This article has been accepted for inclusion in a future issue of this journal.

Content is final as presented, with the exception of pagination.

IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES 1

Quadrature Amplitude Backscatter Modulator for

Passive Wireless Sensors in IoT Applications
Ricardo Correia, Graduate Student Member, IEEE, Alrio Boaventura, Student Member, IEEE,
and Nuno Borges Carvalho, Fellow, IEEE

Abstract Fully passive wireless networks are a key-enabling

technologies for the internet of things. Backscatter radios are a
hypothesis to design these passive wireless networks. Backscat-
ter modulation allows a remote device to wirelessly transfer
information without requiring a traditional transceiver. Instead,
a backscatter device leverages a carrier transmitted by an access
point or base station. Nevertheless, the traditional approach is
to use amplitude shift keying or phase shift keying backscatter
solutions which limits the data rate of the sensors, since it
transfers only one bit per symbol period. We propose a novel
modulator, which employs a Wilkinson power divider with a
phase shift in one of the design branch and two transistors
acting as switches in order to generate M-quadrature amplitude
modulation backscatter modulation. The design strategy for high
level order backscatter modulations will be explained and a Fig. 1. Application of WPT in wireless sensors.
design approach will be presented in this manuscript.
Index Terms Backscatter radio communication, Internet of combined can actually be a solution to build up totally passive
Things (IoT), quadrature amplitude modulation, wireless power wireless sensor network [8].
transmission, wireless sensor network.
In Fig. 1, a potential solution for totally passive sensor
networks is presented, where the nodes use a combination
I. I NTRODUCTION of WPT and amplitude shift keying (ASK) backscatter trans-
ceivers, as was presented in [8]. This solution can achieve a
I NTERNET of Things (IoT) is transforming the everyday
physical objects that surround us into an ecosystem of
information that will enrich our lives. While the IoT repre-
continuous operation with large number of sensors powered
by fixed wireless power transmitters used for both wireless
sents the convergence of advances in miniaturization, wireless charging and transfer data.
connectivity, increased data storage capacity and batteries, The conventional backscatter communication, used in radio
the IoT would not be possible without sensors. The growth of frequency identification (RFID) and existing passive sensor
the devices will be made possible only if the sensors battery solutions, is an attractive solution in these systems because
needs are eliminated or reduced significantly. For low-power the tags do not need to actively transmit any radio signal.
sensors and devices, careful power management and power Instead, they reflect a radio signal transmitted by the reader,
conservation are critical to device lifetime and effectiveness. and modulate the reflection by controlling their own reflec-
One of the possible solutions is to change completely the tion coefficient [9][11]. The reflected portion of the carrier
paradigm of the radio transceivers in the wireless nodes of the signal incident on the tag antenna is impressed with an enve-
IoT system. The new paradigm should be able to communicate lope (altering amplitude, phase, or both) carrying the binary
and possibly to power up sensors using only electromagnetic backscatter data. The growth of backscatter-radio technology
waveforms transmitted over the air. This gave rise to important is due to the applications with little-to-no power consumption.
concepts as backscatter radio solutions [1][4] but also to Using backscatter communication allows for extremely low
wireless power transfer (WPT) approaches [5][7] that when power dissipation for remote devices, which enables new areas
of sensor development. The backscatter tags, by receiving the
Manuscript received July 16, 2016; revised November 7, 2016 and local oscillator over the air, do not need their own RF oscil-
January 6, 2017; accepted January 22, 2017. This work was supported in part lator or phase-locked loop. By removing these circuits one
by the National Funds through Fundao para a Cincia e a Tecnologia under
Project PEst-OE/EEI/LA0008/2013 and Project UID/EEA/50008/2013 and reduces the power consumption and the cost of a tag chip [12].
in part by the Instituto de Telecomunicaes, CReATION, under Project Nevertheless, the backscattered signal received by the readers
EXCL/EEITEL/0067/2012. This paper is an expanded version from the is weak, thus limiting communication range. In conventional
2016 Wireless Power Transfer Conference, Aveiro, Portugal, May 56, 2016.
The authors are with the Departamento de Electrnica, Telecomuni- backscatter communication, the tags must harvest enough
caes e Informtica, Instituto de Telecomunicaes, Universidade de Aveiro, power from the reader to turn on and modulate data, and the
3810-193 Aveiro, Portugal (e-mail: rjoao@ua.pt). readers must receive strong backscattered signal to operate.
Color versions of one or more of the figures in this paper are available
online at http://ieeexplore.ieee.org. Some recent work [8] show the possibility of using one
Digital Object Identifier 10.1109/TMTT.2017.2661262 frequency to continuously power the wireless sensor and other
0018-9480 2017 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
2 IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES
Fig. 2. Block diagram of the solution presented in [8].
to transfer data by backscatter means with ASK modulation.
