Electric Field Intrabody communication ChannelModeling With Finite Element Method

Electric-Field Intrabody Communication Channel Modeling With Finite-Element Method

Ruoyu Xu*,Student Member,IEEE,Hongjie Zhu,Student Member,IEEE,and Jie Yuan,Member,IEEE

Abstract—Electric-?eld intrabody communication(EF-IBC)is a promising new scheme for the data exchange among wearable biomedical sensors.It uses the body as the signal transmission https://www.360docs.net/doc/ba15535293.html,pared with existing body area network(BAN)schemes, EF-IBC can achieve higher data rate with less transmission power. Until now,the detailed EF-IBC channel mechanism is not well understood.In this work,?nite-element method(FEM)is utilized for the?rst time to investigate the EF-IBC channel.A circuit-coupled FEM model is established for the EF-IBC channel.The FEM model is extensively veri?ed by experimental measurements. The new physical model enables the revelation of characteristics and effects of different components in the EF-IBC channel.The FEM investigation?nds that the capacitive return path is critical to the characteristics of the EF-IBC channel.Parameters of the capacitive return path are quantitatively measured.The investiga-tion also?nds that the body plays an important role to the return path capacitance.The forward body path can be well modeled by a cascade ofπ-shaped circuits.Based on the FEM model of the EF-IBC channel,a simpli?ed circuit model is derived to provide an ef?cient tool for the transceiver design.

Index Terms—Body area network(BAN),capacitive return path, electric-?eld intrabody communication(EF-IBC),?nite-element method(FEM),intrabody communication(IBC).

I.I NTRODUCTION

P ERSONAL health care demands new designs of portable and wearable biomedical sensors,as well as new meth-ods of collecting distributed information from sensors.To pro-vide solutions for the biomedical sensor information exchanges around the human body,the IEEE802.15.6working group has been set up to de?ne new physical and media access control lay-ers for body area network(BAN)[1].Majority of BAN research attempts to apply existing wireless communication technolo-gies,such as ultrawideband techniques[2],[3]and Zigbee[4], to the air channel around the body,which is usually called wire-less BAN(WBAN).

Manuscript received March19,2010;revised July27,2010and October15, 2010;accepted November9,2010.Date of publication November18,2010;date of current version February18,2011.This work was supported by the Hong Kong University of Science and Technology under Grant RPC07/08.EG08. Asterisk indicates corresponding author.

*R.Xu is with the Department of Electronic and Computer Engineering, Hong Kong University of Science and Technology,Clear Water Bay,Kowloon, Hong Kong(e-mail:xuruoyu@ust.hk).

H.Zhu and J.Yuan are with the Department of Electronic and Computer Engineering,Hong Kong University of Science and Technology,Clear Water Bay,Kowloon,Hong Kong(e-mail:zhuhongjie@ust.hk;eeyuan@ust.hk). Color versions of one or more of the?gures in this paper are available online at https://www.360docs.net/doc/ba15535293.html,.

Digital Object Identi?er10.1109/TBME.2010.2093933

Recently,intrabody communication(IBC),which uses the body as the signal transmission medium,has been arising as an alternative solution for BAN[5].It has several advantages over WBAN schemes.Due to its near-?eld-coupling operation, most of the signal from the transmitter(TX)is con?ned to the body and has little interference with other electronic devices[6]. Therefore,data exchange is secure and there is almost no band limitation.Most of all,the attenuation of the body channel can be much lower than that of the air channel up to1GHz[7]. IBC can be categorized into two types:waveguide IBC and electric-?eld IBC(EF-IBC).Waveguide IBC uses a pair of elec-trodes for both the TX and the receiver(RX)to propagate the electromagnetic wave[7],[16].Waveguide IBC links gener-ally have low(kilobits per second)data rate because the body effectively shorts the TX signal with frequency higher than 1MHz[8].For EF-IBC,only the signal electrodes of TX and RX are attached to the body while the ground(GND)electrodes are left?oating in the air.The conductive body forms the for-ward path.The signal loop is closed by the capacitive return path between TX and RX GND electrodes.The TX establishes the quasi-static electric?eld,while the RX detects the electric po-tential at the remote end in this application.Hence,it is usually called the EF-IBC[7],[16].EF-IBC has achieved much higher data rate than waveguide IBC.A10-Mb/s data link over1.8-m distance with4.6-mW transceiver power consumption has been reported[9].The high data rate is desirable in new medical sen-sor applications,such as wireless capsule endoscopy[10]and neural recording[11].

