Proceedings of the 2009 IEEE 9th Malaysia International Conference on Communications

15 -17 December 2009 Kuala Lumpur Malaysia

Vehicular Communication System for Vehicle

Safety Using RFID
Abdullah Al Masud1, Md. Nazrul Islam Mondal2, Kazi M. Ahmed3
1
Department of Computer & Communication Engineering, International Islamic University Chittagong
Chittagong, Bangladesh
2
Department of Computer Science & Engineering, Rajshahi University of Engineering & Technology
Rajshahi, Bangladesh
3
Department of Telecommunications, Asian Institute of Technology
Bangkok, Thailand
1
iammasud@yahoo.com
2
nimbd@yahoo.com
3
kahmed@ait.ac.th

Abstract— Vehicular communication system is a key part of
intelligent transportation system (ITS) while vehicle safety
communication (VSC) is a major target of vehicular
communication. The use of radio frequency identification
(RFID) system for vehicular communication has been proposed
for pedestrian-safety. In this paper, an extended RFID system
and infrastructure for vehicle safety communication through co-
operative routing of information on vehicles’ sudden motion and
direction changes and warning messages for post-accident
scenarios is proposed. It also provides a demonstration on the
structure of warning codes and flow of information within the
system and the vehicular ad-hoc network (VANET) for the
vehicle safety. This paper also includes the simulation results for
line-of-sight (LOS) communication and a special non-line-of-
sight communication (NLOS) to observe the co-operative
distance covered by co-operative routing and the related bit
error rates (BERs). Doppler shift effect is also considered in the
simulation.
Keywords— RFID, Vehicular Communication System, V2V,
V2R, Co-operative Routing, VANET
I. INTRODUCTION
T O enhance the vehicle safety, recent wireless
communication standards are deployed in vehicular
communication systems. Wi-Fi [1], Bluetooth [2], radio
frequency identification (RFID) [3] and dedicated short range
communication (DSRC) [4] are suitable for vehicular
communication and already deployed in several countries.
Wi-Fi is mostly used for internet access, entertainment
content download through an access point while the vehicle is
stationary or parked at some places. In case of multipath
propagation, Wi-Fi may work suitably for high packet error
rate as demonstrated in [5]. Bluetooth is used for a very short
range communication and mostly used in in-vehicle
communications for entertainment and some vehicle
equipment controlling purposes.
RFID is already being used for some vehicular
communication services [3]. There are several RFID products
in the market for vehicle identification at higher speeds even
at 200 km/h [6].
DSRC is mainly based on IEEE 802.11a and it still needs
some kind of standardization [7] and market products should
also be available to get its benefit.
RFID is chosen for enabling vehicular communication to
achieve vehicle safety as it is a common market product with
application potential and extendibility. It is a cheaper
technology also. Another important reason to use RFID is that
the pedestrian-safety scheme described in [8] which uses
active RFID tags in Ultra High Frequency (UHF) range is
extended here. In [8], authors illustrate that tags carried by the
pedestrians and wheelchairs can be used for providing nearer
vehicles information on the pedestrians’ movement at the
intersection point of any road. Some Low Frequency (LF)
signal generators positioned at the roadsides are used to
activate the tags within its activation range. Activated tags
transmit the tag ID and the carrying persons’ positional data
generated by the global positioning system (GPS) component
interfaced with the tags. A tag reader/repeater is placed in the
roadside unit (RSU) to gather tag information and then to
broadcast it to the approaching vehicles. The information is
displayed on the on-board display unit of the vehicles. Thus
pedestrian-involved accidents are avoided utilizing the
provided information. Tags can transmit over several tens of
meters. Reader/repeaters route this information co-operatively
over a distance of 200 to 300 m which is considered feasible.
Our extension to the above system also includes vehicular
information. So, tags are needed in the vehicles also. Vehicles
carry a tag reader as stated in [8]. So, both the tag and the
reader are considered in the on-board unit (OBU) to enable
inter-vehicle communication as well as intra-vehicle
communication.
Chapters are organized as follows: System models of the
proposed OBU and RSU are given in Section II. Section III
contains the flow and structure of warning messages. Section

