Journal of Global Positioning Systems (2004)
Vol. 3, No. 1-2: 32-39
Using RFID for Accurate Positioning
Hae Don Chon , Sibum Jun, Heej ae J ung, Sang Wo n A n
Samsung Electronics Co. , LTD
Received: 15 Nov 2004 / Accepted: 3 Feb 2005
Abstract. In Korea, telematics is regarded as the
technology to enhance and make everyday-driving
experience more comfortable and safer. An essential part
of the telematics is navigation and it is mainly based on
GPS as the choice of positioning technology. The
accuracy of GPS, however, is approximately ten to
twenty meters. Combining with map-matching
technologies, most navigation systems guide drivers with
a best effort manner. In addition to telematics, RFID
(Radio Frequency Identification) is an old but newly
emerged technology. In this paper, we incorporate RFID
technology into a navigation system to improve the
accuracy. The skeleton of the idea is as follows: install
RFID tags on roads in a certain way, store very accurate
location information along with other necessary
information in the tags, add an RFID reader module to the
navigation system, and use this new location information
along with GPS and a gyroscope to produce highly
accurate location information. With this scheme, the
accuracy of positioning can be dramatically improved,
especially in tunnels and in downtown areas. Preliminary
results show that this idea is feasible.
Key words: Positioning, RFID, Navigation, Telematics.
1. INTRODUCTI ON
Telematics can be defined as the technology that enables
a vehicle to exchange data via wireless communication in
the form of services or applications. Core components of
a telematics system include positioning technology,
Human-Machine Interfaces (HMI), and a navigation
module. Over the last few years, telematics has drawn
much attention in major indu strial countries. Korea is one
of the nations that end eavor to capitalize on telematics. In
fact, the Ministry of Information and Communication of
Korea (MIC) has recently released its master plan called
the IT839 Strategy to help the IT industry develop new
technologies (MIC, 2004). The IT839 Strategy is a plan
for activating eight services (WiBro, DMB, Home
Network, Telematics, RFID, W-CDMA, Terrestrial
Digital TV, VoIP), implementing three infrastructures
(Broadband Convergence Network, Ubiquitous Sensor
Network, IPv6), and developing nine new growth engines
(Next-Generation Mobile Communications, Digital TV,
Home Network, IT SoC, Next-Generation PC, Embedded
SW, Digital Contents, Telematics, Intelligent Service
Robot). Telematics is one of the eight new services and
also one of the nine new growth engines in the IT839
Strategy.
As mentioned above, positioning technology is a core
component of telematics and most of telematics devices
sold today rely on GPS as a choice of the positioning
technology. With GPS, it is inherently difficult, if not
impossible, to get a position in a tunnel or in downtown
areas surrounded by skyscrapers. This issue in GPS
would not be resolved for many years to come. To
enhance the accuracy to a level at which lane-by-lane
route guidance can be possible, using RFID technology
for positioning is proposed.
RFID is not a new technology, rather its origin da tes back
to World War Two. Recently, RFID comes under the
spotlight for its potential for changing the world by
replacing barcodes and providing many services that are
unknown today. In addition to the replacement of
barcodes, RFID has gaining interests in mobile phone and
consumer electronics industry (www.nfc-forum.org).
The RFID positioning can be divided into four steps: in
the first step, install RFID tags on ro ads in a certain way,
store very accurate location information along with other
necessary information to the tags, add an RFID reader
module to the navigation system, and use this new
location information. Apart from the RFID system, we
also propose to use a tag database.
Due to the memory constraint on the tag and the data size
that needs be written in a tag, the use of a database for

Chon et al.: Using RFID for Accurate Positioning 33
tags is a necessary condition . In addition , the speed of th e
RFID communication also makes the use of the tag
database indispensable.
Preliminary results show that the RFID communication
speed does not solely depend on the bit data rate in the
specification. Nonetheless, it will be shown that
retrieving ID from a tag can be done fast even at a high
velocity (by the high velocity we mean 150km/h).
Another performance study on the tag database access
time indicates that the access time is marginal.
In the RFID and telematics literature, there is not much
research on RFID positioning pertinent to vehicle
navigation. Kubitz et al. developed a technique for robot
navigation using RFID (Kubitz et al., 1997). They,
however, did not take the velocity of robots into account,
whereas in vehicle navigation fast RFID communication
would be crucial. Penttilä et al. presented a performance
study on an RFID system, in which they achiev ed reliable
identification accuracy at up to 40 km/h (Penttilä et al.
2004).
