission line.It's a ratio between the incident power and the reflected power.Generally, the return loss of any antenna is expressed in dB using the equation 11 as follows : -10 log 10 Inc power Ref power (11)   It's clearly observed from figure 8.a and figure 8.b that the antenna exhibit improved impedance matching, as indicated by a return loss measurement that falls below -10 dB line at 2.4 GHz in free space and in human body environments.In free space, the return loss of the proposed antenna is equal to -43.01 dB.However, the return loss of antenna N1, antenna N2, antenna N3, antenna N4, and antenna N5 is significantly increased in human body, S 11 becomes equal to -23.97 dB, -19.24 dB, -23.33 dB, -24.47 dB, and -25.52 dB respectively for N1, N2,N3,N4, and N5.In general, the increase of return loss for different antennas is due to the highest values of physiological parameters of the human body such as the conductivity of the muscle tissue (σ = 1.705S/m).Also, the power absorbed by each different tissue layers (Skin, fat, Muscle).

Radiation pattern Analysis
The analysis of 3D radiation patterns is necessary and essential to study the impact of electromagnetic phenomena such as the influence of energy absorption on the gain and directivity of wearable antenna sensors in free space and especially in the presence of a human body.It is, therefore, necessary to understand the behavior of these devices in free space and in close proximity to the human body, as well as the influence of human body power absorption on the radiation performance of the wearable antenna.
In the Fundamental mode of excitation, the proposed antenna N has the highest directivity at the direction perpendicular to the patch where the angle θ = 90°in the vertical plan, as well as a minimum move away in the direction of endfire where the angle θ = 0°in the horizontal plane.In the free space environment, as shown in figure 9, the proposed antenna has a directivity of 6.9 dBi and a gain of 2.61 dBi at 2.4 GHz.
However, according to the obtained results when the antenna is placed in different human body parts, we notice a slight decrease of the antenna gain and approximate stability of the directivity for different antenna positions.Therefore, during transition, from free space to human body environment, as observed in figure 11, figure ??, figure 12, figure 13, and figure 14 , the peak gain is reduced to 1.22 dBi, 2.04 dBi, 0.943 dBi for antenna sensors N1,N2, and N3 respectively.The mean peak gain of the patch antenna N4 is approximately similar to the mean peak in free space environment.Moreover, the radiation patterns analysis shows that the peak directivity of the designed wearable is slightly reduced to 6.27 dBi, 5.98 dBi, and 6.71 dBi for antennas N1, N3, and N4 respectively.The mean peak of directivity Antennas N2 and N5 is equal to 7.06 dBi and 7 dBi respectively.Consequently, the slight decrease and attenuation in radiation patterns at the angle θ = 90°, is due to the impacts of biological parameters of the human body such as the high conductivity of the muscle tissue (σ = 1.705S/m), also considering the conductivities of the skin and fat tissues (1.44079 S/m and 0.1 S/m respectively).Generally, the overall radiation patterns of gain and directivity confirm that the wearable designed antenna is suitable for free space and on-body medical applications.

