An UWB capsule endoscope antenna design for

biomedical communications
Qiong Wang, Klaus Wolf, Dirk Plettemeier
Chair for RF and Photonics, Communication Laboratory
Dresden University of Technology,
Dresden, Germany
[email protected] [email protected]

Abstract—This paper proposes a conformal trapezoid strip (640×480×24-bits×10-frames/s) for uncompressed VGA video
excited broadband hemispherical dielectric resonator antenna image transmission capability [6]. Both ZigBee and Bluetooth
(DRA) design in ultra wide-band (UWB) low band 3.1-4.8 GHz support communications with data rates typically below 1
for medical capsule endoscope applications. The proposed Mbps.
antenna can be conformal to the capsule dome and is also
characterized by light weight, low profile, low cost, small size. A Ultra-wideband (UWB) solution has emerged for wireless
broadband trapezoid monopole receiver antenna is developed for implanted communication interface with high data rate in the
transmission measurements. The UWB transmission loss future telemedicine systems [7]. An UWB capsule endoscope
characteristics in body tissue are exploited based on the has also been proposed in [1] to ensure real-time video
measurement solution. Moreover, a successful video transmission monitoring of the internal parts of a patient’s body. Some
from inside to outside body has been carried out based on multi- research work has been done for the radio propagation inside
band orthogonal frequency-division multiplexing (MB-OFDM) the human body within the FCC allocated UWB frequency
scheme using the proposed antennas. Additionally, a preliminary band of 3.1-10.6 GHz [8]. Highly frequency-dependent signal
research based on a binary phase-shift keying (BPSK) carrier attenuation has been concluded within the UWB band, which
transmission has been done to investigate the potential capsule brings a big challenge for the UWB capsule system design. [9]
endoscope communication performance. has suggested an impulse radio (IR) UWB wireless implanted
communication in 3.4-4.8 GHz, the UWB low band. This paper
Keywords-body area network; capsule endoscope; implanted
antenna; dielectric resonator antenna; ultra wideband;
proposes an UWB capsule endoscope antenna design operating
biomedical communication in the UWB low band of 3.1-4.8 GHz. Based on this capsule
antenna design, the in-body real-time transmission and the
I. INTRODUCTION communication performance are exploited.
The wireless capsule endoscopy first appeared in 2001 [1]. The paper is organized in the following way. Section II
The ingestible capsule can take pictures during its course describes the antenna design, optimization as well as
through the digestive tract after being swallowed. It can validation. In Section III, transmission measurements are
therefore transmit the real-time biological data from inside the carried out for transmission loss, video transmission and
body to exterior medical instruments, which pushes forward the communication performance evaluation. Conclusions are
non-invasive diagnosis. The typical biotelemetric imaging summarized in Section IV.
capsule consists of the five main parts as shown in Figure 1:
CMOS imager, batteries, radio telemetry transmitter and
antenna. The antenna design for this ingestible capsule is very (a) Optical dome
(b) Lens holder
challenging due to a very limited capsule volume with (c) Short focal-length lens
dimensions 11mm x 26mm [1]. Only a few antenna designs (d) Light-emitting diodes
(e) CMOS chip camera
have been reported for capsule endoscope biomedical (f) Two silver oxide batteries
communications. A spiral microstrip antenna design and a (g)
(h)
Radio telemetry transmitter
Antenna
cylindrical helical antenna design in Medical Implant
Communication Services (MICS) band 402-405 MHz have
been proposed in [2] and [3], respectively. A conformal
chandelier meandered dipole has been designed in the Figure 1. Biotelemetric imaging capsule components
with dimensions 11x26mm [1]
frequency band of 1395-1400 MHz allocated by the FCC for
Wireless Medical Telemetry Services (WMTS) in [4]. In fact, II. ANTENNA DESIGN
MICS band 402-405 MHz and WMTS band 1395-1400 MHz
essentially belong to narrow band systems which make them A. Human tissue dielectric characteristics
low data rate systems. However, the imaging capsule system is Antenna design for implant medical communications is
required to transmit real-time biological image data, which greatly affected by the body tissues which act as the
implies a high-speed transmission. In fact, at least 40 MHz of surrounding medium. Classical antenna design theory describes
bandwidth are required to support 10 Mbps in implant sensors antennas in a non-conducting surrounding with a relative
[5]. Capsule endoscope sensors require up to 73.8 Mbps permittivity of 1, that is, usually in vacuum or air. The capsule

