2011 IEEE International Conference on RFID

Reference Modulation for Calibrated Measurements

of Tag Backscatter
Daniel G. Kuester∗†, Student Member, IEEE, David R. Novotny∗, Jeffrey R. Guerrieri∗ ,
Randal H. Direen∗, and Zoya Popović†, Fellow, IEEE
∗
U.S. Department of Commerce, National Institute of Standards and Technology, Boulder, CO, USA 80304
{daniel.kuester, david.novotny, jeffrey.guerrieri, randy.direen} at nist.gov
†
University of Colorado, Department of Electrical, Computer, and Energy Engineering, Boulder, CO, USA 80309
{daniel.kuester, zoya.popovic} at colorado.edu

Abstract—This paper presents an approach for calibrating Existing test standards leave significant unresolved chal-
backscattering measurements from 860-960 MHz Ultra-High Fre- lenges, particularly in the return link. They do not yet ad-
quency Radio Frequency Identification (UHF RFID) tags. An S- dress either reader sensitivity or interference rejection, though
parameter model is formulated to relate diode switch and antenna
input circuit parameters with the scattering performance of the previous work has suggested that reader sensitivity is be-
calibration device. Measurements of modulated backscattered coming a more significant system range constraint [2], and
power agree with the model to within ±0.1 dB. Tag backscatter that readers are more vulnerable to interference than tags
measurements can then be calibrated by comparing them to the [1]. Standard ISO 18047-6 [5] prescribes a tag backscattering
reference signal. In an example testbed, the expanded uncertainty performance test characterized as the difference between the
of these measurements is estimated to be ±0.4 dB, compared
with uncertainties worse than -0.9 dB, +1.2 dB for methods that radar cross section values between its two load modulation
calibrate against radar cross section (RCS) standards in the same states. The prescribed test method calibrates measurements of
testbed. tag backscattering against the change in received power from
adding a thin rod to the test environment.
I. I NTRODUCTION The measurement uncertainty of one test performed this
way was estimated to be approximately 2 dB [6]. The current
Continued technical advances in RFID systems are allowing ISO/IEC 18047-6 method introduces further systemic error by
passive transponders (“tags”) to operate with increasingly directly subtracting power quantities, neglecting phase effects,
lower powers received from transceivers (“readers”) in UHF though existing works show how phase can be considered
860-960 MHz bands. This trend enables operation at greater correctly [7][8][9]. Adding and removing the entire thin rod
range, but also means that backscattered replies from tags are calibration standard introduces systemic error via structural-
becoming fainter. If this trend continues, existing interference mode modulation, which interacts with multipath in the test
effects [1] will be made worse, and reader detection of environment differently [2] than tags’ antenna-mode modula-
tags may become a dominant constraint on communication tion [7][8]. The use of such an electrically small calibration
in practical deployments [2]. Unfortunately, performance test target requires faith in the accuracy of the analysis used to
standards are immature for tags, and do not yet exist for compute its RCS, which makes the measurement result not
readers. The lack of strong test standards may add effort and traceable to fundamental physical standards of any national
expense to the designing robust inventory tracking systems. metrology laboratory. These errors may make measurement
System operation, in contrast, has been more extensively results challenging to repeat between different testbeds, and as
standardized. Typical commercial systems comply with the a result some parties may choose not to undertake the expense
EPC Global “UHF Class 1 Gen 2” and ISO 18000-6C stan- of running the tests.
dards, which outline the operation and parameters of the half- To address these shortcomings, a new model and calibration
duplex protocol used for UHF RFID. In the forward link, a device for traceable, more accurate measurements of backscat-
reader transmits a modulated carrier towards a tag field, which tered tag power are proposed in this paper. The contributions
harvests the incident energy to supply power for communi- are organized as follows:
cation and processing circuitry. Each tag responds by time- • Section II introduces the model, based on S-parameters;
varying its antenna load impedance, encoding information in • Section III shows how a calibration device can be real-
the backscattered carrier. This modulated backscattering im- ized with a horn antenna and a laboratory switch, and
plementation, known as load modulation, forms the return link. characterized with the developed model parameters;
Similar processes have been used for antenna measurements • Section IV presents the technique for calibrating tag
[3] and espionage [4]. backscatter with the calibration device;
• Section V shows a calibrated tag backscatter measure-
U.S. GOVERNMENT WORK — Not subject to U.S. copyright. ment with the new model and device.

