632 A. LAZARO, J. LORENZO, R. VILLARINO, D.

GIRBAU, BACKSCATTER TRANSPONDER BASED ON FREQUENCY

Backscatter Transponder Based on Frequency Selective
Surface for FMCW Radar Applications
Antonio LAZARO, Javier LORENZO, Ramon VILLARINO, David GIRBAU

Dept. of Electronics, Electrics and Automatic Control Engineering, Rovira i Virgili University,
Av.Pasos Catalans 26, 43007 Tarragona, Spain
[email protected], [email protected], [email protected], [email protected]

Abstract. This paper describes an actively-controlled
frequency selective surface (FSS) to implement a back-
scatter transponder. The FSS is composed by dipoles
loaded with switching PIN diodes. The transponder ex-
ploits the change in the radar cross section (RCS) of the
FSS with the bias of the diodes to modulate the backscat-
tered response of the tag to the FMCW radar. The basic
operation theory of the system is explained here. An ex-
perimental setup based on a commercial X-band FMCW
radar working as a reader is proposed to measure the
transponders. The transponder response can be distin-
guished from the interference of non-modulated clutter,
modulating the transponders RCS. Some FSS with differ-
ent number of dipoles are studied, as a proof of concept.
Experimental results at several distances are provided.
Keywords
Transponder, FMCW radar, Frequency Selective
Surface (FSS), RFID.
1. Introduction
FMCW (Frequency Modulated Continuous Wave)
radar differs from pulsed radar in that an electromagnetic
signal is continuously transmitted [1]. The FMCW radar
emits a RF signal that is usually swept linearly in fre-
quency. The received signal is then mixed with the emitted
signal and due to the delay caused by the time of flight for
the reflected signal, a frequency difference, called beat
frequency, is produced. The range from the radar to the
target is proportional to the beat frequency. Due to their
ability to determine range, FMCW based systems are
commonly used for measuring distances in applications
such as tank level gauging, where high resolution non-
contact measurements in harsh conditions are required [2].
Other common applications include automotive collision
avoidance radars [3], [4], altimetry [1], or marine radars
[5]. It is also proposed for wall-penetrating radar with the
object of imaging and detection applications, security sen-
sors against intrusion [6], and human vital-sign detection
and measurement [7]. In recent years, FMCW radar has
been proposed in the literature for wireless local
positioning systems [8]-[10]. It has also been used as
reader for long distance transponders, integrating different
types of sensors such as pressure [11], [12] or temperature
sensors [13].
The major challenges for wireless local positioning
systems are disturbances caused by the multipath reflec-
tions [14]. Often, parasitic reflections are several orders of
magnitude greater than the target. Thus, clutter contamina-
tion is an important trouble in several applications such as
the determination of the radar cross section of pedestrians,
which is lower than that of cars or ground. In other cases,
tank walls can produce errors in level gauges [2]. Further-
more, FMCW radars have a minimum measurement dis-
tance limited by the antenna coupling and phase noise
power that produce a strong interference in the low-fre-
quency part of the output spectrum [1].
In order to mitigate these common problems,
a modulated backscatter transponder has been proposed in
[14-17]. In this case, the spectrum at the output of the
mixer is shifted by the modulated frequency of the trans-
ponder or tag, f
tag
[15]. The range measurement is per-
formed by analyzing the spectrum of the beat signal around
f
tag
and by verifying the presence of a couple of peaks. The
frequency difference between peaks is proportional to the
distance between the radar and the tag. Several approaches
have been employed to implement the backscatter trans-
ponder. The simplest is to match/mismatch the output of
the antennas transceiver achieving a modulation of the
backscattered field [14], [16]. Retroactive modulated Van
Atta arrays [17] have been employed both to reduce multi-
path effect by using directive antennas and to increase the
received signal strength. Another active solution is based
on a secondary radar, in which a transponder responds to
an interrogating radar signal [18]. Here the main challenge
is the delay synchronization between the radar and the
transponder [18]. One solution proposed is based on
a switched injection locked oscillator transponders [19].
Also, when using a secondary radar, the power consump-
tion of an RF oscillator is notably higher than the power
needed to modulate a semi-passive system whose operating
principle is based on matching or mismatching the output
of the antenna [19]. Therefore, these active solutions are
limited by battery life-time. The existing transponders
RADIOENGINEERING, VOL. 23, NO. 2, JUNE 2014 633
based on backscattering are limited to a range under 10 m,
whereas secondary radars transponders can typically reach
a hundred meters at the cost of higher power consumption,
more complexity and higher cost.
It is known that the range resolution in a FMCW ra-
dar is only limited by the sweep bandwidth [1]. This means
that resolution does not depend on the frequency of the RF
signal exactly. Thus, the backscatter transponder should
support the FMCW radar sweep bandwidth. Sometimes the
transponders presented in the literature are narrow band
[15], [17], [19].
The frequency selective surfaces (FSS) are periodic
structures in each direction of the plane which act as
a filter. This filter property has been exploited in several
applications [20], [21]. FSS can be frequency tunable,
introducing some switching elements (e.g. PIN diode) [22]
or tuning elements (e.g. varactors) [23], [24]. Recently, the
FSS modulating properties are also proposed in millimeter-
wave communications [25], Terahertz Applications [26]
and Radio Identification (RFID) tags design [27], [28]. In
[27] FSS are used to design narrow-band transponders.
However, wide bandwidth of the FSS will be required for
the purpose of this work to achieve high resolution. In
a recent work [28], the authors show that FSS can be used
to modulate the amplitude of time-domain UWB signals.
A tunable FSS based on loaded dipoles was used as a vari-
able filter in [28]. In this work, the UWB pulse was filtered
and the amplitude of the backscattered UWB pulse was
a function of the integrated backscattered field over the
entire pulse frequency spectrum. This fact is a fundamental
difference between time and frequency-domain backscat-
tered transponders in which the amplitude depends on the
differential radar cross section [29] at the interrogating
frequency. The aim of this work is to study the feasibility
to use FSS as low-power consumption and wide-band
backscatter transponders. Here, the large bandwidth that
can be achieved with FSS is used to modulate the FMCW
interrogating signal. The work demonstrates that the re-
quired bandwidth can be achieved using simple FSS based
on loaded dipoles with PIN diodes like those used in UWB
signals [28]. Depending on the diode state, the frequency
response of the FSS is shifted. Therefore, a large differen-
tial radar cross section (RCS) can be achieved over a large
frequency bandwidth. This frequency bandwidth is often
difficult to reach if backscatter transponders based on low-
profile antennas such as patch antennas are used. The typi-
cal single-layer patch printed on a dielectric substrate is
a narrow-band element. This well-known fact is mainly
due to the limitations imposed by the dielectric substrate.
For instance, to achieve the bandwidth, thicker substrates
must be employed or enhancement bandwidth techniques
must be considered [30]-[32]. Another challenge is the
transponder detection. This can be improved increasing the
differential radar cross section (RCS). To this end, an array
of antennas can be used. However, increasing the number
of FSS elements, the differential radar cross section can be
easily increased as will be shown in this work. The cost
that must be paid is an increase of the power consumption
due to the large number of diodes during the forward state.
However, reasonable values of RCS can be achieved with
a small number of dipoles, for example with 10 dipoles
with a current consumption of 2 mA assisted by a 3 V
battery.
The paper is organized as follows. Section 2 deals
about the basic theory of the backscattered transponder,
using the active-controlled FSS. Section 3 describes
a transponder designed as proof of concept. The design of
the active reflecting surface is based on electromagnetic
simulations. This section includes the experimental results
yielded. Section 4 draws the conclusions.
2. Tag Based on Frequency Selective
Surface
2.1 Range Measurement using Modulated
Transponder
A block diagram of the proposed system is shown in
Fig. 1. It is composed by the tag and the reader. In this
case, a FMCW radar is used as a reader enabling the meas-
urement of the tag to reader distance. The radar interro-
gates the tag which answers modulating its RCS. Depend-
ing on the application, data from sensors or identification
codes can be sent back to the reader. The tag consists of
a FSS loaded with PIN diodes that modulate the backscat-
tered field of the tag (see Fig. 2).
VCO
t T
V
0
A/D
LNA
PA
Z1 Z2
f
tag
FMCW RADAR
BACKSCATTER
TAG
CLUTTER
LPF
XT(t)
XR(t)
XIF(t)

