Ultra Wideband Wireless Propagation Channel Characterizations for Biomedical Implants

Size: px

Start display at page:

Download "Ultra Wideband Wireless Propagation Channel Characterizations for Biomedical Implants"

Alfred Davidson
6 years ago
Views:

1 Ultra Widebad Wireless Propagatio Chael Characterizatios for Biomedical Implats Bao-Li Wei, Chu Xiog, Hog-Wei Yue, Xue-Mig Wei, Wei-Li Xu, Qia Zhou, ad Ji-Hai Dua Abstract I order to ispect the feasibility ad safety of wireless commuicatio operated i GHz bad betwee the devices located i vivo ad o body or off-body, the path gai ad specific absorptio rate (SAR) were ivestigated through embeddig a high-resolutio 3D electromagetic model of huma body ito a umerical electromagetic (EM) simulator which is based o fiite itegratio techique (FIT) to solve the Maxwell equatios. Based o the electromagetic (EM) simulatig results, a chael umerical statistical model depictig the i vivo distace-depeded chael path gai was proposed. The experimetal results idicate that it is feasible ad safe for the wireless commuicatio of implatable devices i GHz bad. The i vivo distace-depeded path gai ca be modeled by a modified classical power law fuctio, ad the averaged root-mea-square error (RMSE) betwee the computatioal results of umerical statistical model ad EM simulatio is 9.8. Idex Terms electromagetic model of huma body, electromagetic radioactive, ultra widebad (UWB), fiite itegratio techique (FIT), implatable devices W I. INTRODUCTION ITH the developmet of medical techology, more ad more sophisticated biomedical devices are used for diverse applicatios, such as telemedicie systems, remote health moitorig, body implatable devices, ad so o[1-3]. The physiological data of patiets ca be measured with the aid of electroic devices located o or iside the huma body ad trasmitted wirelessly to a exteral receiver for processig, thereby ehacig the health moitorig, diagosis, ad therapy of the patiets. I-body medical implatable devices are typically battery-operated, ad oce implated, they should operate for several years ad cosume as little power as possible. Furthermore, more ad more advaced implatable devices eed high data rate, low complexity, ad small size. For istace, a multi-chael eural recordig system eeds more tha 1 Mbps data rate [4], ad a capsule edoscope eeds more tha 1 Mbps data rate [5]. The Medical Implat Commuicatio System (MICS) which is i the bad of MHz, has bee applied to implatable medical devices by the US Federal Mauscript received Ja. 14, 214; revised Aug. 1, 214. This work was supported by the Natioal Natural Sciece Foudatio of Chia (611664, , ), ad by the Guagxi Natural Sciece Foudatio (214GXNSFAA118386, 213GXNSFAA19333). B.-L. Wei, C. Xiog, H.-W. Yue, X.-M. Wei, W.-L. Xu, Q. Zhou, ad J.-H. Dua are all with School of Iformatio ad commuicatio, Guili Uiversity of electroic techology, Guili 5414 Chia (Correspodig author: B.-L. Wei, Phoe: , Fax: , guiliwxb@163.com). Commuicatios Commissio (FCC), ad it has suitable propagatio characteristics of goig through huma tissues. However, MICS oly provides a maximum data rate of 8 Kbps, ad it belogs to arrow-bad stadard ad eeds power hugry compoets like voltage cotrolled oscillator (VCO), phase locked loop (PLL) ad aalog to digital coverter (ADC). Due to the iheret characteristics of simple electroics ad high trasmissio speed, ultra widebad (UWB) techology has great potetial to satisfy the eeds of the ext geeratio implatable devices aforemetioed characteristics, ad becomes a promisig cadidate [5]. I particular, impulse radio UWB (IR-UWB), the simplest form of UWB, is a promisig low complexity solutio. A IR-UWB trasmitter which is always implated i the body is simple to desig ad has lower power cosumptio. Cotrarily, complexity is more o the receiver side which is always located o the body or off the body. Moreover, the power cosumptio of IR-UWB scales dow with data rates. The desigs ad architectures of commuicatio system are iflueced by the trasmissio rates, the wave propagatio chael, modulatio ad demodulatio schemes, ad the system power cosumptio [5]. To provide a clear uderstadig of the commuicatio lik betwee the implated device ad the trasceiver located outside the huma body, it is essetial to aalyze the wave atteuatio i high-loss huma tissues ad propagatio i the viciity of the huma body. Moreover, accurate kowledge of the propagatio chael ad strategies to optimize the commuicatio lik quality is ecessary for efficiet desig of implatable UWB wireless commuicatio systems. However, curretly, there is o stadard model of the UWB chael properties for i-body medical implats due to the uavailability of measuremets ad the complexity of simulatios for the frequecy-depedet tissue properties of the huma tissues or orgas [5]. Cosiderable research efforts have bee devoted to characterizig the propagatio characterizatio of radio waves for o-body or i-body devices [2, 6-8]. The characterizatio of the UWB commuicatio lik of chest-to-waist ad chest iside-to-outside was preseted i [6], but the frequecy bad oly covered the lower bad of UWB from 3.4 GHz to 4.8 GHz. The wave propagatio for lie-of-sight (LOS) scearios i the frequecy bads of 1-1 MHz ad 1-5 GHz were evaluated i [2] by usig a model of the huma torso ad, a secod-order polyomial was used to approximate the permittivity of huma tissues calculated from the Gabriel s four-pole Cole-Cole equatio, which simplified the calculatio. However, the permittivities of huma tissues are frequecy-depedet complicated

