Upload
independent
View
1
Download
0
Embed Size (px)
Citation preview
Telecommunication Systems Enabling Real Time Navigation
Alessandro Bazzi, Barbara M. Masini, Gianni Pasolini, Paolomaria Torreggiani
WiLab, IEIIT-BO/CNR, DEIS-University of Bologna,
V.le Risorgimento 2, 40136 Bologna, Italy.
Email: {alessandro.bazzi, barbara.masini, gianni.pasolini, paolomar.torreggiani}@unibo.it
Abstract— Traffic is one of the most important problemsin industrialized countries: roads are getting more and morecrowded, especially during rush hours, with heavy consequenceson people’s quality of life, accidents, pollution, and fuel con-sumption. The transmission of up-to-date traffic information toon-board navigators would allow the redistribution of trafficflows over alternative routes, thus reducing road congestionsand slowdowns. In this paper we investigate the possibilityto send the information on the average-speed in any road ofinterest to on-board navigators by means of current broadcastand unicast communication technologies. A discussion on themost convenient transmission strategy will also be given withreference to real scenarios.
I. INTRODUCTION
The impact of road traffic on both the health of citizens
and the economy of industrial countries is becoming almost
unsustainable: the mix of private and commercial vehicles
that everyday jams urban and suburban roads reduces the
transportation efficiency and increases the travel times as
well as the number of accidents, the fuel consumption and
the air pollution. The road transport system of modern cities
is, in particular, very fragile and prone to collapse when
unpredictable events (such as car accidents) or deterministic
yet disruptive ones (such as rush-hour periods) occur.
Several approaches have been adopted in order to coun-
teract traffic congestions and improve vehicular mobility.
Intelligent traffic lights, for instance, are able to adaptively
modify their timing, travel time. Unfortunately they cannot
suggest alternative routes and their effectiveness is reduced
when a large amount of cars choose the same path.
More effective solutions are based on the provision of real-
time traffic information by means of variable message panels.
Also this solution, however, has a limited effectiveness, since
only few main roads are monitored and the information is
very localized, thus useless for optimal long-distance routing.
It is evident that the real breakthrough for services aimed
at supporting the vehicular mobility is the adoption of a
suitable communication technology transmitting up-to-date
traffic information directly to in-vehicle navigators.
This kind of service, aimed at supporting users’ mobility
by means of information and communication technologies,
falls within the framework of “infomobility” services, deeply
investigated by both the researchers’ community and car
manufacturers [1], [2], [3], and is gaining an increasing
interest. Some solutions are, in fact, becoming available.
Work performed in the framework of the national project Pegasus, fundedby the Italian Ministry for the Economical Development
In several Countries, for instance, on-board navigation
devices receive updated traffic information by means of the
general packet radio service (GPRS) technology [4]. Another
example is represented by the traffic message channel (TMC)
[5], [6], aimed at delivering traffic and travel information
using the radio data system (RDS) on conventional frequency
modulation (FM) radio broadcasts. These technologies and
applications have however some severe limitations:
• the data rate reserved for these services is very limited,
thus allowing the traffic conditions update for a very
limited number of roads;
• only the main events are described, with few approxi-
mate descriptions. We are informed, for instance, that a
given road segment is congested due to road-works and
that the consequent car queue is increasing. We are not
given, however, the average-speed in the nearby roads,
which is crucial to trigger a possible path re-routing.
Indeed, the knowledge of the average-speed of vehicles in
a given road is already a matter of fact: let us consider, for
instance, that at the present time about one million vehicles
in Italy are equipped with an on board unit (OBU) integrating
both a global positioning system (GPS) receiver and a GPRS
transceiver. The main task of this device, which is fostered by
insurance companies, is to continuously collect and locally
store information about the vehicle status (position, speed,
acceleration, etc.), to be examined in case of accident. In
addition, however, this device periodically sends both its
position and speed to a control center, which can therefore
derive the average vehicular speed in any given road travelled
by equipped cars. Providing this information back to the
vehicles (including those without the OBU) would allow the
on-board navigators to always choose the optimal route to
the destination, thus minimizing inconveniences.
