Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017
Radio frequency identification
enabled wireless sensing for
intelligent food logistics
Zhuo Zou1 , Qing Chen1 , Ismail Uysal2 and
rsta.royalsocietypublishing.org
Lirong Zheng1,3
1 iPack Vinn Excellence Center, School of Information and
Review
Cite this article: Zou Z, Chen Q, Uysal I,
Zheng L. 2014 Radio frequency identification
enabled wireless sensing for intelligent food
logistics. Phil. Trans. R. Soc. A 372: 20130313.
http://dx.doi.org/10.1098/rsta.2013.0313
One contribution of 10 to a Theme Issue
‘Intelligent food logistics: decrease waste and
improve quality by new technologies and
advanced warehouse management’.
Subject Areas:
electrical engineering
Keywords:
radio frequency identification,
wireless sensing, food monitoring,
Internet of Things
Authors for correspondence
Zhuo Zou
e-mail: [email protected]
Qing Chen
e-mail: [email protected]
Ismail Uysal
e-mail: [email protected]
Lirong Zheng
e-mail: [email protected]
Communication Technology, Royal Institute of Technology (KTH),
Isafjordsgatan 39, 164 40 Kista, Stockholm, Sweden
2 RFID Lab for Applied Research, Electrical Engineering, University of
South Florida, 4202 E. Fowler Avenue, ENB118, Tampa, FL 33620, USA
3 School of Information Science and Technology, Fudan University,
No. 220, Handan Road, Shanghai, People’s Republic of China
Future technologies and applications for the Internet
of Things (IoT) will evolve the process of the food
supply chain and create added value of business.
Radio frequency identifications (RFIDs) and wireless
sensor networks (WSNs) have been considered as the
key technological enablers. Intelligent tags, powered
by autonomous energy, are attached on objects,
networked by short-range wireless links, allowing
the physical parameters such as temperatures and
humidities as well as the location information to
seamlessly integrate with the enterprise information
system over the Internet. In this paper, challenges,
considerations and design examples are reviewed
from system, implementation and application perspectives, particularly with focus on intelligent
packaging and logistics for the fresh food tracking
and monitoring service. An IoT platform with a twolayer network architecture is introduced consisting
of an asymmetric tag–reader link (RFID layer)
and an ad-hoc link between readers (WSN layer),
which are further connected to the Internet via
cellular or Wi-Fi. Then, we provide insights into the
enabling technology of RFID with sensing capabilities.
Passive, semi-passive and active RFID solutions
are discussed. In particular, we describe ultrawideband radio RFID which has been considered
as one of the most promising techniques for
2014 The Author(s) Published by the Royal Society. All rights reserved.
Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017
ultra-low-power and low-cost wireless sensing. Finally, an example is provided in the form of
an application in fresh food tracking services and corresponding field testing results.
— the development of the IoT platform that can early adopt novel RFID and wireless sensing
technologies while seamlessly integrating with existing ICT infrastructures;
— advancing wireless sensing systems and intelligent tags with extended functionalities
and reduced cost; and
— facilitating the integration of RFID-enabled wireless sensing systems into global
information systems.
This review presents challenges, considerations and design examples of the future RFID-enabled
wireless sensing for IoT from system, implementation and application perspectives. An IoT
platform with a two-layer network architecture is introduced consisting of an asymmetric
tag–reader link (RFID layer) and ad-hoc links between readers also known as master nodes
(WSN layer), which are further connected to the Internet via cellular (3G/GPRS) or Wi-Fi. The
review focuses on the state-of-the-art and the future trend of the RFID technologies for sensing
and tracking applications. Three different technical approaches of (i) passive RFID with ultrawideband (UWB) radio, (ii) active RFID using 2.4 GHz radio, and (iii) semi-passive ultrahigh
frequency (UHF) RFID are reviewed. Finally, we show a deployment example of the system in a
global fresh food tracking service.
2. IoT platform for fresh food tracking and monitoring applications
(a) Application requirements and system considerations
IBM has reported that over 50% of food globally ends up going to waste [8]. Another statistic
from Billerud [9] indicates that approximately 10% of the fresh fruits and vegetables coming from
.........................................................
The Internet of Things (IoT) is a revolution of the information and communication technology
(ICT), with a vision of bridging the physical world with the Internet cloud. Radio frequency
identification (RFID)-enabled wireless sensing has been considered as the key technological
enabler to realize the IoT [1]. The future RFID tags will be equipped with sensors and other
peripherals for interaction with their carriers, monitoring their surrounding environment as well
as their geographical information. Such intelligent tags with unique addressable identification
number or IP address will be attached on objects, networked by short-range wireless links to
existing ICT infrastructures, allowing the connections between the virtual world on the Internet
and the physical world of things.
This technological revolution brings many emerging applications and services, creating added
value in the marketplace. Numbers of early-bird applications of IoT in logistics, tracking
and healthcare have already been deployed [2]. In particular, there have been several papers
that introduced the RFID-enabled wireless sensor networks (WSNs) in fresh food tracking for
intelligent logistics, aiming to improve the process of the fresh food supply chain and create added
value of business [3–7]. For example, RFID-enabled temperature loggers are used to monitor
and record the transportation and distribution temperature for perishable food products. This
information is used for shelf-life prediction to identify and rank items with respect to their
remaining shelf-lives which can enable the necessary insights for intelligent supply chain logistics
such as first-expired-first-out (FEFO) instead of more traditional first-in-first-out (FIFO). From
those research and early industrial deployments, three major issues are concerned:
rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 372: 20130313
1. Introduction
2
Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017
As discussed in §2a, an IoT platform that can adopt the emerging technologies while keeping
the compatibility with existing ICT infrastructure is crucial. Figure 1 illustrates the network
and system architecture of an intelligent tag system which is ideal for fresh food tracking and
monitoring applications. A two-layer network hierarchy is used which is connected to the Internet
through an IP gateway [14]. Layer 1 is essentially an RFID system where tags are coordinated by
a reader also known as the master node. In layer 2, ad-hoc networks are self-organized between
master nodes as a WSN. Layer 1 (RFID layer) is a heterogeneous network with asymmetrical
links providing wide range selections of RFIDs. Intelligent tags are ideally embedded in food
packaging, such as paper boards or other flexible substrates, communicating with a reader by
micropower wireless links. The intelligent tag usually consists of a power harvesting block, sensor
interfaces, a digital processor and memory and a radio transceiver. The radio interface can be
standard RFID by means of near-field coupling or far-field backscattering in high frequency
(HF) or UHF bands, or active tags in the industrial, scientific and medical (ISM) radio bands.
Meanwhile, the RFID layer can embrace non-standard RFID which usually presents attractive
features and unique advantages for sensing applications. For instance, an RFID tag incorporating
UWB technologies enables high-speed identification, accurate positioning and time-domain
sensing. Another example is the printed chipless tag with inherent sensing characteristics through
RF measurements of the impedance changes as a response to the varying environment, such
.........................................................