2 variables ad the same tissue has differet permittivity i differet frequecy. Therefore, it was cursory to evaluate the average power desity i a wide frequecy bad over 1-5 GHz i [2]. I [7], several differet huma tissues electrical parameters were used to defie each layer as a specific dielectric material i a multi-layer model which was implemeted i HFSS (High Frequecy Structure Simulator) to ivestigate the chael path loss of implatable wireless lik. However, the multi-layer model was too simple to well characterize the complicated real huma body. The i-body propagatio characterizatio preseted i [8] oly covered 2.5 GHz ad 3.5 GHz. To more accurately characterize the propagatio chael betwee i-body implated devices ad o-body or off-body devices i the frequecy bad of GHz, this paper performed electromagetic simulatio based o a high resolutio electromagetic model of huma body which was comprised of 85 kids of tissues or orgas ad costructed from livig huma body s computed tomography (CT) ad magetic resoace imagig (MRI) images. The average power desity of electromagetic field i the body ad the specific absorptio rate (SAR) absorbed by the huma body were also explored to ivestigate the validity of implatable commuicatio i this bad. Fially, the distace-depedet path gai for iside body was modeled based o the average power desity. II. ELECTROMAGNETIC HUMAN BODY MODEL AND SIMULATION SCENARIO A. Electromagetic Model of Huma Body Differet from typical wireless commuicatios through the free space, the mai difficulty for i-body UWB chael propagatio modelig is the frequecy-depedet material properties of differet huma tissues ad orgas which must be cosidered i the calculatios. The various tissues ad orgas withi the huma body have their ow uique coductivity, dielectric costat, thickess, ad complex geometry. O the other had, the electromagetic properties of differet huma tissues, such as the permittivity ad coductivity, deped largely o the frequecy ad therefore ca oly be cosidered costat over a arrow rage of frequecies. The complex relative permittivities as a fuctio of agular frequecy for huma body tissues have bee made available based o the four-pole Cole-Cole equatios, by Gabriel [9]: i ˆ( ) (1 ). (1) 1 ( j ) j This equatio was derived from the well-kow Debye expressio, i which is the permittivity at field frequecies with 1, is agular frequecy, i is the static ioic coductivity ad is the permittivity of free space, the distributio parameter,, is a measure of the broadeig of the dispersio, ad is the relaxatio time for each dispersio regio. For a kow relative permittivity, the coductivity ca be obtaied by equatio: 2 2 r 2. (2) f I this equatio, is atteuatio coefficiet, r is relative permittivity, is magetic permeability of free space, ad f is frequecy. It is show from (1) ad (2) that the permittivity ad coductivity are complex fuctios of agular frequecy respectively. Therefore, they ca oly be cosidered costat over a arrow rage of frequecies. I order to ivestigate the characteristics of UWB wireless commuicatio lik i/through the huma body, the body has to be characterized as a medium for wave propagatio, ad a high resolutio electro-magetic model of huma body would be costructed. The accuracy of EM calculatio is determied by the resolutio of the electromagetic huma vowel model. This paper has costructed a electromagetic model of huma body based o CT ad MRI slices data provided by Yale Uiversity [1]. The model was costructed from 498 slices, ad the slices images of torso ad head were take from a livig huma male, the slices of arms ad legs were from the Visible Huma Project (VHP) of Natioal Library of Medicie [11] by Stuchly. A sample of the cut view image is illustrated i Fig. 1(a). The origial images were recostructed i a matrix with a resolutio of 1 mm i the x, y plaes; the z-axis resolutio is 1 cm from eck to mid-thigh ad.5 cm from eck to crow of the head. The proposed 3D electromagetic model of huma body is show i Fig. 1(b). It is with mm 3 resolutio. The more huma tissues or orgas are cosidered, the more the accuracy of the model is achieved. The dielectric ad physiological parameters of 85 kids of differet huma tissues or orgas were cosidered i the electromagetic huma body model of this work. The dielectric properties, i.e., permittivity ad coductivity, ca be foud from website of the Natioal Research Coucil i Italy Istitute of Applied Physics, which was prepared by Adreuccetti et al. Some average characteristics of the tissue materials such as skull, brai, ad muscle come from the website of FCC. The physiological parameters are thermal coductivity ad desity. The thermal coductivity values are directly available for almost all the cosidered tissues from [12]. All these parameters are idexed by image gray value ad mapped to the tissues or orgas i the electromagetic model of huma body. (a) Cut view (b) 3D model Fig. 1. Electromagetic model of huma body. B. Simulatio Sceario The electromagetic model of huma body ca be embedded i umerical electromagetic (EM) simulator which is based o time-domai fiite itegratio techique (FIT) to perform electromagetic computatio. The simulatio sceario withi the electromagetic model of huma body take from livig males is illustrated i Fig. 2.