978-1-4244-5532-4/09/$26.00 ©2009 IEEE 697

IV contains the parameters of simulation. Results of the
simulation and the corresponding discussion are given in
Section V. Section VI concludes the paper.
II. SYSTEM MODELS FOR OBU AND RSU
A. OBU System Model
For V2V communication, each vehicle is equipped with an
active 950 MHz RFID tag along with a tag reader/repeater
interfaced with the in-vehicle storage and display system. The
tag is connected to the GPS module to transmit positional
value of the vehicle. At the same time, the tag is also
interfaced with the in-vehicle monitoring system (IVMS).
This interface should be able to instruct the tag to transmit a
warning code relevant to any particular warning situation and
the tag should be able to transmit data whenever there are data
to be sent. The whole system for an OBU is shown in Fig. 1.
Fig. 1. The OBU system
The interface along with the tag or IVMS is not shown as a
separate module as it is assumed that it is embedded in either
a tag or an IVMS. The interface is a major feature of the OBU
in the proposed system. It is also assumed that the interface
can be produced by the tag manufacturers or IVMS designers.
Use of both the tag and the reader in the OBU is to make
room for future extension of the proposed system for
applications such as vehicle security (anti-theft measure),
login to some access points etc. which are under investigation.
B. RSU System Model
For R2V communication, the RSU contains only the tag
reader/repeater and it is connected to the central server. The
connection provides a way for central database management.
This link is to be suitable for both-way communication. The
system model for the RSU is shown in Fig. 2.
Fig. 2. The RSU system
III. DATA FLOW FOR THE VEHICLE SAFETY AND THE
STRUCTURE OF THE WARNING CODE
For vehicle safety, message transmission is considered for
two types of situations. The first situation is for the case when
the vehicles are taking left or right turn at intersection points
or turning points. And the second situation considers that
698
there is an accident occurred at any point of the road and the
OBU of the crashed vehicle is transmitting its accident
information. The information from any of the above cases is
considered as a warning message. The followings are the brief
on how messages flow in a particular case and the structure of
the warning code is also shown here.
A. Vehicle and Pedestrian Information Flow for Avoiding
Accident
Pedestrian information flow at intersection point is already
defined and described in [8]. So, for this option, pedestrians
should carry tags and LF signal generators should be present
in the roadside to activate pedestrian’s tags. But the proposed
system is also targeted to transmit the vehicles’ warning
messages when they are taking turn. So, tag in a vehicle
transmits a warning message when the IVMS senses a turn-
taking activity in the vehicle. The message consists of a
warning code and the positional data of the transmitter. The
warning code is actually a short code which represents the
warning information to be displayed on the display, such as ‘a
left turn is being taken’ or ‘a hard brake situation is happened’
etc. Fig. 3 shows the flow of information for the V2V
communication.
Key:
IVMS = in-vehicle monitoring system
wc = warning code
A(x,y) = position of A
Fig. 3. Information flow in V2V communication
B. Information Flow in Post-accident Message Routing
Post-accident warning message routing is proposed in this
proposed system to avoid a mass crash which may occur due
to hard brake or a sudden track change and such other
activities from the first vehicle observing an accident in its
front. In such a case the vehicles following the first observer
vehicle may not be aware of its abnormal actions. The flow of
information in this case is the same as the one shown in Fig.3.
Just the warning code should be relevant to the situation. An
IVMS is assumed to be capable of sensing all these situations.
It is to be noted that in post-accident case, the first observer
vehicle routes the accident information as well as its own
abnormal actions (e.g. sudden brake or track change). Again,
for suitable IVMS device, the crashed vehicle may also
transmit warning messages.
 C. Structure of the Warning Code TABLE I
SIMULATION PARAMETERS
In this system, warning codes are used against a database of
warning information. Codes are defined based on a tree-based Sample Road
structure. The architecture of the warning code tree is Road structure Divided into several phases
described below. for a lane and the road
If there are messages related to Road, Road-side, Human consists of 2 lanes in a
particular direction
and Vehicle, a 2-bit code can be defined for each of these four Length 1 km (max)
types. If there are N types of objects (e.g. animal, trees) for Lane Width 10 m (min)
which messages to be generated, each object is represented No of lanes 1-2 (varying for scenarios)
Vehicle parameter
with a log2(N)-bit code. For the four types stated here, the No. of vehicles 15 (Scenario-1) & 30
codes are shown in Fig.4. (Scenario-2 & 3)
Maximum speed of 60 km/h
the vehicles
RFID tag,
transponder and
communication
parameters
Frequency 950 MHz
Bandwidth 100 kHz
RF link range 150 m
Tx output power 1 W (max.)
Fig. 4. Tree-based structure of the warning code Tx output gain 6 dBi
Source data rate 100 kbps
Modulation in tag BPSK
For a pedestrian, the warning code for the ‘human’ is to be SNR at 150 m 8 dB
stored in the memory of the tag carried by him and the Transmission data
Warning ID may represent his movement along the road. Length of 50000 bits (needed for getting
transmission data BER for more SNR)
Actually, for a pedestrian, his position is the main information Propagation
and the Warning ID is just to match particular information to through channel
be displayed in the vehicle. For the Road information, there Propagation/ LOS - Free space and two-ray
Channel model propagation(direct and
should be a Road ID for every road in a country. The Warning ground-reflected rays) and
ID corresponds to a specific information on the road such as direct link (with directional
“Road is blocked by an obstacle” or “Road is damaged”. If antenna) with AWGN
NLOS- Rayleigh
the OBU or RSU is extended with a road sensor, the road Doppler shift May be or may not be
information can be generated. RSU may decode the message effective
storing information in its internal memories or getting
connected with the main server where a database is framework [9]. The simulation model and the scenarios are
maintained. demonstrated below:
The other types of objects have the same types of ID (e.g. 1) Scenario-1: In scenario-1, there are 15 vehicles along a
Road-side ID, Vehicle ID) and a Warning ID (related to a road with 4 turns (4 phases). A phase is a straight road
particular warning message). Depending on the standards without any turn. The road forms a rectangular structure and
followed for the messages and the messages defined for the only one turn is possible from a phase. In all cases, right turn
warning scenarios, code lengths may vary. Using the tree is allowed in any direction along the road. There is only one
architecture speeds up the search mechanism which helps in lane in the road. In a lane, a vehicle moves towards a
reducing the processing delay of the warning message for particular direction. The road is a closed loop and it does not
displaying it on the on-board display unit. affect this paper’s simulation work as it is only for a small
portion of the whole time of vehicle movement in the
IV. SCENARIOS AND PARAMETERS FOR SIMULATION scenario. Fig. 5 shows the road of scenario-1. The whole area
of the scenario is 1000 m × 1000 m. Each of the phases is 10
A. Parameters for the Simulation
m wide. The inner rectangle shows an open space with 980 m
Table I shows the parameters with their respective values to × 980 m area. Vehicles are represented by black circles
be used in the simulation. Power and gain-related values for within a rectangle.
the tags and readers are determined according to Japan’s 2) Scenario-2: In scenario-2, there are 30 vehicles along
Radio Law [10], [11]. almost the same road structure with 4 turns for 4 phases. This
B. Scenario Files for the Simulation road also forms a rectangular structure. But there are two
Three different scenarios are used for the simulation lanes in the road. In the two lanes, vehicles move opposite to
purposes. Scenarios are generated by the IMPORTANT each other and a vehicle cannot change a lane. Fig. 6 shows
the road of scenario-2. The whole area of the scenario is 1000

