The Detuning Effects of a Wrist-Worn Antenna and
Design of a Custom Antenna Measurement System
John Buckley1, Kevin G. McCarthy2, Brendan O’Flynn1, Cian O’Mathuna1
1Tyndall National Institute, Lee Maltings, Prospect Row, Cork, Ireland
1john.buckley@tyndall.ie
2Department of Electrical and Electronic Engineering, University College Cork, Ireland
2k.mccarthy@ucc.ie
Abstract—This paper investigates the effects of antenna detuning
on wireless devices caused by the presence of the human body,
particularly the wrist. To facilitate repeatable and consistent
antenna impedance measurements, an accurate and low cost
human phantom arm, that simulates human tissue at 433MHz
frequencies, has been developed and characterized. An accurate
and low cost hardware prototype system has been developed to
measure antenna return loss at a frequency of 433MHz and the
design, fabrication and measured results are presented. This
system provides a flexible means of evaluating closed-loop
reconfigurable antenna tuning circuits for use in wireless mote
applications.
I. INTRODUCTION
Wearable wireless technology is seeing a rapid emergence
in recent times in areas such as healthcare and activity
monitoring, driven by the need for small and reliable
monitoring devices. This technology can be used in
applications such as out-patient monitoring and providing inhome
lifelines for the elderly [1-2]. Wireless sensor motes [3]
are an ideal platform for these types of applications.
When wireless devices including motes, are placed close to
the human body, the electromagnetic coupling between
antenna and body can affect antenna parameters such as
impedance, efficiency, radiation and polarization
characteristics [4-5]. In this work, the target system is a
modular wireless mote that operates in the 433MHz frequency
band which has favourable on-body propagation
characteristics. The investigation focuses on the degree of
antenna impedance variation caused by the presence of the
human wrist. Antenna impedance mismatch causes a
reduction in radiated power and therefore, the transmitter
power needs to be increased to maintain the original radiated
power level, leading to increased power consumption and
reduced battery lifetime. One solution to this problem is to
employ an impedance matching network to correct for the
variations in antenna impedance. The design of a suitable
matching network requires accurate knowledge of the range of
antenna impedances that are possible in practice. It is difficult
to obtain consistent antenna impedance measurement in
practice using real human bodies due to problems of
movement, repeatability and traceability. Accurate phantoms
that mimic the human body are commercially available but are
expensive. An alternative approach is presented here in the
development and characterization of a low cost human
phantom, using readily available materials. A low cost
hardware prototype system has been developed to measure
antenna return loss in real time and display the results on a
GUI. The measured results for the prototype system are
compared to those using traditional vector network analyser
(VNA) measurements. The system provides a flexible
platform for design and testing of reconfigurable antenna
tuning circuits.
II. ANTENNA DETUNING
The detuning effects of the human wrist were first
investigated by measuring the return loss of a chosen test
antenna in free-space and then on the human wrist. The test
antenna is a commercial 433MHz, 50= planar type [6],
mounted on a 25 x 25mm printed circuit board as shown in
Figure 1. The antenna is fed by an SMA connector and 50=
co-planar waveguide, printed on an FR-4 substrate of
thickness 1.52 mm, permittivity 4.5, and loss tangent of 0.02.
The measured results are shown in Figure 2.
Fig. 1 Test antenna assembly and antenna on human wrist
Fig. 2 Antenna detuning caused by human wrist
It can be seen that in free-space, the antenna is correctly
tuned at a frequency f0 of 433MHz with a return loss of close
to 30dB. When the antenna is then placed on the human wrist,
touching the skin, the resonant frequency decreases to a value
of 363MHz with almost 30dB decrease in return loss at f0 and
this represents a very large impedance mismatch with a
VSWR of almost 8:1. This detuning is attributed to the change
in antenna impedance from approximately 43.81-j24 = in
free-space to a value of 45.7+j122.1= when the antenna is
placed on the human wrist.
III. PHANTOM ARM DESIGN & CHARACTERIZATION
Different materials have been used to simulate human
tissue materials at microwave frequencies [7-8]. One approach
uses a sucrose (sugar) and saline solution where the
ingredients are inexpensive and readily available. These
liquids can provide an accurate equivalent of the dielectric
properties, particularly the relative permittivity and
conductivity of human tissue over a certain frequency range.
The sucrose ingredient is used to control the permittivity of
the solution while sodium chloride is used to control
conductivity. For this study, the proportions (by weight) of
water, sucrose and sodium chloride were selected as 52.4%,
46.2% and 1.4% respectively, based on similar work at
frequencies of 100MHz to 1GHz reported in [7-8]. A hollow
fibreglass mannequin shell of thickness 3mm was selected to
contain the liquid and has a physical size and shape that
closely resembles that of a human arm as shown in Figure 3(a).
