System Architecture for High-speed Close-proximity Low-power RF Memory Tags and Wireless Internet Access

System Architecture for High-speed Close-proximity Low-power RF Memory Tags
and Wireless Internet Access
Iiro Jantunen
Department of Applied Physics
University of Eastern Finland
Kuopio, Finland
iiro.jantunen@uef.fi
Joni Jantunena
, Harald Kaajaa
, Sergey Boldyrevb
a
Nokia Research Center
b
Nokia Location & Commerce
Helsinki, Finland
{joni.jantunen, harald.kaaja, sergey.boldyrev}@nokia.com
Le Wang, Jyri Hämäläinen
Department of Communications and Networking
Aalto University
Espoo, Finland
{le.wang, jyri.hamalainen}@aalto.fi
Abstract — We have developed an open architecture platform
for implementing passive radio-frequency identification
(RFID) tags with a mass memory for close proximity environ-
ment. Purposes for such mass memory tags are, e.g., multime-
dia files embedded in advertisements or logged sensor data on
a low-power sensor node. In the proposed architecture, a mo-
bile phone acts as the reader that can read or write the
memory of these RFID tags. The architecture also enables
creation of a new type of wireless internet access suitable for,
e.g., internet kiosks. The architecture is designed so that devel-
opment path to a full Network on Terminal Architecture (No-
TA) is feasible. The wireless reading speed of the mass memory
tags, demonstrated to be 112 Mbit/s, is in the range that a 3-
minute 640×320-pixel video can be loaded from the tag to the
phone in less than 10 s. Our solution supports Nokia’s Explore
and Share concept.
Keywords — memory architecture; multimedia systems;
RFID; telephone sets; RF memory tags; Internet connection
I. INTRODUCTION
Today’s mobile phones contain music and video players,
which make it possible for consumers to enjoy entertainment
while on the move. Acquiring new multimedia content by
downloading or streaming, however, is hampered by the high
cost and slow speed of Internet connections, as well as by the
fact that commonly used physical multimedia formats, such
as optical disks, cannot be read with a mobile phone. Thus,
to make acquiring new content easier, cheaper and less pow-
er-consuming, we propose a new technology based on radio
frequency (RF) memory tags readable and writable by mo-
bile phones [1].
RFID tags are increasingly a part of our life; transport,
traceability, and secure access are some of the main uses of
this close proximity technology today. Conventional ma-
chine-readable wireless tags, e.g., Near Field Communica-
tion (NFC) tags, normally have a very small memory in the
range of hundreds of bytes or kilobytes. Some RFID stand-
ards include an option to have a flexible-use memory, but the
capacity is low compared to factory-set fixed-content
memory. Tag selection is based on reading the content in a
selected tag memory address (e.g., tag or manufacturer ID).
As the memory capacity of these tags is small, the amount of
data to be transferred is also small and power consumption of
RF communication is, thus, not a critical issue.
Various research groups have developed improvements
to the commercially available RFID technologies. To over-
come the storage capacity limitation of passive tags, Wu et
al. increase effective tag storage sizes with proposed distrib-
uted RFID tag storage infrastructure (D-RFID stores) [2].
Tags would be distributed in space and time in this architec-
ture. Ahmed et al. focus on RFID system unreliability and
improvements in middleware for object tracking and object
location with moving readers or tags [3]. As a result of their
research, a virtual reader system architecture was introduced.
Ying described a verification platform for RFID reader that
utilized Ultra High Frequency (UHF) frequency [4]. This
platform is applicable for customization with different RFID
standards. Pillin et al. have developed a passive far-field
RFID tag using the 2.45 GHz Industrial, Scientific and Med-
ical (ISM) band, with a data rate of 4 Mbit/s on the range of
5.5 cm [5]. As an example of a proprietary solution, HP’s
Memory Spot tag also works on the 2.45 GHz ISM band and
has demonstrated 4 MB memory and 10 Mbit/s data rate but
only allowing a touch range [6].
The problems of low data reading rate and small memory
size provided by contemporary RFID tags become empha-
sized if one considers mobile users reading multimedia files
from tags embedded on, e.g., paper media. The attention
span of a mobile user is about 10 seconds [7]. Within this
period, the user could get a single multimedia content file
from a memory tag. Considering a movie trailer, the file size
for a 2-minute 640×320-pixel 30-fps (3 Mbit/s), encoded
with H.264, would be around 50 MB [8]. The required min-
imum data transfer rate from the user point-of-view is thus
50 Mbit/s. This exceeds the maximum data rate available by
the 13.56 MHz NFC technology, 848 kbit/s, by a factor of
60. Even the maximum data rate for NFC demonstrated on a
laboratory set-up, 6.78 Mbit/s [9], is not enough. When users
are getting used to NFC, the speed and storage capacity be-
comes quite easily a limiting factor. Thus, there is a need for
a new high-speed touch-range RFID radio interface.

The aim of our research has been to develop a high-
capacity memory tag, which is wirelessly readable with a
mobile phone and suitable for consumer markets in ubimedia
applications [10][11]. The mobile phone acts as the user
interface for reading and writing passive RF memory tags
that contain a high-capacity memory (0.1–1 GB). The rea-
soning for the proposed technology was justified by modern
trends in non-volatile memory technologies, according to
which the power consumption, physical size, and price of
memory are continuously decreasing. The technology pre-
sented in this paper supports Nokia’s Explore and Share
concept, a new way of transferring content (e.g., multimedia,
maps, and applications) to a mobile phone [12].
Another use case for the technology presented in this pa-
per is an Internet Kiosk, i.e., a short-range hotspot providing
access to the Internet. Traditionally, an Internet kiosk is
nearly always a computer connected to core network via a
backhaul link that can be wired or wireless broadband. The-
se Internet Kiosks are typically pay-as-you-go (credit card
or pre-paid) or funded by advertisements presented within
the kiosk screen.
Nowadays, accessing the Internet freely and securely has
become critical for business and recreational needs. Personal
wireless terminals (such as mobile phones or laptop comput-
ers) are, thus, associated with modern Internet hotspots based
on wireless local area network (Wi-Fi) technology. These
Wi-Fi hotspots are either free or paid by credit card when
logging in and are widely deployed. Another kind of wire-
less Internet hotspot is femtocell-based [14], providing a
local connectivity through cellular technology such as wide-
band code division multiple access (WCDMA). Convention-
ally, femtocells are designed for use in a home or small busi-
ness to improve indoor wireless coverage. As Bluetooth is a
competitive wireless solution especially from energy saving
point-of-view, it has also been considered as one technology
to enable wireless kiosks [15]. Moreover, NFC featured
kiosks allow users to be connected and download multimedia
content via NFC-enabled mobile phones [16].
All the wireless internet kiosk or hotspot types mentioned
above share the demand of internal power within the phone,
the power usage of the connection depending on the technol-
ogy and status of the hotspot (distance, amount of devices
connected etc.). If draining the battery of the phone is not
acceptable (e.g., while travelling), a wired connection to a
power outlet is needed.
In this paper, we describe and specify a network architec-
ture, which enables mobile phones to read and write passive
RF memory tags and use a RF memory tag based Internet
connection. The architecture has been developed and demon-
strated in the EU's 6th
Framework Programme (FP) “Micro-
Nano integrated platform for transverse Ambient Intelligence
applications” (MINAmI) project [17], and thus the architec-
ture is referred to as MINAmI Architecture. Important archi-
tecture requirements include openness, modularity, scalabil-
ity, and energy efficiency. Openness and modularity are
needed to support creation of novel applications and services
by 3rd
parties. Scalability of data rate is needed to enable
evolution of the technology along with evolution of multi-
media services. Energy efficiency is essential to enable pas-
sive operation of the tags as well as to save the phone’s bat-
tery.
The paper is organized as follows. In Section II, we in-
troduce the system architecture, along with a key component
of the architecture, RF memory tags. In Section III, we intro-
duce a novel dual-band radio subsystem and its hardware and
software implementation. In Section IV, we introduce a new
type of power-saving short-range wireless internet access for
mobile devices and compare its power use to other available
wireless internet access technologies. In Section V, we pre-
sent the current status of implementation of the architecture,
discuss possibilities for future development, and draw some
conclusions.
II. MINAMI ARCHITECTURE
The proposed MINAmI architecture makes use of the
mobile phone’s capability of running software and providing
several radio interfaces (Fig. 1). The architecture is modular,
enabling simpler and faster development of new technical
extensions (e.g., RF memory tags). Our architecture focuses
on utilization of modularity on component level (e.g., where
to plug memory tag functionality) and on communication
level (e.g., how the available memory tags are utilized). At
close proximity domain (range < 1 m), different tags are
communicating locally with a mobile phone. In the present
work we have concentrated on the RF memory tags. The
sensor parts of the architecture (RFID sensor tags and Blue-
Figure 1. MINAmI Architecture [11]

