Marc Simon Wegmueller, Student Member, IEEE, Sonja Huclova, Juerg Froehlich,Michael Oberle, Student Member, IEEE, Norbert Felber, Niels Kuster, and Wolfgang Fichtner, Fellow, IEEE
Abstract—Galvanic coupling has been shown to be the bestmethod for low-power on-body data transmission. This approachhas been extended toward wireless communication between im-planted devices, building a real intrabody sensor network. Toshow the feasibility, a test system has been developed and im-plemented. The measurements are compared with the numericalsimulations. The prototype offers four concurrent channels with athroughput of 4.8 kb/s. The main focus is the future implantabilityof such a miniaturized system for the medical long-term surveil-lance of patients. To achieve this goal, small circuit size, low powerconsumption, and electrical safety have to carefully be considered.
Index Terms—Biomedical sensor, body area network, bodycharacterization as transmission medium, galvanic coupling,intrabody communication, measurement system.
I. INTRODUCTION
CONTROL of the implanted devices by a vital parameterrecorded at different locations within the body requires
communication between different sensor nodes. Due to thedielectric properties of tissue, such a device has to operate in anonhomogeneous highly attenuating environment. In addition,the patient’s safety has to be ensured, e.g., the induced currentsshould not cause nerve stimulation and should not interfere withthe body signals located in the frequency range of operation.Therefore, signals at frequencies below 10 kHz have to beavoided. In [1], galvanic-coupled units have been tested in tis-sue. Recent studies have focused on intrabody communication,with the implanted miniaturized pills operating as transmitterand receiver by deploying coils [2] and volume conduction(galvanic coupling) [3]–[5]. Different coupling strategies, e.g.,galvanic-coupled electrodes, dipole antennas, and coils, havebeen evaluated and compared in [6].
This paper investigates the technology of galvanic-coupledimplantable pills, as shown in Fig. 1. The differential electrodesthat build the two ends of the pills are used to couple thealternating current into the human tissue. Since the conductivity
Manuscript received July 5, 2007; revised November 25, 2008. First pub-lished April 24, 2009; current version published July 17, 2009. The AssociateEditor coordinating the review process for this paper was Dr. Robert Gao.
M. S. Wegmueller, N. Felber, and W. Fichtner are with the IntegratedSystems Laboratory (IIS), Swiss Federal Institute of Technology (ETH) Zurich,8092 Zurich, Switzerland (e-mail: [email protected]).
S. Huclova and J. Froehlich are with the Electromagnetic Fields and Mi-crowave Electronics Laboratory (IFH), ETH Zurich, 8092 Zurich, Switzerland.
M. Oberle and N. Kuster are with the Foundation for Research on Informa-tion Technologies in Society (IT’IS), ETH Zurich, 8092 Zurich, Switzerland.
Color versions of one or more of the figures in this paper are available onlineat http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TIM.2009.2015639
Fig. 1. Distributed implanted sensors: miniaturized sensor pills communicatewith each other. The implantable cardioverter defibrillator and its lead couldserve as a particular sensor of the network.
of the muscle tissue is higher compared with the conductivityof the skin, lower communication losses will result comparedwith skin-mounted electrodes. According to previous feasibil-ity measurements [7], the frequency range between 100 and500 kHz has been investigated. The attenuation in tissue ismeasured in a phantom used for testing body-mounted wirelessdevices filled with tissue-simulating liquid. The measurementresults are compared with numerical simulations. Based onthese results, a simple model of a human torso is computedto estimate the attenuation in a more realistic scenario. Basedon the experimental results, a hardware prototype was realized,and the achievable bandwidth has been assessed. The patient’ssafety is ensured by limiting the peak current to 1 mA.
The design is targeted to allow for medical long-term sur-veillance of patients. Therefore, the devices shall be designedin view of a low-power consumption, miniaturization toward animplantable pill size, and flexibility for signal recovery counter-vailing the high-channel attenuation. The system aims for datacommunication between one central master and multiple slavesensors. The data link to external monitoring devices will beorganized from the master sensor node with a standard wirelesscommunication link. However, the specification of that remotecommunication interface is beyond the feasibility presented inthis paper.