IBC channel modeling is important to the understanding of the communication mechanism and to the transceiver design.The waveguide IBC channel has been modeled by simpli?ed circuits [12],?nite-element method(FEM)[13],?nite-difference time domain(FDTD)method[14],respectively.However,the EF-IBC channel is only modeled by the distributed RC circuit so far[6].Although the circuit model helps the transceiver design, quantitative characteristics of the critical components in the channel are not known.The mechanism of the EF-IBC channel is yet to be studied.

In this work,we study the EF-IBC channel with FEM for the?rst time.FEM is able to simulate electrical behavior with a physical https://www.360docs.net/doc/ba15535293.html,pared to circuit models,it can match the real physical system better.The FEM model of the EF-IBC channel is thoroughly veri?ed by experimental measurements. Both the measurements and simulations are conducted in the 1–180MHz frequency range,which covers the optimal band for EF-IBC[6].Through comprehensive FEM simulations,char-acteristics and effects of different components in the EF-IBC channel are quantitatively investigated.The body in?uence on

0018-9294/$26.00?2011IEEE

Fig.1.EF-IBC system.(a)Application.(b)Simpli?ed circuit model.

the capacitive return path is studied.To facilitate the transceiver design,a simpli?ed circuit model of the EF-IBC channel is derived based on the FEM model.

The paper is organized in six sections.EF-IBC is introduced in Section II.Section III describes the experimental setup.FEM simulation setups are elaborated in Section IV.In Section V, measurement results,simulation results,and analyses are given. Finally,Section VI draws the conclusions.

II.E LECTRIC-F IELD I NTRABODY C OMMUNICATION

An EF-IBC application is shown in Fig.1(a).Both TX and RX have a signal electrode(Sig TX/Sig RX)and a GND electrode (GND TX/GND RX).Sig TX and Sig RX are put on the body,while GND TX and GND RX are left?oating.The conductive body tis-sues form the forward path between Sig TX and Sig RX.GND TX and GND RX are capacitively coupled to each other through the air or through the external GND as the return path to close the signal loop.The coupling between the body and the external GND causes signal leakages.A simpli?ed circuit model of the system is shown in Fig.1(b).The EF-IBC channel is composed of the conductive forward body path and the capacitive return path.For the quasi-static electric?eld,the EF-IBC channel gain can be roughly calculated by

V RX V TX =

Z RX

Z TX+Z RX+Z body+Z return

(1)

where Z RX is the RX impedance,Z TX is the TX impedance, Z body is the impedance of the body path,and Z return is the impedance of the return path.As the coupling capacitances between GND TX and GND RX(C return TX and C return RX) are generally small,they are the most critical components in the EF-IBC channel at low frequencies.When the frequency is within kilohertz range,the impedance of the body path can be neglected.So in the?rst EF-IBC study,the whole body was treated as one node[15].When the frequency is higher than 10MHz,impedances of the body path and the return path are comparable.The body impedance cannot be ignored any more. EF-IBC applications should use signal frequency lower than150 MHz to avoid signi?cant channel variation and radiation

[6].Fig.2.Experimental TX board.(a)Front side.(b)Back side.(c)TX and RX signal electrode(Sig T X/Sig R X).

III.E XPERIMENTAL S ETUP

A.TX Board

It is critical to preserve the capacitive return path in the mea-surements.For earth-grounded instruments such as signal gener-ator,spectrum analyzer and network analyzer,their port shields are connected to the external GND.If the measurements are performed with both TX and RX electrodes connected to these instruments,the capacitive return path will be shorted[6].