3 The body was exposed to a icomig plae wave from the back. A total of 324 electric ad magetic field probes were placed outside ad iside the body from 1 mm to 14 mm depth, with space of 1 mm depth ad 2 mm height, respectively. They were located i the middle from left to right side, ad embedded i differet tissues or orgas such as boes, cartilages, blood, ad heart, fat, liver, lug ad muscles. The icidet plae wave was excited with the UWB Gaussia sigal for the correspodig frequecy bads cosidered, herei. All of the field probes are ideally frequecy idepedet with a specified polarizatio, ad do ot have ay couplig amog them. Ope boudary coditio is used for the simulatios, ad therefore the body eviromet reflectios are cosidered. plae wave probes z x y Fig. 2. Simulatio sceario of huma body with electric ad magetic field probes iside III. NUMERICAL SIMULATION Performed EM simulatio by utilizig the aforemetioed simulatio sceario, electric ad magetic field characteristics i the body over GHz bad were obtaied, ad the the average power desity ad specific absorptio rate (SAR) characterizatio were ivestigated, ad the feasibility ad security of wireless commuicatio for implatable devices deep iside the huma body were validated. A. Average Power Desity The US FCC has declared a maximum trasmit power desity of dbm/mhz for UWB. Therefore, eve with the full bad (i.e. 7.5 GHz) of GHz, the maximum trasmitted power is oly -2.6 dbm [13], so the sigal stregth at the receiver becomes importat for achievig high data rate trasmissio, ad the strategies to optimize the UWB commuicatio lik quality are importat too [2]. Average power desity is a measuremet of sigal stregth ad lik quality i the chael. The electromagetic power desity ca be computed by the electric ad magetic field. The Poytig vector (S) ca be used to describe both the power desity ad the directio of wave propagatio, it is depicted as: SEH (3) where E (V/m) ad H (A/m) are the electrical ad magetic field desity, respectively. For icidet plae wave propagatig alog y-axis directio with a horizotal (x-axis) or vertical (z-axis) polarizatio, the power flux desity ca be calculated from the Poytig vector alog the domiat directio (S y (t)) ito the body. Therefore, the average power desity for oe observatio probe is computed by itegratig the Poytig vector alog the domiat directio (S y (t)) over the observatio time [2]. The observatio time was chose to iclude 99% of the sigal power, ad it was approximately 2 s i this work. The average power desity for differet depths was computed by averagig the power desity experieced at each probe o the same plae. The average power desity versus distace from i vitro to back iside is depicted i Fig. 3. It shows that the power desity is reduced by distace from the source due to the high loss of the huma tissues ad orgas, ad the power loss icreases sigificatly with the icrease of trasmitted frequecy ad, with the icrease of propagated depth iside body. Average power desity (dbw/m 2 ) GHz 6. GHz GHz 6.5 GHz 4. GHz 7. GHz 9. GHz 4.5 GHz 7.5 GHz GHz 5. GHz 8. GHz 1. GHz 5.5 GHz 8.5 GHz 1.5 GHz Distace (mm) Fig. 3. Average power desity iside body versus distace for GHz. It ca be see from the results i Fig. 3 that the higher frequecies lead to the greater sigal atteuatio i huma body, makig their receptio almost prohibitive at a deeper distace. However, the UWB commuicatio is still possible for short distaces. For istace, whe implatable devices are implated subcutaeously or the implated depths are less tha 3.5cm ad they have to commuicate with other devices which are either o-body or at short distaces outside, the UWB commuicatio is possible. The typical examples are retial ad cochlear i which implat distaces are a few cetimeters. Assumig a UWB trasceiver with a bad width of 5 MHz ad a maximum trasmitted power desity limit of -41.3dBm/MHz is adopted, the maximum trasmit power is dbm. It ca still bear a path loss of 8 db ad make UWB wireless commuicatio possible for a typical receiver whose sesitivity is -95 dbm. O the other had, sice the huma body is a high lossy medium ad the trasmitter for a implatable device is always embedded iside the body, it should be able to trasmit more tha dbm/mhz while it still accommodates the FCC requiremet outside the body [14]. For istace, based o Fig. 3, the trasmit power level ca be arraged for a typical receiver with a sesitivity of -95 dbm, as illustrated i Fig. 4. The deeper the peetratio depth is (i.e., the deeper the trasmitter is implated ito the huma body), the higher the trasmissio power are required. B. Specific Absorptio Rate (SAR) Recetly, the most recogized radio-frequecy (RF) protectio stadards adopt the SAR as the basic parameter to establish the safety of huma body exposure i the high