tooth sensor devices) have been studied in an earlier project
[18][19].
The main RF memory tag architectural design challenges
include target platform performance obstacles, such as avail-
able bus operations (read/write) and power requirements,
especially when drawing the line for autonomous operations
in the described MINAmI architecture. The other challenge
is minimizing changes to the existing system communication
layering, only to the external memory stack block. The tech-
nological choices in the MINAmI system architecture were
able to support both existing standard radios for low-rate
sensors, and the high-rate high-capacity memory tags.
A. Network-on-Terminal Architecture (NoTA)
NoTA is a modular service-based system architecture for
mobile and embedded devices offering services and applica-
tions to each other [20]. The concept is being defined in an
open initiative. NoTA is also known as an open device dis-
tributed architecture, which allows direct connections be-
tween different nodes, within subsystem or between subsys-
tems, supporting several physical interfaces within the device
or between devices [21]. This architecture also supports both
messaging and streaming services. The beauty in the archi-
tecture resides in modularity and transport independency.
Direct connection between subsystems improves the effi-
ciency as they do not necessarily require any processor in-
volvement, when subsystems have all the needed functionali-
ties available for their independent operations. Transport-
specific portion is hidden underneath NoTA communication
layering.
NoTA communication layering is built around transport-
independent parts and it provides interfaces towards
transport-specific parts (Fig. 2). Extension architecture with
Device Interconnect Protocol (DIP) enables flexible open-
source architecture for different hardware platforms. DIP
provides logical links between a requesting subsystem and
other subsystems or within a subsystem [21]. DIP is a de-
vice-level communication protocol that can be implemented
for various physical interfaces ranging from MIPI (Mobile
Industry Processor Interface) high speed serial interfaces and
Universal Serial Bus (USB) to wireless interfaces, such as
Bluetooth [22][23]. Another example of utilization of DIP is
the Open Modem Interface Protocol [24].
NoTA host subsystem and neighboring subsystems are
connected via the high speed physical interface. DIP adapts
physical interfaces to the upper layers. It is the lowest layer
that is common for all subsystems (i.e., also for MINAmI
subsystem) and hides the physical dependencies underneath.
Above DIP there is a common service interface used for
resource management, file systems, and system boot-ups.
Middleware frameworks, e.g., for multimedia, USB, and
other applications, use a common service interface or exten-
sion Application Programming Interface (API). The architec-
ture also takes into account vertical solutions, which may
require an optimized protocol design for certain requirements
that are tied to HW-specific applications.
NoTA subsystem structure takes into account possibility
to add different types of independent service or application
subsystems to the architecture, and the MINAmI subsystem
forms one high data rate high capacity subsystem. When
properly designed, the modular NoTA-based subsystem
specification involves clear distinction of the system design-
er/integrator and vendor views of subsystem description and
scenarios. Based on a provided subsystem specification the
vendors test and validate their subsystem implementation
and deliver them to the product designers for integration
[25].
The MINAmI subsystem offers memory tag read/write,
storage and local connectivity services to other subsystems
within mobile device, and its architecture is compatible with
NoTA communication layering. The MINAmI subsystem
includes both the mobile phone (Mobile Reader/Writer) and
the tag and all the relating hardware and software resources.
Mobile Reader/Writer sees the contents of the memory of a
Figure 2. NoTA extension architecture for MINAmI subsystem, where HSI = High Speed Serial Interface; HS-MMC = High Speed MultiMediaCard: SPI =
Serial Peripheral Interface bus.