Fig. 2 presents a potential solution presented in [8] for this
implementation.
In most RFID systems and passive sensors, the reader to
tag communication is an ASK or phase shift keying (PSK)
that modulates either the amplitude, or both the amplitude
and phase, of the readers transmitted RF carrier. The use of
this technology entails a number of advantages over barcode
technologies such as tracking people, items, and equipment
in real time, nonline of sight requirement, long reading
range, and standing harsh environment. However, a recent
work [13] has shown that modulated backscatter can be
extended to include higher order modulation schemes, such as
four-state quadrature amplitude modulation (4-QAM). While
ASK and PSK transmit 1 b of data per symbol period,
4-QAM based can transmit 2 b per symbol period, thus
increasing the data rate and leading to reduced on-chip power
consumption and extended read range. The work presented
in [13] and [14] refers to a 4-QAM backscatter in semipassive
systems, by using a coin cell battery as a power source
for the modulator and a microcontroller that needs 3 V of
supply. This way, the authors proved the QPSK modulator
and battery powered system, by using an approach with
four lumped impedances connected to an RF switch that is
controlled by a microcontroller. The same authors developed
a 16-QAM modulator for UHF backscatter communication
with a consumption of 1.49 mW at a rate of 96 Mb/s only
in the modulator (not the overall system with data gener-
ation logic feeding the modulator) [15].This modulator was
implemented with 5 switches with lumped terminations as a
16-to-1 Mux to modulate the load between 16 different states.
The proposal in this work focuses on a backscatter modula-
tor with high order backscatter modulation format, by propos-
ing an RF circuit design that is able to be replicated to
include more modulation levels. In [16], a 4-QAM backscatter
RF circuit design was presented, which in this work will be
improved to a 16-QAM RF circuit design with a different
design strategy approach, that uses the transistor model to
create different backscatter impedances. A fully passive system
was implemented in [17], and is presented in Fig. 2, combining
WPT with ASK modulation. To obtain a fully passive system
with the modulator proposed in this paper, a similar approach
of the work presented in [8] and [17] should be followed.
A similar approach as the one used in this paper was
presented in [18]. The authors presented an I/Q backscatter
modulator that use bias currents to change the impedance of
two p-i-n diodes. The circuit comprises a Wilkinson power
divider, two filters (low pass and high pass) to guarantee
symmetrical paths on the board, one in each branch, and one
p-i-n diode for each branch. The bias consumption of the
circuit is 80 mW (excluding DACS and FPGA logic), which
imposes high power consumption and data rate limitations.
The use of this circuit for a low power sensor is not feasible.
Another approach by using the p-i-n diodes was presented
in [19]. Pozar [19] presented a reflection type phase shifter
using a quadrature hybrid and two p-i-n diodes. By biasing
the diodes to the ON and OFF state it was possible to change
the total path length for both reflected waves, producing a
phase shift at the output. However, the consumption of the
p-i-n diodes is not suitable for low power sensors.
One different solution to obtain a BPSK modulation was
presented in [20]. The authors presented a phase-shift modu-
lator with two switches that are connected to each other by
a 90 delay line or 0 delay line. The phase-shift modulator
was implemented as a two-port device that selectively delays
the signal by 90 between port 1 and port 2, or passes the
signal from port 1 to port 2 with no delay, achieving a
BPSK modulation.
In [21], two multiantenna technologies were used. The
authors presented an energy harvester [staggered-pattern
charge collector] that has two independent antenna arrays and
harvests power to supply a microcontroller. They also present a
retrodirective array phase modulator (RAPM) that backscatters
the signal from the reader to the reader. The RAPM comprises
two switches that are controlled by a microcontroller. The four
switching states are connected through coplanar waveguides of
different wavelengths, each separated by 90. The RAPM can
be used for QPSK modulation by calibrating the phase offsets
in the switches. The solution presented in our manuscript is
more compact and more capable to be expanded to higher
modulation orders when compared with the one presented
in [21].
In this paper, we first provide an overview of state-of-
the-art in WPT and backscatter modulation in order to face
the above technical challenges. Section II shows the dif-
ferent configurations used to implement high order modu-
lations and the approach used for the 16-QAM RF circuit
design. Section III describes the approach used to generate
the constellation diagram impedance based on the model
of the E-PHEMT. Section IV shows the solution for the
QAM backscatter implementation. Simulated and measured
results from the implementation are presented in Section V.
Finally, we conclude this paper in Section VI.
II. QAM BACKSCATTER M ODULATION
D ESIGN S TRATEGY
Backscatter radios are based on the reflected waveform
that should switch between a fully matched and short circuit
solution built on a switch that is directly connected to the
receiving/transmitting antenna of the sensor. For a higher
level modulation format, the number of different backscattered
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