In this paper we aim at investigating the possibility to
send the information on the average-speed in any road of
interest to on-board navigation systems by means of current
broadcast and unicast communication technologies; as far
as broadcast systems are concerned, in the following we
will investigate multimedia broadcast and multicast services
(MBMS) over GPRS, MBMS over universal mobile telecom-
munication system (UMTS), digital video broadcasting-
terrestrial (DVB-T), and digital video broadcasting-handheld
(DVB-H). Actual cellular systems (both GPRS and UMTS)
will be investigated as unicast solutions.
2010 13th International IEEEAnnual Conference on Intelligent Transportation SystemsMadeira Island, Portugal, September 19-22, 2010
TB4.2
978-1-4244-7658-9/10/$26.00 ©2010 IEEE 1057
Fig. 1. Example scenario: the best route with real time traffic conditions.
II. THE ENVISIONED SERVICE
Let us clarify the infomobility service envisioned in this
paper, that will be hereafter denoted as “smart navigation”,
discussing its effect from the final user’s point of view.
Suppose the driver enters his car and turns on his smart
navigator. Then, he selects his home as the final destination;
his problem is not how to get there, he perfectly knows the
possible routes; his problem is how to minimize the time
spent in his car.
Once turned on, the smart navigator receives information
on the average-speed in each road segment from its current
position to the destination and calculates the optimal route;
on a real time basis, the smart navigator updates its data and,
in case, modifies the route in order to avoid any slowdown.
A representation of the expected benefit is given in Fig.
1: the navigation system must select the best route to the
destination; although under normal conditions the route A
is the shortest, both in terms of distance and time, the
route B may be preferred taking into account the actual
traffic conditions (that is, the average-speeds of cars already
travelling along route A and route B).
To investigate how communication systems can support
this service is the very objective of this paper.
III. THE TPEG STANDARD
Independently on the communication technology, we as-
sume the adoption of the transport protocol experts group
(TPEG) technology [7], [8] at the highest layers of the
protocol pillar.
TPEG has been developed by a number of broadcasters
and consumer industries to overcome the limitations of the
RDS-TMC technology, allowing to provide a wider range of
services to a wider range of users and devices.
TPEG specifications define in details few main applica-
tions but do not limit the use of TPEG to them; among
these it defines the road traffic message (RTM) application
to transmit updated traffic information.
Particular attention is given to localization. RDS-TMC, in
fact, is based on a static road-list defined by national trans-
port ministries, that includes only a small fraction of the
road network. Although it seems possible to overcome the
Fig. 2. TPEG frame structure.
limited extension of the road-list by simply identifying any
road segment of interest with the coordinates of its starting
and ending points, in practice this is not so easy due to
possible mismatch among different digital maps [9].
TPEG specifications provide a method to solve this prob-
lem (which is called TPEG-loc) but it also allows the
adoption of any other localization method chosen by the
service provider (e.g., TMC or AGORA-C).
TPEG RTM data organization is shown in Fig. 2. The
highest level structure is the TPEG transport frame, which
carries a 7 bytes header and a service frame. The service
frame includes 4 overhead bytes and a variable number of
component frames. Each component frame includes one mes-
sage (Component Data) preceded by 5 bytes of component
header. Each message is finally divided into three parts:
the message management container, which includes general
information, such as the message identifier and the instant
of generation, the event container, which includes the event
to be communicated, and the localization container, which
determines the position of the event.
The size of the message is not fixed: it depends on the
content of the message management container, on the specific
event description in the event container, and on the way
the position is indicated in the localization container. In an
experimental implementation performed in the Korean city
of Daejeon by Sammo Cho et al. [10], an average value of
60 bytes was observed for TPEG RTM messages. Hence a
component frame of 60 bytes plus 5 overhead bytes for each
travelling direction will be considered henceforth. Note that
a high number of messages should be multiplexed into one
service frame, so that the transport frame and service frame
overheads would be negligible with respect to the useful data.
In the following we will show that for the envisioned
service this size requires a significant resource utilization
for most of actual transmission technologies. For this reason
we also considered, for comparison purposes, a different
solution, denoted as short messages, that requires a less
redundant container.