passive RF memory tag only when there is an established
connection, i.e., power field and data connection exists.
B. RF Memory Tags
The focus of our research has been on mobile-phone-
operable memory tags suitable for consumer markets and
ubimedia applications. The tag is developed as a part of our
mobile-phone-centric architecture. Our memory tag devel-
opment targets improving both transfer speed and storage
capacity. These improvements give direct benefit for
ubimedia users.
The target memory capacity of our memory tag has been
in the range of gigabits and mobile reader/writer transfer
speed to and from memory tag in excess of 10 Mbit/s. The
same design platform is usable for both ends, for mobile
phone platform reader/writer and for tag implementation.
When designing the platform, various important design pa-
rameters, such as the selection of the used radio technology,
were considered to provide an efficient and low-power solu-
tion for mobile reader/writer and tags.
It was important to make sure that connectivity technolo-
gy is simple enough for the user, e.g., it should facilitate easy
content selection (see Section III.D). Memory tag content
selections should be based on metadata (e.g., filenames, file
content types, file content keywords). Due to the large
memory size, power consumption for memory access is a
critical design issue, both for reading and writing the
memory tag. To be successful on the market, RF memory
tags for ubimedia must be passive to make them as small
(size) and cheap as possible, and to achieve autonomous
usage with minimum maintenance (e.g., usage without
charging of battery). This severely limits the power budget.
On the other hand, a short communication range (even touch)
is sometimes preferable to make it easier for the user to
physically select the tag. An RF memory tag will be read
many times by different users, but written more rarely – in
some cases, only once. The memory unit must work reliably
even with several consecutive read cycles. A limited write
throughput due to power constraint is not an issue, as data is
rarely written by the users.
III. UWB LOW END EXTENSION
As memory tags have high data storage capacity, a high-
speed radio is needed for communication to enable reading
even all the contents of the tag in an acceptable time. Cur-
rently available mobile phones contain several radio trans-
ceivers, such as cellular, Bluetooth, and Wi-Fi, along with
NFC. Most of the technologies are made for well-established
communication between active devices, consuming a rela-
tively large amount of power. These technologies are also
not inherently designed for ad-hoc, possibly one-time, con-
nections between devices that have not communicated with
each other before, resulting in long latency in establishing
the communications. For example, in an environment with
many unknown Bluetooth devices, the Bluetooth connection
setup latency can be over 10 seconds [26]. NFC enables
communications between an active and a passive battery-less
device and is physically more selective; its communication
range is almost in touch. However, it has severe limitations
in data transfer speed.
To provide higher data rates, a wider frequency band
available on higher frequencies needs to be used. On the
other hand, the efficiency of wireless power transfer (WPT)
decreases as a function of center frequency. To solve the
problem of providing high-speed communication (high fre-
quency needed) while simultaneously providing power wire-
lessly to the tag, a dual-band radio interface has been pro-
posed [27]. One narrowband signal on RFID frequencies
(e.g., RFID frequency bands globally available between 860–
960 MHz) is used to power the tag and to provide a mutual
clock reference for both ends of the communication link,
whereas the communication link itself is based on impulse
Ultra-Wideband (UWB) technique to provide a high com-
munication bandwidth and scalability for even higher data
rates.
As the selected RFID frequencies are approximately in
the same frequency range as Global System for Mobile
Communications (GSM) or WCDMA 900 MHz, in the read-
er there is a possibility of integrating the WPT function to the
existing Phone Radio Subsystem, as presented in Fig. 3. In
that case, Phone Radio Subsystem is designed so that the
WPT Physical (PHY) Layer function may request a direct
access to control the activation of the narrowband transmit-
ter. Especially, the time-domain interleaving of different
Figure 3. MINAmI Architecture on phone
PHONE
UI/APP
MIPI BUS
WPT
MINAmI CONTROLLER
POWER
SUBSYS
MEMORY
SUBSYSTEM
BATTE
RY
PHONE
RADIO
SUBSYS
MINAmI
SUBSYSTEM
POWER
MANAGEMENT
I-UWB
HSI
MEMORY
CONTROL LAYER
MIPI mPHY
PCM
NETWORK
UWBLEE MAC
UWBLEE PHY
I-UWB WPT
CONTROL BLOCK
Figure 4. MINAmI Architecture on a RF memory tag

functions is important to support co-existence of
GSM/WCDMA and WPT signaling.
The architecture of the RF memory tag (Fig. 4) is similar
to the MINAmI subsystem on the mobile phone. For simple
RF memory tags, no network layer implementation is needed
to take care of the point-to-point communication between the
reader and the tag, and therefore is handled on Medium Ac-
cess Control (MAC) layer.
As an option for use-cases like data-logging sensor de-
vices, the memory control layer provides a sensor interface.
During the sensing, the sensor data is stored to the Phase-
Change Memory (PCM) block and the low data-rate data
capturing is powered from a battery or with energy harvested
from the environment. For fast downloading of the logged
data, the reader powers the sensor tag wirelessly.
A. Hardware architecture
The hardware, both the radio front-end and the memory,
of the RF memory tag needs to run on the energy scavenged
from the UHF transmission of the mobile phone. This sub-
section describes the enabling technologies: low-power
high-data-rate radio front-end and low-power high-capacity
high-speed non-volatile memory.
1) Radio Front-end
As presented in [27], a very simple super-regenerative
transceiver architecture can be used in impulse UWB com-
munication to achieve required data-rates over short distanc-
es. In contrast to conventional impulse UWB transceivers
[28], there is no need for multipath recovery over the dis-
tances below 30 cm. This decreases the requirements set for
the UWB transceivers. This is used to minimize complexity
and power consumption of the transceivers. In the aforemen-
tioned super-regenerative transceiver one super-regenerative
oscillator is used alternately both to generate transmitted
pulses and to amplify received pulses, and no linear amplifi-
ers are needed. Thus, the architecture utilizes the inherently
low duty cycle of the transmitted impulse UWB signal also
in reception the receiver being fully active only exactly dur-
ing the detection of incoming pulses.
Synchronization is often problematic in impulse UWB
systems because of the low duty cycle and pseudo-random
timing of pulsed signal, and due to frequency drift and dif-
ferences of reference clocks between the transceivers. In the
proposed system the frequency synchronization between the
reader and tag is achieved thanks to the mutual narrowband
WPT signal, which is also used as the reference clock. The
phase synchronization of impulse UWB transceivers is also
easier to achieve due to decreased need for pseudo-random
time-coding of pulse patterns.
The transceiver structure supports simple On-Off-Keying
(OOK) modulation. The data-rate and power consumption is
also scalable depending on the power level available for the
wirelessly powered tag. Due to the simplified transceiver
structure, targeted ultra-low power consumption and partial
exploitation (500 MHz) of full UWB band (3.1–10.6 GHz)
authorized by Federal Communications Commission (FCC)
for unlicensed use, the impulse UWB system referred here is
called UWBLEE (UWB Low End Extension).
Altogether, the optimized transceiver architecture makes
it possible to achieve required high data-rates with a low
power consumption performance (a few mW) suitable for
WPT. As a proof-of-concept, a complete wirelessly powered
RF front-end implementation of the super-regenerative trans-
ceiver is presented in [29] and [30] by using a single super-
regenerative oscillator for transmission and reception. The
front-end implementation supports data-rates up to 112
Mbit/s with the energy consumption of 48 pJ/bit in reception
and 58 pJ/bit in transmission. The feasibility of the ultra low
power consumption in high data-rate two-way communica-
tion is verified with an integrated RF front-end implementa-
tion based on the symmetrical transceiver architecture pro-
posed earlier [27]. A 900 MHz WPT signal is used as a mu-
tual clock reference and the communication is done over an
impulse UWB link at 7.9 GHz center frequency. The scala-
ble data-rate of UWB link up to 112 Mbit/s has been demon-
strated as well as robustness against narrowband interfer-
ence.
2) Non-Volatile Memory (NVM) technology
The main reason to pick up PCM in favor of any other
memory technology [31] were the benefits of PCM technol-
ogy, e.g., the estimated high number of read/write cycles as
1×10^6, which consequently results in need of no or just a
lightweight wear leveling algorithm, and the bit alterability –
lack of need of block erase cycles (as with flash memory)
when data should be stored. From the perspective of technol-
ogy lifecycle PCM stands now between a pure innovative
technology and early adopters’ stage. There are several 90
nm products [32] on the market already and more to come.
Aggregating main memory characteristics in comparison
with NAND/NOR flash technology and dynamic random-
access memory (DRAM) execution memory, PCM stands
between those two in terms of cost per die. It is characterized
as 5.5 F^2 factor in cell size having the same wafer complex-
ity as DRAM technology. Currently only Single Level Cell
(SLC) PCM is available, though Multi-Level Cell (MLC)
PCM is on the way out, which can substantially extend the
density and, justify the cost structure [31]. Thus, the applica-
tion range can be quite wide from external usage (cards,
keys) and wireless applications (RF memory tags) to high
performance computing applications (caches, code execu-
tion, data storage). Considering reliability characteristics it is
important to note that PCM technology gives more than 10
years retention ratio that can be extended even further, if
necessary, by proper bit error management.
PCM has performance characteristics such as read &
write latency and read & write endurance almost as good as
DRAM, while giving clear benefits through the non-volatile
nature of PCM technology. PCM has a low system-wise
energy consumption (~0.2 mW/pF read, <1.25 mW write)
~<1 mW/GB of idle power, access time comparable to
DRAM (~85 ns), with read latency 50–100 ns, write band-
width from 10 to 100+ Mbit/s/die, write latency 500 ns – 1
µs, various packaging and die stacking solutions, high-speed
low-pin-count low-power interface solutions, and maturity of
the technology as such.
The PCM technology highlights provide clear reasoning
for the selection of such technology for the RF memory tag