U.S. Government work not protected by U.S. copyright 154

 mission are S23 = S32 = b!3 /a!2 and S31 = S13 = b!3 /a!1 ,
and can be measured conveniently with a network analyzer if
they are connectorized. Port 3 bisects the reference backscatter
standard. The reflection coefficients of the backscattering
antenna and the modulator are S33 = b!3 /a!3 and ΓL = b!L /a!L ,
respectively. These coefficients are measurable with a network
analyzer by disconnecting the reference backscattering antenna
from the modulator.
When the two networks are connected (as in Fig. 1b),
an incident carrier wave from port 1 arrives at the load
modulator, which is time-varying its impedance between two
states. Reflections off of this interface arrive at port 2 with two
corresponding states that have in-phase and quadrature (IQ)
voltage components illustrated in Fig. 2.
The modulated signal ∆Vbs is related to a transmit signal
Vtx in an arbitrary environment as follows. The traveling
1/2
waves a1 and b2 in Fig. 1b are a1 = Vtx /Z0 and b2 =
1/2
Vbs /Z0 , with purely real port impedances Z0 . An expression
for the received wave b2 in this system with port 3 loaded by
the tag is [10]
Fig. 1. Reference backscattering model illustrating S-parameters with the ! " ! "
modulator (a) disconnected and (b) connected to the transmission network. S31 S23 ΓL S32 S23 ΓL
When connected, a transmitter delivers a traveling wave a1 into port 1 of b2 = S21 + a1 + S22 + a2 .
an arbitrary transmission system S, which is described in (a). The reflection 1 − S33 ΓL 1 − S33 ΓL
coefficient presented to waves arriving at port 3 switches between ΓL → ΓL1 (1)
(the ZL1 state) and ΓL → ΓL2 (the ZL2 state). Reflection back to port 2 Assuming that Port 2 is not transmitting (a2 = 0), the
is the traveling wave b2 . The transmit-to-receive transmission coefficient in a transmission coefficient τ = b2 /a1 between the transmit and
tag state is τ = b2 /a1 .
receive ports becomes
S31 S23 ΓL
τ = S21 + . (2)
1 − S33 ΓL
This relationship has been observed for other two-port mod-
ulated scattering models [11] [12]. If the modulator load
alternates between {ZL1 , ZL2 }, corresponding to reflection
coefficients ΓL → {ΓL1 , ΓL2 }, then the change in the trans-
mission coefficient ∆τ will be

∆Vbs ΓL2 − ΓL1

∆τ = = S31 S23 . (3)
Vtx (1 − S33 ΓL2 )(1 − S33 ΓL1 )
Fig. 2. Illustration of demodulated ∆Vbs on the complex (IQ) plane received The term on the right gives a convenient definition for a
from a reference backscatterer. The signal is shown in rectangular coordinates
as Re(∆Vbs ) + jIm(∆Vbs ), and in polar form as magnitude and phase. return modulation depth
ΓL2 − ΓL1 ∆τ
M= = , (4)
(1 − S33 ΓL2 )(1 − S33 ΓL1 ) S31 S23
II. L INK AND LOAD MODULATION MODEL
relating system behavior through the transmission network
A. Derivation (∆τ , S31 , and S23 ) with measurable parameters inside the
The proposed model for transmission and backscattering backscatter reference (ΓL1 , ΓL2 , and S33 ).
between an interrogator, and modulator (realized in subsequent The relationship in (3) expressed with received modulation
sections as a calibration device) is illustrated in Fig. 1. The power Pbs between the two switched states at the reader,
linear and reciprocal transmission network is shown as S, transmission link losses |S31 |2 and |S23 |2 , and the carrier
where interrogation signals are incident at port 1, backscatter- transmit power Ptx is†
ing is measured at port 2, and the backscatter modulator loads
port 3. This network can incorporate any cables, antennas, Pbs |∆Vbs |2 /2Z0
= = |∆τ |2 = |S31 |2 |S23 |2 |M |2 . (5)
linear propagation effects, or other linear circuit elements in Ptx |Vtx |2 /2Z0
the testbed. † Power quantities given in this model are available power. Finding power
If the two networks are disconnected (Fig. 1a), transmission delivered into a receiver or from a source must additionally account for
coefficients between the loaded ports in the reciprocal trans- mismatch between the impedances on either side of the network interface.