This paper is organized as follows. In Section II, the ap-proach of galvanic coupling is introduced. Section III showsthe numerical simulations, and Section IV discusses the mea-surement setup and results. Furthermore, Section V presents thedeveloped prototype, which is a reliable low-power data linkbetween four transmitter units and one receiver unit. Section VI
WEGMUELLER et al.: GALVANIC COUPLING ENABLING WIRELESS IMPLANT COMMUNICATIONS 2619
Fig. 2. Photo of the two pill models on the right-hand side, and schematics of the construction details on the left-hand side. The two versions used had thedimensions: A = 0.4 cm, B = 5.0 cm, C = 1.0 cm for the longer electrodes, and C = 0.0 cm for the short electrodes.
Fig. 3. Numerical model of (top left) the elliptical and (top right) the torso phantom including (bottom) the distribution of the electric potential for theconfiguration of the long transmitter and receiver pills.
summarizes the conclusions of this paper and discusses possibleapplications.
II. GALVANIC COUPLING WITH IMPLANTS
The measurement system described in [8] has been applied toinvestigate the feasibility of communication between implants.
The experimental setup couples a sinusoidal output current upto 1 mA in a frequency range of 10 kHz to 1 MHz into a muscle-tissue-simulating liquid (MSL).
For the initial experiments, an oversized pill model withconnectors on both ends has been realized. As shown in Fig. 2,two versions of electrode pairs have been investigated: longelectrodes of a cylindrical copper (length: 1 cm and diameter:
2620 IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 58, NO. 8, AUGUST 2009
TABLE IMATERIAL PARAMETER USED FOR SIMULATION AND MEASUREMENT [10]
4 mm) and short electrodes with only the circular surface toprovide contacts (4 mm). For better comparison with earlieron-skin electrode measurements [9], the distance between theelectrode contacts was set to 5 cm. The electrode cables areshielded. The shields are connected to the reference potentialon the measurement system units.
III. NUMERICAL SIMULATION
Simulations of the pill–pill communication have been per-formed using the commercial Comsol Multiphysics finite-element-based simulation platform. The dimensions of thenumerical model are much smaller than the wavelength of thefrequency range considered; therefore, the quasi-static simula-tion mode was used.
The following steps were carried out by numericalsimulation.
1) Validation of the experimental setup including the twoversions of electrode pairs.
2) Evaluation of the dependence on tissue dielectric parame-ters within the frequency range from 10 kHz to 27 MHz.
3) Evaluation of attenuation using a realistically sized pill(B = 1 cm).
4) Evaluation of the influence of phantom size on attenua-tion using a phantom having “torso-like” dimensions.
The numerical model of the elliptical phantom (Fig. 3)reflects the measurement setup regarding the dimensions(60 × 40 × 10.5 cm) and the dielectric parameters of theliquid (Table I). The muscle tissue parameters given byGabriel et al. [10] were used for the simulations at discretefrequencies. The pills were modeled as two copper elementsplaced on a dielectric rod. A current flow was used as excita-tion. The boundaries of the model are electrically insulating.Different distances (5, 10, 15, and 20 cm) between the two pillswere computed. Simulations for two long (C = 1 cm) and twoshort (C = 0 cm) electrode pills were separately performed.The scaled pills are a factor of 5 smaller compared with the longelectrode pills. The attenuation was calculated via integrationof the potential over the corresponding surfaces. In Fig. 4,the numerical simulations are compared with the measurementresults of the two electrode configurations at 300 kHz. Inaddition, the attenuation of a homogeneous elliptical cylinder(32 × 26 × 58 cm) having the same dielectric parameters rep-resenting a simplified torso model is shown. The dimensionsof the elliptical cylinder are approximate values based on themeasurements taken from Dutch adults of age between 31 and60 years [11].
Fig. 4. Comparison of simulated results and typical measurements of theattenuation for galvanic coupling of long, short, and scaled pill types at 10 kHz,300 kHz, 1 MHz, and 27 MHz. In addition, the simulated attenuation in thesimplified torso model is shown.
Fig. 5. Image of the measurement setup with the long electrodes inserted intothe measurement phantom.
Fig. 4 shows the simulation results. The influence of thedielectric parameter variation is not significant for the inves-tigated frequencies at 10 kHz, 300 kHz, 1 MHz, and 27 MHz.However, the different active sizes of the long and short elec-trodes result in an attenuation difference of more than 10 dB.Furthermore, the scaled pills perform significantly lower. Theattenuation in the torso model suggests a higher attenuationwithin the body compared with the experimental phantom. Thishas to further be investigated by using an anatomical model thatrepresents the different organs and tissues.