978-1-4244-8132-3/10/$26.00 ©2010 IEEE

antenna, placed inside a human body, will be surrounded by the simulated with a thin capsule polyethylene shell so as to
body torso tissue, which is a lossy material with high insulate the conducting strip from direct contact with human
permittivity. In a material with high permittivity, the wave tissues.
propagation speed will be lowered and therefore the
wavelength in the material will be shorter. The shortened A tissue-simulating fluid based on the C12H22O11 recipe
wavelength in body tissue will be beneficial for a small-size solution [13] is employed to simulate the average tissue
antenna design implementation. dielectric characteristics in 3.1-4.8 GHz. The antenna is
positioned in a foam container filled with the tissue-simulating
The dielectric characteristics of human body tissues are fluid as shown in Figure 2. The tissue-simulating fluid phantom
highly frequency-dependent. The UWB body tissue dielectric is with dimension of 200x200x200mm, which is about twenty
characteristics can be found on the FCC official website [10], times the effective wavelength of the central frequency 4.0
which provides the dielectric parameters in 10-6000 MHz for GHz in body tissue. Figure 3 compares the simulated and
average tissue as well as all body tissue types. The human body measured return losses. As can be seen, the measured result
torso tissue can be approximated using average muscle tissue. covers slightly broad 10-dB bandwidth compared with the
Table I shows the average dielectric parameters for human simulated result. The disagreement gap between simulated and
torso tissue in 3.1, 4.0, 4.8 GHz. Therefore, by averaging the measured results is due to the prototype fabrication of the small
body torso tissue dielectric characteristics in 3.1-4.8 GHz, the antenna as well as the approximate phantom configuration.
surrounding medium of the capsule antenna can be
approximately determined as one homogeneous layer with
average dielectric constant of 51.5 and average conductivity of
3.2 S/m.
TABLE I. AVERAGE DIELECTRIC PARAMETERS
OF HUMAN TORSO TISSUE
Dielectric parameters
Frequency
Relative Conductivity Density(app.)
[GHz]
permittivity [S/m] [kg/m3]
3.1 52.7 2.3 1040.0
4.0 51.5 3.2 1040.0
4.8 50.4 4.0 1040.0
Figure 2. Conformal taper strip excited hemisphere DRA and return loss
measurement setup in tissue-simulating fluid phantom
B. Conformal taper strip excited hemisphere DRA
Dielectric resonance antenna (DRA) is characterized by one 0
of the main advantages of its high degree of flexibility and
versatility. This can allow for designs to suit a wide range of -5
physical or electrical requirements. Moreover, high radiation
Return loss [dB]

-10
efficiency can be maintained by selecting a dielectric material
with low-loss characteristics due to minimal conductor losses -15
associated with the DRA [11]. Hemispherical DRAs can make
use of the capsule dome volume allocated for the antenna. -20
Based on the hemispherical DRA theory, the Q-factor will be simulation
as low as 2 for the hemispherical DRA operating in 3.1-4.8 -25 measurement
GHz with the fractional bandwidth of 43%. Then a dielectric
material with low permittivity value will be needed to maintain -30
the low Q-factor and wide band characteristics. Meanwhile the 2 3 4 5 6
Frequency [GHz]
dielectric material with low permittivity value can keep off the
high loss induced by high permittivity dielectric materials. A
Figure 3. Simulated and measured return losses of the conformal taper strip
kind of PVC material with dielectric constant of around 3 is excited hemisphere DRA in tissue-simulating fluid phantom
chosen to build the hemisphere. A conformal strip, instead of
the traditional probe, is employed to excite the hemispherical C. Antenna validation in free space
DRA. It allows the whole electric current to flow on the DRA
The effective wavelength of the antenna in body tissue is
surface and is therefore more efficient in energy coupling than
much shorter compared with the wavelength in free space due
the probe-feed counterpart [12]. In order to broaden the
to the change of the permittivity and conductivity of the
impedance bandwidth, a conformal taper copper conducting
surrounding medium. The complex permittivity ߝ௖ of the lossy
strip is employed to excite the DRA and connected with a 50
body tissue is
RG405 feeding coaxial cable (d=0.51 mm, D=1.7 mm,
epsilon_r=2.1) as shown in Figure 2. The taper conducting strip ఙ೐
ߝ௖ ൌ ߝ௘ െ ݆ (1)
is conformal and mounted on the external surface of the ఠ