(b) System architecture
3
rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 372: 20130313
different parts of the world into the European market are damaged during the transportation
process. This causes not only a loss of 10 billion euros per year, but also a big threat to
public food safety. The cases of food spoilage have been intensively studied in [10,11], and
the desired sensing target has been summarized in [6]. The main causes of fresh food damage
during transport include microbial infections, biochemical changes (e.g. owing to respiration
or ripening), physical damages and mishandling [12]. Early food tracking systems usually
focused on the ability to follow the movement of a food through specified stages of production,
processing and distribution [13]. Introduction of RFID to replace the barcode for automated
identification aimed at reducing labour costs and process time. Recent advances in RFID with
sensing and networking capabilities allow continuous monitoring of simple parameters such
as temperature and humidity in a real-time manner, then further control the environmental
conditions [3]. More recently, Pang et al. [6] suggested an income-centric fresh food tracking
system leveraged by the latest IoT technologies, creating additional values through shelf-life
prediction, sales premium, precision food production and insurance cost reduction. Beyond
simple RFID systems with temperature and humidity sensors, in the study, a sensor portfolio was
carefully selected based on the value creation, availability analysis and cost assessment, including
the GPS coordinates, temperature, relative humidity, carbon dioxide (CO2 )/oxygen (O2 )/
ethylene (C2 H4 ) concentration and three-axis acceleration.
In order to realize a fresh food tracking system that is able to continuously monitor and
trace the food quality and safety during the whole food supply chain from farm to table, there
are several requirements that need to be considered in addition to traditional RFID systems.
Visibility, traceability and controllability of food quality and safety by monitoring the surrounding
environments and smartly governing the parameters should be available at all levels of packaging
or carriers in the food supply chains. Therefore, RFIDs with different capabilities in terms of
sensing functionalities, energy source, processing and storage capabilities, and wireless interfaces
will be deployed in containers, packaging boxes or even attached to food itself at an item level.
The RFID tags can be either disposable such as printed chipless tags, or reusable such as active
wireless sensor nodes, for a most cost-effective solution. Moreover, the system installation and
maintenance should be simple without the need of specialized staff or knowledge. To this end,
solutions simply extended from a traditional RFID system cannot cope with all of these issues,
and therefore a networked platform that embraces varieties of sensors and different classes of
tags with heterogeneous radio interface and functionalities is expected.
Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017
4
Internet
energy
scavenging
...
sensor
interface
sensors
IP gateway
display
digital
processor
radio
(RFID)
memory
tag integrated circuit
intelligent tag on paper or flexible substrate
tag
asymmetric links (layer 1)
reader/master node
peer-to-peer multi-hop links
(layer 2)
Figure 1. A generic IoT platform consisting of intelligent RFID tag and reader with a hierarchical two-layer network. (Online
version in colour.)
as temperature and humility. Tags and sensors in this layer should be extremely low-cost and
energy-efficient under energy autonomy conditions. The master node serves as a reader to collect
the sensory data from the tag and link to the Internet cloud through standard air interfaces such as
WiFi, GSM/GPRS and 3G. Also, the master node can be a superior wireless sensory station, which
is usually equipped with sophisticated sensors such as chemical sensors and imaging sensors as
well as GPS for geographical tracking.
3. RFID technologies for future IoT
Here, we focus on the RFID technology as a key enabler for fresh food tracking and monitoring
applications towards the context of the IoT. After a brief review of the principles of RFID,
technical challenges of the future of RFID for sensing applications will be discussed. Afterwards,
two design examples for passive and active RFID technologies and a summary of the hybrid
semi-passive technology to address those challenges will be introduced in §3c–e, respectively.
(a) RFID basics
An RFID system comprises readers and tags, also named interrogators and transponders.
Basically, tags can be divided into two major categories based on how they communicate with the
reader and their power source: active and passive.1 Active RFID tags contain an internal power
source (e.g. battery). Thus, a tag’s lifetime is limited by the stored energy. Passive tags, on the other
hand, usually harvest operating energy from the reader’s signal. Passive RFID is more attractive
for applications where it is best for tags to not require batteries. In a typical passive system, the
tag is powered up in the interrogating field and transmits the data to the reader [15]. Operation
principles and frequency bands essentially dominate the performance and potential applications
of an RFID system in terms of operating range, data rate as well as the cost. Frequencies of RFID
range from a few kHz of low frequency (LF) to microwave band (e.g. 2.4 GHz), and further extend
to the UWB. Among them, 125 kHz (LF) tags, 13.6 MHz (HF) tags and 900 MHz (UHF) tags are
the most common ones for passive systems.
Conventional passive RFID technologies work in either magnetic/electrical coupling or
electromagnetic coupling (backscattering), corresponding to the near-field and the far-field,
respectively. The far-field backscattering RFID tags at the UHF band (860–960 MHz) are getting
more attractive for wireless sensing and monitoring applications. Such tags are more powerful
1
In some literatures, the grey area between active tag and passive tag is further defined, i.e. semi-passive tag: a battery for
the tag’s circuit and backscattering for communication; semi-active tag: power-harvest for energy and active transmitter for
communication.
.........................................................
power
management
rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 372: 20130313
external
energy
Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017
Beyond simple identification function in traditional RFID, the new paradigm of intelligent tags
with multiple functionalities of sensing, computing, networking is to keep or reduce the cost
compared with today’s RFID. The performance of the enabling RFID technology needs to be
improved in several aspects, mainly including (i) development and integration of sensors and
interface circuitry into RFID tags, (ii) improved link performance in terms of data rate and reading
distance, and (iii) system integration of integrated circuit (IC) chip and non-silicon blocks such as
antennas and sensors on RFID tags, etc.
(i) Sensors and interface circuits
The sensing function determines the application of the intelligent tags. For packaging and
transportation applications for food logistics, the sensing targets of interest include temperature,
humidity, CO2 /O2 /C2 H4 concentration and three-axis acceleration. Existing wireless sensor
nodes are composed of multiple components on a printed circuit board, resulting in a milliwattpowered system with bulky battery for limited lifetimes, and the sensors in most cases dominate
the overall cost of the node [16]. The advances in RFID are creating the opportunity to combine
the sensing functionalities with ID tags, which significantly reduces the cost and size. Siliconbased sensors integrated on the IC chip are the most feasible solution providing the best
trade-off in terms of cost and performance. For example, on-chip complementary metal–oxide–
semiconductor (CMOS) temperature sensors are usually implemented using bipolar junction
transistors with delta–sigma analogue-to-digital conversion (ADC) readouts, as demonstrated
in [17]. Law et al. [18] demonstrated a CMOS temperature sensor based on serially connected
subthreshold MOS operation for passive RFID food monitoring applications. However, it is still
challenging to integrate sensors and their interface circuits into wirelessly powered RFID tags.