IV. MEASUREMENTS
A. Measurement Setup
Feasibility measurements have been performed in an ellipti-cal phantom (60 × 40 × 10.5 cm) that is used for compliancetesting of body-mounted wireless devices (Fig. 5, [12]) filledwith a tissue-simulating liquid MSL27. The dielectric propertiesof the liquid correspond to muscle tissue at 27 MHz. No liquidwas available representing the muscle tissue for the lowerfrequency range. The pills were placed 5 cm below the liquidlevel by specially designed plastic holders.
WEGMUELLER et al.: GALVANIC COUPLING ENABLING WIRELESS IMPLANT COMMUNICATIONS 2621
Fig. 6. Measured gain of the galvanic-coupled electrodes with long and short electrodes.
TABLE IIPILL ORIENTATION MEASUREMENTS: TRANSMITTER AND RECEIVER PILL
ANGLES AND RESULTING ATTENUATION
Measurements of the attenuation were conducted with in-creasing distance between the transmitting and receiving pills,i.e., starting at 5 cm and ending at 45 cm in steps of 5 cm. Tokeep the environmental noise level to a minimum, the tests wereperformed inside an anechoic chamber.
B. Measurement Results
The firmware of the transmitter was configured to generatesinusoidal waves at the following frequencies: 100, 200, 250,300, 400, and 500 kHz. The measurement results for optimallyaligned electrodes are shown in Fig. 6. The difference betweenthe two electrode types is 15 dB; the transmission characteris-tics for different distances have been shown to be similar.
Additional measurements have been conducted to determinethe worst-case condition of transmission for this specific sce-nario of galvanic coupling. Table II shows the combination oftransmitter and receiver rotations. The rotations in Z-directionby 0◦ or 90◦ were chosen to simulate an unfavorable alignment.These measurements were conducted with the long electrodes.The results are summarized in Table II. When only the transmit-ter electrode is rotated, a signal loss of about 15 dB is observed,whereas a rotation of only the receiver electrode did not allowfor the detection of signals above the noise level because theincoming E-field is normal to the receiving gap.
Finally, the bit error rate (BER) has been calculatedaccording to
BER =12
erfc
(√Eb
N0
), Eb =
A√
Tb√2
Fig. 7. BER estimations for galvanic coupling with the long electrodes.
where Eb is the signal energy per bit derived from measureddata, A is the signal amplitude, Tb is the duration of 1 bit, andN0 is the mean noise power spectral density of the receivedsignal.
Fig. 7 shows that the BER is below 10−4 up to 35-cm mea-surement distance for all frequencies. It must be considered thatthe calculation of the BER depends on the chosen modulationscheme and the duration of one symbol. Therefore, the BERplot for BPSK presented here can only be used as a roughestimate.
V. SYSTEM DESIGN
The almost constant attenuation within the frequency rangeconsidered allows for the easy implementation of a frequency-division multiple-access (FDMA) system without additionalband equalization. The system approach considers a unidi-rectional communication system with four transmitters and acentral receiver. This choice offers more freedom for trans-mitter separation compared with time-division multiple access(TDMA). The latter has not been chosen due to the fact that slot
2622 IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 58, NO. 8, AUGUST 2009
TABLE IIIGENERATION OF DIFFERENTIALLY ENCODED DATA: THE ENCODED
SEQUENCE IS INITIALIZED BY 1
negotiation and synchronization would require bidirectionalcommunication and additional effort in synchronization. Thecarrier frequencies of the four channels in this application arelocated at 100, 150, 200, and 250 kHz.
For demonstration purposes, the electrical decoupling of thesensor from the power net is realized with optoelectric couplersand battery powering. The power for the transmitter and thereceiver units is supplied by a battery at 3.3 V.
A. Digital Communication Methods
Differential binary phase-shift-keying (DBPSK) modulationhas been chosen due to the high immunity to amplitudevariations and the high hardware efficiency. It incorporatesdifferential encoding, including an error correction code andconsecutively binary phase-shift modulation.
Differential Encoding and Error Correction Code: For dif-ferential encoding, the transmission is independent of the ab-solute carrier phase. The transmitted information is encoded inthe phase difference between two consecutive symbols. Hence,the polarity of the implants does not affect the transmission.Therefore, much more freedom is offered to the surgeon forplacing the units in the body.