Fig. 1. Schematic block diagram of a FMCW radar using
a modulated backscatter tag.
The determination of range of modulated backscatter
transponders has been proposed in several works, see for
example [8], [14], [17]. In this section, the basic theory of
operation is summarized for completeness. The transmitted
frequency is modulated by a saw teeth pulse of duration T.
The transmitted signal is given by:

2
1
( ) cos 2 ( )
2
T c
x t A f t t t
| |
= +
|
\ .
(1)
where A is the amplitude, f
c
is the carrier frequency,
= B/T, T and B are the sweep slope, duration and
bandwidth respectively.
The received signal is the transmitted signal
attenuated, delayed by the time of flight between the
transmitter and the tag, and modulated by the tag, with tag
frequency f
tag
:
634 A. LAZARO, J. LORENZO, R. VILLARINO, D. GIRBAU, BACKSCATTER TRANSPONDER BASED ON FREQUENCY
( )
2
0
1
( ) ' cos 2 ( ( ) ( ) cos 2
2
R c tag
x t A f t t f t t t t t |
| |
= + +
|
\ .

(2)
where ( = 2d/c), d is the transmitter to tag distance and c is
the speed of light, A is the received amplitude, |
0
is the
phase associated to tag oscillator.
The IF-signal at the output of the mixer and low-pass
filtered can be expressed as:

1
2
( ) '' cos(2 ( ) )
'' cos(2 ( ) )
IF tag
tag
x t A f t
A f t
t t |
t t |
= + +
+ +
(3)

where A is the amplitude at the IF-output, and |
1
, |
2
, are
phase constants generated from the mixing process that do
not influence in the result.
To determine the mixing frequency, the time signal of
the IF-signal is sampled. Then, a Fourier Transform and
a peak search are carried out to obtain a raw frequency esti-
mation. In practice, Fast Fourier Transform (FFT) with
zero-padding or chirp Z transform can be used to improve
the frequency resolution. In order to reduce side lobules
effect due to time-windowing, the Hamming Window can
be used instead of a Rectangular Window.

(a)

(b)
Fig. 2. (a) Diagram of a designed FSS (dimensions are
included). (b) Transponder and reflector used in the
measurement setup as reference.
From (3), the distance can be determined from the
frequency offset between the couple of peaks in the
spectrum around f
tag
.