First, sensors usually have high power consumption. For instance, the reported temperature
sensors for RFID applications still consume a few tens to hundreds of µW, whereas chemical
and gas sensors are even more problematic, which usually consume mW-level power and (or)
require extra heating of the sensing film [19]. Second, data from analogue sensors need to
be digitalized before being processed, stored and transmitted. Therefore, extremely low-power
ADCs are essential to cope with µW-level power budget for a wirelessly powered application.
In order to address these issues, µW successive approximation register ADCs with a figure of
merit down to 4.4 fJ per conversion step have been designed for RFID applications, providing
promising energy efficiency and good scalability for power and performance [20]. Nevertheless,
the additional power consumed by the ADC still limits the reading distance of the tag.
Over the past a few years, advances in large-area electronics and printing technologies
have enabled the realization of printed electronics and sensors on various substrates including
such mediums as papers and organic substrates. Some exciting research focuses on various
printed sensors, including biosensors, pressure sensors and temperature sensors. More recently,
.........................................................
(b) Technical challenges and design trends
5
rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 372: 20130313
than near-field RFID tags, ensuring a larger operating distance (up to several metres), a higher
data rate (up to hundreds of kb/s) and a smaller antenna size. The tag captures the energy of
continuous waves (CWs) from the reader. A power converter rectifies the alternating potential
difference (electromagnetic energy) across the antenna. The scavenged energy is used to power
up the circuitry on the tag. The tag sends data to the reader using a backscattering mechanism.
The modulation is performed by changing the antenna’s impedance over time, so the tag can
reflect back more or less of the incoming signal in a pattern that encodes the tag’s ID. By
contrast, the active RFID is usually associated with short- or medium-range wireless in ISM band
such as ZigBee and Bluetooth. It usually uses traditional CW modulation, providing superior
performance to passive RFIDs, in terms of operating distance and network throughput, at the
expense of higher power consumption and complexity.
Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017
(a)
6
Z2
(b)
RFID reader
sensing
materials
Figure 2. (a) Conventional backscattering RFID. (b) Principle of ‘RFID as sensor’. (Online version in colour.)
a significant amount of research work has focused on ‘RFID technologies as sensors’ or ‘RFIDenabled sensors’ [21]. By integration of sensing materials such as water absorbing materials
for humidity sensors and carbon nanotubes for chemical sensors on RFID tags, the changes in
electrical characteristics such as capacitance or impedance can be observed by the reader through
the backscattered signal, as shown in figure 2. The humidity sensors and the temperature sensors
are most relevant for food monitoring. An inkjet-printed humidity sensor tag has been presented
in [22] for UHF RFID systems, and a passive wireless temperature sensor based on a chipless
UWB RFID tag was demonstrated in [23].
(ii) Wireless transmission
Extended operating distance and high data rate are highly desired for the future of
RFID technology. The wireless sensing applications present transmitter-dominated asymmetric
communications, i.e. tag-to-reader communication dominates the traffic, and the RF transceiver
dominates the power budget. The digitized sensor data not only complicate the circuit design
of the tag, but also substantially magnify the need for wireless transmission, e.g. greater than
10 metres and greater than 1 Mbps [24,25]. The UHF backscattering RFID provides superior
performance among passive RFID systems, i.e. a few hundreds of kbps and less than 10 m
coverage, but it is still insufficient for sensing applications where data from multiple sensors need
to be sent wirelessly. On the other hand, the active transceivers provide ideal performance, but
are either too complex and/or have high power consumption to cope with the µW-level power
budget of passive RFID tags.
The UWB impulse radio is a promising solution for future RFID systems to overcome most of
the limitations of current narrow bandwidth RFID technology [24,26]. It uses ultra-short pulses
in time domain instead of CW modulation, spreading the spectrum up to a few GHz [27]. The
wideband signal exhibits strong advantages for RFID to achieve both high-speed identification
and precise tag localization. Dari & D’Errico proposed a passive UWB RFID based on backscatter
modulation [28] achieving more than 20 m operating range. A UWB transmitter integrated in a
conventional passive UHF tag has also been proposed [29], with −18.3 dBm downlink sensitivity
and 10 Mbps uplink data rate. More recently, chipless RFID incorporating the UWB technology
has been developed for ultra-low-cost item-level tracking [30]. Such research has demonstrated
the feasibility of UWB to provide both reliable identification and high-definition localization of
tags, in spite of the increased complexity of the reader design [31]. Another approach to extend
the reading distance is to use semi-passive UHF tags. Semi-passive tags include a battery, much
like the active tags; however, the battery is used only to ‘assist’ communications in the presence of
a reader rather than supporting fully active transceiver circuitry. Semi-passive tags communicate
.........................................................
Z1
rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 372: 20130313
RFID reader
Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017
Intelligent tags attached on objects are intended to be small in geometry. Seamless integration
of intelligent devices into food packaging, the human body and various everyday objects calls
for novel approaches. Beyond existing system-on-chip integrations, future sensor tags could
be a multi-module system mounted on different materials or substrates. Therefore, it creates
new demands for interconnecting and packaging solutions, such as flexible electronics and
system-in-package. Meanwhile, as the antenna and its integration are up to 50% of the total tag
cost nowadays, innovative manufacturing processes need to be exploited. Technologies such as
printed antennas or on-chip antennas are expected to facilitate system miniaturization as well
as cost reduction. Although inkjet printing with functional inks enriches the functionalities and
lowers the cost of RFID tags, there is still a need to embrace silicon-based chips in most application
scenarios. Heterogeneous integration [36] offers a solution of faster adoption of new technologies
without sacrificing the overall system performance. It enables assembly of silicon circuits and
non-silicon materials on a flexible substrate, such as sensors, antennas, displays and IC chips. This
not only benefits from the advancement of new technologies for additional functionalities and
cost reduction, but also keeps the advantages of the current technologies such as performance,
robustness, availability, compatibility and more importantly time-to-market.
(c) Passive RFID with ultra-wideband
(i) Chipless UWB RFID
In contrast to conventional RFID tags with ICs, chipless RFID uses the radar principle, with the tag
information embedded into the electromagnetic signature of the structure. It does not need any
IC and communication protocol, and avoids the connections between IC and antenna. It features
high reliability and ultra-low cost, and facilitates fully printable tags through an inkjet printing
process on various substrates such as paper [37].