To emulate a real EF-IBC channel,a battery-powered signal generator can be used as TX[6].In this work,a versatile TX board based on the AD9953direct digital synthesizer(DDS) is designed with programmable frequency capability controlled by the AT89S52microcontrol unit(MCU),as shown in Fig.2. Its size is11×9cm2.The output sinusoid signal can sweep from1to180MHz.Output1is used in the measurements and output2is for testing.Except the few passive elements,the back side of the TX board is covered with a solid GND plane. It has been found that the metal material of electrodes has little effect on the signal coupled on the body[16].In our design,a 2×2cm2copper electrode is used to couple the signal to the body.It is connected to output1of the TX board with a short cable,as shown in Fig.2(c).The current-mode output of the DDS is terminated with a50-Ωresistor to match the SMA port and the cable.The designed TX source voltage is0.186V rms. With50-Ωloading,the maximum output power is?7.6dBm, which is well within the human safety range[17].

B.Experimental Setup

The experimental setup is shown in Fig.3.Apart from the battery-powered TX board,an Agilent E4404B spectrum ana-lyzer is used as the RX to measure the received signal magnitude. It is connected to Sig RX with a coaxial cable.Sig RX is identical as Sig TX.The arm is70cm above the external GND. Different measurement con?gurations are listed in Table I and described in Fig.4(a)and(b).Measurements are performed with

XU et al.:ELECTRIC-FIELD INTRABODY COMMUNICATION CHANNEL MODELING WITH FINITE-ELEMENT METHOD

707

Fig.3.EF-IBC experimental setup.

TABLE I

M EASUREMENT AND FEM S IMULATION C

ONFIGURATIONS

Fig.4.Measurement con?gurations.(a)M1(d =10cm;sep =4.5cm),M2(d =30cm;sep =4.5cm),and M3(d =100cm;sep =4.5cm).(b)M4(d =30cm;sep =0.5cm),M5(d =30cm;sep =∞),and M6(d =10cm;sep =?4.5cm).

different Sig TX ?Sig RX distance (d ),and different separation (sep)of GND TX above or beneath the arm.Sig RX is ?xed at the wrist while Sig TX is moved to different locations on either arm depending on the distance.The negative separation in Table I indicates GND TX is placed beneath the arm.

M1,M2and M3are performed to investigate the in?uence of body path length.The distance between Sig TX and Sig RX is set to 10,30,and 100cm,respectively.In these measurements,the TX board is placed above the arm with 4.5-cm separation.M4,M5and M6are performed to investigate the effect of body on the return path.GND TX is placed very close to the body with 0.5-cm separation in M4.GND TX is moved far away from the body in M5.In M6,GND TX is put beneath the arm with the 4.5-cm separation.

In experiments,the return path capacitor at the RX side is shorted by the earth-grounded spectrum analyzer.Nonetheless,the return path capacitor at the TX side is well preserved

because

Fig.5.Circuit-coupled FEM model of the EF-IBC channel.

of the ?oating GND TX .In real EF-IBC applications,TX and RX are symmetrical because their return path coupling situations are almost identical.Therefore,the real EF-IBC channel character-istics can be known if the return path coupling at the TX side is well studied with aforementioned experimental setups.

IV .FEM S IMULATION S ETUPS

Three simulation setups are established in this work.The ?rst one models the overall EF-IBC channel in the measurement setup.In order to study the forward body path and the capacitive return path in Fig.1(b)individually,the next two simulation setups are developed to investigate the return path,and to form a circuit model for the body path,respectively.A.EF-IBC FEM Model

ANSYS is used for the FEM investigation [18].The circuit-coupled FEM model of the EF-IBC channel is shown in Fig.5.The system includes air,body,TX circuit,and RX circuit.

The environment is divided into three regions:near-?eld re-gion,transition region and far-?eld region.The outer boundary of the far-?eld region is set to in?nity,and the bottom of the entire model is set as the external GND.The body is in the near-?eld region.The height of the arm above the external GND is 70cm.The body is separated from the external GND with 3cm.Since the air in the near-?eld region is close to the TX and the body path,the electric ?eld is relatively strong.Therefore,the near-?eld air is densely meshed.In the far-?eld region,the elec-tric ?eld gradually decays to zero.Low mesh density is used in this region.The transition region is between the near-?eld region and the far-?eld region.It is meshed with a medium density.The body model is composed of arm,chest,abdomen,and leg,as shown in Fig.6.Relative permittivity and conductivity of different tissue layers can be found in [8]and are plotted in Fig.7for the EF-IBC frequency band.Both arm and leg include concentric layers of skin,fat,muscle,cortical bone,and bone marrow.Cortical bone and bone marrow are excluded for chest and abdomen.Heart is chosen to ?ll the torso cavity because it has conductivity and relative permittivity close to the average of the organs [8].The thicknesses of the tissue layers in different body parts are listed in Table II.