2
2
d c
f d f
c
t

A = = = A
(4)
where f is the mixed frequency.
For a non-modulated target, f
tag
= 0, the distance to the
target is obtained using the well known FMCW radar
equation:


2 2
c cT
d f f
B
= A = A (5)

Assuming that the target is scanned during the entire
sweep, the signal is windowed by a rectangular time-win-
dow of duration T. Therefore, the precision in the determi-
nation of the mixed frequency is about 1/T. The range
measurement resolution of the system can be mainly esti-
mated as follows:

2
c
d
B
A =
. (6)

Hence it is proved that the range measurement resolution
depends only on the sweep bandwidth.
The two cases (non-modulated tag and modulated tar-
get) are schematically shown in Fig. 3. In Fig. 3a the tag is
difficult to be distinguished due to the phase noise and it is
interfered by the strong clutter reflections. Whereas when
the tag is modulated (Fig. 3b) a pair of peaks appears
around f
tag
whose separation is proportional to the distance
(4). In this last case, the tag can be easily detected. Fur-
thermore as the phase noise power decreases proportionally
with the frequency, the noise floor is limited by the re-
ceiver noise figure. In consequence the signal to noise ratio
is increased. To avoid aliasing, the maximum distance is
limited by the modulated tag frequency and the sampling
frequency:
0
tag
f t > , (7)
/ 2
tag s
f f t + < . (8)
Thus, the maximum non ambiguous distance is given by

/ 2
max ,
2 2
tag s tag
cf f f
d c



<
`

)
. (9)

S
p
e
c
t
r
u
m
S
p
e
c
t
r
u
m
(a) (b)
Fig. 3. Spectrum after low-pass filter in the IF stage:
(a) Non-modulated tag. (b) Modulated tag.

RADIOENGINEERING, VOL. 23, NO. 2, JUNE 2014 635
2.2 FSS as a Backscatter Modulator
When the transmitted signal hits the tag, a portion of
the power is backscattered towards the receiver. The tag
modulates the incident field by switching the PIN diodes
that load the FSS. In this work, the FSS is made as in [28]
by an array of dipoles. Fig. 2a shows a manufactured pro-
totype of 10 dipoles and a photography is shown in Fig. 2b.
Then, an active FSS can be viewed as an array of antennas
loaded with the impedance of the PIN diodes. Hence, the
backscattered field can be studied using fundamental an-
tenna scattering theory [33], [34].
The backscattered field of an antenna can be split as
a sum of two terms: a structural mode and an antenna
mode, or a load-independent term and a load-dependent
term, respectively:
( ) s est m
L
E Z E E = + I ,
(10)

where E
S
(Z
L
) is the scattered field by the tag connected to
the load Z
L
, E
est
is the scattered field when the tag is
connected to a reference load Z
L
= Z
a
*
, where Z
a
is the
antennas impedance. is the power reflection coefficient
given by [33]: = (Z
L
- Z
a
*
)/(Z
L
+ Z
a
).
When the load is the conjugate impedance of the
antenna impedance, all the incident power is transferred to
the load and the antenna only reflects the structural mode.
The structural mode arises from the induced current on the
antenna conducting surface by the incident wave, and it
does not depend on the load. The structural mode depends
on characteristics such as the antenna type, geometry, and
material. Thus, the structural mode is independent of the
load reflection coefficient, whereas the antenna mode is
proportional to it.
By switching the PIN diodes that load the dipoles of
the FSS, the reflect coefficient is modulated. As a first
approximation the reflection coefficient can be approxi-
mated by a square waveform with amplitude . It can be
developed in a Fourier series:
( ) ( ( ))
n c tag
n
f c f f nf o
+
=
I = +

(11)
where c
n
are the Fourier coefficients, and f
c
is the input
frequency of the incident signal that illuminates the tag.
For a square waveform with duty cycle , the coefficients
c
n
are given by:

, 0
sin
, 0
avg
n
n
c
n
n
n
to
o
to
I =

=

| |
AI =
|
\ .
(12)
where
avg
is the average power reflection coefficient
between on and off states.
The backscattered field can be expressed as:

( )
0
( )
( ( )
s est m
avg c
m
n c tag
n
E E E f f
E c f f nf
o
o
=
= + I
+ +

(13)
Therefore, the first term in (13) results in the non-
modulated term, and it depends on the structural mode of
the tag. The second term represents the modulated side-
bands that are function of the antenna mode. Considering
the coefficients of higher amplitude in the Fourier expan-
sion (n = 1), results in the components at the frequencies
f
c
f
tag
. Therefore the received signal at the radar can be
expressed by (2) after taking into account the propagation
delays. The radar cross-section of a target is a far-field
quantity that can be expressed as:

2
2
2
lim 4
s
d
i
E
RCS d
E
t

=
. (14)
The differential RCS [29] is the RCS due to antenna
mode that depends on the load reflection coefficient differ-
ence between the two modulated states (diodes on and
off). The differential RCS can be expressed as [29]:

2
2
1
2 2 2
2
lim 4
4
m
dif
x
in
E c
RCS d G m
E

t
t
= = AI (15)
where is the wavelength, G is the tag gain, and m is a
modulating factor that can be obtained from (12) as
m = |c
1
|
2
/||
2
.
In contrast with narrow-band RFID systems, here the
differential RCS must be maximized in the entire radar
frequency band.
3. Measurements and Results
3.1 FSS Design
Some prototypes of FSS have been manufactured
following the guide lines described in [28]. The FSS are
designed using Rogers 4003 substrate (relative permittivity