dielectric hemisphere with a diameter of 10 mm. The The relationship between the wavelength in body tissue and in
hemisphere DRA is mounted on a finite circular ground plane free space is shown as follows [14]:
with the same diameter as the hemisphere. The design is

978-1-4244-8132-3/10/$26.00 ©2010 IEEE

 ఒబ ఒబ excitation gives better return loss performance with 10-dB
ߣ௘௙௙ ൌ ഄ
ൌ ഑
(2)
ோ௘ሾට ೎ ሿ ோ௘ሾටఌೝ ି௝ ೐ ሿ bandwidth from 3 to 5 GHz. The radiation pattern in tissue is
ഄబ ഘഄ బ
shown in Figure 7. A broad beam radiation can be obtained at
in which ߣ௘௙௙ is the effective wavelength in body tissue, ߣ଴ is the excitation side. Besides, this antenna design produces two
the wavelength in free space, ߝ௥ and ߪ௘ are the relative linear polarized components which can be employed to
permittivity and the effective conductivity of the body tissue, implement polarization diversity in the receiver.
଴ ൌ ͺǤͺͷͶ ൈ ͳͲିଵଶ is the permittivity of vacuum, ߱ is the
angular frequency and Re[ ] denotes real part. Table II
compares the wavelengths in free space and in body tissue in
3.1, 4.0, 4.8 GHz respectively. A shorting factor of 7.2 is
concluded at the central frequency 4.0 GHz.
TABLE II. COMPARISONS OF THE WAVELENGTHS
IN FREE SPACE AND IN BODY TISSUE
Frequency Shortening
ࣅ૙ [m] ࣅࢋࢌࢌ [m]
[GHz] factor
3.1 0.097 0.013 7.3
4.0 0.075 0.010 7.2 Figure 5. Optimized conformal trapezoid strip excited hemisphere DRA and
the fabricated prototype
4.8 0.063 0.009 7.2

The proposed conformal strip fed hemispherical DRA 0

operates in the quasi-TM101 mode in free space assuming a -5
infinite ground plane. Characteristic equations are given in [11]
to determine the resonant frequency and absolute bandwidth of -10

Return loss [dB]

the hemispherical DRA with the radius and dielectric constant -15
known. Keep in mind that the equations do not take into
-20
account the effects of the feeding mechanism and moreover
assume an infinite ground plane. Based on the characteristic -25
equations, a hemispherical DRA with a radius of 5 mm and a -30 original taper excitation
dielectric constant of 3 will resonate ideally in free space at
-35 optimized trapezoid excitation
24.5 GHz with an impedance bandwidth of 6.3 GHz. The
simulated and measured results in Figure 4 validate the theory -40
in free space even though there is some shift in the resonance 2 3 4 5 6
frequency due to a finite ground plane and the effects of the Frequency [GHz]
feeding.
Figure 6. Simulated return losses of original taper excitation in Figure 2 and
0 optimized trapezoid excitation in Figure 5 in tissue-simulating fluid phantom

-10
Return loss [dB]

-20

-30
simulation
-40 measurement

-50
0 5 10 15 20 25 30
Frequency [GHz]

Figure 4. Simulated and measured return losses of the conformal taper strip Figure 7. Radiation pattern of the optimized trapezoid strip excited
excited hemisphere DRA in free space hemisphere DRA at 4 GHz in tissue-simulating fluid phantom