application, preserving the opportunity to justify it even
further when some other application should be designed.
B. Software Architecture (protocol stack)
The MINAmI software architecture (protocol stack) is
designed to be modular and scalable. The protocol stack is
based on three layers: Network Layer, MAC Layer, and PHY
Layer. The APIs of the layers are open for 3rd
parties. These
layers will be presented in the following sections. The proto-
col stack has been developed taking into account future
compatibility with NoTA architecture. Care should be taken
to have a clear implementation path towards the final archi-
tectural (NoTA) solution.
1) Network Layer
Network Layer will first only provide point-to-point con-
nections regardless of state. In future, also applications using
multiple targets could become feasible when MINAmI Sub-
system is in active mode. If a point-to-multipoint network
protocol is needed, nanoIP is easily implementable [18][33].
However, to get full internet support classical IP protocol
may be valid, and more common in networking devices. In
the final architecture (NoTA) solution, the network layer will
consist of Device Interconnect Protocol (DIP), as a middle-
ware, which guarantees the compatibility with NoTA. In DIP
protocol, it is possible to select, which transport mode and
network is used. For example DIP TCP L_IN (transport
selected) is ready to be used within one device and between
several devices in a sub-network as such. Multicasting must
be enabled in IP interface in order for device discovery to
work. Nodes, which are in different sub-network, cannot be
detected [21].
Packet size is an important parameter and depends on
what is feasible for MAC and PHY layers. Upper layer pack-
ets are segmented and reassembled and this is dependent on
what kind of packet sizes the system supports.
2) MAC Layer
The MAC of the novel dual-band radio interface has
three different operational modes: the passive mode, where
no internal power source is available or used; and the active
and semi-passive modes, where internal power source is
available. Tags on battery-less objects without power wire
connection (e.g., implanted on paper) are passive.
In the active mode, the mobile phone actively searches
and selects the target tags, sends the targets the WPT signal
for powering and for frequency synchronization of the com-
munication link, reads/writes data on the tags, and closes the
connection to the target when active connection is no longer
required. This operation can be an automatic feature, or ena-
bled by the user (initiating the application for reading and
writing the tag). In the semi-passive mode the phone receives
data sent by an outside device, but powers itself, allowing a
longer communication range, which would otherwise be
limited by the WPT link. In semi-passive mode, however, the
initiator device takes care of the synchronization of the I-
UWB communication link.
Active mode states are used by battery-powered mobile
devices, whereas passive mode states are applied for passive
devices and tags. In passive mode, possible connections are
powered by an outside device with WPT. In the passive
operating mode, the default state (when powered by an out-
side device) is P-IDLE, i.e., ready to receive any data, after
the boot-up sequence.
The main operational states of UWBLEE MAC are
shown in Figures 5 and 6. In addition to the shown directions
of movement from state to state, there need to be possibility
of built-in error recovery operation from any operational
state to the corresponding idle state (A-IDLE or P-IDLE).
For the applications requiring higher security, a suitable
security protocol can be applied for the ongoing data trans-
mission.
3) Physical Layer
UWBLEE PHY layer controls both the I-UWB commu-
nications and Wireless Power Transfer (WPT) transmission.
Depending on the operational mode (active or passive) WPT
link is used to send (or receive) power and/or to provide the
clock reference signal.
UWBLEE PHY is divided to two sub-blocks: I-UWB
PHY and WPT PHY. I-UWB PHY controls the Impulse-
UWB radio interface and WPT PHY controls the Wireless
Power Transfer interface. I-UWB PHY and WPT PHY are
coordinated by UWBLEE PHY so that I-UWB transmission
is synchronized with the WPT transmission.
The function performed by UWBLEE PHY is defined by
UWBLEE MAC, as shown in Table 1.
Figure 5. Active (and semi-passive) UWBLEE MAC states on a mobile
phone. Active states denoted with A, semi-passive with S.
Figure 6. Passive UWBLEE MAC states on a RF memory tag.
TABLE I: UWBLEE PHY IN DIFFERENT MAC STATES
MAC mode
Passive Semi-passive Active
I-UWB Transmit / receive
WPT synch Receive Receive Transmit
WPT power Receive Transmit
Power source WPT
reception
Battery Battery
Remarks Being read / written Reading / writing
other devices
ACTIVE MODE
OFF A-IDLE A-SCAN
A-CONNECT A-CONNECTED
S-CONNECT S-CONNECTED
MOBILE
PHONE
PASSIVE MODE
OFFP-IDLEP-CONNECT
P-CONNECTED
PASSIVE MEMORY TAG