r
= 3.54, loss tangent tan = 0.003, and height of 32 mil).
The FSS are designed to cover the frequency band between
9.25 GHz and 10.75 GHz, which is the frequency band of
the radar used in the experimental results (Siversima model
RS3400X). Low-cost NXP BAP51-03 PIN diodes [35] are
selected to reconfigure the response of the FSS.
In order to understand the operation of a tunable FSS,
Fig. 4 shows schematically the reflectivity for the two ideal
diode states: ON, where the diode is forward-biased, and
OFF, where all the diodes are reverse-biased. It is known
[20] that the FSS presents a high reflectivity when the
frequency of the incident wave is close to the dipoles reso-
nant frequency (when the dipoles become approximately
half a wavelength). Therefore, the FSS can be modeled
with a LC series equivalent circuit (Fig. 4) whose resonant
frequency depends on the diode state. When the diodes are
in state ON, their impedance theoretically should be a short
circuit for an ideal diode, hence the FSS has a high reflec-
tivity and RCS around the resonant frequency of the di-
636 A. LAZARO, J. LORENZO, R. VILLARINO, D. GIRBAU, BACKSCATTER TRANSPONDER BASED ON FREQUENCY
poles. When the diodes are in state OFF, their impedance
should theoretically be an open circuit. Under this load
condition, the FSS is like an array of dipoles which are half
long than the forward-biased case. As a consequence, the
resonant frequency of the FSS loaded with an open circuit
is about the double that the case in which the dipoles of the
FSS are loaded with a short circuit. Therefore in this situa-
tion, the differential RCS should be higher between the two
resonant frequencies.
The next rules are followed to design the FSS:
1) The dipole width chosen is much lower than the
wavelength at the center frequency. In this case a 2 mm
width is selected allowing a smooth diodes weld.
2) Once the width of the dipole arms is chosen, the
length of the dipoles is computed to resonate at the low
frequency edge of the radar band. Thus, the frequency
band with high differential RCS covers the entire radar
frequency band. A first approximation, assuming that the
ideal diode in ON state is a short circuit, occurs when the
length is a half-wavelength. As the FSS grid is printed on
a dielectric then the resonance frequency becomes lower,
especially for thick substrates. The effective permittivity
can be estimated using results given in [20] (Appendix E).
However, due to the end open capacitance effect, the reso-
nance frequency is lower (about 10% depending on the
dipole width). Due to the effect of diode parasitics, the
resonance frequency of the FSS when the diodes are OFF
is lower than in the ideal open circuit [28]. Therefore, the
frequency band is reduced compared to an ideal diode case.
These parasitic effects change the current distribution
along the dipoles [20]. The influence of some of the above
factors can be quantified theoretically leading to generic
rules. The influence of the rest of the factors must be de-
termined numerically or experimentally. Therefore elec-
tromagnetic simulations must be performed to study the
bandwidth reduction and for the calculation of the dipoles
length. Several simulation electromagnetic methods such as
[20] can be employed to simulate. In this work, to perform
the electromagnetic simulations a Finite Element Method
in frequency domain implemented in the electromagnetic
full-wave Ansoft HFSS software is used. The diode is
taken into account in the simulator using a simplified
lumped RLC equivalent circuit [28]. Fig. 4 shows the elec-
trical circuit model of the BAP51-03 PIN diode [35] used
in the simulation for the two-states. In the forward-biased
case, the diode mainly represents a small resistance, which
has small effect on the desired response of the FSS. Be-
cause of its small value, the series self inductance of the
diode in this case should be considered for simulations.
However, when it is reverse biased, the parasitic capaci-
tance considerably deviates the position of the surface
stop-band by altering the total effective capacitance of the
unit cell. Therefore, it is necessary to consider its effect in
the design process.
3) Small spacing between the dipoles generates
a strong coupling between them and a reduction of the
effective permittivity and consequently of the resonance
frequency is produced. Moreover the effective area and the
RCS are smaller. However, the spacing cannot be arbitrar-
ily increased because it must be less than about 0.4 to
avoid grating lobes [20]. Here, it is fixed to about a quar-
ter-wavelength at the center frequency.
4) Metallic frames surrounding the FSS window may
influence on its behavior. Therefore, the bias lines must be
carefully designed. 10K bias resistors are used to set the
bias current to the diodes and they present high impedance
that allows to isolate the bias lines and the dipoles. The
bias lines between diodes are designed to be shorter than
the length of the arms of the dipoles in order to resonate at
a higher frequency, outside the frequency band of the ra-
dar. The other lines are orthogonal and isolated by the
resistors. It has been checked that the bias lines have
a small effect on the RCS within the radar frequency band.
Reflectivity
Frequency
f
0
2f
0
f
res
DIODES
ON
DIODES
OFF
f
res

Fig. 4. Equivalent circuit model of the FSS. Reflectivity for
the two diode states: State ON (diode is shorted). The
waves whose frequencies are close to the resonance
frequency are reflected. State OFF (diode is open-
circuit), the FSS is transparent for these.

Fig. 5. High-frequency equivalent circuit for the PIN diode
used.
Fig. 6 shows the simulation of the complex RCS
(magnitude and angle) for the two diode states. FSS with
the same spacing and dipole lengths but with different
number of dipoles (3, 4 and 5), and the case of an FSS with
10 dipoles in two columns is considered. The length of the
arm of the dipoles is 6.25 mm, the width of the dipoles is
2 mm, and the horizontal and vertical spacing are 7.5 mm
and 9.25 mm, respectively (see Fig. 2a, a schema of a 10
dipoles prototype). Fig. 7 shows the differential RCS ob-
tained subtracting the complex RCS for the two states. This
figure shows that the level of the differential RCS increases
because the effective area of the FSS increases, but also
that the bandwidth increases with the number of dipoles.
A peak of differential RCS is obtained when the phase
differences between the two diode states are maximum,
RADIOENGINEERING, VOL. 23, NO. 2, JUNE 2014 637
which happens approximately at the middle frequency
between the two resonances (when diodes are ON and
OFF). For practical applications, the number of the FSS
dipoles may be limited by its power consumption. In this
case, it is limited to 2 mA, considering an FSS made with
10 dipoles. From (13), the differential RCS of a backscat-
tering antenna can be computed. Using a 10-dipole FSS,
the differential RCS is the same as modulating an antenna
with a gain of 13.5 dB at 10 GHz, assuming the same
modulating factor m. To achieve this gain using a low-
profile antenna, an array of wideband patches should be
used.