699
m × 1000 m. Each of the lanes is 10 m wide. The inner
rectangle shows an open space with 960 m × 960 m area.
Fig. 5. Simulation scenario-1
Fig. 6. Simulation scenario-2
3) Scenario-3: In scenario-3, there are also 30 vehicles
along a different road structure with 4 turns for 4 phases. Fig.
7 depicts scenario-3. It has two lanes with a width of 10m.
The rectangle surrounding the scenario is 100m × 100m in
area. The lengths of the lanes are not important in simulation.
Fig. 7. Simulation scenario-3
700
V. RESULTS AND DISCUSSIONS
In the result, the cumulative BER means the BER for the
last vehicle that takes part in the co-operative routing of the
warning message. It is considered here that the received bits
are redirected by each vehicle without any change. The BER
values and the co-operative distances calculated here gives an
average of the BER and co-operative distance values for six
consecutive transmissions of 50000 random bits at an interval
of 100 µs. The six samples are considered as enough for the
averaging. For more samples simulation time may become too
high. The results in the simulation may vary for the
transmission of higher bits. 50000 bits are used in the
simulation as lower amount of bits in transmission may cause
insufficient BER results in some links. Average SNR and
average BER are taken to get the average value of the
performance. It is assumed that taking average does not lead
to an error as the average SNR vs. average BER values match
the theoretical values.
Here the term ‘co-operative distance’ for a particular
vehicle refers to the distance covered by the co-operative
routing of a warning message generated by that vehicle.
Cooperative distance 0 means there are no neighboring
vehicles (vehicles within 150 m distance) for the transmitting
vehicle. So, for the vehicles with nonzero co-operative
distance, if the intermediate vehicle pairs possess a shorter
inter-vehicle distance (< 100 m), their links face lower SNRs.
But if any of these pairs possesses a longer distance (near 150
m or higher), SNR is higher. This is due to high attenuation of
message signal at high distance. The followings are the
simulation results for the three scenarios:
A. BERs and Co-operative Distances for Scenario-1
In scenario-1, the average co-operative distance for LOS
case shows that the highest average co-operative distance is
262.95 m where 6 vehicles takes part in co-operative routing.
Fig. 8 shows the co-operative distances for LOS links in
scenario-1.
Fig. 8. Average BER and co-operative distance results for LOS case in
scenario-1
The LOS co-operative links for message transmission are
found for vehicles 1,2,5,8,9,10 and 12. The BER results show
that there are enough intermediate SNRs for all cases to
produce any BER for the current amount of bits transmitted in
the simulation.
Again for NLOS case, the average co-operative distance for
vehicle 5 is 145.65 m which is better than the other co-
operative distances. The average BER for this vehicle is
0.025.
Fig. 9 shows the co-operative distances and BERs for
NLOS links of the scenario-1. It is seen that the other average
BER values are less than 0.025.