CORREIA et al.: QUADRATURE AMPLITUDE BACKSCATTER MODULATOR FOR PASSIVE WIRELESS SENSORS IN IoT APPLICATIONS 3

Fig. 4. Photograph of the 4-QAM backscatter circuit.

Fig. 5. Model of the QAM solution.

large and even larger if we extend to higher order modulation,

such as 16-QAM.
In order to minimize the dimensions of the circuit and to
increase the modulation order to a 16-QAM solution, a new
implementation was evaluated. Fig. 5 presents the design for
Fig. 3. Three types of design for two different backscatter modulation
schemes. (a) ASK backscatter modulation scheme. (b) 4-QAM backscatter the 16-QAM solution. This model employs a Wilkinson power
modulation scheme without the possibility to increase the modulation order. divider and each branch is terminated with a line and an ideal
(c) Our proposed scheme for 4-QAM backscatter modulation with the possi- impedance. The lines present a 45 phase shift respect to each
bility to increase the modulation order.
other, so as to allow that the reflected wave from each branch
waves should change among several states, thus the switch has 90 phase difference from the other. Equation 1 shows
combinations should clearly create a specific impedance for the reflected wave as a sum of both reflected waves from each
that symbol combination. To determine the most suitable branch. Using (2) and (3) is possible to determine the reflection
impedances we had to test several implementations with coefficient of the model for different impedances.
different number of transistors (E-PHEMTs). Fig. 3 presents
three different schemes, for ASK [Fig. 3(a)] and 4-QAM b1 b2 a1 1 e j 2 a2 2
b= + = + (1)
[Fig. 3(b) and (c)] backscatter modulation. 2 2 2 2
Fig. 3(a) presents the traditional ASK backscatter, R1 Z 0 R2 Z 0
1 = 2 = (2)
a ON or OFF switch will impose two different reflected waves R1 + Z 0 R2 + Z 0
by changing the impedances. The switch is made using a
b 1 1 2 1 2 1
2 j 2
simple transistor (ATF54143, from Avago) and biasing it to  = = e + 2 = + j . (3)
obtain an open or a short circuit. Fig. 3(b) presents a higher a 2 2 2 2
order level backscatter modulation with two transistors. This Considering two different values for each resistance
implementation create a 4-QAM solution with four stages in (0 or ) we obtain four different combinations with four
the switches: ( ON, ON), (ON, OFF), (OFF, ON), and (OFF, OFF). different reflection coefficients as follows.
This particular solution is very difficult to implement since
1 1
the two states (ON, ON) and (OFF, OFF) are not completely R1 = 0 and R2 = 0  = j
complementary, which imposes that the impedances seen 2 2
1 1
by each of them, from the antenna point of view, are not R1 = 0 and R2 =  = j
complementary and create a complex matching network to be 2 2
1 1
designed to achieve the 4-QAM modulation format. To expand R1 = and R2 = 0  = + j
this solution into higher order modulations, such as 16-QAM, 2 2
1 1
the complexity would increase and it would be very difficult R1 = and R2 =  = + j .
to implement. 2 2
Fig. 3(c) presents a solution for 4-QAM, where a four- Fig. 6 presents the previous reflection coefficients in the Smith
switch approach was used [16]. Each branch of each switch chart. A similar reasoning can be used to create any type
is exactly equal and by changing the line and/or the stub of multilevel higher order modulation. In order to do that,
in the terminal of each switch, it is possible to achieve arbitrary real impedances could be used, instead of the 0 and
the impedance or the scattering parameters required. Fig. 4 impedance. Thus different impedances are implemented
presents the photograph of this solution. Despite the very good using an active load based on a transistor, one example is