Fig. 6. Complex RCS (magnitude, top figure) and phase
(bottom figure) in two different tags when the diode is
shorted or open-circuit.
D
i
f
f
e
r
e
n
t
i
a
l

R
C
S
(
d
B
)

Fig. 7. Differential RCS of different tags as function of the
number of dipoles.
3.2 Experimental Results
In order to characterize the frequency behavior of the
differential RCS, the experimental setup described in Fig. 8
has been used. A 20 dB standard horn antenna is connected
to a signal generator (Rhode SMF-100) and it was used to
illuminate the tag with a CW. The FSS is biased with
a low-frequency sinusoidal generator that generates
a 50 kHz 3Vpp sinusoidal signal (Agilent 33521A). This
signal is used to switch the diodes of the FSS. A receiver
horn antenna is connected to the spectrum analyzer (Rohde
FSP-30). Microwave absorbers are used to reduce multi-
path interference. Fig. 9.a shows an example of the spec-
trum measured by the spectrum analyzer when a 3-dipole
FSS is illuminated at 10 GHz. A strong peak at the trans-
mitted frequency due to the coupling between transmitter
and receiver antenna can be observed. The sideband peaks
are originated by the modulation of the tag and the fre-
quency space between each other is the modulation fre-
quency (50 kHz). Due to the clipping effect of the diodes,
the spectrum is very similar to that of a square signal (sinc
function). Therefore, the amplitude of the sideband is pro-
portional to the differential RCS (15). Fig. 9.b compares
the received power of different RCS after a correction of
the antennas gain is done. This figure shows that the re-
ceived power and the bandwidth increase with the number
of dipoles. The measurements show a frequency depend-
ence of the differential RCS, in agreement with simulations
of Fig. 7. The small difference is probably due to parasitic
effects that have not been taken into account in the diode
model during the simulations. For instance, it can be ob-
served that the frequency response has a flatness of 2 dB
for the case of a FSS with 10 dipoles.
Antennas
Transponder
Agilent 33521A
Rohde&Schwarz FSP
Rohde&Schwarz SMF100

Fig. 8. Photography of the system used for modulation
characterization.
P
o
w
e
r

(
d
B
m
)

(a)
P
o
w
e
r

(
d
B
m
)

(b)
Fig. 9. a) Measurement of the spectrum at 10 GHz. b) Com-
parison of the received power for different RCS after
the gain of the antennas was corrected.
Commercial radar from Siversima, Sweden (model
RS3400X/00) is used to demonstrate the viability of the
proposed solution. The RS3400X/00 is a synthesizer, X-
band, FMCW radar front end. The nominal transmitted
power of the radar module is 0 dBm+/-5 dB. The radar
sweeps the 9.25-10.75 GHz frequency band with a sweep
638 A. LAZARO, J. LORENZO, R. VILLARINO, D. GIRBAU, BACKSCATTER TRANSPONDER BASED ON FREQUENCY
time of 75 ms and it is connected to a 20 dB standard py-
ramidal horn. The IF signal is sampled at 20 kHz. Then the
chirp Z transform (a variant of FFT that does not require
zero padding to increase the frequency resolution) is ap-
plied to compute the IF spectrum. A Hamming Window is
applied before computing the spectrum. A 90 rectangular
dihedral reflector (20 cm by 29 cm) is located close to the
tag (see Fig. 2b). This reflector presents a high RCS
(19.7 dB oriented to the radar) and it is used as a reference
for comparison purposes. Therefore, the reflector intro-
duces a strong interference to the backscattered signal from
the tag. The FSS is modulated at 5 kHz. Fig. 10 shows
some results at different distances (at 2 m and 10 m). Pho-
tography of the experimental setup is shown in Fig. 11.
Without modulation, the tag signal cannot be detected
because it is masked by the strong reflection of the reflec-
tor, clutter and phase noise interference. However, when
the tag is modulated, two peaks appear at the spectrum of
the IF signal around the frequency modulation. Further-
more, some clutter removal can be achieved if the non-
modulated signal is subtracted. When the distance in-
creases, the separation of the frequency sidebands in-
creases according to the theory presented. The received
power and the signal to noise ratio decrease if the distance
increases as it can be viewed in the case of 10 m. Fig. 12
shows the estimated distance obtained from the measure-
ment considering the presence of the reflector that is also
employed as a reference and derived from the spacing
between the sideband peaks. A systematic offset between
the measured and the real distance from the radar to the
transponder is due to the delay produced by the length of
cables, which connect the radar and the antenna. This sys-
tematic offset has been subtracted. Good agreement is
obtained between the two measurements. It can be noticed
that the standard deviation error between the measurements
of the reflector distance (17 cm) and the tag distance
(11 cm) fall within the radar resolution given by (6) (in this
case, 10 cm).
0 1 2 3 4 5 6 7 8
0
50
100
150
P
o
w
e
r

(
d
B
)
Frequency (kHz)
FSS 10 dipoles - 2 m
Modulator OFF
0 1 2 3 4 5 6 7 8
0
50
100
150
P
o
w
e
r