The encoding process is illustrated in Table III. Every binary“0” data bit flips the state of the last encoded bit, every “1”leaves it unchanged, and the same value is encoded again. Thus,the information is encoded in the phase shift of the carrier waverather than the phase itself.
In addition, a hamming code for the detection and correctionof one faulty received bit per data byte has been realized. Perdata byte, four additional hamming bits have to be included intothe data stream. Therefore, a transfer rate of 7.2 kb/s is requiredto achieve the desired data throughput of 4.8 kb/s.
Modulation Scheme: In addition to its high immunity toamplitude variations, BPSK also offers low complexity formodulation and demodulation. The carrier frequency of thecorresponding transmitter unit is modulated with the encodedinformation sequence. In DBPSK, the transmission of an en-coded “1” results in no phase shift, whereas a “0” shifts thephase by 180◦.
Pulse Shaping: The signal-to-noise ratio of each channel canbe improved by shaping the modulated signal by a windowfunction. Ideally, the sinc function
Asincpulse =sinc(x)=sin(π · x)
π · x for −∞ ≤ x ≤ ∞ (1)
transforms to a rectangle in Fourier space. Due to hardwarelimitations, the filter length of the pulse-shaping window islimited. With a sampling rate of 1 MHz and a bit rate of 7.2 kb/s,one bit pulse contains 140 samples. The sinc window lengthshall match the number of samples per bit. Therefore, the
Fig. 8. Normalized spectrum of the BPSK-modulated signal with and withoutpulse shaping using a center frequency of 100 kHz, a bit rate of 7.2 kb/s, and asample rate of 1 MHz.
Fig. 9. Block diagram of the transmitter. Serial data are DBPSK modulatedand differentially coupled to the pill electrodes by a current amplifier.
sinc function is cut at x = −1, . . . , 1. The spectrum of sinc-shaped BPSK is not ideally rectangularly shaped, but possessessidelobes. Regarding a signal bandwidth of 50 kHz per channel,the spectrum of the pulse-shaped signal at fcenter ± 25 kHz ismore than 20 dB lower than the unshaped BPSK spectrum, asshown in Fig. 8. Therefore, the pulse shaping will result in asignificant improvement of the signal-to-noise ratio if multipletransmitters are simultaneously active.
B. Functional Blocks of the Transmitter
The transmitter is composed of a minimal number of oper-ational units, as shown in Fig. 9. A complex programmablelogic device (CPLD) contains all the digital signal-processingunits of the transmitter. That is, CPLD is perfectly suitablefor the transmitter complexity. It will be replaced by a futureapplication-specific integrated-circuit (ASIC) integration.
As shown in Fig. 9, a serial interface serves to configurethe digital unit and provides data for transmission. The highestcarrier frequency of 250 kHz defines the system clock of1 MHz, which provides a four-times oversampling. The dig-ital part implemented in CPLD feeds data through an 8 bitdigital-to-analog converter (DAC) to the analog output driver. Aparallel-input DAC is used since a serial input would require aneight-times-higher system clock frequency for data transmis-sion, which results in higher power dissipation. To save energywhile no data are sent, the entire analog part can be shut down.
WEGMUELLER et al.: GALVANIC COUPLING ENABLING WIRELESS IMPLANT COMMUNICATIONS 2623
Fig. 10. Schematic of the differential current amplifier: (left part) current regulators driven by the DAC and (right part) constant current sources.
Fig. 11. Block diagram of the receiver: received signals are bandpass filtered, amplified, analog–digital converted, digitally demodulated, and decoded for serialdata transmission.
Fig. 12. Simplified input amplifier stages for implant communication: (left part) instrumentation amplifier structure, (middle, top: high pass; middle bottom:low pass) second-order Sallen–Key filters, and (right part) final noninverting amplifier.
The analog part is presented in Fig. 10. The two symmetricalbranches include the current regulators and the constant currentsources. The architecture is a simplified version of the designpresented in [8]. The signal filtering in the current regulators isomitted, and the discrete elements are adapted.