As shown in Fig.6,circuit elements are added to apply the ex-citation and loading conditions.The 0.186-V rms voltage source in series with a 50-Ωresistor is used as TX.RX is modeled as a 50-Ωresistor to the external GND.The metal electrodes are realized by setting nodes of a certain region in the air or the skin

708IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING,VOL.58,NO.3,MARCH

2011

Fig.6.Body model and TX and RX

circuits.

Fig.7.Relative permittivity and conductivity of the tissues at different fre-quencies [8].

TABLE II

T HICKNESSES OF THE T ISSUE L AYERS

(mm)

to the same potential.Sig RX is ?xed at the distance of 20cm from one end of the arm model,which is similar to the wrist position in the experiments.Sig TX is set to different locations on the arm.The distance between Sig TX and Sig RX and the separation from GND TX to the arm are adjusted to match the experiments listed in Table I.

B.Simulation Setups for the Return Path

In this simulation,we aim to ?nd the parasitic coupling ca-pacitance in the return path.An FEM setup with two metal electrodes is established as shown in Fig.8.Two identical metal electrodes are placed to emulate GND TX and GND RX .The cou-pling capacitance to the external GND (C GND )and the cross coupling capacitance (C cross )can be directly extracted by the FEM tool.The in?nity boundary is the extension of the external GND.Here,C GND includes capacitance to the in?nity boundary (C GND1)and capacitance to the ground (C GND2).The distance between the electrodes,the height of the electrodes above the external GND,and the electrode size are the three parameters that affect the return path capacitances.Their effects are studied through

simulations.

Fig.8.Dual-electrode FEM setup for return path capacitance

investigation.

Fig.9.FEM setups to investigate the in?uence of the body on C G N D (d =0).(a)CS1(sep =∞).(b)CS2(sep =4.5cm).(c)CS3(sep =?4.5cm).(d)CS4(sep =0.5cm).

It is important to notice that the involvement of the body can affect the capacitive return path.FEM setups,CS1–CS4,are established to investigate the capacitance change due to different body positions relative to GND TX ,as shown in Fig.9.Sig TX and Sig RX are shorted to remove the body path and study only the return path.In CS1,the body is not involved.GND TX is placed at the height of 82.9cm above the external GND,which is the same as in M1/S1–M3/S3shown in Figs.4and 5.In CS2–CS4,the arm model is inserted.The separation between GND TX and the arm is 4.5,?4.5,and 0.5cm for CS2,CS3,and CS4,respectively.

C.Simulation Setup for the Body Path

In this simulation,we aim to form a circuit model for the body path with unit length.Then,multiple circuit models can be cascaded to model the body path with any length.With the two-port network in Fig.10,Y-parameters (admittance parameters)of the body path can be extracted.Because of its passiveness and symmetry,Y-parameters can be converted to a πcircuit model [19].The body is put above the external GND with the same height as in S1–S6shown in Fig.5.Following this methodology,circuit models of arm and torso are formed in Fig.11(a)and (b),respectively.A 10-cm segment is used as the unit length for the arm model in this work.Indeed,the πmodels in Fig.11have their physical meanings.Since the tissues are capacitive

XU et al.:ELECTRIC-FIELD INTRABODY COMMUNICATION CHANNEL MODELING WITH FINITE-ELEMENT METHOD

709

Fig.10.The two-port network including the body path and the external

GND.

Fig.11.πcircuit model of the two-port network with the body path and the external GND.(a)Arm.(b)

Torso.

Fig.12.RX power in M1/S1(d =10cm;sep =4.5cm),M2/S2(d =30cm;sep =4.5cm),and M3/S3(d =100cm;sep =4.5cm).

and conductive,the distributed RC network of the body path can be modeled by a lumped resistor (R arm /R chest )parallel with a lumped capacitor (C arm /C chest )[6],[20].The coupling from the body to the external GND can be modeled by two lumped leakage capacitors (C leak arm /C leak torso )shunting from the two ends of the body segment.For the torso,the torso impedance (R torso /C torso )is also involved in the shunting path.