This paper is organized as follows: in Section 2 a brief
overview of the RFID technology is given as background
information. In Section 3, the idea of RFID po sitioning is
described along with the tag database. The results of
performance study are provided in Section 4. Conclusions
and future work are described in Section 5.
2. RFID BACKGROUND
In this section a brief overview of RFID technology in
general is given. An RFID system consists of tags, a
reader with an antenna, and software such as a driver and
middleware. The main function of the RFID system is to
retrieve information (ID) from a tag (also known as a
transponder). Tags are usually affixed to objects such as
goods or animals so that it becomes possible to locate
where the goods and animals are without line-of-sight. A
tag can include additional information other than the ID,
which opens up opportunities to new application areas.
An RFID reader together with an antenna reads (or
interrogates) the tags. An antenna is sometimes treated as
a separate part of an RFID system. It is, however, more
appropriate to consider it as an integral feature in both
readers and tags since it is essential for communication
between them. There are two methods to communicate
between readers and tags; inductive coupling and
electromagnetic waves. In the former case, the antenna
coil of the reader induces a magnetic field in the antenna
coil of the tag. The tag then uses the induced field energy
to communicate data back to the reader. Due to this
reason inductive coupling only applies in a few tens of
centimeter communication. In the latter case, the reader
radiates the energy in the form of electromagnetic waves.
Some portion of th e energy is absorbed by the tag to tu rn
on the tag’s circuit. After the tag wake up, some of the
energy is reflected back to the reader. The reflected
energy can be modulated to transfer the data contained in
the tag.
Three frequency ranges are generally used for RFID
systems: low (100~500 kHz), intermediate (1 0~15 MHz),
and high (850~950 MHz, 2.4~5.8 GHz). Detailed
characteristics of these three frequency ranges along with
examples of major applications can be found in RFID
Handbook (Finkenzeller, 2002). The communication
range in an RFID system is mainly determined by the
output power of the reader to communicate with the tags.
The field from an antenna extends into the space and its
strength diminishes with respect to the distance to tags.
The antenna design determines the shape of the field so
that the range is also influenced by the beam pattern
between the tag and antenna. Although it is possible to
choose power levels for different applications, it is
usually not allowed to have complete freedom of choice
due to legislative constraints on power levels as in the
case of the restrictions on carrier frequencies. More
interested readers are referred to RFID Handbook
(Finkenzeller, 2002) and an M.S. thesis from MIT
(Scharfeld, 2001).
3. RFID POSITIONING
As mentioned in Section 1, we propose to use RFID
technology for positioning. This technique, however,
would not replace GPS rather it is a complementary
technique. In this section, we give an overview of the
RFID positioning and describe the feasibility of the idea.
3.1 Overview
The RFID positioning can be explained as follows: First,
RFID tags need to be installed on a road in a manner
which could maximize the coverage and the accuracy of
positioning. This dep loyment scheme is discussed later in
this subsection. Upon installatio n, necessary information
such as coordinates of the location where the tag is
installed needs to be written on each tag. The accuracy of
this position information is v ery critical for this technique
to be successful. The position information can be
acquired by using DGPS or some other methods, which
would take much longer time to compute the location.
Contrary to GPS in navigation systems where real time
positioning is necessary, the time for getting the accurate
information would be tolerated since this computation
would take place once.
Vehicles, then, need to be equipped with an RFID reader
that can communicate with the tags on a road. No matter

34 Journal of Global Positioning Systems
how accurate the RFID positioning is, it only gives the
position where the tags are. Therefore the vehicles need
also to be equipped with a GPS receiver and inertial
sensors such as a gyroscope for positioning when there
are no tags around. While driving, the vehicles constantly
monitor the presence of a tag. On detection, the reader
retrieves the information from the tag including
coordinates of the location, which are supposedly very
accurate information. Figure 1 shows a scheme of
deploying RFID tags on a road. In the figure, circles (blue
dots) on each lane represent a tag and the tag itself is
enclosed in a special purpose cat’s eye. A cat’s eye is a
device built into a sturdy housing and placed on a road as
Figure 1. RFID Tags on a Road
a lane marker.
The deployment should be done step by step: places such
as tunnels from which getting GPS signal is not an option
should be the first, intersections the next, urban areas, and
then nationwide. Due to this nationwide scale,
governmental actions are necessary. As part of the IT839
Strategy, the government of Korea funds a research
project called Development of Telematics Test Bed being
done by Electronics and Telecommunications Research
Institute (Lee, 2004). Among other things, the research
includes the development of RFID positioning.
3.2 Feasibility
In this subsection, the issues of feasibility of the RFID
positioning are discussed. There has been work on using
RFID for autonomous robots (Kubitz et al., 1997).