results presented in [16], the dimensions of the circuit are very presented in Fig. 7, where different impedances were used in
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
4 IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES
Fig. 6. Simulation of the model for 4-QAM solution.
Fig. 7. Simulation of the model for 16-QAM solution.
Fig. 8. Block diagram for different impedance terminations.
the scheme presented in Fig. 6, the impedances were 0, 20,
60, and 200 .
In a generalized scheme, the last approach can be extended,
where the load presents different impedance values, consider
for instance a, b, c, and d real impedances. The impedances
are then matched with the overall circuit using a two branch
approach and a microstrip line scheme, were each branch
imposes a specific delay. By combining these two degrees of
freedom, several impedance levels can be created. Equation (4)
was deduced, where R represents the four different impedances
in one branch and R represents the four impedances in
the other branch. It was based on Fig. 8 and presents the
16 possibilities of impedances
X = R  + Q  .
III. C ONSTELLATION D IAGRAM I MPEDANCE G ENERATION
In order to create different impedances for the constellation
diagram creation, a transistor was used as an active load,
Fig. 9. Model for the characterization. (a) Design of the circuit.
(b) Photograph of the proposed characterization circuit.
Fig. 10. Simulated, measured, and selected measured S11 for different
voltage levels at the gate of transistor for 2.45 GHz at 0 dBm.
in this case the transistor was matched so that it presents
an impedance for different control voltages. The approach
followed in this case was to use a single transistor as is
presented in Fig. 9(a). The gate voltage in this case can be
changed from 0 to 0.6 V. The matching network was optimized
for the extreme cases 0 and 0.6 V, where the resistive value
of the resistor is 4.7 + j 0.5  and 379.6 j 630.8 , respec-
tively. By changing the gate voltage other impedances can
be generated as is presented in Fig. 10. For this paper and in
order to create a QAM approach, the selected impedances were
4.7 + j 1.4 , 14.9 + j 1.8 , 61.0 j 1.2 , 566.6 + j 1.4 ,
as is presented in Fig. 10.
In order to validate the system, we implemented and
designed a simple circuit with only one transistor and an
impedance matching. The objective in the simulations was
to achieve an open circuit when the gate voltage is 0.6 V
and a short circuit with 0 V at the gate of transistor.
Nevertheless, we added two different voltages at the gate
of transistor, 0.2 and 0.4 V, to characterize the system to
four different voltage levels. We measured the circuit for
26 voltages levels (from 0 to 0.6 V) and we selected the
levels more approximated with the simulations (Fig. 10). Since
the sensitivity of the circuit is higher between 0.3 and 0.5 V,
the measurement step selected, within this range, was 0.01 V
instead of 0.1 V. Fig. 9(a) and (b) shows the design of
the circuit and the photograph of the implemented circuit,
respectively. Fig. 10 presents the real impedances in the Smith
chart. The purpose of the matching network is to achieve
real impedances in all levels at the drain of transistor, from
(4) 0 to 0.6 V. Fig. 11 shows the simulation of total reflection
coefficients possibilities of the solution. The voltages at the
gate of transistor were chosen from 0 to 0.6 V with a step
of 0.01 V. Using this approach with a two branch network
allows the creation of a multitude of impedance arrangements,
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

CORREIA et al.: QUADRATURE AMPLITUDE BACKSCATTER MODULATOR FOR PASSIVE WIRELESS SENSORS IN IoT APPLICATIONS 5

Fig. 11. Simulated S11 for different voltage levels (from 0 to 0.6 V with a
step of 0.01 V) at the gate of each transistor for 2.45 GHz at 0 dBm.