(
d
B
)
Frequency (kHz)
Modulator ON
0 1 2 3 4 5 6 7 8
0
50
100
150
P
o
w
e
r

(
d
B
)
Frequency (kHz)
Background subtraction
0 1 2 3 4 5 6 7 8
0
50
100
150
P
o
w
e
r

(
d
B
)
Frequency (kHz)
FSS 10 dipoles - 10 m
Modulator OFF
0 1 2 3 4 5 6 7 8
0
50
100
150
P
o
w
e
r

(
d
B
)
Frequency (kHz)
Modulator ON
0 1 2 3 4 5 6 7 8
0
50
100
P
o
w
e
r

(
d
B
)
Frequency (kHz)
Background subtraction


Fig. 10. Measurement comparison of the tag and the reflector at 2 m and 10 m. Tag measurements in case of no modulation, or with f
tag
= 5 kHz
modulation. The background subtraction technique permits to recover the peaks clearly (black circles).


Ref. Technology
Central Frequency
(GHz)
Bandwidth
(MHz)
Radar EIRP (dBm) Measured Range Comments
18
Synchronized
VCO
5.8 150 10 550 m outdoor
Very large power consumption
and complexity
19
Switched Injection-
Locked Oscillator
5.8 140 14
25 m indoor
200 m outdoor
Large Power Consumption
(60 mW)
16
Modulated Backscatter
antenna
76 1000 20 3.8 m anechoic chamber
22 mA/5 V (110 mW)
GaAs PIN MMIC switch
17
Switched Van Atta
array
5.8 100 14 20 m indoor
Moderate power consumption
8 PIN SPSTs. Complex antenna
array design
This work
Active
modulated FSS
10 1500 15 12 m indoor
2 mA/3 V (6 mW), Low Cost PIN
diodes. Easy and scalable design
Tab. 1. Comparison with existing transponder technologies.
RADIOENGINEERING, VOL. 23, NO. 2, JUNE 2014 639

Fig. 11. Photography of the measurement system.
D
i
s
t
a
n
c
e

m
e
a
s
u
r
e
d

(
m
)