The simplest approach is to encode the ID number by variations of the impedance over
the transmission line, resulting in the on–off-keying modulated data by means of UWB pulse
reflections in time domain. The tag receives an interrogation signal from a reader via a UWB
antenna. The signal is then sent onto a transmission line and propagates forward until reflected by
a train of codes encrypted in the tag. The reflected signal is received by the reader and compared
with the original interrogation signal to pick up codes in sequence. The tag as illustrated in
figure 3a mainly consists of a microstrip line (ML) and a set of capacitors. The received UWB pulse
propagates over the ML until it encounters a shunt capacitor, which will introduce an impedance
discontinuity and reflect part of the incident signal back, denoting the coding of the digit ‘1’.
Meanwhile, the left signal will propagate forward, until it arrives at the next coding point, where
.........................................................
(iii) Integration and miniaturization
7
rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 372: 20130313
with the reader, as do the passive tags, but their battery can be used to increase the sensitivity
of the tag to reader signals in the environment and to boost the backscattered signal power to
increase the communication range.
In addition to enhancing the link capacity in terms of the data rate and the communication
range, reducing the amount of data to be transmitted is an alternative solution to relax the
communication requirements. Sensors gather large amounts of data about the conditions of their
surrounding environment, but not all of these data contain useful information. Compressed
sensing has been widely studied in recent years [32,33]. The principle is to use sub-Nyquist rate
to sample sparse events, thus reducing the amount of data to be stored or transmitted. Moreover,
technology scaling of the CMOS allows more complex microprocessors on the tag to perform
signal processing and extract useful data before transmission. The concept of ‘on-tag intelligence’
has been proposed [34], aiming to leverage more digital circuits on the RFID tag to relax the
analogue and RF circuits as well as the communication overhead. There have been a few ultralow-power processors with near-zero leakage power featuring less than 100 µW per MHz, which
demonstrate feasibility in RF energy harvesting applications [35].
Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017
(a)
8
interrogation
UWB signal
(b)
ML
(transmission line)
…...
9.5 cm
C1
C2
Rmatch
4.9 cm
Cn
…
...
Figure 3. (a) Principle of chipless UWB tags. (b) An implementation example of a fully inkjet-printed tag [38]. (Online version
in colour.)
(a)
UHF
ant.
storage cap.
conventional UHF tag
(b)
PPM-UWB modulator of the tag
power managment unit
matching
power
scavenging
voltage
sensor and
switches
low drop
regulator
analogue input
from sensors
Vdd
triangle
gen.
power on
reset
envelope
detector
data and clock
recovery
logic
control
data Rx
global clock
active
TX
Tx enable
UWB
ant
data Tx
UWB Tx
.
UWB TX
UWB RX.
ToA esimator
TDC
digitized
data
triangle
gen.
PPM
waveform
HF clock
injection divider
analogue input
from sensors
one-bit
comp
UWB reader as wireless TDC
PPM-UWB
output
RFID chip
Figure 4. (a) Schematic of UHF-powered UWB RFID tag [29]. (b) The system concept of time-domain sensing using PPM
UWB RFID. (Online version in colour.)
if no capacitor exists, it means the coding of ‘0’ and otherwise a coding of ‘1’. An implementation
example of the chipless UWB tag on photo paper with fully inkjet printing of metallic inks is
shown in figure 3b [38].
(ii) Passive UHF RFID with active UWB radio transmitter
The chipless tag shows strong potential for ultra-low-cost applications such as item-level tracking.
However, its performance such as reading distance and the system (coding) capacity is limited. In
order to overcome the shortcomings of the chipless solution to achieve extended reading distance
and system capacity, an alternative approach using an active UWB transmitter powered by UHF
RF energy harvesting has been presented [29].
Previous studies have found that a UWB receiver has high power consumption and is too
complex to be implemented in wirelessly powered systems such as RFID. On the other hand, a
UWB transmitter can be extremely simple and low-power. Approximately 10 to approximately
100 pJ per pulse energy consumption with an aggressive duty-cycling scheme is able to cope
with µW-level energy scavenged by RF waves as conventional passive UHF RFID. We also have
identified the similar asymmetric nature of passive identification and sensing applications, in
terms of the traffic load and the hardware complexity, i.e. tag-to-reader transmission dominates
the traffic, and therefore an energy-efficient and low-complexity transmitter at the tag side is
highly necessary.
Combining the asymmetric natures of RFID systems and UWB systems, a UHF/UWB hybrid
tag has been designed and implemented, as shown in figure 4a [29]. In the downlink (reader–
tag), the tag is powered and controlled by UHF signals as conventional backscattering tags,
whereas it uses a UWB transmitter to send data for a short time at a high rate in the uplink (tag–
reader). Such an innovative architecture has the benefits of UWB transmissions, whereas the tag
avoids the complex UWB receiver by shifting the burden to the reader. The silicon tape-out and
.........................................................
time-coded
reflected signal
rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 372: 20130313
discontinuity,
reflected signal
Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017
As discussed in the previous sections, the ADCs to sample the sensing data not only complicate
the tag IC, but also significantly increase the amount of data to be transmitted. A sensor
sampled at fs with Nb resolution ADC requires a transmission rate at Rdata = fs ∗ Nb . Passive
backscattering tags with a few hundred kb/s limit the quantization accuracy or sampling rate,
whereas active transmitters are power-consuming. There have been a number of sensors that
can easily be represented by time-domain information (pulse durations), because capacitive and
resistive sensor interfaces can directly convert the analogue signal into time-domain information.
Some ADCs for sensor interfaces use a time-to-digital converter to digitize data for wireless
transmissions [40].
As an attractive feature of UWB, ultra-wide signal bandwidth with fine resolution in time
domain provides high-precision time-of-arrival (ToA) estimations. Making use of this advantage,
we suggested an ADC-free time-domain sensor interface with pulse-position-modulated (PPM)
UWB signals. As shown in figure 4b, the analogue value is first converted into the time-domain
wave and modulated by UWB pulses.2 The time interval of the pulses represents the value of the
sensors which is recovered on the reader side by a ToA estimator. It eliminates the ADCs on tags
and, more importantly, it relaxes the amount of bits to be transmitted from the tag to the reader.
Using this approach, Bao et al. [25] presented a UWB sensor tag achieving 15 MS/s sampling and
transmission with 85 µW power consumption that is nearly three orders of improvement over
traditional approaches.
(d) Active RFID with hierarchical sensor portfolio
In spite of the need for battery replacement and relatively high cost, active RFID is still the primary
choice for most large-scale industrial level deployments for wireless sensing and monitoring.