C. Packet-level Communication
The MINAmI subsystem communication between active
mobile reader/writer and passive RF memory tag consists of
periods shown in Fig. 7. In the beginning, there are no tags
within the mobile reader/writer local connectivity coverage.
If the mobile reader/writer detects a tag during the powering
period, it tries to scan all tags available (in the polling peri-
od) and – based on the current selection criteria – choose one
with whom to communicate (in the activation period). The
right tag is found by scanning the coverage area, synchroniz-
ing communications with the tags, and selecting the right tag.
After this selection, connection and device configuration is
executed in the initialization period to set communication
parameters, to specify packet level parameters (e.g., length,
memory allocation). The connection period is initiated when
connection between mobile reader/writer and selected tag is
established. This is followed by the data transmission period,
reading and/or writing selected content from/to tag. After
successful data transmissions, in the termination period,
connection is closed or continued with another read/write
operation to the tag.
Basic connection procedure between a mobile reader and
a tag is described in Fig. 8, which also identifies affected
internal entities, e.g., MINAmI server, memory management,
and communication entity (MAC and PHY layers). For the
air interface, the data from/to the non-volatile storage
memory (PCM) is buffered into a DPRAM buffer memory
equal to the maximum packet size transferred over the air.
D. File System Design
The mobile phone can read tags and with writeable tags
the phone can also write all or parts of their contents. The
communication capacity between the mobile terminal and
the RF memory tag is targeted to exceed 50 Mbit/s (as dis-
cussed in Section I). Plug-in software (External memory
stack in Fig. 1) is required to facilitate seamless use of the
tag memory for mobile phone applications.
The memory tag can be used as an extension to the local
file system of the reader (e.g., mobile phone). The memory
tag can be either a passive and cheap one (Fig. 9) or an active
Figure 7. Mobile reader/writer to RF memory tag communication sequence
Memory Tag
“touch the tag”
Polling period
Activation period
Powering period
Initialization period
Connection period
Data Transfer period
Termination period
Scan
Select
Wakeup
Active/Passive tag
Configuration
Connection establishment
Read from tag / write to tag
Close / Exit / Malfunction
Memory Tag
Figure 8. Basic MINAmI subsystem communication setup sequence
Memory Tag
Memory MAC
SW
Reader/Writer
PHY
Request the
service
Configuration
PHY cmds
Read tag
memory content
Data
transfer
(R/W) PHY cmds
Transfer
completed
PHY cmds
Retransmission
over the air
Retransmission
from buffer
buffer buffer
buffer buffer
Release the memory
MINAmI
Server
GUI
Reader PHY
MAC
SW
Memory
Close the tag
connection
Close the connection
Scan & select
(tag 1, tag X)
Response to
polling and activate
the tag
Read / Write Tag
memory content
Start closing the
connection
Release the memory
Retransmission
from buffer

one, including an own power source and thus being more
expensive (Fig. 10) [11]. Plug-in software in the file system
of the reading device handles the connection to the memory
tag. Storage space on the memory is mounted on the local
file system in the same way as any detachable storage. The
volatile nature of the connection causes overhead in main-
taining the file system view in the reader/writer device. This
kind of RF memory tag would be suitable for e.g., a concert
ticket containing implanted multimedia available to be read
with a mobile device.
Adding a processing element to the memory simplifies
the connection. An ultra low-power processing element can
process the access requests independently and even provide
some more advanced services like metadata-based queries
[34]. A service proxy relays the service interface of the
memory directly to the applications running on the accessing
device. The volatile nature of the connection is not a problem
if the server is made stateless and transactions atomic. This
type of RF memory tag will be able to support more complex
use cases.
Device internal modules need to support NoTA to get full
benefit of subsystem independency and still give a fast con-
nection between subsystems. This interconnect architecture
allows future extensions for modules within one device.
IV. WIRELESS INTERNET ACCESS
Service providers and device manufacturers are continu-
ally challenged to deliver value and convenience to users by,
for example, providing compelling network services. These
services can include selling and distributing content. More
effective and efficient way is needed to distribute the con-
tent. As a complementary solution to address the issue, RF
memory tag based internet kiosk can provide the ultra-fast
and power-efficient connectivity for downloading multime-
dia content, which is beneficial especially when ad-hoc
downloading large amount of content.
The multimedia content such as audio and /or visual con-
tent can be ordered by users and/or pre-downloaded by ser-
vice providers. The content is preloaded into a radio frequen-
cy memory tag installed in the Internet kiosk as shown in
Fig. 11. When a user stays in the range of UWB, a request is
generated for the content stored in the memory tag. The
Internet kiosk initializes wireless transmission to push the
content from the memory tag to the user’s terminal in re-
sponse to the request via UWB [35]. The Internet kiosk can
be deployed in public spaces for users to access ubimedia
applications, e.g. downloading magazines, newspapers or
audio/movie multimedia for recreation in airports.
As discussed in Section III, most of the technologies,
namely Wi-Fi, cellular, Bluetooth and NFC are designed for
well-established communication between active devices,
which result in relatively high power consumption and a long
connection establishment time. In our RF memory tag based
solution, MINAmI subsystem is within mobile devices to
provide necessary connectivity to the Internet kiosk, memory
tag and storage for downloading ubimedia content as illus-
trated in Figs. 3 and 4. The subsystem can be standalone and
operational without maintaining from main application bus
in mobile devices and consuming extra system resources.
Figure 9. File system view of a mobile phone reading a passive memory
tag: a cheap tag without its own processor [11].
Figure 10. File system view of a mobile phone reading a passive memory: a
more expensive tag with its own processor [11].
Figure 11. Mobile phone operating in passive or semi-passive mode to
download data through an Internet Kiosk.
Data flow
Radio transmission
Phone
Tag
Memory
Memory
Power
Data Tag
Memory
Internet
Kiosk
Cache
Internet
Server