4 frequecy bads. Trasmit power level (dbm/mhz) I-body UWB levels Exteral UWB level required by FCC Peetratio distace (cm) Fig. 4. Power level arragemet for trasmitted UWB sigal i body Specific absorptio rate (SAR) is the uit of measuremet for the amout of radio-frequecy eergy absorbed by tissues or orgas whe exposed to electromagetic field. It is defied as the time derivative of the icremetal eergy (dw) absorbed by (dissipated i) a icremetal mass (dm) cotaied i a volume elemet (dv) of a give desity (): d dw d dw SAR. (4) dt dm dt dv Aother mathematical expressio which relates SAR to the electric field is show below: 2 2 p E J SAR, (5) 2 2 where p is power loss desity, J is curret desity, is coductivity, ad E is electric field amplitude i tissue. Values of SAR deped o the icidet field parameters ad the characteristics of the exposed body, i.e., its size ad exteral ad iteral geometry, ad the dielectric properties of the various tissues. The huma body model allows the evaluatio of local SAR as averaged over 1 g ad 1 g all alog the body. The SARs i differet frequecies were evideced by performed EM simulatios. The max SAR, average SAR over 1 g ad average SAR over 1 g absorbed by huma body i GHz bad are listed o Table I. As the guidelies from the Iteratioal Commissio o No-Ioizig Radiatio Protectio (ICNIRP) [15]: the TABLE I SUMMARY OF SAR BY HUMAN BODY IN GHZ BAND Frequecy (GHz) Max. SAR (W/kg) 1g SAR (mw/kg) 1g SAR (mw/kg) localized SARs of head ad truk for huma body i 1 MHz-1 GHz frequecy bad are uder 2 W/kg. It ca be see from Table I that the SAR values absorbed by huma body i GHz bad are uder the limit of ICNIRP, so usig this bad for implatable devices wireless commuicatio is secure. IV. CHANNEL MODELING The path gai is a measure for the average atteuatio of the sigal from trasmitter to receiver. It obviously depeds o the distace betwee receiver ad trasmitter ad, together with the trasmit power, determies the rage of a wireless system: for distaces beyod a certai rage, the received sigal becomes so week that proper receptio is ot possible aymore [16]. The path gai modelig is eeded to estimate the sigal eergy at differet distaces from the source. The distace depedet path gai for both idoor ad free space were modeled with the classical power law (the logarithmic fuctio) [16] give as Gp( d) Gp 1log d 1 d, (6) where, G p is a power scalig costat, d is the distace betwee the ed ad source, ad correspods to the path gai factor (propagatio expoet). The path gai factor depeds o the eviromet i which the system operates, for free space propagatio, =-2 is valid, the typical values for lie-of-sight (LOS) are o the order of -1.5, ad for o-los o the order of -3 to -4. Due to the ifluece of lossy tissues ad orgas, the path gai for both outside ad iside the huma body calculated from (6) does ot fit with the simulated data i Fig. 3. Hece, the classical power law is modified as ( d d) Gp( d) Gp 1log 1 d. (7) The EM simulatio results ad umerical computatio results from (7) are depicted i Fig. 5. It ca be see that the proposed umerical model produces a good approximatio. The miimum ad average root-mea-square error (RMSE) is 4.3 ad 9.8, respectively. All of these fitted path gai factor (), power scalig costat (G p ) ad d for bad of GHz are tabulated i Table II. It ca be see that the path gai factors () i huma body are -6.5 to -11. for GHz bad, which idicates that the higher frequecy is trasmitted, Average power desity (dbw/m 2 ) EM simulatio results Modelig results 3. GHz GHz 6. GHz 3.5 GHz 6.5 GHz 4. GHz 7. GHz 9. GHz GHz 7.5 GHz 9.5 GHz 5. GHz 8. GHz 1. GHz 5.5 GHz 8.5 GHz 1.5 GHz 1.5 GHz Distace (mm) Fig. 5. EM simulatio results ad modelig results of average power desity iside body versus distace