V .R ESULTS AND A NALYSES

A.EF-IBC FEM Simulation and Measurement Results Measurements M1,M2,M3and FEM simulations S1,S2,S3are performed to investigate the EF-IBC channel with different body path lengths.The measured and simulated RX power is shown in Fig.12.The difference between the measurement and simulation results at high frequencies is possibly caused by discontinuous interfaces of the cable [19].

The EF-IBC channel gain has a high-pass pro?le.Below 4MHz,the EF-IBC channel gain is almost independent of the body path length and increases by 20dB/dec.This is because the channel is dominated by the capacitive return path.Beyond 4MHz,longer body path leads to higher attenuation due to the higher impedance in the body path and the larger leakage

from

Fig.13.RX power.(a)M2/S2(d =30cm;sep =4.5cm),M4/S4(d =30cm;sep =0.5cm),and M5/S5(d =30cm;sep =∞).(b)M1/S1(d =10cm;sep =4.5cm)and M6/S6(d =10cm;sep =?4.5cm).

the body to the external GND.Measurements on three different persons show little variation,which indicates that the return path impedance plays an important role in the channel.

Measurements M4,M5,M6and FEM simulations S4,S5,S6are performed to show the body effect to the return path.The measured and simulated RX power in M2/S2,M4/S4,and M5/S5is shown in Fig.13(a).When GND TX is moved far away from the arm in M5and S5,the RX power is enhanced by around 14dB nearly in the whole frequency range,compared to M2/S2with the 4.5-cm separation.When the separation is reduced to 0.5cm in M4/S4,the channel gain has a rising slope of 25dB/dec instead of 20dB/dec below https://www.360docs.net/doc/ba15535293.html,pared to M2/S2,the channel loss is about 5dB less at 20MHz with smaller separation.Since the body path is kept the same in all these measurements and simulations,the return path is the only active factor.Therefore,the presence of the body affects the return path capacitance.In M6/S6,GND TX is placed beneath the arm instead of above the arm in M1/S1.As shown in Fig.13(b),the tiny difference between M1/S1and M6/S6indicates that the relative position of GND TX to the body hardly affects the EF-IBC channel characteristics.

B.Investigation on the Capacitive Return Path

The coupling capacitance of the return path is investigated and extracted with the FEM simulation setups in Section IV-B.1)Return Path Capacitance Without Body (sep =∞):The capacitive return path without the body is investigated with the dual-electrode FEM setup in Fig.8.Here,C GND decreases as the electrodes are moved higher above the ground.It is re-duced to a value (C ?nal )when electrodes are higher than 10cm.Without the body,C ?nal(sep=∞)of different electrode sizes are extracted and shown in Fig.14.It is proportional to the side length of the GND plane instead of the area.For the 11×9cm 2GND TX /GND RX ,C ?nal(sep=∞)is about 4.2pF.Also,C cross is less than 0.2pF when the distance between GND TX /GND RX electrodes is larger than 30cm.Therefore,in real EF-IBC appli-cations,the return path is mainly coupled through the external GND.

2)Body Shielding Effect:The body in?uence to the return path is studied with the FEM simulations CS1–CS4in Fig.9.

710IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING,VOL.58,NO.3,MARCH

2011

Fig.14.C ?n al for different electrode sizes with and without body,the C ?n al reduction ratio R with the

body.

Fig.15.Potential distributions in the C G N D setup (d =0).(a)CS1(sep =∞)at 5MHz.(b)CS2(sep =4.5cm)at 5MHz.(c)CS3(sep =?4.5cm)at 5MHz.(d)CS4(sep =0.5cm)at 5MHz.(e)CS4(sep =0.5cm)at 1MHz.(f)CS4(sep =0.5cm)at 10MHz.