155
 If the input propagation environment (an antenna or trans- antenna reciprocity (|S31 |2 = |S23 |2 in (5)) and the Friis
mission line) is matched to an instrument’s characteristic transmission equation, the backscattered power is
impedance Z0 (S33 = 0), then the modulation index in (5) ! "4
expressed in terms of power |M |2 is bounded by |M |2 < 4. Pbs λ
= G2 G2bs |M |2 , (6)
This is the case of two modulator states with reflection Ptx 4πR
coefficients of with magnitude 1 that are out of phase. If where R is the range between the measurement and backscat-
S33 #= 0, |M | can become quite large or small, but energy tering antennas and λ is the free-space carrier wavelength.
is conserved because the mismatch reduces the magnitude of This makes the same assumptions as the radar equation, and
the transmission coefficients S31 and S23 . has the same form, so it can be related to the modulated radar
This relationship applies most directly to bistatic mea- cross section σmod (see for example [14]) used for tags:
surement antenna configurations, with separate transmit and
receive antennas corresponding to ports 1 and 2 of Fig. 1. It λ2
σmod = G2bs |M |2 . (7)
can also apply to monostatic configurations if S incorporates 4π
a circulator. Relating the S-parameter network model to modulated RCS
in free space therefore requires additional information about
B. Comparison with other scattering models the gain of the backscattering antenna.
The best-known modulated tag scattering characteristic is
III. A REFERENCE SOURCE OF BACKSCATTER
known as either a “differential” (e.g. [7] [13]) or “modulated”
MODULATION
([14]) radar cross section (RCS). These terms are also some-
times used interchangeably with the terms “delta RCS” or The model in the previous section suggests that a carefully
“∆RCS” used in ISO 18047-6, though this term may cause characterized source of reference modulation with a known
confusion over the role of phase in tag scattering. modulation depth (|M |) and measurements of link losses
The modulated RCS model is formulated with power waves (|S31 | and |S23 |) are sufficient to predict the backscattered
[7][13][14][15], which should not be confused with the S- reference power. This section introduces a device that can be
parameter traveling waves used in this paper. The power wave characterized to serve this purpose, and how it can be validated
parameters are convenient for chip-to-antenna matching in tag as an accurate reference for calibrating tag backscatter.
design. In metrology, however, power waves are not suit- A. Calibrated signal generation
able for extracting physical parameters like impedance from
measured data [16], making them unsuitable for calibration The modulator shown in Fig. 3 realizes the model in Fig. 1.
purposes. The switch has a nominal 20 ns rise/fall time to within 10%
of steady state, which is fast enough to emulate the maximum
A more recent model of tag scattering based on reciprocity
640 kHz symbol rate by tags compliant with ISO/IEC 18000-
and impedance (Z-) parameters was introduced in [8], and can
6C.
be shown to be equivalent to the work presented here. Like
A standard gain horn with a measured gain of 6.69 dBi at
ours, this model requires knowledge of a chip and antenna
1 GHz is the reference backscattering antenna. It has return
circuit parameters. The choice of S-parameters instead Z-
loss greater than 10 dB across 860-960 MHz, corresponding
parameters was motivated by practical because far-field prop-
to |S33 |.
agation effects between antennas are understood in terms of
Instruments with 50 Ω input impedance serve both as a
traveling waves (rather than mutual impedances), and because
matched load modulation state (with |ΓL2 | ≈ 0), and to allow
microwave instrumentation uncertainty is expressed in terms
measurements of interrogation signal link losses. A network
of S-parameters.
analyzer makes a convenient matched load for characterizing
This paper is not intended to offer a “black box” characteri-
the device, and a power meter is subsequently used to measure
zation like RCS for tag testing. Direct application of the model
received interrogation power. An additional 3 dB pad between
presented here to tag characterization would require chip and
the switch output and the matched load attenuates reflections
antenna parameters either by prior knowledge or destructive
between the horn and the instrument. The other switch load is a
disassembly of a tag.
short, for |ΓL1 | ≈ 1, though the actual |ΓL1 | is approximately
2 dB smaller because of switch insertion loss.
C. An idealized free-space example
With |ΓL2 − ΓL1 | ≈ 1, and the backscatter antenna approx-
Though the goal of this work is not to present RCS mea- imately matched for small |S33 |, the anticipated |M | is near
surements, the relationship between the S-parameter scattering 0 dB.
model and modulated RCS can give some insight. As a
simple test case, consider an idealized monostatic system in B. Measuring modulation depth
an anechoic environment. The backscattering and detection To become suitable for calibrations with reference backscat-
antennas are well-matched and co-polarized, and we assume ter, the modulation depth must be measured accurately. Equa-
there are no near-field effects. Measurement and backscattering tion (4) gives a choice between two sets of parameters
antennas have gains G and Gbs towards each other. With that can be measured with a network analyzer: “wireless”