Specific Absorption Rate Analysis
We present in this section a Specific Absorption Rate (SAR) analysis of five wearable antennas at 2.4 GHz for on-Body medical applications.A Human body is a complex propagation medium, it is highly conductive.As a result, human body tissue layers adds an additional effect to propagation waves such as diffraction, reflection, shadowing and power absorption.Considering the losses due to the human tissue frequency absorption and the complexity of this propagation environment, one of the most critical challenges in BAN is the design of efficient wearable antennas and the analyses of the Specific Absorption Rate (SAR) to protect the human body from radio frequencies and to ensure human body safety.The absorbed power per unit mass of human body is analysed by the SAR.Moreover, electromagnetic waves radiated by wearable antennas and penetrated in human body tissues can cause harmful and devastating effects of human body.Thus, It is possible to identify the nature of these emitted waves and their impacts by measuring the Specific absorption rate using CST microwave studio simulator.
The specific Absorption Rate (SAR) quantifies electromagnetic energy radiation absorbed by tissues and represents the amout of energy or power deposition per unit mass of biological tissue.The standard unit for SAR is watt per kilogram (W/kg).Thus, The local SAR is expressed as follows : Where σ is the conductivity of the human body tissue (S/m), ρ is the mass density of the tissue (Kg/m 3 ), and E is the electric field strength (V/m).SAR may also be expressed as a function of the rate of temperature elevation in body tissues 33 .According to the study conducted in 34 , the international community has standardized and regulated the SAR limitations for cell phones and similar devices, the maximum safety limit of SAR specified by the federal Communications Commission (FCC) is 1.6 W/kg for 1 g of tissue.A maximum SAR limit of 2 W/kg for each 10 g of tissue has also been established by the International Commission of Non-Ionization Radiation Protection (ICNIRP) in Europe according to the European Standard (IEEE C95.3).Moreover, in study conducted in 35 , authors have indicated that the FCC has determined SAR limits of a maximum of 2W/kg for 10 g of tissue, with a lower limit of 1W/kg for 10 g of tissue as the higher and lower limits, respectively.For these reasons, in the work, we consider that the safer SAR threshold for 1g is 1.6 w/k for 1g of tissue and 2w/kg for 10 of tissue.
Moreover, Some authors have defined the SAR, as the time derivative of incremental energy (dW) dissipated in an incremental mass (dm) in a volume (dV) of a given density ρ 36 .
Where, dW, is the energy absorbed by the human tissue; dm is the mass; and dV is the volume element.Thus, In our work, we calculate the SAR over 1g and 10g of human body tissue.As shown in table 7 and figures 15, 16, 17, 18, and 19, SAR varies with the variation of the node position on the human body.For 1g tissue, antenna 5 has the highest SAR value of 0.47 w/kg at 2.4 GHz.However, antenna 2 has the lowest value of 0.0514 w/kg.
On the other hand, we notice the same interpretation at 2.4 GHz, for 10g tissue, antenna 5 has the highest SAR of 0.201 w/kg and antenna 2 has the lowest value of SAR of 0.0188.Generally, for 1g of tissue and 10g of tissue the transmitter antennas N1, N2, N3, N4, and N5 produce SAR values below the limits defined by the FCC.This makes the simulated antennas suitable for Wireless Body Area Networks and for wearable applications.By simulations in both free space and on-body environments, we have noticed that our antenna is more efficient than several other antennas proposed in the literature.For a free space environment, for example, Compared to 28 , our antenna is more efficient in terms of all performance metrics.Also, the proposed antenna is more effective in terms of return loss and VSWR than the antenna designed in the study conducted in 10 .The human body is a complex propagation environment.Thus, to approach a realistic on-body channel modeling between wearable sensors with Tx and Rx antennas, on the voxel body, we have considered in our simulations the real dielectric properties of biological human body tissues (ε r : Permittivity and σ : conductivity) for the skin, Fat and Muscle layers.The parameters utilized in our work have been established based on the U.S. Air Force Report AFOSR-TR-96, as defined by the Federal Communications Commission (FCC).These parameters were determined using the body tissue dielectric parameters tool.To evaluate the behavior of the wearable antenna sensors, we placed the proposed antenna on the human body waist.We have chosen this position not by coincidence because a significant part of the power is absorbed by biological tissues, especially the fat layer in this region.As mentioned in table 10, for the On-body propagation environment, our wearable antenna is more efficient in terms of return loss and VSWR than other studies illustrated in this table.We notice a significant increase of return loss and VSWR of the antenna placed on the human body waist in the two studies 28 and 10 .In studies 28 and 10 , authors have focused only on muscle tissue, however in our simulations, we have included the skin, muscle, and fat as part of our simulated model.Nevertheless, we achieved favorable performance results in both environments, despite the significant power absorption by human body tissue layers.
In our work, despite the power absorption by the three human body tissues (Skin, Fat, and Muscle), the return loss and VSWR consistently remain within acceptable ranges.As shown in table 10, we notice a slight increase in the two parameters.Consequently, the return loss increases from -43.01 dB to -26.71 dB., the S 21 parameter serves as the channel parameter employed to measure the path loss between the wearable antennas.Therefore, our initial focus is on studying the S 21 channel parameter for path loss modeling between a coordinator and each antenna sensor node of the Body Area Network (BAN) in various on-body antenna distances and positions within the frequency domain.Within our simulated model, the S 21 parameter signifies the path loss between the coordinator (N0) and each individual node (Ni) within the topology of the Body Area Network (BAN).
As it can be seen from figure 21, We can notice a clear increase in path loss as a function of distance and antenna sensor positions.At 2.4 GHz, which is the focus of this paper, when N1 is positioned on the left wrist, the path loss is lower at a distance of 0.279 m, in this case, S 21 is equal to -64.71 dB.The second weaker channel is obtained when the antenna N2 is positioned on the left chest at a distance of 0.315 m, resulting in an S 21 value of -58.11 dB.At the same frequency, when the antenna N5 is situated on the head at a distance of 0.601 m, a substantial increase in path loss is observed, with the S 21 value reaching -39.07 dB.In this particular scenario, the S 21 parameter indicates the highest level of signal attenuation.In all remaining on-body antennas, it is also evident that path loss increases with increasing distance between Tx and Rx antennas.The two proposed channel models share some common points and factors.These models use the same frequency range, and the same distances between nodes and also they consider the biological tissues of the human body.Furthermore, these two models are used to estimate the channel attenuation between wearable antennas in BAN networks.To compare the CM3A model and mathematical voxel-based human on-body body channel model, we have transformed the S 21 channel results to obtain the simulated path loss values in decibels using the equation 15.This equation is used to transform the magnitude of S 21 to the decibel unit.
S  In the study conducted in 41 , authors have calculated the S21 channel parameter using a Vector Network Analyzer.However, in this paper, we used CST simulator to calculate the path loss characteristic and we have demonstrated that this simulator is a solid tool for path loss modeling between wearable antennas despite the problem of signal absorption by biological tissues of Voxel human body models.Moreover, Simulated results demonstrate the effectiveness of the main objective channel model for path loss modeling between wearable sensors.