Contrary to the robo t case, an RFID tag in our application
will be installed on a road, where the operation
environment for the tag is very harsh; high temperature in
summer, low in winter, dusts, rain, snow, etc.
Furthermore, vehicles equipped with an RFID re ader that
is compatible to the tags on the roads can move very fast;
though speeding is illega l, some cars (Porsche) can go as
fast as 300 km/h. To be more useful the tags should
contain the information about the road property (number
of lanes, which lane it is on, how far to the nearest
intersection, etc) other than the location information.
More data decreases the communication speed and
requires more memory, which leads to high cost.
In summary, there are issues to be addressed before full-
fledged deployment of RFID tags nationwide. The first
issue is making RFID tags that can withstand a harsh
environment. The second one is fast communication
speed between readers and tags. The third one is the data
size. Besides these three issues, there are other issues
such as standards (frequency, data format, air interface,
etc), which are not covered in this paper. Although it is
very important, the environmental problem is beyond the
scope of this paper. We only address the communication
speed and the data size issues in this and the next section.
3.3 Tag Database
While there would be location information in a tag, it
would be almost impossible to embed all the necessary
information in a tag due to memory constraints and the
dynamic nature of some information. Information such as
absolute coordinates of the location will not be changed.
On the contrary, relative coordinates and the property of
the road on which the tag is could change some time
(unlikely, though). Moreover, we can embed more useful
information such as nearest museums, restraints, and gas
stations. However, the contents of the information vary
all the time. The data size as well as the dynamic nature
of it prevents from writing all the information at the
installation time. To address this issue, we devise a tag
database which stores information corresponding to the
tags available on the roads in a region (country for
instance). The information stored in the tag database is
whatever information on real tags and more such as point
of interests.
Another reason for the necessity of the tag database
comes from the speed of the RFID communication.
Although we do everything we could to speed up the
communication, it may not be fast enough to get all the
information from a tag while driving at, for instance,
150km/h. As we show in Section 4, however, getting only
identification (ID) is very feasible even at such a high
velocity. Once the ID is retrieved, it can be efficiently

Chon et al.: Using RFID for Accurate Positioning 35
searched the tag database and extracted whatever
information necessary.
The tag database is a collection of tags and a part of the
digital map that a navigation system may carry.
Generally, a digital map consists of cells each of which
contains network information for route guidance. The
network information is a graph with nodes and links.
Figure 2 shows a class diagram of the digital map. In the
diagram, TagDB is an aggregation of Tag objects which
represent tags in a real world.
Each cell has links to the collections of nodes, links, and
Figure 2. Tag Database Class Dia gr am
tags. For simplicity, we only show the attributes of the
Tag object. As in the diagram, a Tag object includes ID,
absolute coordinates X and Y, relative coordinates RX
and RY, link ID where the tag is, and the property field.
This last field is for the number of lanes of the link, type
of the road (highway, local, etc), and so on. In Java
language and most of other programming languages, type
long is 8 bytes, type float and int are 4 bytes, and type
short is 2 bytes. Summing up, the data size of a tag is 30
bytes. Therefore even with a million tags on the roads,
thereby in the database, the size of the tag database is
approximately 30MB. Since more and more embedded
devices have large flash memories and even gigabytes of
hard drives, the sheer size of the tag database would not
be a big issue.
4. PERFORMANCE
We are in a very early stage of a research project and
constructing right test equipments and environment.
Nevertheless, preliminary results of experiments on
communication speed and tag database access time are
provided in this section. Authors of this paper did their
best to produce accurate results.
4.1 Communication Spe ed
Since there were no RFID tags and readers for the
purpose of RFID positioning on a road, we have used a
commercial one, KIS900RE from Kiscomm Co.
(www.kiscomm.co.kr). Table 1 shows the specification of
KIS900RE.
According to the specification, the bit data rate is 64
Kbps which would be fast enough to communicate
several hundred bits at a high velocity. Figure 3 shows
the distance needed to read 64 bits at various data rates
with respect to vehicle velocities. In the figure, the data
rate is assumed to be the sole factor of RFID
communication. Only the distance for 64 bits is drawn
since the data size of a tag ID is 8 bytes in our program.
The purpose of the experiments explained in this section
is to measure the time that actually takes to complete a
reading transaction.
For the experiment, the radiation pattern of the antenna of
KIS900RE is measured in a Satimo chamber (See Figure
4). Based on the radiation pattern, it can be assumed that
the reliable reading angle from the center of the antenna
is roughly 68 ° (assuming 3dB beam width).