as can be seen in Fig. 11, for several gate voltages. Despite

this is a simulation value, it already shows that for a real
implementation a calibration procedure is fundamental for the
development of a true higher order modulation.
Fig. 12. 16-QAM backscatter modulation scheme. (a) Design of 16-QAM
backscatter modulator. (b) Photograph of the 16-QAM backscatter circuit.
IV. QAM BACKSCATTER I MPLEMENTATION
Fig. 12(a) and (b) shows the design of the proposed
16-QAM scheme and the prototype implemented that includes
the Wilkinson power divider, two matching networks, and
two transistors, respectively. This prototype was optimized for
0 dBm at 2.45 GHz and it was simulated in advanced design
system. In Fig. 12(b) is possible to view the difference of
line length in each branch, and it is related with the 45
phase shift. As it was previously mentioned we used two
transistors, in order to achieve the 16-QAM modulation. Fig. 8
Fig. 13. Photograph of the measurement setup.
presents the block diagram approach to explain the different
impedances obtained. With this approach the design illustrated
in Fig. 12(a) was implemented in order to obtain the 16-QAM
modulation. Each transistor is switched for different voltage
levels to achieve different reflection coefficients.

V. R ESULTS
The measured results were obtained by using a power
supply and the performance network analyzer (E8361C, from
Agilent Technologies) that was calibrated at 0 dBm for the
frequencies from 2 to 3 GHz. Fig. 13 shows the photograph
of the measurement setup used to acquire the reflection coef-
ficients from the circuit shown in Fig. 12(b). The switches
were changed according to the simulated results, by biasing
the gate of each transistor. The results of the simulations
and measurements are presented in Fig. 14 and in Table I
are presented the different reflection coefficients for different
Fig. 14. Simulated and measured S11 for different voltage levels at the gate
voltages at the transistor. Table I shows also the simulated of each transistor for 2.45 GHz at 0 dBm.
and experimental voltage levels used to obtain the matching
between the reflection coefficients. As can be seen in Table I, A. Demodulation and Achievable Data Rates
the voltages measured and simulated are different, due to the The maximum data rate of the communication on the
models used for the transistor. The obtained results clearly modulator link was studied by using the measurement setup
showed that the 16-QAM modulation is viable and as it was presented in Fig. 15 to view the received constellation and eye
mentioned previously, the points can be better adjusted in the diagrams of the I and Q signals at multiple transmission rates.
Smith chart by biasing the gate of each transistor with more We used a vector signal generator (ROHDE&SCHWARZ
precision. SMJ 100A) to generate the transmitter signal at 2.45 GHz.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
6 IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES
Fig. 15. Measurement setup for demodulation and achievable data rates.
(a) Block diagram. (b) Photograph.
Fig. 16. Received 16-QAM constellation for a center frequency of 2.45 GHz
for different bit rates. (a) 4 Mb/s with 8.19% of EVM and 21.74 dB of SNR.
(b) 40 Mb/s with 10.96% of EVM and 19.20 dB of SNR. (c) 120 Mb/s with
16.76% of EVM and 15.52 dB of SNR.
An arbitrary waveform generator (TEKTRONIX AWG5012C)
was used to generate the voltage levels at the gate of each
transistor mimicking the modulated data rate. A coupler was
used to measure the reflected signal from the circuit, by the
spectrum analyzer (ROHDE&SCHWARZ FSW8).
To analyze the achievable data rates, the voltages presented
in Table I were applied to the gate of each transistor, creating
16 reflected impedances or symbols. The received constel-
lations acquired from the signal and spectrum analyzer are
shown in Fig. 16 for bit rates of 4, 40, and 120 Mb/s,
respectively. The constellation points lie in the appropriate
quadrants, and each point is clearly visible. The constellation