CONCLUSIONS
In this work, first, we have proposed a rigorous analytical voxel-based path loss model for on-body channel modeling between wearable antenna sensors in both free space and human body environments.The proposed model can be applied for channel modeling in the WBAN field.Also, we have proposed a design flow and performance analysis of the wearable antenna in different parts of the voxel human body model.The performance of the proposed antenna in terms of return loss, Gain, and directivity in both environments has been studied.Moreover, a SAR analysis has been performed to ensure human body safety.Second, we have suggested a path loss performance evaluation and comparison of the IEEE 802.15.6 CM3A channel model with our proposed mathematical on-body voxel-based channel model for on-body communications.Furthermore, we have studied the behavior of the deployed antenna sensors in free space and in the presence of a voxel human body considering the realistic biological properties of the human tissues.According to the obtained simulation results, we have examined the channel attenuation between the coordinator antenna placed on the body waist and antenna sensors placed on different positions using S 21 parameter.In addition, we have improved the effectiveness of our mathematical voxel-based channel model in both free space and human body environments and to evaluate on-body channels around the body.
tunable RF MEMS capacitor/inductors and their wireless applications.He is currently an associate professor in the High Institute of Computer Engineering and Communication Technology, in Hammam Sousse, part of the University of Sousse, Tunisia.He is a researcher member in NOCCS research Lab of ENISo, university of Sousse, Tunisia.He has served as a consultant with several international companies in Industry 4.0 and Biomedical Engineering, where he has been actively involved in many industrial projects.
Aref Meddeb : obtained his Engineer degree from ENIT, Tunisia, in 1992, and both his M.S. and Ph.D. degrees from Ecole Polytechnique, Montreal, Canada, in 1995 and 1998, respectively.He worked with Alcatel, INRS-Telecom, Teleglobe, and Nortel.He was associate professor and deputy director at ISITCom, Tunisia, where he also headed the Telecommunications Department.He is currently full Professor at the National Engineering School of Sousse (ENISo), University of Sousse where he also was Director and Director of Study.He also founded the Networked Objects, Control and Communication Systems (NOCCS) research Laboratory.His research interests include Internet of Things, Wireless Sensor Networks, RFID with focus on Security, QoS, Routing, and Design.