The phantom arm was placed on a non-conductive fibreglass
tripod stand for subsequent VNA measurements as shown in
Figure 3(b).
Fig. 3 (a) Human and phantom hand/wrist comparison, (b) VNA test setup
During measurements, the test antenna was first placed on
the phantom wrist. Several ferrite-beads were placed on the
50= cable to help suppress cable shield currents to improve
measurement accuracy. The test antenna was then placed on
the human arm with a 3mm plexiglass insulator used to model
the phantom arm shell thickness and prevent contact of the
back side of the antenna with the skin as skin contact has a
pronounced effect on the measurement and is difficult to
simulate. The impedance and return loss of the test antenna
were then measured on both the phantom and human wrists
and the results compared. In Figure 4(a) and (b), it can be seen
that the measured antenna resistance and reactance for the
human and phantom wrists are in good agreement at
frequencies close to the nominal antenna resonant frequency
of 433MHz, denoted f0 in Figure 4(a).
Fig. 4 (a) Antenna resistance, (b) Antenna Reactance, (c) Return Loss,
(d) Variation of antenna impedance with distance to phantom wrist
Note that two resonances (A and B) were observed in the
antenna resistance and reactance plots occurring at frequencies
of approximately 370 and 490MHz as shown by Figure 4(a)
and 4(b). The antenna impedance at these frequencies is much
greater than the system characteristic impedance of 50= and
therefore the resonant effects are not reflected in the return
loss measurements. Figure 4(c) compares the measured
antenna return loss for the human and phantom cases and the
results are in very close agreement. Figure 4(d) plots the
antenna impedance profile when the separation between
antenna and phantom wrist is varied. It can be seen that the
reactance of the antenna is capacitive for separations of
greater than 20mm and becomes inductive for distances less
than 20mm approximately. The inductive component of the
antenna impedance increases dramatically for separations less
than 5mm. For this particular test antenna, the real component
does not vary greatly and remains in the range 42 to 46=
approximately. A number of other antenna types were
measured on the human and phantom wrist and the impedance
profile (both antenna resistance and reactance) was found to
vary considerably depending on the type of antenna. From the
above measurements, it can be seen that the matching circuit
for this antenna needs to able to match any impedance in the
range of approximately 44-j24= to 46+j122=.
IV. CUSTOM ANTENNA MEASUREMENT SYSTEM
In this section, the design of a custom antenna measurement
system is presented. The purpose was to develop a flexible,
low-cost, experimental platform to investigate the
effectiveness of using the mote itself as part of a return loss
measurement system rather than the use of an extremely
accurate but expensive VNA instrument. The other aim was to
develop a flexible system that would allow the evaluation of
various types of closed-loop reconfigurable antenna tuning
circuits and tuning algorithms for use in wireless mote
applications. The block diagram of the system is shown in
Figure 5 and the blocks within the dotted section are the
subject of this discussion.
Fig. 5 Mote antenna characterization system
The radio transceiver (mote) is used as the source of RF
power. During measurements, the transceiver is placed in
transmit mode and is programmed to output a +10dBm single
tone, continuous-wave (CW) signal at a fixed frequency of
433MHz. The transceiver is connected to the input port of a
commercial bi-directional coupler [9]. The output port of the
coupler (denoted Port 1 in Figure 5) is connected to a
matching circuit or the input of an antenna. The coupler also
provides a sample of the coupled forward power Pfwd as wellas the coupled reflected power Pref from Port 1. The chosen
coupler for this application has a bandwidth of 10 to 600MHz,
a low insertion loss of 0.3dB at 433MHz with coupling and
directivity of 20dB and 25dB respectively and a voltage
standing wave ratio (VSWR) of 1.05:1. The coupled forward
and reflected powers are measured using two commercial
power detectors [10] with temperature compensated DC
outputs. These detectors have a wide bandwidth of 10MHz to
8GHz, a sensitivity of -55dBm to +20dBm with a low VSWR
of 1.05:1.
The system controller is implemented using a low-power,
high-performance, 8-bit RISC based Microcontroller [11].
The microcontroller is used to digitise the power detector
outputs as shown in Figure 6. The power detectors were precalibrated
from physical measurements to generate a lookup
table to accurately compute the values of forward and
reflected power denoted Pfwd (mW) and Pref (mW).
Fig. 6 Power detectors and system controller
The system controller then calculates the magnitude of the
antenna reflection coefficient at port 1 as
Knowledge of the antenna reflection coefficient is then
used to calculate the return loss and VSWR at Port 1 as well
as allow the tuning algorithm to determine when the matching
network has been adjusted for minimum antenna reflection. A
photograph of the fabricated prototype antenna
characterization system is shown in Figure 7.