D. Optimized trapezoid strip excited hemisphere DRA III. TRANSMISSION MEASUREMENTS

In order to further improve the resonance performance of
A. Transmission loss measurement
the proposed antenna, the feeding structure is modified as
shown in Figure 5: a trapezoid strip is connected to the coaxial The UWB trapezoid monopole in Figure 8 is first proposed
cable by a small rectangular conducting strip. Moreover, the to be used as the receiver antenna outside the body in order to
hemisphere diameter has been reduced from 10 mm to 8 mm, form a data telemetry link together with the transmitter antenna
which is more favorable for the integration of the antenna into in the body torso. This trapezoid monopole can work efficiently
the 11mm x 26mm capsule. Figure 6 compares the return loss in the 3-5 GHz band. Figure 8 also shows the far field radiation
performance for the taper excitation and the optimized pattern at 4 GHz. The maximum gain is around 5.8 dB. As can
trapezoid excitation. As can be seen, the optimized trapezoid be seen, due to the finite ground plane, there is a radiation

978-1-4244-8132-3/10/$26.00 ©2010 IEEE

spillover at the back lobe which results in a lowering of the
gain. A larger ground plane can alleviate this, however.
Transmission loss measurement was set up as shown in
Figure 9. The receiver trapezoid monopole antenna is fixed
outside the tissue phantom at a distance of 5 cm to the phantom
border, which can simulate the situation in practical telemetry
application. Besides, the distance of 5 cm to the phantom
border did not put a huge influence on the radiation gain and
the gain deterioration towards the tissue direction is only within
2 dB. The transmitter trapezoid strip excited DRA antenna is
located in the middle of the tissue phantom and the maximum
radiation beam is towards the RX antenna. The transmission
loss measurement results are shown in Figure 10 by changing
the distance of the TX DRA antenna to the tissue border. An
amplifier with 30 dB gain was employed in case the received
Figure 9. Transmission loss measurement setup diagram
power is significantly attenuated by the highly lossy body (TX antenna: optimized DRA shown in Figure 5;
tissue. As can be seen, the lossy tissue brings high transmission RX antenna: trapezoid monopole shown in Figure 8)
loss and the transmission loss increases for higher frequencies.
Notice that the transmission loss is around 60 dB at 4 GHz with
0
d=5 cm without amplifier. When the distance becomes larger,
the transmission loss will be very high and the received signal -10 d=0.5cm with Amplifier
may fall below the noise level which is a huge challenge for the -20
S21 measurement. -30 d=5cm with Amplifier

S21 [dB]
B. Video transmission measurement -40 d=0.5cm without Amplifier
The UWB capsule is required to transmit real-time video -50
data of the internal parts of a patient’s body. A video -60
d=5cm without Amplifier
transmission system, based on a wireless USB to VGA kit -70
Wireless UWB Technology, was set up as shown in Figure 11.
-80
The transmitter and receiver operate at the WiMedia band
group 1 from 3.168-4.752 GHz, based on the MB-OFDM 3 3.5 4 4.5 5
Frequency [GHz]
scheme. The bandwidth is divided to three sub-bands and the
time-frequency-coding is employed. The data rate in the Figure 10. Measured transmission loss for the setup in Figure 9
physical layer is 480 Mbps. The maximum transmitted power
complies with the FCC regulated -41.3 dBm/MHz. Table III
shows the signal quality variance when we move the TX
antenna away from the RX antenna. Within 10 cm buried in the
body torso, there is favorable video transmission available
based on the current measurement setup. When the distance is
larger than 10 cm, the transmission will break off. It can be
concluded that real-time video transmission from inside the
human body within 10 cm is successful using the current
trapezoid strip excited hemisphere DRA antenna. Alternatively,
advanced receiver antenna system with diverse polarization,
rather than the linear polarized monopole, can strengthen the
receiving gain to compensate the high loss in the body tissue
and to further enlarge the transmission distance.