Fig. 12. Estimated distance obtained from the reflector meas-
urements (solid line) and distances derived from the
spacing between the sideband peaks (dashed line).
In Tab. 1 a comparison between the results obtained
above and the ones obtained applying other technologies
based on FMCW radar are made. The first two rows de-
scribe two technologies based on different secondary radar
solutions [18], [19]. The third row is based on a modulated
backscatter transponder [16]. The main characteristic of
Van Atta architecture [17] is described in the fourth row.
Finally, the last row shows the results reached in this work.
It can be seen that the power consumption of the active
secondary radar solutions is notably higher than the other
technologies and it is also a more complex solution. On the
other hand, the range in secondary radar systems is higher
because the signal is regenerated. Therefore the attenuation
ideally decreases with the square of distance in free-space
channel, whereas in the other transponders the received
power is inversely proportional to the distance raised to the
fourth. In the case of modulated solutions (the last three
cases), the range depends on the frequency, the transmitted
power and receiver sensitivity. When the frequency in-
creases, the free-space path loss increases and thus the
range is reduced. The results of this work evidence that
a moderate range and high resolution (high bandwidth) can
be achieved with a low power consumption and a low cost.
Although several diodes are used, it does not need to flow
large currents, because the impact in the differential RCS
of the FSS is small. The main impact in the differential
RCS is due to the OFF state of the diodes that is deter-
mined by the parasitic elements, but in this situation the
current is nearly zero and does not affect to the power
consumption. In Van Atta architectures, the power con-
sumption of the PIN switch is generally high in order to
reduce the insertion losses. On the other hand when the
frequency of operation increases, the cost and power con-
sumption of the switching elements also increases. The
proposed FSS is easily frequency scalable adjusting the
length and the spacing of the dipoles, if the design rules
given above are followed. The designs at millimeter fre-
quencies require diodes with low capacitance and parasitic
inductances, thus the cost and mounting complexity in-
creases if flip-chip diodes or chip MMIC soldered with
wire bonding are required.
4. Conclusions
This work has studied the feasibility to use actively
controlled frequency selective surfaces (FSS) with FMCW
radars. The communication between the tag and the reader
is produced using the backscattering technique. The basic
operation theory of the system is explained. In order to
increase the tag detection the maximization of the differen-
tial RCS is required. These types of radars are used in
applications in which a good range resolution proportional
to the bandwidth must be guaranteed. Hence, the FSS must
support this bandwidth. The simulations and experimental
results show that the bandwidth of a standard FMCW radar
is enough to exploit FSS in these conditions. Given that the
RCS is modulated, interference from non-modulated clutter
can be removed using a simple background subtraction
technique that consists on subtracting the modulated and
not modulated tag measurements. This technique is essen-
tial to be applied in indoor scenarios or in applications in
which strong clutter from reflectors are presented. An ex-
perimental setup based on commercial low-power FMCW
radar working as a reader is proposed to measure the
modulated radar cross section of the tags. Good results are
obtained up to 11 m using the Fig. 1 configuration in spite
of the presence of a reflector.
Acknowledgements
This paper was supported by Spanish Government
Project TEC2011-28357-C02-01 and BES-2012-053980.
References
[1] SKOLNIK, M. I. Introduction to Radar Systems. 2
nd
ed. Singapore:
McGraw-Hill, 1981.
[2] BRUMBI, D. Low power FMCW radar system for level gauging.
In IEEE MTT-S International Microwave Symposium Digest.
Boston, MA (USA), 2000, vol. 3, p. 15591562.
[3] ROHLING, H., MEINECKE, M.-M. Waveform design principles
for automotive radar systems. In Proceedings CIE International
Conference on Radar. Beijing (China), 2001, p. 14.
[4] JEONG, S.-H., YU, H.-Y., LEE, J.-E., OH, J.-N., LEE, K.-H.
A multi-beam and multi-range radar with FMCW and digital beam
forming for automotive applications. Progress in Electromagnetics
Research, 2012, PIER 124, p. 285299.
[5] HUANG, Y., BRENNAN, P. V., PATRICK, D., WELLER, I.,
ROBERTS, P., HUGHES, K. FMCW based MIMO imaging radar
640 A. LAZARO, J. LORENZO, R. VILLARINO, D. GIRBAU, BACKSCATTER TRANSPONDER BASED ON FREQUENCY
for maritime navigation. Progress in Electromagnetics Research,
2011, vol. 115, p. 327342.
[6] RALSTON, T. S., CHARVAT, G. L., PEABODY, J. E. Real-time
through-wall imaging using an ultrawideband multiple-input
multiple-output (MIMO) phased array radar system. In IEEE
International Symposium on Phased Array Systems and
Technology (ARRAY). Waltham (MA, USA), 2010, p. 551558.
[7] ANITORI, L., DE JONG, A., NENNIE, F. FMCW radar for life-
sign detection. In IEEE Radar Conference. Pasadena (CA, USA),
2009, p. 16.
[8] VOSSIEK, M., ROSKOSCH, R., HEIDE, P. Precise 3-D object
position tracking using FMCW radar. In 29th European Micro-
wave Conference. Munich (Germany), 1999, vol. 1, p. 234237.
[9] HUANG, Y., BRENNAN, P. V., SEEDS, A. Active RFID location
system based on time-difference measurement using a linear FM
chirp tag signal. In IEEE 19th International Symposium on
Personal, Indoor and Mobile Radio Communications (PIMRC).
Cannes (France), 2008, p. 1518.
[10] BRUGGER, M., CHRIST, T., KEMETH, F., NAGY, S.,
SCHAEFER, M., PIETRZYK, M. M. The FMCW technology-
based indoor localization system. In Ubiquitous Positioning
Indoor Navigation and Location Based Service (UPINLBS).
Kirkkonummi (Findland), 2010, p. 16.
[11] THAI, T. T., CHEBILA, F., MEHDI, J. M., PONS, P., AUBERT,
H., DEJEAN, G. R., TENTZERIS, M. M., PLANA, R. Design and
development of a millimetre-wave novel passive ultrasensitive
temperature transducer for remote sensing and identification. In
European Microwave Conf. Paris (France), 2010, p. 4548.
[12] AUBERT, H., CHEBILA, F., JATLAOUI, M., THAI, T.,
HALLIL, H., TRAILLE, A., BOUAZIZ, S., RIFAI, A., PONS, P.,
MENINI, P., TENTZERIS, M. Wireless sensing and identification
of passive electromagnetic sensors based on millimetre-wave
FMCW radar. In IEEE International Conf. on RFID-Technologies
and Applications (RFID-TA). Nice (France), 2012, p. 398403.
[13] PONS, P., AUBERT, H., MENINI, P., TENTZERIS, M. Wireless
