Optmized Patch-Like Antennas for Through The Wall
Radar Imaging and Preliminary Results with
Frequency Modulated Interrupted Continuous Wave
Francesco Fioranelli, Sana Salous
School of Engineering and Computing Sciences
Durham University
South Road, DH1 3LE, Durham, UK
sana.salous@durham.ac.uk
Abstract— This paper presents optimized patch-like antennas for
Through The Wall Imaging (TTWI) radar applications in the
frequency range 0.5-2 GHz, and preliminary results using
Frequency Modulated Interrupted Continuous Waveform
(FMICW). Results of numerical simulations using basic models
of the antenna are presented. The antenna optimization was
aimed at making the radiation pattern more directional by
focusing the energy in a single lobe to be directed towards the
wall and the targets to be detected. The optimized antenna was
manufactured and its measured parameters are compared with
the simulated results which show good agreement. Some
preliminary results from the FMICW radar system using this
antenna are presented.
Keywords-component: UWB; printed antenna; through the
wall imaging; radar; FMCW; FMICW
I. INTRODUCTION
Through The Wall Imaging (TTWI) radar techniques have
been attracting interest in the past few years as a way to
provide detailed information about areas which could not be
otherwise seen through conventional methods. This can benefit
police or soldiers operating in an urban environment (e.g.
breaking in a room occupied by hostile agents), or firemen in
search and rescue operations (e.g. entering a burning building,
or locating people buried under rubbles). Microwave systems
using Ultra Wide Band (UWB) signals are a mature technology
for TTWI radar applications. These systems can provide both
bandwidths wide enough to achieve sufficient range resolution
to create an accurate image of the targets, and low enough
central frequency of the spectrum to penetrate through the wall
materials. The range resolution of the radar system is indeed
proportional to the bandwidth, while the through-wall
penetration is inversely proportional to the frequency[1].
In TTWI radar systems antennas are of fundamental
importance and they should have well matched impedance
across the required wide bandwidth, and radiation pattern
suitable to focus most of the energy towards the wall and the
targets to be detected. Constraints on weight and size of the
antennas should be also taken into account, depending on the
kind of radar system (e.g. radar mounted on a vehicle or
handhold). Several design models have been proposed in the
978-1-4673-4455-5/12/$31.00 ©2012 IEEE
Ivan Ndip
Fraunhofer Institut für Zuverlässigkeit und Mikrointegration
Fraunhofer IZM
Gustav-Meyer-Allee 25
13355, Berlin, Germany
literature. Antipodal or exponentially tapered Vivaldi antennas
are widely used for their high gain and narrow beam-width in
their radiation patterns, either as single antenna or in array
configurations [2-3]. Vivaldi and linear-tapered slot design are
sometimes combined to create hybrid antennas [4]. Horn
antennas can be also used to achieve high gain and radiation
pattern directivity similar to the Vivaldi design. Elements
printed on dielectric substrate are another possible choice since
they are generally compact, easy to manufacture and to use in
array configurations. Many different shapes are available in the
literature, for instance simple geometrical square or elliptical
patch which can be used in an array directly worn by users in
the field [5], or more elaborate shapes to achieve desirable
properties like dual polarization or side lobes suppression as in
[6]. Ad hoc designs of antennas may also be used to improve
the performance, for instance the two-flares design for
detection of people buried under rubble as reported in [7].
Among all the aforementioned models, a printed patch-like
antenna design is chosen and presented in this paper for TTWI
radar applications, in the frequency range 0.5-2GHz. Printed
antenna elements are indeed easy to design and to manufacture.
Given a desired operational bandwidth, their overall
dimensions are also relatively compact compared with other
designs, such as the exponentially tapered Vivaldi.
Frequency Modulated Continuous Waveform signals have
long been used in ionospheric sounding and radar applications
as they avoid the peak power limitations of pulse radar.
However, due to the continuous transmission at the transmitter,
in a mono-static operation, the transmitted signal can block the
receiver. To avoid this limitation FMICW signals are switched
on and off at the transmitter to allow for listening intervals at
the receiver which is switched with the complement of the
gating sequence used at the transmitter [8].
In section II of this paper we present a basic version of the
antenna which has been extensively simulated and