156
 0#2

M through antennas (dB)

1
0#1

Error (dB)
0
0#0
!1
!0#1
Measured ΓL2 "ΓL1 "S33
!2 Measured ∆τ"S31 ! S23
(a) !0#2
800 850 900 950 1000
Frequency (MHz)

Fig. 4. Validation of the reference backscatter with a network analyzer in

a semi-anechoic test environment, computed with measurements of the terms
in (4). The curves agree to ±0.1 dB over the 860-960 MHz tag response
bandwidth.

C. Validation
Because the model in equation (4) gives two expressions for
M in terms measurable network parameters, the modulation
depth characterization can be validated by measuring the
parameters for each with a network analyzer and comparing
(b) the results.
The validation tests, for simplicity, were performed is mono-
Fig. 3. Simplified schematic (a) and realization (b) of a modulation switch for
reference backscatter calibrations. The load ZL1 is intended to connect with
static. The transmit and receive antenna is a commercial RFID
a matched 50 Ω instrument such as a power sensor or network analyzer, to patch with at least 10 dB return loss across 895-940 MHz.
measure power delivered to the backscatter reference and serve as a matched The change in reflection coefficient ∆ρ into the antenna
reflection state for modulation. The device is mounted in a 33 cm × 18 cm
× 5 cm shielded box with ±5 V DC biasing inputs, and bias tees to improve
was taken to be equivalent to ∆τ , with the simplification
DC to RF isolation. S23 = S13 . The anechoic environment reduces interference
from outside signals, but the calibration for backscattered
signal levels applies in other, more reflective environments too,
if interference is below a tolerable level.
measurements of S23 , S31 , and ∆τ , or reflection coefficient
measurements of ΓL2 , ΓL1 , and S33 . In either case, the Results are shown in Fig. 4. Across the 860-960 MHz
modulation state of the modulator is fixed with a DC voltage tag response bandwidth, the two measurements of M agree
supply during measurement. within 0.1 dB. Below 860 MHz, detection antenna mismatch
introduces additional noise in transmission measurements S31
One advantage of calibrating M from measurements of and ∆ρ, because received signals are weaker.
ΓL2 , ΓL1 , and S33 directly is that (to first order) measure-
ment dynamic range is not reduced by moving the reference IV. C ALIBRATION WITH REFERENCE MODULATION
backscatter antenna in an anechoic test environment. Further,
ΓL2 and ΓL1 can be measured with phase-stable cables near A. Calibration procedure
the network analyzer more accurately than with the long cables 1) Reference |M | calibration: The validated characteriza-
that are necessary to measure reflection coefficients of objects tion of the reference backscatter modulation depth M from
inside the test chamber. the previous section is used. In this work, tag calibrations use
Calibrating M from ∆τ and measurements of propagation the modulation depth based on measurements of ΓL2 , ΓL1 ,
losses S23 and S31 has different advantages. Transmission and S33 .
measurements of propagation losses can have smaller uncer- 2) Transmission loss measurements: During tag measure-
tainties than the reflection measurements in the first approach, ments, it is impractical to measure transmission coefficients
but any motion in the long cables may introduce additional |S31 |2 = |S13 |2 and |S23 |2 = |S32 |2 with a network analyzer.
phase errors. This calibration also needs fewer measurements, Instead, we use power sensors to measure (1) transmitted
which may reduce the contribution of operator mistakes to interrogation power Ptx available to port 1 or 2 with a coupler,
measurement error. and (2) power received at the output of the modulator switch,
(ref )
Detailed quantitative comparison of uncertainties in these Prx . In this paper, transmit, received (by transmission),
approaches will be left for future work. The following subsec- and backscattered power from an interrogation into port n are
tion will validate that either set of measurements can produce represented as Ptx,n , Prx,n , and Pbs,n . Assuming the network
a valid characterization of |M |. analyzer and power sensors are similarly well-matched, power