The major reasons are the maturity for a guaranteed quality of service, and the well-established
development flow for software and applications. The most popular sensor nodes commercially
available include iMote2, TelosB, WaspMote and Arduino. They use off-the-shelf components
with standard radio and a simple application-programming interface. An active RFID tag consists
of three basic components. (i) Transceiver: IEEE 802.15.4 compliant transceivers are the most
common ones, such as TI CC2420, XBee and Nordic nRF24L. (ii) Processor: the processing
unit, which is generally associated with a small storage unit, performs tasks and executes the
communication protocols. Depending on applications, a variety of processor cores with different
trade-offs on power and performances are available, such as the eight-bit 8051, 16-bit MSP430,
to 32-bit ARM Cortex M. (iii) Energy source: in addition to traditional batteries, there have been
several alternatives, such as fuel cells and thin-film batteries. Moreover, energy harvesting for
active RFID sensors is widely investigated [41], which extracts energy from the environment of
the sensor tag, offering another important way to prolong the lifetime of the tags.
Figure 5a shows a design example of an active tag (sensor node) developed by Fudan
University, China, specifically for smart packaging and food tracking and monitoring systems.
It is in a credit-card-sized enclosure, with a temperature and humidity sensor, an accelerometer
2
PPM UWB signal can be directly generated for time-domain sensors which use the duration of pulses to represent the sensing
value.
.........................................................
(iii) UWB tag with time-domain sensor interface
9
rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 372: 20130313
experimental results have demonstrated that the tag has comparable complexity in terms of the
die size. The sensitivity of the tag is −18.5 dBm (14.1 µW). It corresponds to approximately 13.9 m
power-up distance, assuming 4 W effective isotropic radiated power for the reader, a matched
antenna with 0 dB gain for the tag, and free space propagations. Meanwhile, the UWB transmitter
consumes 91.8 µW instantaneous power at 10 MHz pulse rate corresponding to 9.2 pJ per pulse,
ensuring high data rate transmission (10 Mb/s) and high-precision positioning (submetre to
centimetre) [39]. Low-duty-cycle packet transmissions owing to a high data rate reduce collisions,
thus improving the energy efficiency and overall system throughput (greater than 2000 tag/s).
Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017
(b)
25
chipset
(RF+MCU)
28%
accelerometer
10%
battery
10%
transmission
20
15
processing
10
5
0
sleep
20
sleep
40
60
time (ms)
80
100
120
Figure 5. (a) Prototype of the active sensor tag. (b) Tag cost breakdown. (c) Current consumption of the tag. (Online version
in colour.)
and a magnetic stripe to detect the open/close information of the food packaging. This sensor
tag is to be placed in each first-level packaging, e.g. paperboard box, and communicates with
the master node (reader) installed in the cargo. The chipset of CC2530 is used to integrate an 8051
microcontroller and a ZigBee transceiver powered by a 1500 mAh soft battery. Figure 5b shows the
cost breakdown of the tag, and figure 5c depicts the measured power consumption of the sensor
tag. The tag works in a duty-cycle model and updates the data every minute. The operating time is
29 ms encompassing 24 ms processing time and 5 ms transmission time. The current consumption
of the tag is 11 µA, 9.7 mA and 31 mA in sleep, processing and transmission cycles, respectively.
The corresponding tag lifetime is more than 8 years assuming the battery capacity is automatically
decreased by 15% to account for self-discharge.
The master node, with a large battery capacity, powerful processing unit, GPRS/3G module
and local storages, collects data from the tag. The cost of the master node is strongly dependent
on the sensors and the sensor portfolio as discussed in [6], with respect to the deployment density
and sensing targets. In the prototype presented in §4, the configuration of sensors includes CO2 ,
O2 , C2 H4 , GPS, soil moisture and hydrogen sulfide sensors.
(e) Semi-passive RFID
Both passive and active technologies, as previously discussed in the paper, come with their
unique advantages and disadvantages depending on their application scenario. For example,
passive tags are ideal for high volume, low-cost tracking applications, whereas active tags have
significantly improved communication ranges and auxiliary capabilities for applications such as
remote sensory monitoring. To address some of the technical limitations of passive tags, such
as the lack of auxiliary functions and limited read range without the high cost of active technology,
a hybrid solution in the form of semi-passive RFID was introduced [42]. Instead of relying solely
on passive power harvesting, semi-passive tags use the battery to supply power to parts of
the tag circuitry used for encoding/decoding/processing/memory, etc. Because read range for
passive RFID is mostly limited by the forward link, this significantly increases the read range
for semi-passive tags (assuming comparable antenna sizes).
As far as communication performance of semi-passive tags is concerned, they are subjected to
the same type of performance tests as passive tags, such as tag sensitivity and read range because
they still use passive radio backscatter to communicate with the reader. Tag sensitivity defines
the minimum amount of radio frequency power that needs to be present at the tag location to
power up the tag, so it can transmit information back to the reader. However, for most commercial
applications, a far more useful performance parameter is the communication (or read) range of
the tag, which is directly correlated with the tag sensitivity.
One of the key differences between semi-passive and passive tags in terms of communication
performance is how the battery helps supply power to parts of the tag circuitry used for
encoding/decoding/processing/memory, etc. This means less power needs to be harvested
by the tag to turn on and operate thus easing the power requirements of the forward link
.........................................................
PCB+others
16%
T/H sensor
29%
8.5 cm
10
30
RTC
7%
battery
(c)
rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 372: 20130313
chipset (RF+MCU)
RTC
accelerometer
T/H sensor
current (mA)
(a)
800.0M
(a)
850.0M
900.0M
frequency (Hz)
950.0M
1.000G
(b)
800.0M
850.0M
900.0M
frequency (Hz)
950.0M
1.000G
Figure 6. Forward read range for a fully passive RFID tag—P1 (a) and a semi-passive RFID tag—SP1 (b). (Online version in
colour.)
(reader-to-tag) and increasing the read range for semi-passive tags using comparable antennas.
Because there are relatively few commercially available semi-passive tags compared with passive
tags, it is difficult to match the antenna sizes perfectly for such a comparison. However, antenna
theory tells us that the gain of an antenna is proportional to the effective antenna aperture, which
is a function of the area oriented perpendicular to the direction of the incoming radio wave. In
order to investigate performance differences as they relate to forward read range, we chose and
tested two commercially available RFID tags: a passive tag (P1) with an antenna size of 50 mm by
50 mm, and a semi-passive tag (SP1) with an antenna size of 50 mm × 20 mm. Figure 6 shows the
forward read range for the passive and semi-passive tag, respectively.
Based on the ratio of antenna apertures and assuming similar aperture efficiency, we expect
to see a gain difference of (50 mm × 50 mm)/(50
mm × 20 mm) = 2.5, which should theoretically
√
correspond to a read range increase of 2.5 = 1.6 for P1 compared with SP1. As shown in
figure 6, P1 has a maximum forward read range around 5 m, whereas for SP1 with a much
smaller aperture, the maximum forward read range is also around 5 m. These findings suggest
that a semi-passive tag with a similar antenna aperture will achieve a significantly larger read
range compared with a fully passive tag which supports our original observation that the battery
supplied power should help improve tag sensitivity, which, in turn, increases the true limiting
factor of forward read range.