Fig. 17. Measured EVM of the 16-QAM with a varying data rate.
TABLE I
S IMULATED AND M EASURED S11 FOR D IFFERENT L EVELS
AT THE G ATE OF T RANSISTORS (M1 AND M2)
presented in Fig. 16 is not perfect, but can be calibrated with
more precision by adjusting the voltages at the gate of each
transistor. The received eye diagrams for varying data rates
are shown in Fig. 18, and for the data rates of 256 kb/s,
2.4 Mb/s, and 4 Mb/s the eye is clearly open and the received
signals are clean. For the data rate of 120 Mb/s the received
signals appear with some noise and the eye is beginning to
close. By analyzing Fig. 18(d), it is visible the limit of the
communication rate. Nevertheless, Fig. 16(c) illustrates that it
is possible to demodulate at such higher data rate.
Fig. 17 illustrates the measured EVM for different data
rates. As the data rate increases, the measured EVM also
increases. Nevertheless and despite the EVM seems high,
the process has not passed through any signal processing
equalization at the receiver. These results clearly state the
viability of the proposed solution for a high bit rate backscatter
approach.
B. Sensitivity of Backscatter Modulator
To evaluate the sensitivity of the gate voltage applied to each
transistor, a Monte Carlo simulation was performed. A 1% of
variation on each gate voltage was applied and Fig. 19 presents
the results obtained from the modulator. The red marker in
the figure shows the limit of the possible calibration for this
modulator. By varying the voltages with 0.01 V it is possible
to scale this system to a 64-QAM with adequate calibration.
C. Power Consumption of Backscatter Modulator
In Fig. 20, the measured EVM of the 16-QAM modulator
as function of input power for different data rates can be seen.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
CORREIA et al.: QUADRATURE AMPLITUDE BACKSCATTER MODULATOR FOR PASSIVE WIRELESS SENSORS IN IoT APPLICATIONS
Fig. 18. Eye diagrams of received demodulated data for varying data rates.
(a) 256 kb/s. (b) 2.4 Mb/s. (c) 4 Mb/s. (d) 120 Mb/s.
Since this device was optimized for a 0 dBm (1 mW) RF input
signal, the lowest value of EVM can be seen to be around this
RF power. As can be seen from the figure, the value of energy
spent in this modulator can be as low as 6.7 pJ/bit for a data
rate of 120 Mb/s with EVM of 16.7 %, or 16.7 pJ/bit for a
data rate of 60 Mb/s with an EVM of 10.3 %. The viability
for this type of solution could be improved for lower values
of power also, if the circuit was turned for lower values of
input RF power.
7
Fig. 19. Sensitivity of the modulator with 1% of variation on each gate
transistor. The red marker indicates the limit that is possible to achieve by
calibration of the voltages.
Fig. 20. Measured EVM of the 16-QAM modulator as function of input
power for different data rates.
VI. C ONCLUSION
In this paper, a new approach for higher order modulation
backscatter passive radio solutions was presented. The solution
now presented is actually to our knowledge the one that
achieves higher bit rate with a limited amount of RF power
and with a low count of active devices.
This paper described a modulation technique that enables a
high-bandwidth wireless communication while requiring very
low power demands. This solution can be combined with WPT
for ultralow-power wireless applications requiring high band-
width communications such as remote camera sensors or wire-
less audio. This paper challenges held assumptions concerning
the backscatter link that backscatter communication is limited
to low bandwidth and binary modulation solutions.
R EFERENCES
[1] V. Talla and J. R. Smith, Hybrid analog-digital backscatter: A new
approach for battery-free sensing, in Proc. IEEE Int. Conf. RFID,
Apr. 2013, pp. 7481.
[2] G. Vannucci, A. Bletsas, and D. Leigh, A software-defined radio system
for backscatter sensor networks, IEEE Trans. Wireless Commun., vol. 7,
no. 6, pp. 21702179, Jun. 2008.
[3] S.-N. Daskalakis, S. D. Assimonis, E. Kampianakis, and A. Bletsas,