We will show, therefore, the performance that would be
achieved if 8 bytes messages were used for both directions of
any road segment. This would be possible, for example, using
a predefined table to univocally determine a single link, such
1058
Fig. 3. The number of road segments in a city is evaluated considering acircle centered in the middle of the city with an increasing radius.
as in RDS-TMC (as will be discussed in the following); in
this case, 4 bytes would be enough for link identification, 2
bytes could indicate the average speed for both directions and
2 bytes could be left for message control and error check.
If a large amount of short messages are grouped into the
event container of a single component message, all TPEG
overhead will be negligible.
Although the definition of a new application is required,
the service still remains TPEG compliant. The main draw-
back of this solution is the need for tables updating: Some-
how, an update of the table must be communicated to all
smart navigators anytime a new road segment is created.
IV. AMOUNT OF DATA TO BE TRANSMITTED
Our investigation on the suitability of current broad-
cast and unicast communication technologies to support
the downlink (that is, from the control center to vehicles)
transmission of the average-speeds requires, first of all, the
assessment of the amount of data to be transmitted. Once
the size of each packet related to a road segment has been
defined (in the previous section), it is still not known how
many roads must be updated. Let us assume, to this purpose,
an urban scenario, which is the most challenging since it is
characterized by a high road density. The question is: How
many road segments are present in an urban area?
The answer to this question has been obtained through a
query campaign carried out on the digital-maps of the Italian
road network provided by Tele Atlas1.
The number of roads encompassed in a circle centered in
the middle of the city was evaluated varying its radius (as
represented in Fig. 3). In particular, we considered two
metropolitan Italian cities:
• Rome (2700000 inhabitants, 1285 Km2)
• Milan (1300000 inhabitants, 183 Km2)
and two middle sized cities
• Bologna (377000 inhabitants, 140 Km2)
• Florence (370000 inhabitants, 102 Km2).
1Tele Atlas is the licensing business unit of TomTom N.V., the world’sprovider of location and navigation solutions.
0 2000 4000 6000 8000 100000
1
2
3
4
5
6
7x 10
4
Radius [m]
Ro
ad
se
gm
en
ts
Florence
Bologna
Milan
Rome
Fig. 4. Number of road segments within a circle of variable radius.Comparison among four Italian cities (all FRCs are considered).
The outcomes of this investigation are reported in Fig. 4,
that shows that even considering a small radius (with respect
to the city extension) the number of road segments is very
high; in a circle with a radius of five kilometers (78.5 Km2),
for instance, the number of road segments goes from about
13000 to about 25000, depending on the city.
Of course there is no need to transmit the average speeds
related to roads not affected by any slowdown: the amount of
road segments to be updated is, indeed, only a fraction of the
total. Nonetheless, depending on the adopted communication
technology, the amount of data to be transmitted could still
be unaffordable, especially considering that the time that
elapses between the on-board navigator switching-on and the
completion of the roads’ conditions update (latency time)
should be kept as small as possible, so that to avoid any
waste of time waiting the navigator’s “ready to go”.
Let us observe, however, that within the digital map
database each road segment is given an attribute, denoted
as functional road class (FRC), which is related to the im-
portance of the segment (see the first two columns, topmost
part, of Tab. I).
In order to reduce the latency time we should ask whether
an update of all FRCs within the area of interest is strictly
needed or not. To this end, we focused our attention on the
city of Florence, assumed as case study, and we subdivided
the road segments falling within a circle of a given radius
on the basis of their FRC. Results are shown in Fig. 5.
As we can see, in an urban scenario most of road segments
belong to FRC 6 (local roads) and FRC 7 (local roads of
minor importance). It is easy to understand, however, that,
while FRC 6 is of great importance for the assessment of
local traffic conditions, as it includes road used to travel
within a part of a settlement, FRC 7 is, on the contrary,
substantially useless, since it includes roads that only have a
destination function, such as dead-end roads, alleys, narrow
roads between buildings or in a park/garden, etc.
1059
TABLE I
NUMBER OF SEGMENTS IN TELE ATLAS DATABASE PER EACH FUNCTIONAL ROAD CLASS (FRC) CONSIDERING THREE AREAS OF DECREASING
EXTENSION: ITALY, THE TUSCANY REGION, AND THE FLORENCE PROVINCE.