Fig. 10. Average BER and co-operative distance results for LOS case in
scenario-2

Fig. 9. Average BER and co-operative distance results for NLOS case in
scenario-1

B. BERs and Co-operative Distances for Scenario-2

For scenario-2, the co-operative distance for the vehicle 1 Fig. 11. Average BER and co-operative distance results for NLOS case in
which is 535.74 m is more than those of the other vehicles. 10 scenario-2
vehicles participated in this co-operative routing. Fig. 10
shows the co-operative distances and BER results for all LOS Fig. 12 shows the co-operative distances and BER results
links in scenario-2. In this case, the BER values do not exceed for all LOS links in scenario-2. The highest BER is 2.6e-4 in
2e-5. this case.
The maximum co-operative distance calculated for NLOS The maximum co-operative distance calculated for NLOS
case of scenario-2 is 392.15 m where vehicle 28 is the case is 385.19 m where the message transmitter is vehicle 29.
transmitter. There are 4 vehicles participating in this co- There are 4 vehicles participating in this co-operative routing.
operative routing. The average cumulative BER is 0.044 for The lowest average SNR is 7.72 and the average cumulative
the 50000 bits of transmission in the simulation. BER is 0.053. Fig.13 shows the results for scenario-3.
Fig. 11 shows the co-operative distances and BERs for
NLOS links of the scenario-2. The other BER values are
lower than the BER of vehicle 28.
C. BERs and Co-operative Distances for Scenario-3
For scenario-3, the maximum co-operative distance found
for LOS case is the co-operative distance for the vehicle 19
which is 514.23 m. In this communication, message is
transferred up to 6 vehicles consecutively. This is a random
process. Any no. of vehicles may be found before the message
reaches a vehicle which has no neighbor. And at this point
message transmission stops.