5 TABLE II FITTED PATH GAIN FACTOR (), POWER SCALING CONSTANT (G P ) AND d FOR GHZ Frequecy (GHz) G p d the severer power loss is udergoe. V. CONCLUSIONS By performig electromagetic (EM) simulatio based o a electromagetic huma model, it is foud that the max SAR, average SAR over 1 g ad over 1 g absorbed by huma body i GHz bad are i accordace with the criterios of ICNIRP, substatiatig the security of usig this bad for wireless commuicatio of implatable devices. The loss of average power desity at the differet depths of huma body idicates the feasibility of usig GHz UWB bad for the wireless commuicatio of implatable devices. For devices with implated depth less tha 3.5 cm, the full bad GHz of UWB ca be applied to wireless commuicatio while it still accommodates the requiremet of FCC. For further deep implatable devices, the lower sub-bad of UWB should be applied to wireless commuicatio to accommodate the requiremet of FCC, or the devices i the body should trasmit more tha dbm/mhz to guaratee the proper commuicatio ad make the sigal level accommodate FCC s requiremet outside the body. Fially, the trasmit power level for typical receiver has bee arraged. A modified classical power law fuctio ca be adopted to depict the i vivo distace-depeded path gai. The chael umerical statistical model describig the i-body propagatio chael characterizatios has a simple form, makig it uecessary for the desigers of implatable devices to use a complexity EM simulatio which causes extremely log computatioal time. [4] M. S. Chae, Z. Yag, M. R. Yuce, L. Hoag, ad W. Liu, A 128-chael 6 mw wireless eural recordig IC with spike feature extractio ad UWB trasmitter, IEEE Trasactio o Neural System ad Rehabilitatio Egieerig, vol. 17, o. 4, pp , Aug. 29. [5] A. Khaleghi, R. Chávez-satiago, ad I. Balasigham, Ultrawidebad pulse-based data commuicatios for medical implats, IET Commuicatios,Vol. 4, o. 15, pp , Oct. 21. [6] Q. Wag, K. Massami, ad J. Q. Wag, Chael modelig ad BER performace for wearable ad implat UWB body area liks o chest, i Proc. IEEE Iteratioal Cof. o Ultra-Widebad, New Jersey, 29, pp [7] H. Bahrami, B. Gosseli, ad L. A. Rusch, Realistic modelig of the biological chael for the desig of implatable wireless UWB commuicatio systems, i Proc. 34th Aual Iteratioal Coferece of the IEEE Egieerig i Medicie ad Biology Society, Sa Diego, 212, pp [8] B-L Wei, H-W