Italy Tuscany Florence Province
FRC FRC description 301336 km2 22994 km2 3514 km2
60.0 millions inhabitants 3.71 millions inhabitants 0.98 millions inhabitant
0 Motorways 27.979 1.705 587
1 Major National Roads not Motorways 48.731 2.411 0
2 Other Roads of National Importance 124.701 7.106 2.220
3 Main Regional Roads 217.970 16.411 3.776
4 Regional Connecting Roads 1.044.349 57.108 11.519
5 Local Roads of High Importance 126.312 6.662 2.209
6 Local Roads 1.113.647 39.710 8.304
7 Local Roads of Minor Importance 4.826.062 203.052 43.310
Tot 7.529.751 334.165 71.925
0-2 201.411 11.222 2.807
3-5 1.388.631 80.181 17.504
6 1.113.647 39.710 8.304
All roads in Italy from FRC 1 to FRC 6: 2.703.689
All roads in Italy from FRC 0 to FRC 2: 201.411
Progressive coverage (0-2 Italy, 3-5 Tuscany, and 6 Florence Province): 289.896
It follows that, even considering a high road density
scenario, traffic updates related to roads belonging to FRC
7 can definitely be avoided, while traffic updates related to
FRC 6 must be transmitted, at least in a small geographical
scale (the urban area or the province, for instance).
When planning, on the other hand, a long distance trip,
the on-board navigator is mainly interested in the information
related to the most important roads at national level (from
FRC 0 to FRC 2) and regional level (from FRC 3 to FRC
5). Moreover, in this case the updates must consider a large
geographical scale. It follows that a convenient transmission
strategy must be conceived in order to meet the requirements
of both short and long distance planning.
Of course the choice of a particular transmission strat-
egy depends also on the adopted transmission technology.
Focusing our attention on some of the major current com-
munication infrastructures allowing full coverage in mobility
conditions, two alternatives will be discussed in the following
sections:
1) Broadcast technologies, with particular reference to
RDS-TMC, MBMS over GPRS, MBMS over UMTS,
DVB-T, and DVB-H;
2) Unicast technologies, with particular reference to
GPRS and UMTS.
V. BROADCAST TRANSMISSION
Broadcast transmission allows to send the same data
stream to all receivers avoiding useless duplicates. For the
smart navigation service, a broadcast approach may represent
a suitable solution, allowing to efficiently exploit radio
resources by cyclically transmitting traffic data to many users
over a common radio channel.
Many telecommunication technologies allowing to broad-
cast information to mobile terminals are available nowadays;
among them, RDS-TMC, MBMS over GPRS, MBMS over
UMTS, DVB-T, and DVB-H will be considered throughout
this work. In the following a brief description of their ad-
vantages and drawbacks will be given and their performance
will be shown in terms of refresh time, defined as the time
to transmit a complete update of the road network situation.
To this purpose, a convenient transmission strategy should
be defined. Henceforth the following cases will be investi-
gated:
1) transmission of updates related to all roads (except
FRC 7) at national level (Italy in the case study);
2) transmission of updates related to all roads at national
level belonging to FRCs from 0 to 2 (major roads);
3) transmission of updates related to all roads from FRC
0 to FRC 2 at national level, to all roads from FRC 3
to FRC 5 at regional level (Tuscany in the case study)
and to all FRC 6 at local level (Florence Province in
the case study). This strategy, denoted in the following
as progressive coverage, follows the remark, reported
in Sec. IV, that users are interested on all roads on a
small geographical scale and on the major roads on a
larger scale.
The number of road segments that correspond to these
solutions in the specific case study under investigation is
highlighted in the bottommost part of Tab. I.
In the numerical results, 10% of road segments are as-
sumed interested by slowdowns (thus 10% of road segments
will be updated); although in a real application this per-
centage strongly varies during time, we think it represents
a reasonable value for a busy hour; in any case, results
corresponding to a higher or lower percentage can be easily
derived from those presented with a simple scaling operation.
Please note that in the case of “progressive coverage” we
are assuming that transmitters located in different provinces
of the same region convey different FRC 6 updates, while
transmitters placed in different regions convey different up-
dates related to FRCs from 3 to 6.