The equal-potential lines of the electrical ?eld in CS1–CS4with signal frequency of 5MHz are shown in Fig.15(a)–(d).The equal-potential lines in CS4with different signal frequencies of 1and 10MHz are shown in Fig.15(e)and (f),respectively.The electric ?eld decays to zero uniformly from GND TX to the in?nity boundary and the external GND if the body is not involved in Fig.15(a).Also,C GND consists of C GND1to the in?nity boundary and C GND2to the external GND.When the height is higher than 10cm,C GND1and C GND2are close to 0.5C ?nal .When GND TX is placed above or beneath the arm with

the 4.5-cm separation in Fig.15(b)and (c),the electric ?eld only concentrates on one side of the arm.Equivalently,only C GND1exists in CS2and only C GND2exists in CS3.Moreover,values of C GND1and C GND2also change,which can be seen from the different densities of equal-potential lines in CS2and CS3compared to CS1.Therefore,C GND decreases when the body is involved,which leads to the smaller RX power in M2/S2than in M5/S5as shown in Fig.13(a).

When the separation from GND TX to the body is reduced to 0.5cm in CS4,the electric ?eld partially penetrates the arm to the other side,as shown in Fig.15(d),(e),and (f).Also,C GND2does not vanish completely in these situations.The return path capacitance is larger than that in CS2and CS3.Moreover,the electric ?eld penetration increases with the frequency,which indicates that C GND increases with the frequency.This leads to the higher RX power in M4/S4than in M2/S2and the 25-dB/dec rising slope in M4/S4below 10MHz,as shown in Fig.13(a).3)Return Path Capacitance With Body:Because of the body shielding effect,C GND reaches its minimum with a medium separation of 2–8cm between GND TX and the body.The mini-mum C ?nal is simulated for different GND TX sizes with 4.5-cm separation,and plotted in Fig.14.For the 11×9cm 2GND TX ,C ?nal(sep=4.5cm)is around 1.2pF,which is nearly a quarter of the capacitance without the involvement of the body.The C GND reduction ratio R ,which is the ratio between C ?nal(sep=∞)and C ?nal(sep=4.5cm)in Fig.14,is calculated and plotted in the ?gure.The reduction ratio is bigger for a smaller GND TX .This is because the electric ?eld will be shielded by the body more effectively for a smaller GND TX .This is also observed in experiments.

C.Circuit Model of the EF-IBC Channel

With the extracted return path capacitance plotted in Fig.14and the circuit model of the body path using the Y-parameters extraction method in Section IV-C,the circuit model of the EF-IBC channel is established in Fig.16.The EF-IBC channel circuit model mainly consists of a cascade of unit-length arm models and the torso model.The arm segments near the TX sig-nal injection node have higher impedance than normal arm seg-ments,due to its nonuniform current ?ow.Further,C return TX and C return RX are the return path capacitors at the TX and RX sides,respectively.C return RX can be bypassed with the earth-grounded RX.V TX is set to 0.186V rms .Both R RX and R TX are 50Ω.

EF-IBC channel simulations from the circuit model are com-pared with the FEM model.With earth-grounded RX ( 1in Fig.16),the circuit and FEM simulation results are shown in Fig.17(a)with different distances between Sig TX and Sig RX .The RX power simulated from the circuit model can match the FEM results well.With ?oating TX and RX ( 2in Fig.16),the circuit and FEM simulation results are shown in Fig.17(b).At a short distance of 10or 30cm,the circuit model produces smaller RX power than the FEM model in the whole frequency range because C cross between GND TX and GND RX is ignored in the circuit model.C cross can enhance the capacitive return path effectively at short distances.The circuit model is more

XU et al.:ELECTRIC-FIELD INTRABODY COMMUNICATION CHANNEL MODELING WITH FINITE-ELEMENT METHOD

711

Fig.16.Circuit model of EF-IBC channel ( 1earth-grounded RX; 2?oating

RX).

Fig.17.RX power simulated from the circuit and FEM model of the EF-IBC

channel.(a)Floating TX and earth-grounded RX.(b)Floating TX and RX.

accurate when C cross is negligible at a long distance of 100cm.