Besides, the subsystem can be powered off until a new re-
quest arrives and it would not suffer from long latency in
establishing a new connection.
In this architecture, the mobile devices could operate in
either semi-passive or passive mode instead of active mode.
The downloading is powered by phone battery in semi-
passive mode and it is powered by the Internet kiosk in
passive mode. By running in passive mode, the memory tag
is powered by the Internet kiosk, and the mobile devices
only consume energy to receive content and write these data
into memory. This yields a great potential of energy saving
on mobile devices especially when downloading large
amount of content.
From the communications point-of-view, under the end-
to-end Transmission Control Protocol – Internet Protocol
TCP/IP layer there is a memory connection (MEMCON)
layer connecting the content of the phone tag and the kiosk
tag (see Fig. 12). The memory within the tags is divided into
at least two parts, of which one is outgoing and one incoming
data area. The master device (in the case of the kiosk, the
kiosk tag) reads automatically the data in the outgoing data
area of the slave device (phone tag), copying it to the incom-
ing data area of its own memory. The data is then read by the
controlling software and possible application level com-
mands (e.g., fetch content from web address) found are then
carried out. The reply data (e.g., data fetched from the web
address) is then written to the outgoing data area of the kiosk
tag, which is automatically copied to the incoming data area
of the phone tag. Thus the memory connection layer auto-
matically reads and copies the outgoing data from each de-
vice to the incoming data area of the other device. The upper
layer control software of each device then moves the data
further to the Internet services (Kiosk) or phone memory
(phone) to be used by the application requesting the data.
From the point-of-view of an application on the phone re-
questing data (e.g., Internet message access protocol (IMAP)
email download) from a service provider in the Internet,
there is a TCP/IP connection available.
The solution also makes possible providing an instant
content download possibility by preloading data from an
Internet service to the phone, e.g., a web page, digital maga-
zine, or email account contents. In that case, the kiosk can be
labeled with the logo of the corresponding service, telling the
user what he/she would get by touching the logo with the
phone. As the data is pre-loaded to the tag, the speed of de-
livery is only affected by the speed of the UWBLEE connec-
tion and data handling and displaying within the phone.
A. Comparisons
Maximized throughput and minimized power consump-
tion are critical requirements in order to extend battery life of
mobile devices. Given the scenario of Internet Kiosks, vari-
ous radio technologies could be utilized to provide Internet
connectivity from mobile phones to the Internet Kiosks.
Normally, Wi-Fi and femtocell focus on local coverage.
They are widely deployed and typically used by mobile
devices. However, they are not always the most viable solu-
tions. When traveling abroad, data roaming over cellular
network, such as 3G (WCDMA) may be very expensive.
3G and Wi-Fi typically drain battery quickly on mobile
devices. Since the fixed overhead of transmission is signifi-
cantly high when the radio interfaces are in communication
state. Once the radio interface is on and operates in active
state, most of the power is consumed on circuits and does not
matter how many data are sent or received over the interface.
Especially in 3G networks, the radio switches to the higher
power states, DCH (Dedicated Channel) or FACH (Forward
Access Channel) from IDLE state, when the network is ac-
Phone stack Embedded tag Kiosk tag Internet
Phone
memory
Content
TCP/IP
Buffer Buffer
HTTP
UI/APP
NET
MAC
PHY
NET
MAC
PHY
MEMCON MEMCON
HTTP
OUT
IN
IN
OUT
UWBLEE
TCP/IP
I-UWB
Figure 12. Schematic layout of the communication system when a mobile phone connects to the Internet via a memory-tag-based Internet kiosk.
TABLE II. COMPARISON OF POWER CONSUMPTION OF DIFFERENT
DATA COMMUNICATION TECHNOLOGIES
Wireless
interface
Max
data
rate
Maximum
application
throughput
Power
consumption
Energy
consumption
Mbit/s mW nJ/bit
3G 7.2 ~5 ~850 ~170
IEEE
802.11g
54 ~20 ~500 ~25
Bluetooth
2.0+ EDR
3 ~2.1 ~60 ~28
NFC 6.78 848 kbit/s ~30 ~35
UWBLEE 112 ~50 ~5.4 50-60 pJ/bit (*
*) The value 50-60 pJ/bit is only for the RF front-end of UWBLEE

tive [36]. Based on our measurements on a Nokia N900
phone, IDLE state is considered as low power states, which
consume only around 30 mW. The state of Cell FACH con-
sumes around 400 mW and the state of Cell DCH consumes
around 800 mW. According to 3GPP standard, there are so
called inactivity timers managed by the radio network con-
troller (RNC). The transitions between the different states are
controlled by inactivity timers. Transitioning from the high
to the low power state immediately after a packet is transmit-
ted, the device transitions only when the network has been
inactive for the length of the inactivity timer. This mecha-
nism serves two benefits: 1) it alleviates the delay incurred in
moving to the high power state from the idle state, and 2) it
reduces the signaling overhead incurred due to channel allo-
cation and release during state transitions. Since lingering in
the high power state also consumes more energy, network
operators set the value of the inactivity timer based on this
performance/energy trade-off, with typical values being
several seconds long. However, these timers result in extra
energy consumption even if there is no data to be sent or
received since the radio has to wait for the timers to expire.
The energy consumption is defined as tail energy [36].
To overcome energy consumption constraint in 3G and
WLAN networks, short-range radio communication could be
used. There are several radio technologies that can be con-
sidered for the use of Internet Kiosks. For instance, Blue-
tooth, NFC and UWBLEE where the mobile phone could
operate in semi-passive mode in which communications are
powered by the Internet Kiosk.
In order to reduce power consumption and extend battery
life of mobile phones, battery-operated devices require be-
ing equipped with a radio technology with high bandwidth
and low power consumption. Therefore, the mentioned radio
technologies are taken into consideration of comparison.
The transmission rate and energy consumption of receiving
data over various radios are benchmarked in Table II.
Considering the scenario shown in Section I, which as-
sumes a mobile user downloads 50 MB movie trailer, Fig. 13
demonstrates the time and energy consumption of using
various radio technologies. In the figure, the value of time is
shown in the left y-axis and the value of energy consumption
in the right. In the results of the case in 3G, data rate follows
High-Speed Downlink Packet Access (HSDPA) Category 10
in 3rd
Generation Partnership Project (3GPP) Release 5 and
we assumed inactivity timer lasts 5 seconds. For the
UWBLEE technology the estimated total power consump-
tion for the complete integrated transceiver with digital parts
in passive and semi-passive operating modes is multiplied
with the factor of 2 in comparison to the power consumption
of RF front-end implementation [30]. However, the power
consumption of digital parts is highly dependent on the total
complexity allowed in passive and semi-passive operating
modes, and on the optimized design of integrated circuit. The
energy consumption does not include the part of writing data
into memory storage in all the cases and only shows the
energy consumption of receiving data via different radios.
The power consumption of writing in our NoTA-based solu-
tion is around 2 nJ/bit for NAND flash and approximate 1–2
nJ/bit for PCM. Both of memories are considered competi-
tive from energy efficiency point-of-view [37].
Based on the calculation, the time spent on downloading
the movie trailer is only 8 s and the energy consumption of
RF front-end is 0.043 J when using UWBLEE. As men-
tioned, in the total power consumption for the complete
integrated transceiver the power consumption of digital
parts must be taken into account. In addition, there is a great
difference in the global power consumption of the system in
the passive and semi-passive modes although the power
consumption of the functions in the mobile phone is equal in
the two modes. The reason for this is that in the passive
mode the energy for communication is transferred wireless-
ly, whereas in the semi-passive mode the energy is taken
from the battery. The efficiency of the WPT link, mandatory
in the passive mode, is highly dependent on the factors such
as transfer frequency, size of antennas, and distance, and it
is obviously lower than in battery-powered case. Neverthe-
less, the estimated total energy consumption of downloading
the trailer remains below 0.1 J (and below 1 J with the
memory access) for UWBLEE in the mobile phone. Com-
paring other technologies with UWBLEE, the energy con-
sumption of using 3G is up to 72.25 J, where the tail energy
accounts for 6% of total energy consumption. Moreover,
time consumed on downloading the movie trailer by using
Bluetooth and NFC is up to 190.5 s and 417.7 s respective-
ly. No matter from speed and energy saving point of view,
our RF memory tag based solution would enable shorter
time of downloading ubimedia content, better user experi-
ence, as well as smaller energy consumption.
V. DISCUSSION AND CONCLUSIONS
The RF memory tag (i.e., mobile reader/writer and tag)
solution was developed and tested in the MINAmI project.
Implementation is shown in Fig. 14. The development of a
RF memory tag sub-system of MINAmI project is based on
a flexible, field-programmable gate array (FPGA) based
hardware platform. The sub-system takes benefit from the
ultra-low power UWBLEE transceiver architecture, which is
suitable for data rates required in RF memory tag applica-
tions. The technical results are promising and useful for the
concept of mobile-phone-readable RF memory tags. The
data-rate of 112 Mbit/s has been achieved over the novel
radio interface in technical demonstrations [30]. This leaves
Figure 13. Time and energy consumed of using different radio technologies
when downloading 50MB movie trailer.
0
25
50
75
100
0
50
100
150
200
3G Wi-Fi Bluetooth NFC UWE
Energyconsumption/J
Time/s
Time
Energy
Incl. energy
tail
Up to 471.7 s
Only 86 mJ