C. Functional Blocks of the Receiver
Fig. 11 shows the main functional blocks of the receiverwith the analog input amplifier and filter, the ADC, and thefield-programmable gate array (FPGA), which provides the dig-ital demodulation and interfaces. The input amplifier of thereceiver must amplify very weak signals to a level resolvable
by the following 14-bit ADC. The amplitude of the input signalbetween the receiver electrodes ranges from about 5 mV at 5 cmdown to about 5 μV at 45-cm distance. The primary goal for thedesign of the amplifier is minimal noise, and the secondary goalis low-power consumption.
The amplifier is built in three stages, as shown in Fig. 12.The analog design is based on the structure of the initial designin [8]. Each of the amplifier sections is designed to amplify theincoming signal by a factor of 10. The instrumentation amplifierof the first stage is simplified by omitting the passive inputfilters.
The bandpass filter in the second stage has been imple-mented as a Sallen–Key structure. The discrete elements are
2624 IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 58, NO. 8, AUGUST 2009
Fig. 13. Transmitter and receiver PCB: four layers, 16 cm2, respectively,29.5 cm2.
TABLE IVTECHNICAL SPECIFICATIONS OF THE IMPLANTABLE TRANSMITTER
AND RECEIVER DEMONSTRATOR SYSTEM
redimensioned for much smaller capacitors. The values of thefilter elements can be calculated according to
C1 =2 · C2 (2)
R =√
24 · π · fc · C2
(3)
where fc is the 3-dB cutoff frequency, and the ratio of C1 andC2 is chosen to be 1 : 2. The cutoff frequencies are 75 kHz forthe high-pass filter and 350 kHz for the low-pass filter.
The third stage contains a noninverting amplifier. The gain isfixed to 10, which reduces the hardware complexity but losesthe reconfigurability of the gain stage. All stages have a totalamplification of 60 dB.
The digital part of the receiver is implemented in FPGA. Asshown in Fig. 11, the FPGA receives the digitally convertedsignal, transmits the demodulated data to the serial interface,and provides a user interface. DBPSK demodulation is imple-mented for up to four concurrently transmitted signals. TheFPGA’s built-in digital clock managers, multipliers, and datastorage structures offer convenient features for data decoding.The FPGA-based prototype allows the easy migration of thearchitecture into a system-on-chip (SoC).
VI. IMPLEMENTATION RESULTS
The functional prototype of the sensor network consists offour transmitter boards and one receiver board (Fig. 13). Thekey parameters of the two units are given in Table IV.
The current output of the transmitter unit drives 1 mA up to aload of 1 kΩ. This is by far sufficient considering the couplingmeasurements of the muscle tissue. The modulated signals arein the frequency range between 100 and 250 kHz. Based on the
TABLE VUNCERTAINTY VALUES
measurements, a constant channel attenuation over the inves-tigated frequency range can be expected because of insignif-icant tissue parameter variations. The data are differentiallyencoded and BPSK modulated. The implemented receiver isconcurrently capable of signal filtering and demodulating allfour channels. Thus, a bit rate of 4.8 kb is provided for all fourtransmitters.
The total transmitter unit has a power consumption of only16.5 mW, and the receiver dissipates 400 mW. The high powerconsumption of the receiver unit is caused by the power dissi-pation of 270 mW by the ADC device.
The size of the transmitter board is 16 cm2, whereas thereceiver occupies 29.5 cm2. Further miniaturization is feasiblethrough integration of the transmitter units as mixed-signalASIC to achieve the implantable size of a pill.
The receiver is able to demodulate the data of a singletransmitter up to a distance of 45 cm between the transmitterand receiver electrodes. Using a signal band filter, a 30-dB-higher channel attenuation of one transmitter compared withthe three other transmitters has still allowed to detect its signalin simulations. Therefore, a distance variation between thetransmitters and the receiver of 35 cm will be feasible.
The observed changes in attenuation given in Table V due tothe differences in tissue-simulating liquid parameters and theuncertainty of the readout electronics are much smaller than theexpected change in attenuation due to deviations in positioning.Therefore, the expected absolute maximum uncertainty of theevaluated attenuation is in the range of ±0.5 dB.
In conclusion, the attenuation between galvanic-coupled pillshas been measured in MSL. The transmitted sinusoidal signalis detectable at distances of up to 45 cm and given differentelectrode orientations. The attenuation depends on the trans-mission distance and the length of the conductive part of theelectrodes. The longer electrodes performed better than theshorter electrodes. Numerical simulation of a simplified modelof a human torso suggests a higher attenuation in this kind ofgeometry. However, this has to further be investigated usinganatomical models. The functional prototype for wireless datatransmission in muscle tissue provides four concurrent channelswith a throughput of 4.8 kb/s. The low system clock frequencyand the low logic complexity allow for an efficient low-powerdesign suitable for implant communication in biomedical mon-itoring applications, e.g., temperature, SpO2, and ECG sensors.