Soil moisture scatter radio networking with low power, IEEE Trans.
Microw. Theory Techn., vol. 64, no. 7, pp. 23382346, Jul. 2016.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
8 IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES
[4] C. Konstantopoulos, E. Koutroulis, N. Mitianoudis, and A. Bletsas,
Converting a plant to a battery and wireless sensor with scatter
radio and ultra-low cost, IEEE Trans. Instrum. Meas., vol. 65, no. 2,
pp. 388398, Feb. 2016.
[5] Z. Popovic, E. Falkenstein, and R. Zane, Low-power density wire-
less powering for battery-less sensors, in Proc. IEEE Radio Wireless
Symp. (RWS), Jan. 2013, pp. 3133.
[6] Z. Popovic, Far-field low-power wireless powering for unattended
sensors, in Proc. IEEE 16th Annu. Wireless Microw. Technol.
Conf. (WAMICON), Apr. 2015, pp. 14.
[7] N. Borges Carvalho et al., Wireless power transmission: R&D activities
within europe, IEEE Trans. Microw. Theory Techn., vol. 62, no. 4,
pp. 10311045, Apr. 2014.
[8] R. Correia, N. Borges Carvalho, G. Fukuda, A. Miyaji, and S. Kawasaki,
Backscatter wireless sensor network with WPT capabilities, in Proc.
Int. Microw. Symp., May 2015, pp. 14.
[9] K. Finkenzeller, RFID Handbook: Fundamentals and Applications in
Contactless Smart Cards and Identification. Hoboken, NJ, USA: Wiley,
May 2003.
[10] C. Boyer and S. Roy, Backscatter communication and RFID: Coding,
energy, and MIMO analysis, IEEE Trans. Commun., vol. 62, no. 3,
pp. 770785, Mar. 2014.
[11] J. D. Griffin and G. D. Durgin, Complete link budgets for backscatter-
radio and RFID systems, IEEE Antennas Propag. Mag., vol. 51, no. 2,
pp. 1125, Apr. 2009.
[12] D. Kuester and Z. Popovic, How good is your tag?: RFID backscat-
ter metrics and measurements, IEEE Microw. Mag., vol. 14, no. 5,
pp. 4755, Jul. 2013.
[13] S. J. Thomas, E. Wheeler, J. Teizer, and M. S. Reynolds, Quadra-
ture amplitude modulated backscatter in passive and semipassive UHF
RFID systems, IEEE Trans. Microw. Theory Techn., vol. 60, no. 4,
pp. 11751182, Apr. 2012.
[14] S. Thomas and M. S. Reynolds, QAM backscatter for passive UHF
RFID tags, in Proc. IEEE Int. Conf. RFID, Apr. 2010, pp. 210214.
[15] S. J. Thomas and M. S. Reynolds, A 96 Mbit/sec, 15.5 pJ/bit
16-QAM modulator for UHF backscatter communication, in Proc.
IEEE Int. Conf. RFID, Apr. 2012, pp. 185190.
[16] R. Correia and N. Borges Carvalho, Design of high order modulation
backscatter wireless sensor for passive IoT solutions, in Proc. IEEE
Wireless Power Transf. Conf. (WPTC), May 2016, pp. 13.
[17] R. Correia, N. Borges Carvalho, and S. Kawasaki, Continuously
power delivering for passive backscatter wireless sensor networks,
IEEE Trans. Microw. Theory Techn., vol. 64, no. 11, pp. 37233731,
Nov. 2016.
[18] M. Winkler, T. Faseth, H. Arthaber, and G. Magerl, An UHF RFID
tag emulator for precise emulation of the physical layer, in Proc. Eur.
Microw. Conf. (EuMC), Sep. 2010, pp. 273276.
[19] D. M. Pozar, Microwave Engineering. Hoboken, NJ, USA: Wiley, 2012.
[20] J. Kimionis and M. M. Tentzeris, Pulse shaping for backscatter radio,
in IEEE MTT-S Int. Microw. Symp. Dig., May 2016, pp. 14.
[21] M. S. Trotter, C. R. Valenta, G. A. Koo, B. R. Marshall, and
G. D. Durgin, Multi-antenna techniques for enabling passive RFID tags
and sensors at microwave frequencies, in Proc. IEEE Int. Conf. RFID,
Apr. 2012, pp. 17.
Ricardo Correia (GS15) received the M.Sc. degree
in electronics and telecommunications engineering
from the University of Aveiro, Aveiro, Portugal,
in 2009, where he is currently pursuing the Ph.D.
degree in electrical engineering.
He was an Automation and Electrical Engineer for
a year and three years as a Researcher involved with
embedded systems and signal processing. Currently,
he is a Researcher with the Institute of Telecommu-
nications, University of Aveiro. His current research
interests include wireless power transfer, energy
harvesting, wireless passive sensors for space applications, and low power
communications.
Mr. Correia is a member of the IEEE Microwave Theory and Techniques