room for up to 50% protocol and memory access overhead
when targeting 50 Mbit/s end-to-end communications. On
the PHY and MAC layers short target distance and point-to-
point communication efficiently minimize the protocol
overhead on packet level. However, efficient pipelining in
buffering of the data is in crucial role in optimization of the
end-to-end system. The third important factor is the memory
access speed. This is relevant when reading data from the
source memory and when writing the data to the target stor-
age memory. As shown in Section III, the continuous devel-
opment of NVM memories will provide power-efficient and
fast solutions for the target applications. Altogether, the
listed factors and the results achieved with the demonstra-
tion platform show that mobile reader/writer and the high
capacity memory tag is implementable.
A. Future development
The UWBLEE wireless connection technology presented
in this paper provides data rates significantly exceeding the
existing NFC technology already in the market. From tech-
nology ecosystem point-of-view there is little sense in devel-
oping UWBLEE as an independent technology. UWBLEE
can thus be seen as a possible future high-speed extension to
existing RFID or NFC technologies.
As the range of this wireless interface is fairly short, in
the range of 10 cm, there exist use-cases similar to the NFC
use cases (range touch to 3 cm). Physical selection [38] by
touching of a service-providing tag is thus possible. In such
use, the tag would be marked with a logo of the correspond-
ing service, such as title or picture of a movie or a magazine,
making selection of the service intuitive and easy.
The possibility of using a mobile phone to read a passive
tag is, naturally, not the only operational combination of
these devices, as shown in Fig. 15, where (a) refers to a
phone reading a RF memory tag, (b) to Internet kiosk based
on a RF memory tag, (c) is a variant of (b), and (d) refers to
data transfer between mobile phones. In a multi-device envi-
ronment one device can work as a proxy for the memory tag
and provide other devices with access to its services [11].
There are also possibilities to have memory tags with their
own power sources, which eliminate the need of wireless
powering. In that case, the reading range can be extended or
power use within the mobile phone can be reduced. The
phone can also communicate directly with other similarly
equipped phones.
Our RF memory tag solution supports Nokia’s Explore
and Share concept, a new way of transferring content (e.g.,
multimedia, maps, and applications) to a mobile phone [12].
RF memory tags feed users appetite for ever increasing local
bandwidth and capacity requirements. Users would, natural-
ly, invent new use cases and ways of utilizing these tags in
the local content delivery domain. These Express Tags can
explore new large content shared by others [39]. NoTA is
well positioned in the transport agnostic technology. It fits to
the many inter-device use cases, such as in ubiquitous world.
Our vision is that there is an ever-increasing need to
move content from the Internet to mobile devices and vice
versa, as well as between devices. The amount of available
energy to support all this wireless traffic is not increasing
correspondingly, however. Thus, possibility to distribute the
energy consumption of wireless connections so that either of
the endpoints takes care of most of the power usage is an
interesting enabler to future applications.
B. Conclusions
The evolution of non-volatile memory technologies gives
the basis for the vision about RF memory tags. However, the
large memory creates a need for a high-speed data connec-
tion that can be used to transfer the contents of the tags in a
timeframe acceptable for the user. The dual-band radio inter-
face, UWBLEE introduced in this paper provides the re-
quired data rate and possibility for future scalability as
memory sizes become larger.
Modular architecture is mandatory in the RF memory tag
system to optimize performance. For example, latencies
common in memory access of centralized systems are not
acceptable. Power consumption of the mobile reader/writer is
efficiently minimized with an independent sub-system keep-
ing the involvement of the main processor at the minimum.
In contrast to conventional radio systems, the main processor
only triggers the communication and the independent sub-
Figure 14. Our UWBLEE implementation
Figure 15. Some possible operational combinations of mobile phones
interacting with RF memory tags.
Power transmission Reference clock
Data R/W Powered unit
(a) (b)
(c) (d)