REFERENCES
[1] D. P. Lindsey, E. L. McKee, M. L. Hull, and S. M. Howell, “A newtechnique for transmission of signals from implantable transducers,” IEEETrans. Biomed. Eng., vol. 45, no. 5, pp. 614–619, May 1998.
[2] H. Sawan, Y. Hu, and J. Coulombe, “Wireless smart implants dedicatedto multichannel monitoring and microstimulation,” IEEE Circuits Syst.Mag., vol. 5, no. 1, pp. 21–39, First Quarter 2005.
WEGMUELLER et al.: GALVANIC COUPLING ENABLING WIRELESS IMPLANT COMMUNICATIONS 2625
[3] J. Schulman, J. Mobley, J. Wolfe, E. Regev, C. Perron, R. Ananth,E. Matei, A. Glukhovsky, and R. Davis, “Battery powered BION FESnetwork,” in Proc. 26th Annu. Int. Conf. IEEE Eng. Med. Biol. Soc., 2004,pp. 4283–4286.
[4] J. H. Schulman, J. Mobley, J. Wolfe, H. Stover, and A. Krag, “A 1000+channel bionic communication system,” in Proc. 28th Annu. Int. Conf.IEEE EMBS, New York, Sep. 2006, pp. 4333–4335.
[5] N. Yao, H.-N. Lee, R. Sclabassi, and M. Sun, “Low power digital com-munication in implantable devices using volume conduction of biologicaltissues,” in Proc. IEEE EMBC, New York, 2006, pp. 6249–6252.
[6] M. S. Wegmueller, M. Oberle, N. Felber, N. Kuster, and W. Fichtner, “In-vestigation on coupling strategies for wireless implant communications,”in Proc. IEEE IMTC, Warsaw, Poland, 2007, pp. 1–4.
[7] M. Hediger and T. Kaufmann, “Wireless implant communications,”Diploma thesis, IIS/ETH Zurich, Zurich, Switzerland, 2006.
[8] M. Wegmueller, A. Lehner, J. Froehlich, R. Reutemann, M. Oberle,N. Felber, N. Kuster, O. Hess, and W. Fichtner, “Measurement systemfor the characterization of the human body as a communication channelat low frequency,” in Proc. IEEE EMBC, Shanghai, China, 2005,pp. 3502–3505.
[9] 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.
[10] S. Gabriel, R. W. Lau, and C. Gabriel, “The dielectric properties of biolog-ical tissues: II. Measurements in the frequency range 10 Hz to 20 GHz,”Phys. Med. Biol., vol. 41, no. 11, pp. 2251–2269, Nov. 1996.
[11] TU Delft, Section Applied Ergonomics and Design, DINED Data Set.[Online]. Available: http://www.dined.nl
[12] Schmid & Partner Engineering AG (SPEAG). [Online]. Available:http://www.speag.com/measurement/phantomsnliquids/eli4.php
Marc Simon Wegmueller (S’05) was born in Bern,Switzerland, in 1977. He received the Diplomadegree in 2002 from the Swiss Federal Instituteof Technology Zurich (ETH), Zurich, Switzerland,where he is currently working toward the Dr. sc.degree.
After an internship in the field of vital monitoringsystems and ASIC integration with Miromico AG,Zurich, he joined as a Research Assistant the Inte-grated Systems Laboratory (IIS), ETH, in 2003. Hisresearch interests include the design of VLSI circuits
and systems and digital signal processing for medical communications.
Sonja Huclova was born in Bratislava, Slovakia, in1980. She received the M.Sc. (Dipl.Chem.) degreein chemistry from the University of Bern, Bern,Switzerland, in 2005. She is currently working to-ward the Ph.D. degree with the Swiss Federal Insti-tute of Technology (ETH), Zurich, Switzerland.
Since 2006, she has been a Research Assistantwith the Electromagnetics in Medicine and Biology(EMB) Group, Institute for Field Theory and Tera-hertz Electronics (IFH), ETH. Her research interestsinclude simulations of spectroscopic devices with
nonbiological and biological target systems from macroscale to nanoscale.