VI.C ONCLUSION

An FEM model is established for the ?rst time to investigate the EF-IBC channel in this work.The FEM model of the EF-IBC channel is extensively veri?ed by the measurements.Both the simulations and the measurements are performed up to 180MHz and 1-m distance.In general,simulation results from the FEM model match the measurements well.The EF-IBC channel shows a high-pass pro?le due to the capacitive return path.Characteristics and effects of different components in the EF-IBC channel are quantitatively studied with comprehensive FEM simulation setups.The capacitive return path is critical to the EF-IBC channel.It is found that the return path capacitance depends on the coupling system among the TX/RX GND planes,the external GND,and the body.The body can shield the cou-pling between the TX/RX GND planes and the external GND to reduce the return path capacitance.A medium separation of 4.5cm generates the minimum capacitance,which is around 1.2pF for an 11×9cm 2GND plane.The forward body path can be equivalent to a cascade of π-shaped unit-length arm circuit models and torso circuit model.

Based on the circuit model of the body path and the return path capacitor,a simpli?ed circuit model of the EF-IBC channel is derived to facilitate the transceiver design.EF-IBC channel simulation results from the circuit model can match the FEM model very well.

A CKNOWLEDGMENT

The authors would like to thank Prof.R.D.Murch,Prof.R.Shu-Kwan Cheng,K.W.Chan and Dr.F.Chi-Yuk Chiu for their assistance in the measurements and technical discussions.

R EFERENCES

[1]R.Kohno,K.Hamaguchi,H.Li,and K.Takizawa,“R&D and standardiza-tion of body area network (BAN)for medical healthcare,”in Proc.IEEE Int.Conf.Ultra-Wideband ,2008,vol.3,pp.5–8.

[2]L.Roelens,W.Joseph,E.Reusens,G.Vermeeren,and L.Martens,“Char-acterization of scattering parameters near a ?at phantom for wireless body area networks,”IEEE https://www.360docs.net/doc/ba15535293.html,pat.,vol.50,no.1,pp.185–193,Feb.2008.

[3]Y .P.Zhang and Q.Li,“Performance of UWB impulse radio with planar

monopoles over on-human-body propagation channel for wireless body area networks,”IEEE Trans.Antennas Propag.,vol.55,no.10,pp.2907–2914,Oct.2007.

[4] E.Monton,J.F.Hernandez,J.M.Blasco,T.Herve,J.Micallef,I.Grech,

A.Brincat,and V .Traver,“Body area network for wireless patient moni-toring,”IET Commun.,vol.2,no.2,pp.215–222,2008.[5]IEEE P802.15Working Group.(2009,Apr.).Channel

model for body area network (BAN).[Online].Available:https://https://www.360docs.net/doc/ba15535293.html,/802.15/dcn/08/15-08-0780-09-0006-tg6-channel-model.pdf.

[6]N.Cho,J.Yoo,S.-J.Song,J.Lee,S.Jeon,and H.-J.Yoo,“The human body

characteristics as a signal transmission medium for intrabody communica-tion,”IEEE Trans.Microw.Theory Tech.,vol.55,no.5,pp.1080–1086,May 2007.

[7]J.A.Ruiz,J.Xu,and S.Shimamoto,“Propagation characteristics of

intra-body communications for body area networks,”in Proc.Consumer Commun.and Netw.Conf.,2006,vol.1,pp.509–513.

[8]S.Gabriel,https://www.360docs.net/doc/ba15535293.html,u,and C.Gabriel,“The dielectric properties of biolog-ical tissues:III.Parametric models for the dielectric spectrum of tissues,”Phys.Med.Biol.,vol.41,pp.2271–2293,1996.

[9]N.Cho,L.Yan,J.Bae,and H.-J.Yoo,“A 60kb/s–10Mb/s adaptive fre-quency hopping transceiver for interference-resilient body channel com-munication,”IEEE J.Solid-State Circuits ,vol.44,no.3,pp.708–717,Mar.2009.

[10]H.Y .Chan and J.Yuan,“Current-mode temporal difference CMOS imager

for capsule endoscopy,”in Proc.IEEE Biomedical Circuits and Syst.Conf.,2007,pp.219–222.

[11] C.A.Chestek,P.Samsukha,M.Tabib-Azar,R.R.Harrison,H.J.Chiel,

and S.L.Garverick,“Microcontroller-based wireless recording unit for neurodynamic studies in saltwater,”IEEE Sensors J.,vol.6,no.5,pp.1105–1114,Oct.2006.