manufactured, achieving good similarity between simulations
and measurements. The design has been then optimized to
improve its effectiveness for TTWI radar applications, and this
new prototype has been manufactured. Comparisons of
measurements with simulated results are presented in section
III. Finally in section IV a measurement example of the
optimized antenna in conjunction with a suitable FMICW
signal for TTWI radar system illustrates the effectiveness of the
technique.
II. THE BASIC ANTENNA MODEL
The basic design of the antenna consists of a rectangular
patch of size 5.68 cm by 8.23 cm printed on common FR4
dielectric substrate with 1.5 mm thickness and 80 µm copper
layer as shown in fig. 1 with the other side of the substrate
forming the ground plane. The patch is fed by a microstrip line
with 12.6 cm length and 0.32 cm width, and the length of the
ground plane length is 12.1 cm. The resulting overall
dimensions of the antenna are 16.4 by 25.5 cm. Prior to
fabrication the antenna was extensively simulated with the
software ANSYS HFSS to get its S11 below -10dB across the
desired bandwidth 0.5-2GHz, while keeping the overall
dimensions as small as possible. A model of a SMA connector
to feed the antenna was included in all the simulations for
realistic applications. The dimensions of the patch, the feed
line, and the length of the ground plane were optimized with
respect to the S11 parameter and a prototype of the final design
has been manufactured. The final design was also simulated
using CST MICROWAVE STUDIO which is a time domain
technique simulator (Finite Integration in Time Domain) to
compare the results with those from HFSS which uses a
frequency domain technique (Finite Element Method). Fig. 2
compares the simulation results from the two methods with the
actual measured data for the S11 parameter. The figure shows
that the S11 parameter is less than -10 dB across the desired
bandwidth and good agreement with the measurements.
Figure 1. Top view of the basic antenna model
S11 Basic Antenna Design: simulations and measurements
0
Measurement
HFSS Simulation
-5 CST Simulation
-10
S11 [dB]
-15
-20
-25
-30
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2 2.25 2.5
Frequency [GHz]
Figure 2. Simulated and measured S11 of the basic antenna model
Fig. 3 shows the simulated radiation patterns of the antenna
for theta = 90° in the azimuth plane at 0.5, 1.25, and 2 GHz,
which represent respectively the start, centre, and stop
frequency of the desired bandwidth. It can be seen that this
antenna presents two lobes in the direction 60° and 120°
degrees, and their value increase with the frequency as is
generally expected. This two lobes structure may be
problematic for TTWI radar applications, since the energy is
not clearly focused towards one direction, and even if one lobe
is focused towards the wall, clutter and undesired singles may
be received through the other lobe.
Radiation Pattern Plane Theta 90° at 0.5 1.25 2GHz
0o5 dB
-30o 30o
-6
-60o 60o
-17
-90o 90o
-17
-120o -6 120o
-150o 5 dB150o
180o
Simulation at 2GHz
Simulation at 1.25GHz
Simulation at 0.5GHz
Figure 3. Simulated radiation pattern on plane theta 90° at 0.5, 1.25 and
2GHz of the basic antenna model
III. THE OPTIMIZED ANTENNA MODEL
To suppress one of the lobes and to enhance the gain of the
antenna in the desired lobe an additional ground plane strip was
added. During the process of optimization the antenna size was
further reduced while maintaining the S11 parameter below the
usual -10 dB threshold across the desired 0.5-2GHz bandwidth.
A small metal patch was also added on the lower side of the
substrate in order to improve the impedance matching of the
antenna. Fig. 4 shows the shape of the optimized model which
has a rectangular patch with dimensions equal to 6.53 cm by
7.44 cm with a feeding line length equal to 6.12 cm and 0.32
cm width. The ground plane length and the width of the added
ground plane strip are equal to 5.83 cm and 2.12 cm,
respectively. The dimensions of the metal patch on the lower
side are 3 cm by 3.5 cm with the overall dimensions of the
antenna being equal to 14 by 17 cm, versus the 16.4 by 25.5 cm
of the basic model. The substrate is common FR4 again, with
thickness 1.5 mm and the thickness of the copper layer 80 µm.
Figure 4. Top view of the optimized antenna model
 This optimized design has been extensively simulated using S11 Optimized Antenna: measurements in different environments
both HFSS and CST to investigate the S11 and the gain 0
radiation patterns. A prototype has been built to measure the
actual values of these parameters. Fig. 5 compares the S11 for -5
the two different simulations with the measured data. The
-10
parameter is below the -10 dB threshold across the required
bandwidth and the curves are quite similar.