Juerg Froehlich received the M.S. and Ph.D. de-grees in electrical engineering from the SwissFederal Institute of Technology (ETH), Zurich,Switzerland, in 1990 and 1997, respectively.
In 1998, he was with the Institute of OperationsResearch, University of Zurich, Zurich, where hedeveloped a simulation platform for multistage sto-chastic programming problems. In 2000, he waswith the Foundation for Research on InformationTechnologies in Society (IT’IS), Zurich, where hewas a Project Leader for computational tools and
risk assessment. Since November 2005, he has been with the Laboratory ofFields and Waves, ETH. His research activities cover computational tools forelectromagnetics, applications of electromagnetics in biology and medicine,and technology and risk assessment of wireless technologies.
Michael Oberle (S’95) was born in Mannheim,Germany, in 1965. He received the degree fromthe University of Karlsruhe, Karlsruhe, Germany, in1993 and the Ph.D. degree from the Swiss FederalInstitute of Technology (ETH), Zurich, Switzerland,in 2002.
He was the Cofounder and CEO of Miromico,which is a Zurich-based ASIC engineering company.Since 2005, he has been with Schmid & PartnerEngineering and the Foundation for Research onInformation Technologies in Society (IT’IS), ETH,
as the Head of the medical technology divisions. His general fields of researchhave been low-power low-voltage analog and mixed-signal integrated circuitsfor biomedical sensors, wireless telemetry, and space technology.
Norbert Felber was born in Trimbach, Switzerland,in 1951. He received the Dipl.Phys. (M.Sc.) andDr. sc. nat. (Ph.D.) degrees from the Swiss FederalInstitute of Technology (ETH), Zurich, Switzerland,in 1976 and 1986, respectively.
He was a Research Assistant with the Laboratoryof Applied Physics, ETH. In 1987, he joined theIntegrated Systems Laboratory (IIS), ETH, wherehe is currently a Research Associate and a Lecturerin the field of VLSI design and test. His researchinterests are in telecommunications, digital signal
processing (digital filters, audio, video, pattern recognition, and image process-ing), optoelectronics, measurement techniques, and device characterization.
Niels Kuster was born in Switzerland in 1957. Hereceived the M.S. and Ph.D. degrees in electricalengineering from the Swiss Federal Institute of Tech-nology (ETH), Zurich, Switzerland.
In 1993, he was a Professor with the Depart-ment of Electrical Engineering, ETH. He was anInvited Professor with the Electromagnetics Labo-ratory, Motorola Inc., Plantation, FL, in 1992 andwith the Metropolitan University of Tokyo, Tokyo,Japan, in 1998. Since 1999, he has been the Directorof the Foundation for Research on Information Tech-
nologies in Society (IT’IS), ETH. His research interests are currently focusedon the area of reliable on/in-body wireless communications (measurementtechnology, computational electrodynamics for the evaluation of close nearfields in complex environments, safe and reliable wireless communication links,development of exposure setups, and quality control for bioexperiments).
Wolfgang Fichtner (M’79–SM’84–F’90) receivedthe Dipl.Ing. degree in physics and the Ph.D. degreein electrical engineering from the Technical Univer-sity of Vienna, Vienna, Austria, in 1974 and 1978,respectively.
From 1975 to 1978, he was an Assistant Profes-sor with the Department of Electrical Engineering,Technical University of Vienna. From 1979 to 1985,he was with AT&T Bell Laboratories, Murray Hill,NJ. Since 1985, he has been a Professor and theHead of the Integrated Systems Laboratory, Swiss
Federal Institute of Technology (ETH) Zurich, Zurich, Switzerland. In 1993,he founded ISE Integrated Systems Engineering AG, which is a company inthe field of Technology CAD, which was acquired by Synopsys, Inc., in 2004.In October 1999, he became the Chairman of the Information Technologyand Electrical Engineering Department, ETH Zurich. His research activitiescover physics-based simulation of semiconductor devices and technologies inmicroelectronics and optoelectronics, physical characterization and electronicmeasurement in deep submicrometer and nanotechnologies, and the design andtest of digital integrated circuits.
Dr. Fichtner is a Member of the Swiss National Academy of Engineering.In 2000, he was the recipient of the IEEE Andrew S. Grove Award for hiscontributions to Technology CAD.