157
 Instead, states are clustered by applying a Gaussian filter
to the I and Q scattering components, and finding sharp
peaks after differentiating the result. These peaks are reported
when they are a local amplitude maximum; this threshold
was set somewhat arbitrarily to the standard deviation of the
filtered signal. The Gaussian filter is convenient here because it
removes noise without shifting pulse edges; it is a simpler one-
dimensional implementation of Canny edge detection [18].
If ∆VI and ∆VQ are the differences between the averaged
two backscattering state measurements of each component,
∆Vbs is computed as (3),

∆Vbs = (∆VI + j∆VQ ) exp jφ, (10)

Fig. 5. A demodulated trace from a transaction at 910 MHz with an ISO/IEC
18000-6C tag received by a spectrum analyzer. It shows leaked interrogation
modulation from the forward link, the tag response from the reverse link, and where the exponential term accounts for the arbitrariness I-
reference backscatter from the calibration device introduced in this paper. In Q plane rotation. For vector voltage quantity ∆Vbs , averaging
use, the reference backscatter is only turned on when it is being measured,
to avoid interfering with the tag. across neighboring state transitions in different frames reduces
measurement noise, potentially to below -100 dBm.
The power in these two states, corresponding to Pbs from
measurements and transmission losses are related with (5), was calculated as
(ref )
2
Prx,1
2 |∆Vbs |2 |∆VI + j∆VQ |2
|S31 | = |S13 | = (8) Pbs = = , (11)
Ptx,1 2Z0 2Z0
and in agreement with previous work [7][9][13][19].
(ref )
Prx,2
|S32 |2 = |S23 |2 = . (9) 4) Correcting the tag backscatter measurement: Assuming
Ptx,2 bandwidth of all backscattered signals are narrow about the
Prx /Ptx,1 and Prx /Prx,2 are taken as |S32 |2 and |S31 |2 . Both interrogating carrier, and that cable and antenna mismatch and
sensors are configured to measure average power during the losses are linear with power, the fractional power lost will be
period after the tag reply while the interrogation power is left the same for both a tag and reference backscatter:
on. This period was set to 1 ms, which is longer than that of
(ref,meas) (tag,meas)
typically deployed readers to reduce noise by averaging. Pbs Pbs
(ref )
= (tag)
. (12)
Loss in the switch reduces the measured power compared Pbs Pbs
(ref )
to the available Prx out of the backscattering antenna. To
(ref )
“back out” Prx , the full two-port scattering parameters of This can be rearranged to find “true” backscattered power
the switch are used to de-embed the power available out of received from the tag,
port 3 with transfer (T-) parameters.
(tag,meas)
3) Backscattered signal measurements and processing: (tag) Pbs (ref )
(tag,meas) (ref,meas) Pbs = (ref,meas)
Pbs . (13)
Backscattered power Pbs and Pbs are mea- Pbs
sured with IQ demodulation on a spectrum analyzer. Example
signal traces are shown in Fig. 5. The interrogation signal is (ref )
Equations (5), (8), and (9) can be substituted for Pbs , so
an ISO/IEC 18000-6C query command, with tag and reference for interrogation through port 1,
responses modulated at 160 kHz.
Tag responses to query requests are measured in a 240 µs (tag,meas) (ref )
(tag) Pbs,1 Prx,2 (ref )
period, and reference backscatter in 1 ms. Connecting the Pbs,1 = (ref,meas) P
Prx,1 |M |2 . (14)
10 MHz frequency reference from the spectrum analyzer to P bs,1
tx,2

the RFID interrogation source, carrier phase drifted by less

or through port 2,
than 1.3◦ per symbol at 160 kHz modulation, introducing
negligible error to measurements of signal levels between (tag,meas) (ref )
(tag) Pbs,2 Prx,1 (ref )
adjacent symbols. Across entire tag or reference signal traces, Pbs,2 = Prx,2 |M |2 . (15)
(ref,meas) P
however, the carrier drifts by up to 13◦ . This made separating P bs,2
tx,1
the signal into the two states in software (“clustering”) more
challenging: slow drift in each of the I and Q components Notice that, like the RCS calibration in ISO 18047-6, effects
was often larger than the backscattered signal, making the of mismatch or cable losses between the receive antenna and
straightforward histogram analysis suggested by IEEE 181 the measurement instrument are removed in the calibration
[17] unsuitable. process.