[12]K.Hachisuka,Y .Terauchi,Y .Kishi,T.Hirota,K.Sasaki,H.Hosaka,and

K.Ito,“Simpli?ed circuit modeling and fabrication of intrabody commu-nication devices,”in Proc.IEEE Int.Conf.Solid-State Sens.,Actuators,Microsyst.,2005,vol.1,pp.461–464.

[13]M.S.Wegmueller,A.Kuhn,J.Froehlich,M.Oberle,N.Felber,N.Kuster,

and W.Fichtner,“An attempt to model the human body as a communica-tion channel,”IEEE Trans.Biomed.Eng.,vol.54,no.10,pp.1851–1857,Oct.2007.

[14]K.Fujii,M.Takahashi,and K.Ito,“Electric ?eld distributions of wearable

devices using the human body as a transmission channel,”IEEE Trans.Antennas Propag.,vol.55,no.7,pp.2080–2087,Jul.2007.

[15]T.G.Zimmerman,“Personal area networks:Near-?eld intrabody com-munication,”IBM Syst.J.,vol.35,pp.609–617,1996.

[16]K.Hachisuka,A.Nakata,T.Takeda,Y .Terauchi,K.Shiba,K.Sasaki,

H.Hosaka,and K.Itao,“Development and performance analysis of an intra-body communication device,”in Proc.IEEE Int.Conf.Solid-State Sens.,Actuators,Microsyst.,2003,pp.1722–1725.

[17]World Health Organization.(1993).Electromagnetic ?elds (300Hz to

300GHz).[Online].Available:https://www.360docs.net/doc/ba15535293.html,/documents/ehc/ehc/ehc137.htm.

[18]ANSYS,Inc.,“Multiphysics,”Canonsburg,PA.(2009).[Online].Avail-able:https://www.360docs.net/doc/ba15535293.html,.

[19] D.M.Pozar,Microwave Engineering ,3rd ed.Hoboken,NJ:John Wiley

&Sons,2005,p.700.

[20] A.E.Hartinger,R.Guardo,V .Kokta,and H.Gagnon,“A 3-D hybrid

?nite element model to characterize the electrical behavior of cutaneous tissues,”IEEE Trans.Biomed.Eng.,vol.57,no.4,pp.780–789,Apr.2010.

712IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING,VOL.58,NO.3,MARCH

2011

Ruoyu Xu(S’08)was born in Shanghai,China,in 1986.He received the B.E.degree in microelectron-ics from Shanghai Jiao Tong University,Shanghai, in2008.He is currently working toward the Ph.D. degree at the Mixed-Signal Bio-Medical Integrated Circuits Laboratory,Hong Kong University of Sci-ence and Technology,Kowloon,Hong Kong.

His research interests include mixed-signal IC de-sign for biomedical applications and low-power wire-

less transceivers for body area

networks.

Hongjie Zhu(S’08)received the B.E.degree(with honors)in electronic and computer engineering from the Hong Kong University of Science and Technol-ogy(HKUST),Kowloon,Hong Kong,in2009.He is currently working toward the M.Phil.degree at the Mixed-Signal Bio-Medical Integrated Circuits Labo-ratory,HKUST.

His research interests include analog and mixed-signal integrated circuit design,and circuit and sys-tem design related to biological and biomedical

applications.

Jie Yuan(S’03–M’06)received the B.S.degree from

Tsinghua University,Beijing,China,in2000,and the

Ph.D.degree from the University of Pennsylvania,

Philadelphia,in2006,all in electrical engineering.

Since Aug.2006,he has been with the Department

of Electronic and Computer Engineering,Hong Kong

University of Science and Technology,Kowloon,

Hong Kong,as an Assistant Professor.His research

interest includes high-performance mixed-signal IC

design for biomedical applications.More speci?c re-

search interests of his group include ultra low-power data converters,high-performance analog front-ends,and CMOS imaging sen-sors for medical diagnostics.

Dr.Yuan is the Technical Program Committee Co-Chair of IEEE Biomedical Circuits and Systems Conference2009.He is a technical committee member of the IEEE Symposium on Circuits and Systems.He also serves as the Vice Chair of the Biomedical Circuits and Systems Technical Committee.He is the Finance Chair of the IEEE VLSI-SOC conference.