701
Fig. 12. Average BER and co-operative distance results for LOS case in
scenario-3
Fig. 13. Average BER and co-operative distance results for NLOS case in
scenario-3
The results here show that LOS links are the most effective
if there are no multipath effects. Using error correcting codes,
the BER values can be improved. The results demonstrate that
vehicular communication utilizing RFID for 150 m range is
suitable for achieving a co-operative distance of around 500 m
to broadcast a warning message.
[3] P. Blythe, “RFID for Road Tolling, Road-use Pricing and Vehicle
Access Control,” in IEE Colloquium on RFID Technology (Ref. no.
1999/123), 1999.
[4] Dedicated Short Range Communications Project. Available:
http://www.leearmstrong.com/DSRC/DSRCHomeset.htm.
[5] A. Ebner, H. Rohling, L. Wischhof, R. Halfmann and M. Lott,
“Performance of UTRA TDD Ad Hoc and IEEE 802.11b in Vehicular
Environments”, the 57th IEEE Semiannual Vehicular Technology
Conference (VTC 2003-Spring), vol. 2, pp. 960-964, April 2003.
[6] Avonwood product details. Available:
http://www.avonwood.com/products/product_details.asp?id=6#44
[7] Practel, Inc., “Intelligent transport systems-dedicated short range
communications in 5.9 GHz band: technologies and markets”,
Marketresearch website, January 26, 2007. Available:
http://www.marketresearch.com/product/display.asp?productid=142968
3&xs=r
[8] S. Kubota, Y. Okamoto and H. Oda, “Safety Driving Support System
Using RFID for Prevention of Pedestrian-involved Accidents,” in
Proceedings of the 6th International Conference on ITS
Telecommunications, 2006, pp. 226-229.
[9] A. Helmy, F. Bai and N. Sadagopan, “IMPORTANT: An Evaluation
Framework to Study the ‘Impact of Mobility Patterns On RouTing in
Ad-hoc NeTworks’,” in IEEE INFOCOM 2003, 22nd Annual Joint
Conference of the IEEE Computer and Communications Societies, vol.
2, pp. 825- 835, 2003.
[10] H. Harada, M. Otani, R. Funada, “A UHF (950 MHz) Band RFID
Reader Compliant with Japanese Radio Law,” in 32nd Annual
Conference on IEEE Industrial Electronics, pp. 4773-4778, November
2006.
[11] K. Morikawa, “RFID using UHF band under Japanese Radio Law,”
OMRON Corp., December 28, 2004.

VI. CONCLUSION
This paper proposes the modifications of the current market
products and proposals of RFID system to make it suitable for
the targeted vehicle safety application. For the OBU system,
both of the tag and reader are used in a vehicle. The tag is
interfaced to the IVMS in the proposed system.
In the simulation work, the co-operative distances and
related average BERs for three scenarios are shown. It is
found that BER values are acceptable for general data
communication for vehicle safety. It is understood that RFID
can be a suitable media for vehicle safety communication.
This paper’s system model can be designed and extended
for 2.45 GHz RFID system which supports more facilities in
terms of data transfer and extra system supports such as for
Wi-Fi extension etc. Simulation can be extended to multipath
fading channel models like Ricean or Nakagami models. Use
of the system can be extended to traffic management, data
communication, vehicle theft protection and such other
applications.

REFERENCES
[1] P. Beckman, R. Rao, S. Verma, “Use of Mobile Mesh Networks for
Inter-Vehicular Communication,” IEEE 58th Vehicular Technology
Conference, vol. 4, pp. 2712-2715, October 6-9, 2003.
[2] “Jaguar Leaps into Luxury of Bluetooth® Wireless Technology with
Motorola Hands-Free Communication Solution,” Motorola Media
Center. Available:
http://www.motorola.com/mediacenter/news/detail.jsp?globalObjectId=
5331_5326_23

702