The most common areas of use for semi-passive RFID tags include sensory applications
where a certain environmental variable, such as temperature or humidity values across time,
is measured and stored in the tag memory to be used for first-, second- and third-order supply
chain logistics. First-order logistics look at the stored raw data for compliance issues, such as to
see whether the cooling equipment in the transporting truck failed, or if the product temperature
remained within its prescribed temperature range. Going one step further, second-order logistics
involve processing raw temperature data for more useful information such as product quality
and remaining shelf-life. Finally, third-order logistics further use product quality and remaining
shelf-life data for intelligent supply chain decisions such as FEFO instead of FIFO to minimize
waste and maximize product quality in the case of perishables. For instance, RFID in any type of
perishable food logistics or monitoring agricultural products from farm to fork are applications
of semi-passive RFID within this domain.
Testing sensory circuitry is similar to measuring the accuracy and performance of any
temperature logger. However, compared with a general use temperature monitor, semi-passive
RFID sensor tags have more uniquely defined application fields in cold chain where temperature
accuracy may not be the most important criteria next to other specifications such as cost, RFID
read range, product durability, etc. For instance, for semi-passive temperature tags used in
cold chain monitoring and shelf-life prediction, researchers have come up with a more tailored
approach to measure the real-life sensory performance with the help of context-based accuracy
metrics instead of other objective measures such as root-mean-square error [43]. A more recent
version of this performance metric can be generalized to both continuous and discrete shelf-life
models as well as any other application involving the processing of recorded temperature data
and/or other environmental variables such as humidity, gas concentrations, etc. [44].
11
.........................................................
5.5
5.0
4.5
4.0
3.5
3.0
2.5
2.0
1.5
1.0
rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 372: 20130313
distance (m)
Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017
Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017
(a)
12
CO2 concentration
(ppm) ×104
20
10
1.5
1.0
0.5 T1
0
0
100
2200
CO2 concentration (%)
tempearture (C)
relative humidity (%)
30
2.0
80
60
40
20
0
0
200
400
600 800 1000 1200
time (h)
phase 1 phase 2
phase 3
phase 4
T2
T3
2000
1800
1600
1400
0
200
400
600 800 1000 1200
time (h)
phase 1 phase 2
phase 3
phase 4
Figure 7. Field test (from [45]). (a) Transportation route and system deployment for the field test and (b) field
testing data. (Online version in colour.)
To conclude, each RFID application naturally comes with its own unique requirements that
might benefit from different technologies within the passive-to-active spectrum. In this context,
semi-passive tags provide an effective trade-off between functionality, performance and cost
especially when it comes to smart supply chain logistics.
4. Global fresh food tracking: an application example
Here, we present an application example of an RFID-enabled wireless sensing system deployed
for fresh food tracking services [45].
(a) Architecture and operation flow of the global fresh food tracker
The sensor tags and master nodes (MSN) were deployed based on the two-layer network topology
as mentioned in the previous sections. The system collects all real-time primary condition
.........................................................
40
rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 372: 20130313
(b)
Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017
A six-week field test was performed for sweet melons shipped from Brazil to Sweden,
including handling of the process from harvest to packaging, storage and pre-cooling to air
and road transportation, shipment, distribution and finally to retail (figure 7a). During the
whole transportation, the sensor tags and master nodes measured conditions in the environment
including O2 , CO2 , temperature, humidity and mechanical stress such as vibrations and shocks.
The measurements are stored and transmitted through the cellular network. The software at the
Web server monitors and analyses the measured samples, and triggers alarms to subscribers with
event time and the location.
Analysis of data in figure 7b reveals that the transportation route can be divided into four
phases. In phase 1, sensor tags and master nodes were delivered to Brazil by air. Then, in phase
2, they were assembled in packages and transported by cargo truck in Brazil, where the hot and
wet environment was evident. In phase 3, melons were transferred onto a container ship, where
the air conditions were much better controlled by a cooling machine. Relative humidity increased
slowly from 90% to 95% for the evaporation of water in melons in a well-sealed container. It took
1 day to cool the melons down to the expected temperature. It took 1 day to cool the melons down
to the desired temperature, because the dense packing in a container degraded performance of the
cooling unit. They were transported again by cargo truck after they arrived at a port in Sweden
where the weather became much drier and cooler than in Brazil (phase 4). Damage threats are
observed at T1, T2 and T3 in figure 7b. Human interferences are the main causes of these threats,
such as bad maintenance of air conditions inside the container, transportation between vehicles,
opening of cargo doors and shocks observed in vehicles.
5. Summary
This paper reviewed state-of-the-art technologies and future trends for RFID-enabled wireless
sensing applications, including fresh food tracking and monitoring. As promising as the
prospects of wireless environmental monitoring are, complete implementation of RFID for such
applications is challenging in more than one area owing to technical limitations of both passive
and active technologies. For example, passive tags are cost-efficient and do not require batteries
to operate but they suffer from reduced communication range, relatively low data transmission
rates and insufficient functionalities for sensing. Even though active tags do not have the same
range and speed limitations as passive tags, they come at higher price levels and require batteries
to operate. In this review, we have highlighted the research which specifically addresses and
overcomes the challenges. For example, a passive tag with an active UWB transmitter was
developed which displayed a more than 10 m power-up distance with a 10 Mb/s uplink data rate.
High time-domain resolution UWB pulses enable ADC-free sensor interfaces and high-precision
tag localization. In addition, the proposed active tag architecture reduces the power requirements
of the tag to have an estimated battery life of more than 8 years. Finally, semi-passive tags act
like a hybrid technology to combine the advantages of both active and passive systems in one
design solution.
.........................................................
(b) Field testing and data analysis
13
rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 372: 20130313
parameters, including GPS, temperature, relative humidity, CO2 /O2 concentration and threeaxis acceleration. Furthermore, user-friendly service access tools for service registration and
specification, real-time data monitoring and tracking, alarm and closed-loop controlling
and reliable information sharing are provided as Web-based services running on the
service-oriented architecture. The service is managed by an operation centre, which controls all
the tags, MSNs and database. Services are accessible through different kinds of terminals from
complicated enterprise resource planning systems to the TCP/IP network. Typical user interface
comprises a Web-based data analysis and visualization tool, a Google Maps-compatible route
tracking tool and a short message service-based alarming and query tool for mobile phone users.
Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017
1. International Telecommunication Union. 2005 ITU internet reports: the internet of things.
Geneva, Switzerland: International Telecommunication Union.