Yue, Q. Zhou, W-L Xu, X-M Wei, ad J-H Dua, Modelig of i-body propagatio characterizatio for 2.5/3.5 GHz implatable devices applicatios, Applied Mechaics ad Materials, Vol. 273, pp , Ja [9] S. Gabriely, R. W. Lau, ad C. Gabriel, The dielectric properties of biological tissues: III, parametric models for the dielectric spectrum of tissues, Physics i Medicie ad Biology, vol. 41, o. 11, pp , Nov [1] I. G. Zubal, C. R. Harrell, E. O. Smirh, Z. Ratter, G. Gidi, ad P. B. Hoffer, Computerized three-dimesioal segmeted huma aatomy, Medical Physics, vol. 21, o. 2, pp , Feb [11] M. J. Ackerma, Viewpoit: The visible huma project, Joural Biocommucatio. Vol. 18, o. 2, pp , Aug [12] K. R. Holmes, Thermal coductivity data for specific tissues ad orgas for humas ad other mammalia species, [Olie]. Available: [13] L. Zhao, S-P Liu, L. Ji, ad J-Y Che, Research o the power spectral desity limit i UWB systems, Telecommuicatios Sciece, vol. 27, o. 2, pp , Feb [14] M. R. Yuce, H. C. Keog, ad M. S. Chae, Widebad commuicatio for implatable ad wearable systems, IEEE Trasactios o Microwave Theory ad Techiques, vol. 57, o. 1, pp , Oct. 29. [15] Iteratioal Commissio o No-ioizig Radiatio Protectio, For limitig exposure to time-varyig electric, magetic ad electromagetic fields (up to 3GHz), Health Physics, vol. 74, o. 4, pp , Apr [16] A. F. Molisch, Ultra-wide-bad propagatio chaels, Proceedig of the IEEE. vol. 97, o. 2, pp , Feb. 29. REFERENCES [1] O. Novak, C. Charles, ad B. R. Brow, A area ad power efficiet I-UWB trasmitter for biomedical applicatios implemeted i 65 m CMOS techology, i Proc. IEEE Cof. o Biomedical Circuits ad Systems, Sa Diego, 211, pp [2] A. Khaleghi, I. Balasigham, ad R. Chávez-satiago, Computatioal study of ultra-widebad wave propagatio ito the huma chest, IET Microwaves Ateas Propagatio, vol. 5, o. 5, pp , May 211. [3] E. G. Lim, J. C. Wag, Z. Wag, T. Tillo, ad K. L. Ma, The UHF bad i-body ateas for wireless capsule edoscopy, Egieerig Letters, vol. 21, o. 2, pp. 72-8, May 213.

A New Space-Repetition Code Based on One Bit Feedback Compared to Alamouti Space-Time Code

A New Space-Repetition Code Based on One Bit Feedback Compared to Alamouti Space-Time Code Proceedigs of the 4th WSEAS It. Coferece o Electromagetics, Wireless ad Optical Commuicatios, Veice, Italy, November 0-, 006 107 A New Space-Repetitio Code Based o Oe Bit Feedback Compared to Alamouti