Figure 11. Video transmission measurement setup diagram

(TX antenna: optimized DRA shown in Figure 5;
RX antenna: trapezoid monopole shown in Figure 8)

TABLE III. SIGNAL QUALITY OF VIDEO TRANSMISSION FOR DIFFERENT

TRANSMITTER DISTANCE TO THE TISSUE BORDER

d Signal quality
d 8cm Good video (delay  1s)
Figure 8. UWB trapezoid monopole radiation pattern at 4 GHz and the
8d 10cm Static image (delay  10s)
fabricated monopole prototype
d 10cm No image

978-1-4244-8132-3/10/$26.00 ©2010 IEEE

C. Preliminary evaluation of communication performance band. A successful video transmission from inside body 10 cm
to outside body 5 cm has been obtained based on MB-OFDM
Beside the transmission loss and the real-time video scheme using the proposed antennas. Finally, although not
transmission measurements, we also used a set of digital based on UWB transmission, communication performance
communication systems setup to perform a preliminary evaluation based on a fundamental BPSK transmission system
investigation of the BER performance as shown in Figure 12. can be used to conclude that a reasonable capsule endoscope
The whole system is based on carrier transmission (carrier communication is highly potential using suitable UWB
frequency 4 GHz) with BPSK modulation. A wideband Vivaldi modulation scheme with required data rates.
antenna (4-40 GHz) with high gain (15 dB) and high directivity
is used to substitute the trapezoid monopole antenna. Besides,
an amplifier with 30 dB gain was employed at the transmitter
end in order to compensate the high loss in body tissue and to
make the error detector device work effectively. Although this
operation brings a tranmitted power around 26 dBm which is
nearly 30 dB higher than the maximum allowed power -10
dBm in 3.1-4.8 GHz according to FCC regulation, it is still
lower than the maximum tansmitted power of mobile phone
which is around 33 dBm under maximum radiation mode. One
constraint of this setup is that the minimal available data rate is
0.5 Gbps which is much higher than the required transmission
rate for a capsule endoscope. However, this can account for the
employed high transmitter power. On the other hand, high data
rates can result in a broad-band transmission even with BPSK
modulation. Then the proposed antenna performance can be
validated in a relatively broad band rather than narrow band.
Figure 13 shows the BER performance with different data
rates, in which the cable reference corresponds to setup system Figure 12. Communication performance evaluation system setup diagram
error without any antenna connected, monopole reference (TX antenna: optimized DRA shown in Figure 5;
corresponds to the BER using two trapezoid monoples RX antenna:wideband Vivaldi antenna with gain of 15 dB in 4-40 GHz)
operating in free space and the TX DRA vs RX Vivaldi 1E-03
cable
corresponds to the system setup in Figure 12 with d=0.5 cm. It reference
can be concluded that a BER of 1E-03 can be obtained with a monopole
1E-05
reference
high data rate of 2 Gbps using BPSK carrier transmission for TX DRA vs
the system setup in Figure 12 with d=0.5 cm. Since d=0.5 cm 1E-07 RX Vivaldi
means nearly close to the tissue border, we enlarge the distance
BER

to investigate the BER performance and the Q-factor of the eye

1E-09
diagram of the waveforms. Figure 14 shows the BER and Q-
factor at data rate of 1 Gbps with d sweeping from 2 cm to 5
cm. As expected, the larger the transmission distances in tissue, 1E-11
the worse the BER and the lower the Q-factor. At d=5 cm,
BER of 1E-5 can be obtained with data rate of 1 Gbps. This is 1E-13
highly desirable since it implies larger communication distance 0.5 1 1.5 2 2.5 3
with acceptable error rates in which error correction techniques Data Rate [Gbps]
can still work efficiently. Moreover, advanced data formats
rather than the fundamental BPSK can be employed to realize Figure 13. BER vs. data rate based on BPSK carrier transmission
larger distance communication (larger than 5 cm) with low 1E-03 16
transmitting power and acceptable data rates (no need up to 1 15
Gbps). From the evaluation results based on the current system 1E-05
14
setup, we can conclude that a reasonable communication
Q-factor