158
 TABLE I
ISO/IEC 18000-6C TAG QUERY PARAMETERS
Reader-to-tag modulation PR-ASK
Tag-to-reader modulation FM0
Tag-to-reader link rate (BLF) 160 kHz
Reader-to-tag link rate 160 kHz (data 0)
91 kHz (data 1)
Anticollision slots (Q) 0 (no slots)
Delay after tag response† (T2) 1 ms
Tari 6.25 µs
† Reference backscattering was measured during this period

0.1
Transmit into 6 dBi antenna
Transmit into 8 dBi antenna

/Ptx (dB)
(re f )
Normalized Prx
0.0
Fig. 6. Setup for calibrated tag backscattering measurements. Measurements
were automated by computer via general purpose interface bus (GPIB).

−0.1
5 10 15 20 25 30
Transmit power (dBm)

Fig. 8. Testbed linearity tests, performed by sweeping the power of the

interrogating wave and measuring the reference backscattered power. The
backscatter reference load-modulated 910 MHz carrier reflections at 160 kHz
with the circuit described in Fig. 3. Deviation from linearity below 32 dBm
input power was less than 0.1 dB.

with a comparison against other measurement approaches

taken in the same testbed. Table IIb, for example, estimates
uncertainty in hypothetical backscattered power measurement
Fig. 7. Measurements of backscattered power comparing raw tag and based on the spectrum analyzer’s built-in alignment, with cable
reference signals with the calibrated measurement of the tag. The reference
was switched at 160 kHz modulation, and the tag was interrogated with signal losses measured and de-embedded. The measurement in Table
parameters described in table I. Signals from each were measured with IQ IIc is meant to reflect the uncertainty of a backscattering
demodulation as in Fig. 8. The calibration used M calculated with equation measurement that has been careful calibrated with a thin
(4) from measurements of ΓL1 , ΓL2 , and S33 .
λ/2 rod as a scattering standard, based in part on results
from [2]. All of these measurements are assumed to use IQ
V. E XAMPLE CALIBRATIONS demodulation on the same spectrum analyzer.
Testbed linearity is estimated from measurements of the
A. Calibrated tag measurements reference backscatter as a function of power, shown in Fig.
With the procedure established, we next present an example 8. This measurement was performed with a bistatic receiver
tag backscatter measurement, calibrated with the backscatter- setup. Results between the two traces agreed within ±0.1 dB,
ing source. The overall test setup is summarized in Fig. 6 and which is then the estimate for linearity and noise uncertainty.
Table I. Instruments are controlled with computer via general Reported values follow the methods for evaluating uncer-
purpose interface bus (GPIB). tainty expressed in [20] as closely as possible. Each error
The parameters from Table I were used in the calibrated source is assumed to have a zero mean normal distribution,
tag backscatter measurement shown in Fig. 7, showing highly except multiple reflection errors that are assumed to have a
nonlinear raw and calibrated measurements of the tag under U-shaped distribution. Supplied error estimates are type B,
test with the reference backscatter for comparison. Reference except the noise and linearity term, estimated from 8, and the
backscatter was turned off during measurements of the tag IQ level measurement errors, estimated by simulation of noisy
response, to prevent interference. bandlimited pulses.
These uncertainty terms, because they are shown in decibels,
B. Measurement errors and uncertainty are expressed relative to the actual measurement value. Each
To help gauge the effectiveness of the reference backscatter is listed as an expanded uncertainty, and corresponds to a
calibration, an uncertainty estimate is presented in Table II, fractional term ui that is the contribution of the individual