system handles the transfer and storage of the data. Thus, the
main processor does not have to be involved in the low level
communication processes.
In addition to the RF memory tag reader/writer capability
in mobile devices, mobile devices can be also equipped with
embedded RF memory tags. This enables a new usage sce-
nario called internet kiosk which can be further used to ena-
ble internet connection seemingly with zero power consump-
tion in the mobile device.
ACKNOWLEDGMENT
We thank all the partners of MINAmI project, who have
contributed to development of this technology. This study is
a part of MINAmI project under EU FP6 contract IST-
034690. This research was also supported by the Finnish
Academy under grant no 129446.
REFERENCES
[1] I. Jantunen, J. Hämäläinen, T. Korhonen, H. Kaaja, J. Jantunen, and
S. Boldyrev, "System architecture for mobile-phone-readable RF
memory tags". Proc. UBICOMM 2010. IARIA, 2010, pp. 310–316.
[2] V. Wu, M. Montanari, N. Vaidya, and R. Campbell, “Distributed
RFID tag storage infrastructure”. University of Illinois at Urbana-
Champaign, IL, USA. Tech. Rep. 2009.
[3] N. Ahmed, R. Kumar, R.S. French, and U. Ramachandran, ”RF2ID: a
reliable middleware framework for RFID deployment”. Proc. IPDPS
2007. IEEE, 2007, pp 1–10.
[4] C. Ying, “A verification development platform for UHF RFID
reader”. Proc. CMC'09. IEEE, 2009, pp. 358–361.
[5] N. Pillin, N. Joehl, C. Dehollain, and M.J. Declercq, “High data rate
RFID tag/reader architecture using wireless voltage regulation”. Proc.
RFID 2008. IEEE, 2008, pp. 141–149.
[6] J.T.E. McDonnell, J. Waters, W.W. Loh, R. Castle, F. Dickin, H.
Balinsky, and K. Shepherd, “Memory Spot: A Labeling
Technology”. IEEE Pervas. Comput., vol. 9, 2010, pp. 11–17.
[7] J. Nielsen, “Usability engineering”. Morgan Kaufmann, 1993.
[8] W. Cui, P. Ranta, T.A. Brown, and C. Reed, “Wireless video
streaming over UWB”. Proc. ICUWB 2007. IEEE, 2007, pp.933–936.
[9] H. Witschnig, C. Patauner, A. Maier, E. Leitgeb, and D. Rinner,
“High speed RFID lab-scaled prototype at the frequency of 13.56
MHz”. Elektrotechnik & Informationstechnik, vol. 124, 2007, pp.
376–383.
[10] J. Jantunen, I. Oliver, S. Boldyrev, and J. Honkola, ”Agent/space-
based computing and RF memory tag interaction”. Proc. IWRT 2009.
INSTICC PRESS, 2009, pp. 27–38.
[11] E. Kaasinen, M. Niemelä, T. Tuomisto, P. Välkkynen, I. Jantunen, J.
Sierra, M. Santiago, and H. Kaaja, “Ubimedia based on readable and
writable memory tags”. Multimedia Syst., vol. 16, 2010, pp. 57–74.
[12] M. Cooper, “Explore and Share – Nokia shows ultra-fast wireless
data transfer concept”. Nokia Conversations, 23 Feb 2010.
[13] “The PayKiosks Opportunity”. PayKiosks Internet Terminals, 2011.
[14] A. R. Brisebois and R.Klein. “Enterprise femto based kiosk”. US Pat.
App. 20100318417. 16 Dec. 2010.
[15] Bluetooth Kiosk System Project. webee.technion.ac.il
/labs/comnet/Info/projects/winter08/cn02w08 12.1.2012
[16] K. Ok. V. Coskun, M.N. Aydin, and B. Ozdenizci, “Current Benefits
and Future Directions of NFC Services”. Proc. ICEMT 2010. IEEE,
2010, pp. 334–338.
[17] www.fp6-minami.org 12.1.2012
[18] I. Jantunen, H. Laine, P. Huuskonen, D. Trossen, and V. Ermolov,
“Smart sensor architecture for mobile-terminal-centric ambient
intelligence”. Sens. Actuators A, vol. 142, 2008, pp. 352–360.
[19] Y. Têtu, I. Jantunen, B. Gomez, and S. Robinet, “Mobile-phone-
readable 2.45GHz passive digital sensor tag”. Proc. RFID 2009.
IEEE, 2009, pp. 88–94.
[20] K. Kronlöf, S. Kontinen, I. Oliver, and T. Eriksson, “A Method for
Mobile Terminal Platform Architecture Development”. Advances in
Design and Specification Languages for Embedded Systems, S.A.
Huss, ed. Springer, 2007, pp. 285–300.
[21] projects.forum.nokia.com/NoTA/wiki 12.1.2012
[22] K. Keinänen, J. Leino, and J. Suomalainen, “Developing keyboard
service for NoTA”. VTT, Espoo, Finland. Tech. Rep. 2008.
[23] www.mipi.org 12.1.2012
[24] “Open Modem Interface Proposal Based on Device Interconnect
Protocol Version 1.0”. Nokia, Espoo, Finland. White paper, 2010.
[25] D. Truscan, J. Lindqvist, J. Lilius, I. Porres, T. Eriksson, J. Rakkola,
and A. Latva-Aho, “Testable Specifications of NoTA-based Modular
Embedded Systems”. Proc. ECBS 2008. IEEE, 2008, pp. 375–383.
[26] S. Asthana, and D.N. Kalofonos, “The problem of Bluetooth pollution
and accelerating connectivity in Bluetooth ad-hoc networks”. Proc.
PerCom 2005. IEEE, 2005, pp. 200–207.
[27] J. Jantunen, A. Lappeteläinen, J. Arponen, A. Pärssinen, M. Pelissier,
B. Gomez, and J.A. Keignart, “New symmetric transceiver
architecture for pulsed short-range communication”. Proc.
GLOBECOM 2008. IEEE, 2008, pp. 1–5.
[28] S.R. Aedudodla, S. Vijayakumaran, and T.F. Wong, “Timing
acquisition in ultra-wideband communication systems”. IEEE T. Veh.
Technol., vol. 54, 2005, pp. 1570–1583.
[29] M. Pelissier, B. Gomez, G. Masson, S. Dia, M. Gary, J. Jantunen, J.
Arponen, and J. Varteva, “112Mb/s full duplex remotely-powered
impulse-UWB RFID transceiver for wireless NV-memory
applications”. Proc. VLSIC 2010. IEEE, 2010, pp. 25–26.
[30] M. Pelissier, J. Jantunen, B. Gomez, J. Arponen, G. Masson, S. Dia, J.
Varteva, and M. Gary, "A 112 Mb/s Full Duplex Remotely-Powered
Impulse-UWB RFID Transceiver for Wireless NV-Memory
Applications," IEEE J. Solid-St. Circ., vol.46, 2011, pp. 916–927.
[31] G.W. Burr, B.N. Kurdi, J.C. Scott, C.H. Lam, K. Gopalakrishnan, and
R.S. Shenoy, “Overview of candidate device technologies for storage-
class memory”. IBM J. Res. Dev., vol. 52, 2008, pp. 449–464.
[32] “Intel, STMicroelectronics deliver industry's first phase change
memory prototypes”. Physorg.com, 6 Feb 2008.
[33] Z. Shelby, P. Mahonen, J. Riihijärvi, O. Raivio, and P. Huuskonen,
“NanoIP: the zen of embedded networking”. Proc. ICC’03. IEEE,
2003, vol. 2, pp. 1218–1222.
[34] J. Arponen, A. Lappeteläinen, J. Jantunen, and O. Tyrkkö, ”Content
storing device query”. US Pat. App. 20080281787. 13 Nov 2008.
[35] S. Boldyrev, I. Oliver, J. Jantunen, J. Arponen, and S. Balandin.
“Method and apparatus for retrieving content via a service endpoint”.
US Pat. App. 20110055351. 3 Mar 2011.
[36] N. Balasubramanian, A. Balasubramanian, and A. Venkataramani,
“Energy Consumption in Mobile Phones: A Measurement Study and
Implications for Network Applications”. Proc. IMC’09. ACM, 2009,
pp. 280–293.
[37] S. Eilert, M. Leinwander, and G. Crisenza, “Phase Change Memory:
A New Memory Enables New Memory Usage Models”. Proc.
IMW’09. IEEE, 2009, pp. 1–2.
[38] P. Välkkynen, M. Niemelä, and T. Tuomisto, “Evaluating touching
and pointing with a mobile terminal for physical browsing”. Proc.
NordiCHI’06. ACM, 2006, pp. 28–37.
[39] S. Constantinescu, “Nokia Research Center working on high speed
NFC enabled file transfers”. Intomobile.com, 9 June 2011.

System Architecture for High-speed Close-proximity Low-power RF Memory Tags and Wireless Internet Access

More Related Content

What's hot (20)

Similar to System Architecture for High-speed Close-proximity Low-power RF Memory Tags and Wireless Internet Access (20)

More from Sergey Boldyrev (6)

Recently uploaded (20)

System Architecture for High-speed Close-proximity Low-power RF Memory Tags and Wireless Internet Access