The setting of the experiment is as follows: first a tag is
dropped (freefall) at the height of 13 meters. Since the
Kiscomm tag is very light and soft, it is attached a piece
of paper that is hard enough. To make air resistance

36 Journal of Global Positioning Systems
marginal heavy coins are attached to the tag. The reader
antenna is then placed at the height of one meter so that
the distance between the center point of the antenna and
the falling point is 12 meters. By the law of energy
conservation, the velocity of the tag at the center of the
antenna is 15.33 m/s or 55.21 km/h. See Figure 5 for the
experiment setting.
2
2
1
8.9 mvmh = Law of Energy Conservation
Frequency 908.5 – 914 MHz
(14 channels, 250 kHz spacing Frequency Hope)
Radio Power Output 1W Conductive (Korean Band)
Read Range 6 meters (Typically read range depends on Reader
environment and the used tag)
Operating Temperature Range -10 deg to +50 C
Reader
Humidity 5 to 95 % non-condensing
Tag ID 128 bits
Tag Bit Data Rate 64 Kbit/s
Gain 6 dBi
Antenna In Band VSWR < 1, 2
Table 1. KIS900RE Hardware specification
0
5
10
15
20
25
5060708090100 120140 160 180 200
Speed (km/h)
Distance
(cm)
16 Kbps
32 Kbps
64 Kbps
Figure 3. Distance for reading 64 bit
When passing the read range, the distance between the
tag and the antenna is about 60cm. Considering the 68° of
reliable reading angle, the distance from the entrance of
the read range to the exit of the read range is 81cm
(°××= 34tan22 dR ). In the experiments, the reader
(KIS900RE) was able to read a tag either two or three
times in a freefall. Suppose that the tag passed the center
line of the circle as in Figure 5 when it was read three
times. Then the distance needed to complete one reading
transaction can be computed as 27cm. At this time, the
tag was moving at a velocity of 55km/h and the time
actually taken for one reading transaction is 18
milliseconds. In the unit of bps, the reading speed is
about 7.11Kbps since the data size of the tag is 128 bits.
Note that this is very different from 64Kbps of the
specification. Generally, RFID system has three phases of
operation. The first is charge-up phase where the reader
transmits energy and the tag absorbs and stores it until the
internal supply voltage is high enough to turn on the
circuitry. The second is communication phase which can
be either inductive or back scatter coupling. The last is
discharge phase where the whole circuit of the tag is
reset and the supply capacitor is discharged. Therefore a
possible explanation for the big difference is that 64 Kbps
is measured only for communication phase and the
charge-up and discharge phases take much more time
than the communication phase.

Chon et al.: Using RFID for Accurate Positioning 37
Layer Max value Position Min value Position BeamWidth Average
908.5(MHz) 6.43 -2.86 deg -22.36 174.28 deg 71.00 deg - 0.22
911(MHz) 6.31 -2.86 deg -23.08 174.26 deg 69.75 deg -0.41
914(MHz) 6.24 -2.86 deg -23.50 171.45 deg 69.13 deg -0.50
Figure 4. Radiation pattern of the antenna of KIS900RE
Speed: 55km/h
81cm
1m
60cm
RFID
Antenna
27cm
Three reading
transactions
Tag
12m
68 ˚
Speed: 55km/h
81cm
1m
60cm
RFID
Antenna
27cm
Three reading
transactions
Tag
12m
68 ˚
Figure 5. Experimentation Setting
The result of the experiments shows that the bit data rate
is not the only contributor to the RFID communication
speed. In fact, if the data rate was the sole factor in our
experiment, then it would have been just 3cm for a
reading transaction. As it turned out, however, it took
nine times longer than that. This means that if we were to
improve the communication speed, we need to focus on
somewhere else such as capacitor charging time rather
than the data rate. Nonetheless, it can be inferred from the
experiment that it could need 81cm to read 128 bits at the
velocity of 165k m/h. Notice that the read er we tested and
most of others have read ranges longer than 81cm.
Therefore it is feasible to read an RFID tag at a high
velocity up to 165km/h. However, this does not
automatically mean that RFID readers can read tags on a
road at a high speed. There are issues to address such as
making tags withstand harsh conditio n. As we mentioned

38 Journal of Global Positioning Systems
earlier, this is a preliminary result and we intend to
construct test equipments and environment for more
accurate experiments.
4.2 Tag Database Access Time
In this section, we provide a result of the experiment on
the tag database access time. The necessity of the tag
database is explained in Section 3. What our main
concern here is how fast the information corresponding to
a tag in the tag database can be accessed. Therefore we
concentrate on the tag database out of three databases in
the map shown in Figure 2.