S11 [dB]
-15
S11 Optimized Antenna Design: simulations and measurements
0 Measurement -20
HFSS simulation
-5 CST Simulation
-25
-10 Anechoic chamber
-30 10cm from thin wall
S11 [dB]

-15 10cm from thick wall

-20 -35
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2 2.25 2.5
Frequency [GHz]
-25

-30 Figure 8. S11 of the optimized antenna measured in different environments

-35
0 0.25 0.5 0.75 1 1.25 1.5 1.75
Frequency [GHz]
Figure 5. Simulated and measured S11 of the optimized antenna model
The S11 parameter has been measured also when the
antenna is placed in a real environment very close to the wall,
as in the case of its practical application in TTWI radar
systems. Fig. 6 and 7 show the environments of these S11
parameter measurements: the antenna is placed at 10 cm from
two different kinds of wall, respectively a thin partition wall
and a thick concrete wall. Fig. 8 compares the resulting S11
with the measurement taken inside the anechoic chamber. The
presence of the thick wall raises the S11 slightly above the -10
dB threshold in segments between 0.9-1.25. This increase is
less than 2 dB. For the thin wall the antenna’s performance still
gives the required -10 dB threshold performance. Thus, the
antenna is still usable even very close to walls with some minor
degradation of its performance.
2 2.25 2.5
The gain radiation patterns have been measured using the
standard gain method, i.e. deriving the gain of the antenna
under test from a measurement of the S21 with an antenna with
well-known and calibrated gain, which in this case is a horn
antenna for the frequencies above 1.25GHz and a log-periodic
between 0.5 and 1.25GHz. Figures 9 and 10 show the patterns
in the azimuth plane at theta equal to 90° at 1.25GHz and
2GHz respectively, while figures 11 and 12 show the patterns
at the same frequencies, but in the elevation plane at phi 0°. All
the figures show good similarity between the simulated and the
measured data. In fig. 9 and 10 the single lobe radiation pattern
can be seen as a result of the optimization with the gain of this
main lobe being around 5dB at 1.25GHz and 7dB at 2GHz.
Radiation Pattern Plane Theta 90° 1.25GHz
0o 5 dB o
-30o 30
-3
o
-60 60o
-11

-90o 90o
-11
Figure 6. Antenna close to a thin partition wall
-120o -3 120o

-150o 150o
o 5 dB
180
Measurement
HFSS Simulation
CST Simulation

Figure 9. Simulated and measured radiation pattern on plane theta 90° at

Figure 7. Antenna close to a thick concrete wall 1.25GHz of the optimized antenna model
 Radiation Pattern Plane Theta 90° 2GHz IV. PRELIMINARY RESULTS OF THROUGH THE WALL RADAR
0o7 dB o
-30o 30 An experiment to verify the operation of the FMICW
-2 technique in conjunction with the optimized antenna presented
-60o 60o in this paper was set up in the large anechoic chamber at
-11 Durham University. As shown in Fig. 13 the set up consists of
two arbitrary waveform generators for the generation of the
-90o 90o FMCW signal with and without gating and two antenna
-11
elements have been manufactured to be used at the transmitter
-120o -2 120o and at the receiver. The received signal is mixed with a
reference signal which is a copy of the transmitted signal, and
-150o 7 dB 150o the resulting beat-note signal is filtered, amplified and recorded
180o by an acquisition unit. For this preliminary experiment a
Measurement wooden panel of thickness 3 cm was used as wall and placed
HFSS Simulation between the antennas and the targets area, at roughly 30 cm
CST Simulation
distance from the antennas.