159
 TABLE II
E XPANDED UNCERTAINTY ESTIMATES ( COVERAGE FACTOR 2) standard in the measurement of Table IIc is performed by (very
(a)
slowly) moving the standard by hand, the ±13◦ phase error
Backscattered tag signal measurement becomes significant, and corresponds to a worst-case ±0.23 dB
(with reference backscatter calibration) uncertainty.
M calibration ±0.25 dB
Power measurements ±0.25 dB
The measurement technique presented in this paper has
Multiple reflections ±0.15 dB some unique uncertainty terms. Errors introduced by calibra-
IQ level measurements ±0.05 dB tions of M are a caused by reflection coefficient measurement
Noise and nonlinearity ±0.1 dB
Expanded uncertainty ±0.4 dB
uncertainty of the network analyzer and the short-open-load
calibration standards. The corresponding error estimate is
(b) based on measuring verification load standards on the network
Backscattered tag signal measurement
(without scattering calibration)
analyzer after calibration, and on analysis of manufacturer un-
Transmit power measurements ±0.15 dB certainty specifications. The power measurement uncertainty
Measurement cable losses ±0.1 dB ±0.25 dB is also based on analysis of the manufacturer’s
Instrument uncertainty ±0.9 dB power sensor specifications, from analysis of correlated and
IQ level processing ±0.15 dB
Noise and nonlinearity ±0.1 dB uncorrelated error terms between the two power measurements
Expanded uncertainty −1.2 dB, +0.9 dB of (14) and (15). The multiple reflection estimate is a conser-
vative estimate of detuning effects on the reference backscatter
(c)
Backscattered tag signal measurement antenna reflection coefficient, caused by variations in the test
(with ISO 18047-6 λ/2 rod calibration) setup, such as variations between scattering off of different
Reference λ/2 rod RCS (unknown) tags under test. Empirical tests showed that reorienting the
Multiple reflections ±0.9 dB [2]
IQ carrier phase drift ±0.23 dB antenna or moving the tag in the test zone caused errors
IQ level processing ±0.15 dB smaller than 0.1 dB.
Noise and nonlinearity ±0.1 dB
Expanded uncertainty −1.2 dB, +0.9 dB VI. C ONCLUSION
(neglecting rod RCS uncertainty)
The reference backscatter modulation approach demon-
strated here improves signal accuracy compared to radar cross
section methods, with common lab equipment available at low
error source as a fraction of the measured value. The reported
cost from many different vendors.
value in decibels is related to ui with
With additional effort in analysis of the uncertainty of this
technique as part of RCS tag tests, the authors hope it will
ui (in dB) = 10 log10 (1 ± ui ). (16)
become useful for future versions of tag performance test
The expanded uncertainty estimate uc from n terms is com- standards. Further work may also prove the effectiveness of the
puted with technique in small anechoic environments to reduce test costs.
 %
& n
 Reference signals may also become practical as in sensitivity
&( or interference rejection tests.
uc (in dB) = 10 log10 1 ± ' u2i  (17)
i=1
VII. ACKNOWLEDGEMENTS
by the law of propagation of uncertainty [20], which applies The authors are grateful to Prof. E. Kuester for fruitful
here under the assumption that errors are zero-mean and discussions about the circuit model, and Z. Popovic acknowl-
that dominant errors are normally distributed. The sensitivity edges support from the Hudson Moore Jr. professorship at the
coefficients are already included in the itemized uncertainties University of Colorado.
listed in Table II. The United States Department of Homeland Security Sci-
Some errors are common to all of the measurements. The ence and Technology Directorate has, in part, sponsored
noise and nonlinearity uncertainty is estimated conservatively the production of this material with NIST, under contract
from (8). Uncertainty in the IQ measurement processing, HSHQDC-09-X-00305.
caused by ringing or level clustering errors, is quoted as
±0.05 dB for results based on relative backscattering signal R EFERENCES
results (like the method presented in this paper), or ±0.15 dB
[1] M.R. Souryal, D.R. Novotny, D.G. Kuester, J.R. Guerrieri, K.A. Remley,
for absolute measurements, based on simulation of noisy, ban- “Impact of RF Interference between a Passive RFID System and a
dlimited pulses. Spectrum analyzer measurement uncertainty Frequency Hopping Communications System in the 900 MHz ISM Band,”
is contributes to errors in all of these cases, but when a Proc. 2010 IEEE Symp. on Electromagnetic Compatibility, pp. 495-501.
[2] D.G. Kuester, D.R. Novotny, J.R. Guerrieri, “The relative importance of
scattering standard is used, the relative measurement result the forward and reverse links in propagating UHF RFID with passive
from the instrument is so small that it is negligible. This tags,” Proc. 2010 IEEE Symp. on Electromagnetic Compatibility, pp. 680-
contrasts with the result in Table IIb, which is a dominant 685.
[3] J.H. Richmond, “A Modulated Scattering Technique for Measurement of
±0.9 dB source of uncertainty estimated from the manufac- Field Distributions,” IRE Trans. on Microwave Theory and Techniques,
turer’s specification. Because the modulation of the λ/2 rod vol.3, no.4, pp. 13-15, July 1955.