2. Atzori L, Iera A, Morabito G. 2010 The internet of things: a survey. Comput. Netw. 54, 2787–
2805. (doi:10.1016/j.comnet.2010.05.010)
3. Ruiz-Garcia L, Lunadei L, Barreiro P, Robla I. 2009 A review of wireless sensor technologies
and applications in agriculture and food industry: state of the art and current trends. Sensors
9, 4728–4750. (doi:10.3390/s90604728)
4. Abad E, Palacio F, Nuin M, Zárate A, Juarros A, Gómez J, Marco S. 2009 RFID smart tag for
traceability and cold chain monitoring of foods: demonstration in an intercontinental fresh
fish logistic chain. J. Food Eng. 93, 394–399. (doi:10.1016/j.jfoodeng.2009.02.004)
5. Verdouw C, Sundmaeker H, Meyer F, Wolfert J, Verhoosel J. 2013 Smart agri-food logistics:
requirements for the future internet. In Dynamics in logistics (eds H-J Kreowski, B ScholzReiter, K-D Thoben). Lecture Notes in Logistics, pp. 247–257. Berlin, Germany: Springer.
(doi:10.1007/978-3-642-35966-8_20)
6. Pang Z, Chen Q, Han W, Zheng L. 2012 Value-centric design of the internet-of-things solution
for food supply chain: value creation, sensor portfolio and information fusion. Inf. Syst. Front.
(doi:10.1007/s10796-012-9374-9)
7. Heidmann N, Hellwege N, Peters-Drolshagen D, Paul S, Dannies A, Lang W. 2013
A low-power wireless UHF/LF sensor network with web-based remote supervision:
implementation in the intelligent container. In Proc. IEEE Sensors 2013, pp. 1–4. (doi:10.1109/
ICSENS.2013.6688422)
8. IBM Inc. 2013 Smart food. See http://www.ibm.com/smarterplanet/us/en/food_
technology/ideas/index.html.
9. Billerud AB. 2012 The Billerud book of maths. See http://www.billerud.com/.
10. Jedermann R, Palafox-Albarran J, Jabarri A, Lang W. 2011 Embedded intelligent objects in
food logistics technical limits of local decision making. In Autonomous cooperation and control in
logistics (eds M Hülsmann, B Scholz-Reiter, K Windt), pp. 207–228. Berlin, Germany: Springer.
(doi:10.1007/978-3-642-19469-6_16)
11. GDV. 2002 Container handbook. Berlin, Germany: GDV.
12. Angeles R. 2010 Moderated regression: effects of IT infrastructure integration and supply
chain process integration on the relationships between RFID adoption attributes and system
deployment outcomes. In Enterprise information systems (eds JEQ Varajão, MM Cruz-Cunha,
A Trigo). Communications in Computer and Information Science, vol. 110, pp. 406–415. Berlin,
Germany: Springer. (doi:10.1007/978-3-642-16419-4_42)
13. Regattieri A, Gamberi M, Manzini R. 2007 Traceability of food products: general framework
and experimental evidence. J. Food Eng. 81, 347–356. (doi:10.1016/j.jfoodeng.2006.10.032)
14. Zheng L-R, Nejad MB, Zou Z, Mendoza DS, Zhang Z, Tenhunen H. 2008 Future
RFID and wireless sensors for ubiquitous intelligence. In NORCHIP, 2008, pp. 142–149.
(doi:10.1109/NORCHP.2008.4738269)
15. Finkenzeller K, Waddington R. 1999 RFID handbook: radio-frequency identification fundamentals
and applications. New York, NY: Wiley.
16. Chen G, Hanson S, Blaauw D, Sylvester D. 2010 Circuit design advances for wireless sensing
applications. Proc. IEEE 98, 1808–1827. (doi:10.1109/JPROC.2010.2053333)
17. Opasjumruskit K, Thanthipwan T, Sathusen O, Sirinamarattana P, Gadmanee P,
Pootarapan E, Wongkomet N, Thanachayanont A, Thamsirianunt M. 2006 Self-powered
wireless temperature sensors exploit RFID technology. IEEE Pervasive Comput. 5, 54–61.
(doi:10.1109/MPRV.2006.15)
.........................................................
References
14
rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 372: 20130313
Networked intelligent tags with multi-functionalities will lead to RFID being considered a
key element for intelligent food logistics. It is safe to say that the future trend of the RFID
design will focus significantly on extending the functionality while reducing the cost to create
a smart combination of heterogeneous technologies towards a better system solution. As we
have reviewed in this paper, hierarchical network architecture is critical to fast adoption of novel
RFID and sensing technologies while facilitating the integration of RFID solutions into networked
information systems through existing ICT infrastructures.
Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017
15
.........................................................
rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 372: 20130313
18. Law MK, Bermak A, Luong HC. 2010 A sub-µmW embedded CMOS temperature
sensor for RFID food monitoring application. IEEE J. Solid-State Circuits 45, 1246–1255.
(doi:10.1109/JSSC.2010.2047456)
19. Somov A, Baranov A, Savkin A, Ivanov M, Calliari L, Passerone R, Karpov E, Suchkov A.
2012 Energy-aware gas sensing using wireless sensor networks. In Wireless sensor networks
(eds GP Picco, W Heinzelman), pp. 245–260. Berlin, Germany: Springer. (doi:10.1007/978-3642-28169-3_16)
20. Verma N, Chandrakasan AP. 2007 An ultra low energy 12-bit rate-resolution scalable
SAR ADC for wireless sensor nodes. IEEE J. Solid-State Circuits 42, 1196–1205.
(doi:10.1109/JSSC.2007.897157)
21. Kim S, Mariotti C, Alimenti F, Mezzanotte P, Georgiadis A, Collado A, Roselli L, Tentzeris
MM. 2013 No battery required: perpetual RFID-enabled wireless sensors for cognitive
intelligence applications. IEEE Microw. Mag. 14, 66–77. (doi:10.1109/MMM.2013.2259398)
22. Virtanen J, Ukkonen L, Bjorninen T, Elsherbeni AZ, Sydanheimo L. 2011 Inkjet-printed
humidity sensor for passive UHF RFID systems. IEEE Trans. Instrum. Measure. 60, 2768–2777.
(doi:10.1109/TIM.2011.2130070)
23. Girbau D, Ramos A, Lazaro A, Rima S, Villarino R. 2012 Passive wireless temperature sensor
based on time-coded UWB chipless RFID tags. IEEE Trans. Microw. Theory Tech. 60, 3623–3632.
(doi:10.1109/TMTT.2012.2213838)
24. Zou Z, Baghaei-Nejad M, Tenhunen H, Zheng L-R. 2007 An efficient passive RFID system for
ubiquitous identification and sensing using impulse UWB radio. Elektrotech. Inf. 124, 397–403.
(doi:10.1007/s00502-007-0483-y)
25. Bao D et al. In press. A wirelessly-powered UWB sensor tag with time-domain sensor
interface. In Proc. IEEE Int. Symp. on Circuits and Systems (ISCAS), Melbourne, Australia, 1–5
June 2014.