T A B L E 1
Dielectric parameters of a human body SIMULATION MODEL AND DEPLOYMENT SCENARIOIn this study, we propose a 3-dimensional BAN model based on the IEEE 802.15.6 standard.To approach realistic conditions, we chose to deploy five wearable antennas in different medical positions on an anatomical female Voxel human model called Katja.The 3D human model has an average length of 1.65m.We have placed the simulated model at the origin of the three axis (x, y, z).In simulation work, we have chosen the star topology proposed by IEEE 802.15.6 standard, five transmitter antennas communicate directly with one receiver antenna positioned on the body waist of the proposed voxel body.As shown in figure5, we have placed these antennas in these positions not by coincidence because we aim to target the most vital signs of the human body.The proposed WBAN model operates in a narrow band at 2.4 GHz.We present in Figure6the 3D voxel human model called Katja for real female body model with Tx and Rx microstrip antennas placed in diverse positions defined by the standard IEEE 802.15.6 for BAN networks.

F
I G U R E 4 3D human Body model for WBAN system T A B L E 3 Coordinates of antenna sensors and vital signs monitored of each 15, y5 = 0.6, z5 = 0.1 EEG We present in figure 5 the different positions of the on-body antennas on the voxel human model.F I G U R E 5 (a) Voxel body.(b-f) Sensor nodes positions

F I G U R E 7
Antenna geometry optimization T A B L E 4 Design parameters of proposed antenna Parameter Dimension (mm) Description

F
I G U R E 8 (a) Return loss in free space.(b) Return loss in the human body T A B L E 5 Summary of return loss performance for on-body antennas Antenna Placement Return loss (dB)

F
I G U R E 9 Antenna 3D radiation patterns in free space environment F I G U R E 10 (a) N1 3D Gain on left wrist.(b) N1 3D directivity on left wrist F I G U R E 11 (a) N2 3D Gain on left wrist.(b) N2 3D directivity on left wrist F I G U R E 12 (a) N3 3D Gain on left wrist.(b) N3 3D directivity on left wrist F I G U R E 13 (a) N4 3D Gain on left wrist.(b) N4 3D directivity on left wrist F I G U R E 14 (a) N5 3D Gain on left wrist.(b) N5 3D directivity on left wrist

F
I G U R E 15 (a) Antenna N1 SAR for 1g of tissue.(b) Antenna N1 SAR for 10g of tissue F I G U R E 16 (a) Antenna N2 SAR for 1g of tissue.(b) Antenna N2 SAR for 10g of tissue F I G U R E 17 (a) Antenna N3 SAR for 1g of tissue.(b) Antenna N3 SAR for 10g of tissue F I G U R E 18 (a) Antenna N4 SAR for 1g of tissue.(b) Antenna N4 SAR for 10g of tissue F I G U R E 19 (a) Antenna N5 SAR for 1g of tissue.(b) Antenna N5 SAR for 10g of tissue

T A B L E 9 F
Dielectric parameters of a human body tissues I G U R E 20 Voxel human body tissues

F
I G U R E 21 Channel characteristics of the proposed voxel based model at 2.4 GHz

F
I G U R E 22 CM3 and voxel-based path loss characteristics at 2.4 GHz T A B L E 6 Summary of radiation patterns performance for on-body antennas We summarize in table 8 and table 10 a performance comparison of the designed antenna with other antennas designed in several recent studies in the existing literature.
As per the initial report on channel modeling for wearable and implantable Wireless Body Area Networks (WBANs) specified in the IEEE 802.15.6 standard 21(dB) = 20log 10 |s 21 | (15)As expected, we can see a clear remarkable growth in path loss for both the CM3A model and the proposed voxel-based channel model.From the results illustrated in table11and figure22, it's clear that increasing the distance between Tx and Rx sensors increases the losses in both models.As illustrated in table 11 path losses depend on the distance and position of the nodes.As shown in figure22, Our findings indicate exhibit similar path loss characteristics and produce approximately similar values for different distances.Therefore, we can see a good agreement between the two path loss models.Moreover, the performance of Tx/Rx antennas is further verified by finding out of the path loss S 21 obtained by simulation and by the main model as shown in figure22.While path loss remains approximately constant within the frequency range of 2.3 GHz to 2.5 GHz, it reaches a peak value of 38.52 dB for the main model and 36.2 dB for the simulated model at various distance values.