passive autonomous sensors with electromagnetic transduction. In
International Conference on Microwave and High Frequency
Heating. Toulouse (France), 2011.
[14] THORNTON, J., EDWARDS, D. J. Range measurement using
modulated retro-reflectors in FM radar system. IEEE Microwave
Guided Wave Lett., 2000, vol. 10, p. 380382.
[15] KOSSEL, M., BENEDICKTER, H. R., PETER, R., CHTOLD,
W. Microwave backscatter modulation systems. In IEEE Int.
Microwave Symp. Dig. Boston (MA, USA), 2000, p. 14271430.
[16] SCHMID, C. M., FEGER, R., STELZER, A. Millimeter-wave
phase-modulated backscatter transponder for FMCW radar
applications. In IEEE MTT-S International Microwave Symposium
Digest (MTT). Baltimore (MD, USA), 2011.
[17] MARCACCIOLI, L., SBARRA, E. URBANI, L., GATTI, R. V.,
SORRENTINO, R. An accurate indoor ranging system based on
FMCW radar. In IEEE Intelligent Vehicles Symposium (IV).
Baden-Baden (Germany), 2011, p. 981986.
[18] RHR, S., GULDEN, P., VOSSIEK, M. Precise distance and
velocity measurement for real time locating using a frequency
modulated continuous wave secondary radar approach. IEEE
Transactions on Microwave Theory and Techniques, 2008, vol. 56,
no. 10, p. 23292339.
[19] VOSSIEK, M., GULDEN, P. The switched injection-locked
oscillator: A novel versatile concept for wireless transponder and
localization systems. IEEE Transactions on Microwave Theory
and Techniques, 2008, vol. 56, no. 4, p. 859866.
[20] MUNK, B. A. Frequency Selective Surfaces. New York: Wiley,
2000.
[21] MUNK, B. A., KOUYOUMJIAN, R. G., PETERS, L. Reflection
properties of periodic surfaces of loaded dipoles. IEEE Trans. on
Antennas and Prop., 1971, vol. AP-19, no. 5, p. 612617.
[22] IZQUIERDO, B. S., PARKER, E. A., ROBERTSON, J. B.,
BATCHELOR, J. C. Tuning technique for active FSS arrays.
Electron.Letters, 2009, vol. 45, p. 11071109.
[23] CHANG, T. K, LANGLEY, R. J., PARKER, E. A. Active
frequency-selective surfaces. IEE Proceedings - Microwaves,
Antennas and Propagation, 1996, vol. 143, no. 1, p. 6266.
[24] CHE, Y., HOU, X., GAO, Z. A tunable miniaturized-element
frequency selective surfaces without bias network. In IEEE
International Conference on Microwave Technology &
Computational Electromagnetics (ICMTCE). Beijing (China),
2011, p. 7073.
[25] KIANI, G. I., BIRD, T. S., FORD K. L. 60 GHz ASK modulator
using switchable FSS. In IEEE Antennas and Propagation Society
International Symposium (APSURSI). Toronto (Canada), 2010,
p. 14.
[26] KIANI, G. I., BIRD, T. S. Amplitude shift keying modulator based
on switchable FSS for terahertz applications. Journal of Radio
Science, 2011, vol. 46, RS2015.
[27] NEELAKANTA, P. S., STAMPALIA, A. K., DE GROFF, D.
An actively-controlled microwave reflecting surface with binary-
pattern modulation. Microwave Journal, 2003, vol. 46, no. 12,
p. 2235.
[28] LAZARO, A., RAMOS, A., GIRBAU, D. VILLARINO, R.
A novel UWB RFID tag using active frequency selective surface.
IEEE Transactions on Antennas and Propagation, 2013, vol. 61,
no. 3, p. 11551165.
[29] NIKITIN, P. V., RAO, K. V. S., LAM, S. F., PILLAI, V.,
MARTINEZ, R., HEINRICH, H. Power reflection coefficient
analysis for complex impedances in RFID tag design. IEEE Trans.
Microwave Theory Tech., 2005, vol. 53, no. 9, p. 27212725.
[30] TARGONSKI, S. D., WATERHOUSE, R. B., POZAR, D. M.
Design of wide-band aperture-stacked patch microstrip antennas.
IEEE Transactions on Antennas and Propagation, 1998, vol. 46,
no. 9, p. 12451251.
[31] POZAR, D. M., KAUFMAN, B. Increasing the bandwidth of
a microstrip antenna by proximity coupling. Electronics Letters,
1987, vol. 23, no. 8, p. 368369.
[32] POZAR, D. M. Microstrip antennas. Proceedings of the IEEE,
1992, vol. 80, no. 1, p. 7991.
[33] COLLIN, R. E., ZUCKER, F. J. The Receiving Antenna. Antenna
Theory 1. New-York: McGraw-Hill, 1969.
[34] GREEN, R. B. The General Theory of Antenna Scattering.
ElectroScience Laboratory, Columbus (OH, USA), 1963. Rep.
122317.
[35] BAP51-03. General purpose PIN diode (datasheet), NXP, 2004.
[Online] Cited 2014-02-27. Available at http://www.nxp.com/
documents/data_sheet/BAP51-02_N.pdf
About Authors ...
Antonio LAZARO was born in Lleida, Spain, in 1971. He
received the M.S. and Ph.D. degrees in Telecommunication
Engineering from the Universitat Politcnica de Catalunya
(UPC), Barcelona, Spain, in 1994 and 1998, respectively.
He then joined the faculty of UPC, where he currently
teaches a course on microwave circuits and antennas. In
RADIOENGINEERING, VOL. 23, NO. 2, JUNE 2014 641
July 2004 he joined the Department of Electronic Engi-
neering, Universitat Rovira i Virgili, Tarragona, Spain. His
research interests are microwave device modeling, on-
wafer noise measurements, monolithic microwave inte-
grated circuits (MMICs), low phase noise oscillators,
MEMS, and microwave systems.
Javier LORENZO was born in Reus, Spain, in 1978. He
received the Telecommunications Technical Engineering
degree and MS in Electronics Engineering from Universitat
Rovira i Virgili, Tarragona, Spain, in 2009 and 2011 re-
spectively. Since September 2011 he is a researcher in his
PhD at Universitat Rovira i Virgili, Tarragona, Spain. His
research activities are oriented to microwave devices and
systems, with emphasis on UWB, RFIDs, wireless sensors,
radiometry and positioning.
Ramon VILLARINO received the Telecommunications
Technical Engineering degree from the Ramon Llull Uni-
versity (URL), Barcelona, Spain in 1994, the Senior Tele-
communications Engineering degree from the Universitat
Politcnica de Catalunya (UPC), Barcelona, Spain in 2000
and the PhD from the UPC in 2004. During 2005-2006, he
was a Research Associate at the Technological Telecom-
munications Center of Catalonia (CTTC), Barcelona,
Spain. He worked at the Universitat Autonma de Cata-
lunya (UAB) from 2006 to 2008 as a Researcher and As-
sistant Professor. Since January 2009 he is a Full-Time
Professor at Universitat Rovira i Virgili (URV). His re-
search activities are oriented to radiometry, microwave
devices and systems, based on UWB, RFIDs and frequency
selective structures using MetaMaterials (MM).
David GIRBAU received the BS in Telecommunication
Engineering, MS in Electronics Engineering and PhD in
Telecommunication from Universitat Politcnica de Cata-
lunya (UPC), Barcelona, Spain, in 1998, 2002 and 2006,
respectively. From February 2001 to September 2007 he
was a Research Assistant with the UPC. From September
2005 to September 2007 he was a Part-Time Assistant
Professor with the Universitat Autnoma de Barcelona
(UAB). Since October 2007 he is a Full-Time Professor at
Universitat Rovira i Virgili (URV). His research interests
include microwave devices and systems, with emphasis on
UWB, RFIDs, RF-MEMS and wireless sensors.