Figure 10. Simulated and measured radiation pattern on plane theta 90° at
2GHz of the optimized antenna model

Radiation Pattern Plane Phi 0° 1.25GHz

0o6 dB
-30o 30o
-2
-60o 60o
-10

-90o 90o
-10

-120o -2 120o

-150o 6 dB 150o
180o
Measurement
HFSS Simulation
CST Simulation
Figure 13. Block diagram of the radar system

The antennas were placed 30 cm apart on a support and

Figure 11. Simulated and measured radiation pattern on plane phi 0° at oriented so that the main radiation lobe is directed towards the
1.25GHz of the optimized antenna model target area. For demonstration purposes a simple wooden
rectangle covered with aluminium foil at 1.5 m distance from
Radiation Pattern Plane Phi 0° 2GHz
the antennas was used as a target, as it can be seen in fig. 14
0o2 dB where the wall was temporarily removed to make this picture.
-30o 30o

-60o -4 60o

-90o 90o

-4
-120o 120o

-150o 2 dB 150o
180o
Measurement
HFSS Simulation
CST Simulation

Figure 14. Transmitter-receiver antenna set up with target in the anechoic

chamber
Figure 12. Simulated and measured radiation pattern on plane phi 0° at 2GHz
of the optimized antenna model
 In fig.15 two normalized radar range profiles are shown for
both the FMCW waveform and the FMICW waveform. The
FMCW return shows three peaks: the first peak at roughly 0.78
MHz which is caused by the cross-talk between the transmitter
and the receiver antennas, the second peak at 0.87 MHz which
is the backscattered reflection from the wall, and the third peak
at 1.3MHz which is the reflection from the target. The FMICW
return shows considerable reduction (around 20 dB) of the
unwanted components (cross-talk and reflection from the wall)
and an enhancement of the desired return from the target,
which is the strongest component of the FMICW return.
V. CONCLUSIONS
In this paper an optimized design of a patch-like antenna
was presented for UWB applications in the 0.5-2 GHz band.
While the optimization has resulted in a compact antenna
element with a single main lobe, the direction of the peak of the
radiation pattern is seen to move with frequency. A preliminary
test with a FMCW and FMICW signal shows the effectiveness
of the gating technique in reducing unwanted returns. Further
work is currently being undertaken to investigate other antenna
design elements and to optimize the gating waveform for
applications in real environments.
Measured radar range profiles for FMCW and FMICW
0 [7]
Target
Wall
-10 Crosstalk
Normalized Magnitude [dB]
-20
-30
-40
-50
FMCW
FMICW
-60
0.5 1 1.5
Frequency [MHz]
Figure 15. Example of FMCW and FMICW radar range profile from the
TTWI radar system
ACKNOWLEDGEMENTS
The authors would like to thank Her Majesty’s Government
Communications Centre for the sponsorship of this project, and
for the Fraunhofer IZM for hosting Francesco Fioranelli during
the development of the antenna.
REFERENCES
[1] M. Farwell, et al., "Sense through the wall system development and
design considerations," Journal of the Franklin Institute, vol. 345, pp.
570-591, 2008.
[2] N. Maaref, et al., "FMCW ultra-wideband radar for through-the- wall
detection of human beings," in Radar Conference - Surveillance for a
Safer World, 2009. RADAR. International, 2009, pp. 1-5.
[3] Y. Yang, et al., "Design of compact vivaldi antenna arrays for UWB see
through wall applications," Progress In Electromagnetics Research, vol.
82, pp. 401-418, 2008.
[4] T. S. Ralston, et al., "Real-time through-wall imaging using an
ultrawideband multiple-input multiple-output (MIMO) phased array
radar system," in Phased Array Systems and Technology (ARRAY),
2010 IEEE International Symposium on, 2010, pp. 551-558.
[5] R. Yu-Jiun, et al., "Compact Ultrawideband UHF Array Antenna for
Through-Wall Radar Applications," Antennas and Wireless Propagation
Letters, IEEE, vol. 8, pp. 1302-1305, 2009.
[6] K. E. Browne, et al., "Through-Wall Opportunistic Sensing System
Utilizing a Low-Cost Flat-Panel Array," Antennas and Propagation,
IEEE Transactions on, vol. 59, pp. 859-868, 2011.
J. M. Denoual, et al., "Design of two flare UWB antenna dedicated to
the research of alive buried victims," in Radar Conference (EuRAD),
2010 European, 2010, pp. 519-522.
[8] S. Salous, “Weighted sequences for HF FMCW sounding”, in HF Radio
Systems and Techniques, Seventh International Conference on (Conf.
Publ. No. 441), 1997, pp. 423-427.
2