Given its ID, the time to extract position information of a
tag was measured using a Java program. Given the
number of tags (dbsize) in a cell, IDs were randomly
generated and stored in the tag database. For the
performance and memory constraint reasons that are
normal on embedded devices such as telematics
terminals,
0
10
20
30
40
50
60
70
80
135791113151719Data size
(
K
)
Total5200
PC
Time
(μs)
Figure 6. Average Access Time for a Tag
a sorted array was chosen as the data structure of the tag
database. The key of th e sorting was the ID of tags. After
construction of the tag database, some number (idsize =
10000) of IDs were also randomly generated and stored
in a separate array for the fast execution. The time to
extract idsize tags was then measured and the average
was taken. When searching the database, the binary
search algorithm was used. Given dbsize, this execution
was repeated twenty times and dbsize varies from 1000 to
20000.
The program was run on a Total5200 development
platform. Total5200 is a test platform for telematics from
Freescale Semiconductor, Inc (www.freescale.com). It
consists of 400MHz MPC5200, 64 MB SDRAM, 64MB
Flash memory, J9 (IBM’s Java Virtual Machine), and
more (the detailed specification can be found on their
website). For the comparison purpose, the program was
also run on a PC with Intel Pentium4 2.8 GHz and
512MB. The results are shown in Figure 6.
Notice that the running time increases as the data size
increases, which was expected. Notice also that the
running time on a PC is less than 3.5 microseconds and
the running time on Total5200 is less than 70
microseconds regardless of the data size. The 70
microseconds is a negligible time given that 18
milliseconds are for a reading transaction in the
experiment. Therefore we can conclude that once a tag ID
is acquired, extracting necessary information from a
database would not take any meaningful time.
5. CONCLUSIONS and FU T UR E WORK
In this paper, the idea of using RFID technology for
positioning was proposed. Preliminary results on RFID
communication speed and tag database access time were
shown. For the communication speed, the experiment
showed that 18 milliseconds were needed to complete a
reading transaction. W ith the result, it was computed that
the communication speed is 7.11Kbps as opposed to
64Kbps in the specification. It can be concluded that it is
feasible to retrieve bits for ID of a tag at high velocities
provided that operational tags in harsh conditions are
installed. We also argued that a database for information
corresponding to tags on roads is necessary. The
performance study on the tag database showed that the
access time is insignificant compared to a reading
transaction. Therefore once an ID is retrieved by the
RFID reader, getting the necessary information from the
database does no t pose any probl em.

Chon et al.: Using RFID for Accurate Positioning 39
Although it is our conclusion that the idea of RFID
positioning is feasible, there are many issues to resolve
before any full-fledged deployment. Some of the issues
are (1) making a tag to withstand harsh conditions, (2)
testing communication speed at up to 150 km/h, (3)
developing a scheme to combine an RFID reader, a GPS
receiver, and a gyroscope to produce a consistent and
accurate position information, and (4) developing a tag
database management algorithm and program for future
use. As future work, these issues will be investigated.
ACKNOWLEDGEMENTS:
We give our thanks to Yunkyoung Ko for his cooperation
in measuring the radiation pattern of the RFID antenna.
REFERENCES
Finkenzeller K, RFID Handbook: Fundamentals and
Applcations in Contactless Smart cards and Identification
(2002)
Kiscomm co., LTD, KIS900RE View, www .kiscomm.co.kr
Kubitz O, Berger M, Perlick M and Dumoulin R (1997)
Application of Radio Frequency Identification Devices to
Support Navigation of Autonomous Mobile Robots, IEEE
47th Vehicular Technology Conference, 126-130
Lee J (2004) Development of Telematics Test Bed, TTA
Telematics Standards & Technology Seminar
(www.tta.or.kr, in Korean)
Ministry of Information and Communication (2004) IT 839
Strategy Master Plan, www.mic.go.kr
Penttilä K, Sydänheimo L, and Kivikoski M (2004)
Performance development of a high-speed automatic
object identification using passive RFID technology,
Proceedings of the 2004 IEEE International Conference on
Robotics & Automation, 4864-4868
Scharfeld T (2001) An Analysis of the Fundamental
Constraints on Low Cost Passive Radio Frequency
Identification System Design, MIT M.S. Thesis
Siegemund F and Flörkemeier C (2003) Interaction in
Pervasive Computing Settings using Bluetooth-Enabled
Active Tags and Passive RFID Technology together with
Mobile Phones, Proceedings of the First IEEE
International Conference on Pervasive Computing and
Communications, 378-387