1E-07 13
performance can be expected for capsule endoscope system
BER

based on UWB schemes. 12

1E-09 11
IV. CONCLUSIONS 10
1E-11
An UWB capsule endoscope antenna design operating in 9
the UWB low band of 3.1-4.8 GHz is proposed. It is based on 1E-13 8
an optimized conformal trapezoid strip excited broadband
1 2 3 d [cm]4 5 6
hemispherical dielectric resonator antenna. The antenna design
has been validated via theoretical analysis as well as chamber
measurements. Transmission loss measurement has revealed Figure 14. BER and Q-factor of the eye diagram vs. transmission distcance in
the high loss characteristics of body torso tissue in UWB low the tissue based on BPSK carrier transmission with data rate of 1 Gbps

978-1-4244-8132-3/10/$26.00 ©2010 IEEE

 ACKNOWLEDGMENT [7] P. Gandolfo, D. Radovi, M. Savi and D. Simi, “IEEE 802.15.4a UWB-
IR radio system for telemedicine,” in Proc. IEEE Intl. Conf. on Ultra-
The authors would like to thank our colleague Ignacio Wideband (ICUWB 2008), Hannover, Germany, September 10–12, vol.
González Insua for his great help in the BER measurement. 3, pp. 11–14, 2008.
[8] A. Khaleghi and I. Balasingham, “Improving in-body ultra wideband
REFERENCES communication using near-field coupling of the implanted antenna,”
Microwave and Optical Technology Letters, vol. 51, no. 3, pp. 585–589,
[1] W. A. Qureshi, “Current and future applications of the capsule camera,”
March 2009.
Nature Rev. Drug Disc., vol. 3, pp. 447–450, May 2004.
[9] Q. Wang, K. Masami and J. Wang, “Channel modeling and BER
[2] S. Kwak, K. Chang and Y. J. Yoon, “Ultra-wide band spiral shaped
performance for wearable and implant UWB body area links on chest,”
small antenna for the biomedical telemetry,” Asia-Pacific Microwave
in Proc. IEEE Intl. Conf. on Ultra-Wideband (ICUWB
Conference Proceedings, vol. 1, 2005.
2009),Vancouver, Canada, September 9–11, 2009, pp. 316–320.
[3] S. Kwak, K. Chang and Y.J. Yoon, “The helical antenna for the capsule
[10] http://www.fcc.gov/oet/rfsafety/dielectric.html
endoscope,” 2005 IEEE International Symposium of Antennas and
Propagation, vol. 2B, pp. 804 – 807, 2005 [11] A. Petosa, “Dielectric resonator antenna handbook,” 2007.
[4] P.M. Izdebski, H. Rajagopalan and Y. Rahmat-Samii, “Conformal [12] K.W. Leung, “Conformal strip excitation of dielectric resonator
Ingestible Capsule Antenna: A Novel Chandelier Meandered Design,” antenna,” IEEE Trans. on Antennas and Propagation, vol. 48 , issue 6,
IEEE Trans. on Antennas and Propagation, vol. 57 , issue 4, pp. 900 – pp. 961 – 967, 2000.
909, 2009. [13] Y. Nishida, D. Hara, and T. Kobayashi, ‘Development of Ultra
[5] PSSK Proposal for High-data-rate In-body WBAN PHY (ETRI), IEEE Wideband Electromagnetic Phantom Materials,” Technical report of
15-09-0179-02-0006, March 10, 2009. IEICE. WBS2003-105, AP2003-298, RCS2003-321, MoMuC2003-111,
MW2003-267(2004-03).
[6] R. Chavez-Santiago, A. Khaleghi, I. Balasingham and T.A. Ramstad,
“Architecture of an ultra wideband wireless body area network for [14] P.S. Hall, Y. Hao, “Antennas and propagation for body-centric wireless
medical applications,” 2nd International Symposium on Applied communications,” pp. 258, 2006.
Sciences in Biomedical and Communication Technologies, 2009.
ISABEL, pp. 1 – 6, 2009.

978-1-4244-8132-3/10/$26.00 ©2010 IEEE