26. Dardari D, D’Errico R, Roblin C, Sibille A, Win MZ. 2010 Ultrawide bandwidth RFID: the next
generation? Proc. IEEE 98, 1570–1582. (doi:10.1109/JPROC.2010.2053015)
27. Win MZ, Scholtz RA. 1998 Impulse radio: how it works. IEEE Commun. Lett. 2, 36–38.
(doi:10.1109/4234.660796)
28. Dardari D, D’Errico R. 2008 Passive ultrawide bandwidth RFID. In IEEE Global
Telecommunications Conf., GLOBECOM 2008, pp. 1–6. (doi:10.1109/GLOCOM.2008.ECP.757)
29. Baghaei-Nejad M, Mendoza DS, Zou Z, Radiom S, Gielen G, Zheng L-R, Tenhunen H. 2009
A remote-powered RFID tag with 10Mb/s UWB uplink and −18.5 dBm sensitivity UHF
downlink in 0.18 µm CMOS. In IEEE Int. Solid-State Circuits Conf. Digest of Technical Papers,
pp. 198–199, 199a. (doi:10.1109/ISSCC.2009.4977376)
30. Preradovic S, Karmakar NC. 2010 Chipless RFID: bar code of the future. IEEE Microw. Mag.
11, 87–97. (doi:10.1109/MMM.2010.938571)
31. Zou Z, Mendoza DS, Wang P, Zhou Q, Mao J, Jonsson F, Tenhunen H, Zheng L-R. 2011 A lowpower and flexible energy detection IR-UWB receiver for RFID and wireless sensor networks.
IEEE Trans. Circuits Syst. I Regul. Pap. 58, 1470–1482. (doi:10.1109/TCSI.2011.2142930)
32. Donoho DL. 2006 Compressed sensing. IEEE Trans. Inf. Theory 52, 1289–1306. (doi:10.1109/
TIT.2006.871582)
33. Shancang L, Li Da X, Xinheng W. 2013 Compressed sensing signal and data acquisition
in wireless sensor networks and internet of things. IEEE Trans. Ind. Inf. 9, 2177–2186.
(doi:10.1109/TII.2012.2189222)
34. Zou Z. 2011 Impulse radio UWB for the internet-of-things: a study on UHF/UWB hybrid
solution. Doctoral dissertation, KTH, Sweden.
35. Bartling SC, Khanna S, Clinton MP, Summerfelt SR, Rodriguez JA, McAdams HP. 2013 An
8 MHz 75 µA/MHz zero-leakage non-volatile logic-based Cortex-M0 MCU SoC exhibiting
100% digital state retention at V DD = 0 V with 400 ns wakeup and sleep transitions. In
IEEE Int. Solid-State Circuits Conf., Digest of Technical Papers, pp. 432–433. (doi:10.1109/ISSCC.
2013.6487802)
36. Xie L, Feng Y, Mantysalo M, Chen Q, Zheng L. 2013 Integration of f-MWCNT sensor
and printed circuits on paper substrate. IEEE J. Sensors 13, 3948–3956. (doi:10.1109/JSEN.
2013.2260534)
37. Tedjini S, Perret E, Deepu V, Bernier M. 2010 Chipless tags, the next RFID frontier. In The
internet of things (eds D Giusto, A Iera, G Morabito, L Atzori), pp. 239–249. New York, NY:
Springer. (doi:10.1007/978-1-4419-1674-7_23)
Downloaded from http://rsta.royalsocietypublishing.org/ on June 15, 2017
16
.........................................................
rsta.royalsocietypublishing.org Phil. Trans. R. Soc. A 372: 20130313
38. Shao B, Chen Q, Amin Y, Mendoza D, Liu R, Zheng L-R. 2010 An ultra-low-cost RFID tag with
1.67 Gbps data rate by ink-jet printing on paper substrate. In IEEE Asian Solid State Circuits
Conf., pp. 1–4. (doi:10.1109/ASSCC.2010.5716569)
39. Zou Z, Deng T, Zou Q, Sarmiento M, Jonsson F, Zheng L-R. 2010 Energy detection receiver
with TOA estimation enabling positioning in passive UWB-RFID system. In IEEE Int. Conf. on
Ultra-Wideband, pp. 1–4. (doi:10.1109/ICUWB.2010.5615993)
40. Naraghi S, Courcy M, Flynn MP. 2009 A 9b 14 µW 0.06 mm2 PPM ADC in 90 nm digital
CMOS. In 2009 IEEE Int. Solid-State Circuits Conf., Digest of Technical Papers, pp. 168–169.
(doi:10.1109/ISSCC.2009.4977361)
41. Gungor VC, Hancke GP. 2009 Industrial wireless sensor networks: challenges, design
principles, and technical approaches. IEEE Trans. Ind. Electron. 56, 4258–4265. (doi:10.1109/
TIE.2009.2015754)
42. Jedermann R, Lang W. 2007 Semi-passive RFID and beyond: steps toward automated quality
tracing in the food chain. Int. J. Radio Freq. Identif. Technol. Appl. 1, 247–259. (doi:10.1504/
IJRFITA.2007.015849)
43. Uysal I, Emond J, Bennett G. 2011 Tag testing methodology for RFID enabled temperature
tracking and shelf life estimation. In 2011 IEEE Int. Conf. on RFID–Technologies and Applications,
pp. 8–15. (doi:10.1109/RFID-TA.2011.6068608)
44. Hertog MLATM, Uysal I, McCarthy U, Verlinden BM, Nicolaï BM. 2014 Shelf life modelling
for first-expired-first-out warehouse management. Phil. Trans. R. Soc. A 372, 20130306.
(doi:10.1098/rsta.2013.0306)
45. Pang Z, Chen J, Zhang Z, Chen Q, Zheng L. 2010 Global fresh food tracking service
enabled by wide area wireless sensor network. In IEEE Sensors Applications Symp., pp. 6–9.
(doi:10.1109/SAS.2010.5439425)
Radio frequency identification
enabled wireless sensing for
intelligent food logistics
Zhuo Zou, Qiang Chen, Ismail Uysal and Lirong Zheng
rsta.royalsocietypublishing.org
Phil. Trans. R. Soc. A 372, 20130313 (13 June 2014;
Published online 5 May 2014) (doi:10.1098/rsta.2013.0313)
Correction
Cite this article: Zou Z, Chen Q, Uysal I,
Zheng L. 2014 Radio frequency identification
enabled wireless sensing for intelligent food
logistics. Phil. Trans. R. Soc. A 372: 20140209.
http://dx.doi.org/10.1098/rsta.2014.0209
One of the authors’ names was spelt incorrectly. Qing
Chen should be Qiang Chen.
Author list: Zhuo Zou, Qiang Chen, Ismail Uysal and
Lirong Zheng
2014 The Author(s) Published by the Royal Society. All rights reserved.