160
[4] A. Glinsky, Theremin: Ether Music and Espionage. Champaign, IL, USA: 24 May 2006.
University of Illinois Press, 2000, pp. 259-273. [12] J.C. Bolomey and F.E. Gardiol, Engineering Applications of the Modu-
[5] Radio frequency identification device conformance test methods — Test lated Scattering Technique, Boston: Artech House, 2001, pp. 35-59.
methods for air interface communications at 860 MHz to 960 MHz, [13] F. Fuschini, C. Piersanti, F. Paolazzi, and G. Falciasecca, “Analytical
ISO/IEC standard 18047-6, 2006. Approach to the Backscattering from UHF RFID Transponder,” IEEE
[6] A. Pouzin, T.P. Vuong, S. Tedjini, M. Pouyet, J. Perdereau, and L. Dreux, Antennas and Wireless Propagation Letters, vol. 7, 2008, pp. 33-35.
“Determination of measurement uncertainties applied to the RCS and [14] D.M. Dobkin, The RF in RFID: Passive UHF RFID in Practice, Newton,
the differential RCS of UHF passive RFID tags,” Proc. Antennas and MA, USA: Newnes, 2007, pp. 19-22.
Propagation Soc. Int. Symp., 2009, pp. 1-4. [15] K. Kurokawa, “Power Waves and the Scattering Matrix,” IEEE Trans-
[7] P. Nikitin and K. Rao, “Theory and measurement of backscattering from actions on Microwave Theory and Techniques, vol. 13, Mar. 1965, pp.
RFID tags,” IEEE Antennas and Propagation Magazine, vol. 48, 2006, 194-202.
p. 212–218. [16] D.F. Williams and R.B. Marks, “Comments on ‘Conversions between S,
[8] J.C. Bolomey, S. Capdevila, L. Jofre, and J. Romeu, “Electromagnetic Z, Y, h, ABCD, and T parameters which are valid for complex source and
Modeling of RFID-Modulated Scattering Mechanism. Application to Tag load impedances’ [and reply],” IEEE Transactions on Microwave Theory
Performance Evaluation,” Proceedings of the IEEE, vol. 98, 2010, p. and Techniques, vol. 43, Apr. 1995, pp. 914-915.
1555–1569. [17] Standard on Transitions, Pulses, and Related Waveforms, IEEE Standard
[9] S. Skali, C. Chantepy, and S. Tedjini, “On the measurement of the delta 181-2003, 2003.
Radar Cross Section (∆RCS) for UHF tags,” 2009 IEEE International [18] J. Canny, “A Computational Approach to Edge Detection,” IEEE Trans-
Conference on RFID, 2009, pp. 346-351. actions on Pattern Analysis and Machine Intelligence, vol. PAMI-8, Nov.
[10] D.M. Kerns, R.W. Beatty, Basic Theory of Waveguide Junctions and 1986, pp. 679-698.
Introductory Microwave Network Analysis. Oxford: Pergamon Press, [19] A. Pouzin, T.P. Vuong, S. Tedjini, M. Pouyet, and J. Perdereau, “Bench
1967, p. 108. test for measurement of differential RCS of UHF RFID tags,” Electronics
[11] K. Seemann, F. Cilek, M. Schmidt, R. Weigel, “RFID at UHF Frequen- Letters, vol. 46, Apr. 2010, pp. 590-592.
cies: The Passive Transponder Frontend Approach,” Proc. Intl. Conf. on [20] B. Taylor, Guidelines for evaluating and expressing the uncertainty of
Microwaves, Radar & Wireless Communications, 2006, pp.657-661, 22- NIST measurement results, NIST Technical Note 1297, 1994.

161