Fig. 7 Fabricated prototype antenna characterization system
The system controller board is used to perform
measurements, adjust the re-configurable matching circuit and
transfer the measured results to a personal computer for
display and analysis on a simple GUI. The developed GUI is a
simple interface that was implemented in Labview®. An
example of the user interface is shown in Figure 8(a) with
parameters of interest such as measured coupler power levels,
antenna VSWR, reflection coefficient and return loss
displayed in graphical form in real-time.
(a)
（b）
Fig. 8 (a) Screen shot of GUI, (b) Close-up of measured return loss
Figure 8(b) shows a close-up of the measured return loss
when the user’s hand periodically touches an antenna that has
been connected to Port 1.
In order to test the accuracy of the prototype return loss
measurement, a variable load impedance was applied to the
output of the bi-directional coupler at Port 1. The load VSWR
was swept from a value of 1.06:1 to 10:1, simulating a wide
range of possible antenna mismatch conditions. For each
value of load VSWR, the return loss was measured on the
prototype system denoted |S11|A followed by an accurate
measurement using a calibrated VNA denoted |S11|B. All
measurements were performed at a frequency of 433MHz and
Table I tabulates the results.
TABLE I
COMPARING MEASURED RETURN LOSS MAGNITUDE OF THE PROTOTYPE
SYSTEM VERSUS CALIBRATED VNA MEASUREMENTS (AT 433MHZ)
It can be seen that the prototype and VNA return loss
measurements are in very good agreement across a wide range
of load VSWR values. However, for values of VSWR close to
1:1 (30dB return loss approximately), the difference between
the two measurements increases significantly and is most
likely due to the finite directivity of the bi-directional coupler
(25dB) that sets a limit of the return loss that can be measured
[12].
V. CONCLUSIONS
This paper presented the results of an investigation into
antenna detuning on wireless devices caused by the presence
of the human wrist. A low cost human phantom arm was
developed and characterized that accurately simulates human
tissue at 433MHz frequencies. The measured antenna
impedance and return loss results can be used in the design
and simulation of impedance matching circuits. A flexible and
low cost, 433MHz antenna measurement system was also
developed and characterized to measure antenna return loss.
The accuracy of the system has been verified by comparing
the measured results to accurate calibrated VNA
measurements. The developed system is currently being used
as a platform for evaluating various types of closed-loop
reconfigurable antenna tuning circuits and tuning algorithms
for energy efficient wearable wireless mote applications.
VI. ACKNOWLEDGMENT
We would like to acknowledge the support of Enterprise
Ireland for funding this work under Grant PC_2008_324.
REFERENCES
[1] K. D. Wise, “Wireless integrated microsystems: Wearable and
implantable devices for improved health care”, 15th International
Conference on Solid-State Sensors, Actuators and Microsystems, art.
no. 5285579, pp. 1-8, 2009.
[2] Y. Lee, W. Y. Chung, “Wireless sensor network based wearable smart
shirt for ubiquitous health and activity monitoring”, Sensors and
Actuators, B: Chemical, 140 (2), pp. 390-395, 2009.
[3] B. O’Flynn., S. Bellis., K. Delaney., J. Barton., S. C. O’Mathuna., A.
M. Barroso, J. Benson., U. Roedig., and C. Sreenan., “The
Development of a Novel Miniaturized Modular Platform for Wireless
Sensor Networks”, Proc of 5th International Symposium on Info
Processing in Sensor Networks (IPSN/SPOTS), Apr. 2005.
[4] Wang, Z., Chen, X. and Parini, C.G., "Effects of the Ground and the
Human Body on the Performance of a Handset Antenna", IEEE Proc.
Microwave & Antenna Prop., Vol. 151, No. 2, pp. 131-134, 2004.
[5] K.L. Wong and C.I. Lin, “Characteristics of a 2.4-GHz compact
shorted patch antenna in close proximity to a lossy medium”,
Microwave Opt Technology Letters, pp. 480–483, 2005.
[6] AntennaFactor, Product Data Sheet, ANT-433-SP, 2008.
[7] Ogawa, K., Matsuyoshi, T., Iwai, H., Hatakenaka, N., “A highprecision
real human phantom for EM evaluation of handheld terminals
in a talk situation”, IEEE Antennas and Propagation Society, AP-S Int
Symposium (Digest), 2, pp. 68-71, 2001.
[8] P. Liu, C. Rapaport, Y. Z. Wei and S. Sridhar, “Simulated Biological
Materials at Microwave Frequencies for the Study of Electromagnetic
Hyperthermia,” IEEE EMBS Digest, V, pp. 272-273, 1992.
[9] Minicircuits, Product Data Sheet, ZFBDC20-62HP+, 2009.
[10] Minicircuits, Product Data Sheet, ZX47-55+, 2009.
[11] Atmel, Product Data Sheet, Atmega128L, 2009.
[12] R. White, Spectrum and Network Measurements, PTR Prentice Hall,
New Jersey, 1993.
1741
京公网安备 11010802033920号