Society and a COST Action IC1301 WiPE member. He is also a Reviewer
for IET Microwaves, Antennas & Propagation, Wireless Power Transfer, and
the IEEE T RANSACTIONS ON M ICROWAVE T HEORY AND T ECHNIQUES .
He was a recipient of the 2016 URSI/National Authority of Communica-
tions (ANACOM) Prize awarded by the URSI Portuguese Section and the
Portuguese ANACOM.
Alrio Boaventura (S11) was born in Santo Anto,
Cape Verde, in 1985. He received the M.Sc. and
Ph.D. degrees in electrical engineering from the
University of Aveiro, Aveiro, Portugal, in 2009 and
2016, respectively.
From 2008 to 2010, he was with Acronym Lda,
where he designed HF and UHF RFID reader and
antenna products. From 2010 to 2016, he was with
the Portuguese Telecommunications Institute, car-
rying out research in wireless power transmission,
RFID, and backscatter radios. Since July 2016, he
has been with the National Institute of Standards and Technology (NIST),
Boulder, CO, USA. He is currently a Research Associate Faculty Member
with the University of Colorado, Boulder, CO, USA. His current research
interests include on-wafer measurements and characterization of microwave
superconducting circuits. He has co-authored 3 books and over 40 papers in
conferences and journals. He is the co-inventor of one patent.
Dr. Boaventura served as a reviewer for the IEEE T RANSACTIONS ON
M ICROWAVE T HEORY AND T ECHNIQUES , the IEEE J OURNAL OF E MERG -
ING AND S ELECTED T OPICS IN C IRCUITS AND S YSTEMS , and the Cam-
bridge Journal of Wireless Power Transfer. He was the recipient of a
merit scholarship for graduation from the Portuguese Calouste Gulbenkian
Foundation for excellence in high school as the Best 2004 Student of Cape
Verde. In 2011, he was a finalist of the Student Paper Competition of the IEEE
MTT-S International Microwave Symposium. He was an Ex aequo recipient
of the 2011 URSI-Portugal/ANACOM Prize. He was also distinguished
in the publication University of Aveiro, 40 Years, 40 Inventors, and 40
Entrepreneurs. In 2013, he was also a recipient of the IEEE MTT-S Graduate
Fellowship Award.
Nuno Borges Carvalho (S97M00SM05F15)
was born in Luanda, Angola, in 1972. He received
the Diploma and Doctoral degrees in electronics and
telecommunications engineering from the University
of Aveiro, Aveiro, Portugal, in 1995 and 2000,
respectively.
He is currently a Full Professor and a Senior
Research Scientist with the Institute of Telecom-
munications, University of Aveiro. He has been
a Reviewer and has authored over 200 papers in
magazines and conferences. He co-authored Inter-
modulation in Microwave and Wireless Circuits (Artech House, 2003),
Microwave and Wireless Measurement Techniques (Cambridge Univ. Press,
2013) and White Space Communication Technologies (Cambridge Univ.
Press, 2014). He is the co-inventor of four patents. He has recently been
involved in the design of dedicated radios and systems for newly emerging
wireless technologies. His current research interests include software-defined
radio front-ends, wireless power transmission, nonlinear distortion analysis
in microwave/wireless circuits and systems, and measurement of nonlinear
phenomena.
Dr. Carvalho is an Associate Editor for the IEEE T RANSACTIONS ON
M ICROWAVE T HEORY AND T ECHNIQUES , IEEE Microwave Magazine, and
Cambridge Wireless Power Transfer Journal. He is the Co-Chair of the IEEE
MTT-20 Technical Committee and the Past-Chair of the IEEE Portuguese
Section and MTT-11. He is the Vice-Chair of MTT-20 and belongs to the
Technical Commitees, MTT-11, MTT-24, and MTT-26. He is also the Chair
of the URSI-Portugal Metrology Group. He was a recipient of the 1995
University of Aveiro and the Portuguese Engineering Association Prize for the
Best 1995 Student at the University of Aveiro, the 1998 Student Paper Com-
petition (Third Place) of the IEEE Microwave Theory and Techniques Society
International Microwave Symposium, and the 2000 IEEE Measurement Prize.
He is a Distinguished Microwave Lecturer for the IEEE Microwave Theory
and Techniques Society.