A. RDS-TMC
RDS-TMC is a specific application of the FM-RDS used
for broadcasting real-time traffic and weather information.
It was developed to offer dynamic route guidance, thus
it broadcasts messages alerting the drivers of obstacles or
1060
0 2000 4000 6000 8000 100000
2000
4000
6000
8000
10000
12000
14000
Radius [m]
Road s
egm
ents
FRC 0
FRC 1
FRC 2
FRC 3
FRC 4
FRC 5
FRC 6
FRC 7
Fig. 5. City of Florence: number of road segments for each FRC as afunction of the circle radius.
slow downs all over a nation. Data related to traffic flows,
accidents, weather, etc. are gathered from traffic monitoring
systems, motorists’ calls, and collected at a central traffic
information center. They are then passed to the TMC traffic
information service provider, who generates TMC messages
according to the ALERT-C coding protocol [6].
RDS-TMC carries digital data at 1187.5 bps with an
overall bandwidth of approximately 5 KHz inside one FM
radio channel. The RDS-TMC data stream is partitioned into
packets and blocks. A packet consists of four blocks, each
being 26 bits long (i.e., 104 bits per packet); only 16 bits
are for information while the remaining 10 bits convey the
overhead (including the cyclic-redundancy-check sequence).
Considering that each TMC packet carries the information
of a single event (an accident, a congested road, etc.), it
is clear that the data rate is fairly low for the envisioned
service. Furthermore, the high error-rate that occurs in RDS-
TMC data received in areas where FM signal propagation
is impaired, requires a robust system with ample repetition
to ensure that the messages are delivered reliably. However,
the low-price of RDS broadcast stations and the widespread
diffusion of RDS receivers contributed to the adoption of
RDS-TMC for cars navigation systems in recent years, and
must be thus considered, at least for comparison purposes.
In the case of RDS-TMC we did not consider the TPEG
structure discussed in Sec. III; in order to not further reduce
RDS-TMC capacity, its standard use will be supposed.
To have a fair comparison with other communication
technologies, a road table including all roads belonging to
FRCs from 0 to 2 will be assumed, which is the one of the
cases (the least loaded one) considered for more performing
technologies.
The refresh time obtained with RDS-TMC is shown in Fig.
6. From this analysis it is clear that the TMC-RDS is not a
suitable solution for the envisioned service; even neglecting
all roads with an FRC higher than two, the update of 10%
of the road table requires about 1 hour.
RDS−TMC(*) MBMS−2G MBMS−3G 64k MBMS−3G 128k DVB−T DVB−H10
−1
100
101
102
103
104
105
Re
fre
sh
tim
e [
s]
FRC 0−2 Italy
Progressive coverage
All FRC Italy but 7
Progressive coverage, short messages
5 minutes
1 minute
Fig. 6. Refresh time adopting TMC-RDS, MBMS over GPRS, MBMSover UMTS at 64 kbps, MBMS over UMTS at 128 kbps, DVB-H, andDVB-T. Four cases are compared, varying the considered road segments(one case with short messages is depicted for comparison). 10% of roadsare supposed congested. (*) The standard TMC message is used for theperformance evaluation of RDS-TMC.
B. MBMS
Given the success of high bandwidth applications in third
generation cellular systems, especially with a large number of
users receiving the same high data rate services, efficient in-
formation distribution is essential. For this reason 3GPP pro-
vided (from the release 6) the MBMS technology, allowing
multicast and broadcast packet delivery in GPRS and UMTS
networks [11]. Adopting MBMS packets can be forwarded
from one source to many receivers without overloading the
network and wasting the scarce radio resources.
1) MBMS over GPRS: GPRS is the evolution of the
global system for mobile communication (GSM), the second
generation (2G) cellular technology, and represents the first
packet-switched communication technology for mobile users.
GSM and GPRS adopt the time division multiple access
(TDMA), with eight time slots per TDMA frame. While one
time slot per direction per user is reserved and used for each
voice call (silent suppression techniques allow to statistically
half this value), GPRS allows a single mobile station to
transmit on multiple time slots following the effective need.
Different coding schemes (CS) are provided in order to
allow different trade-offs between reliability and data rate.
With 3GPP release 6, GPRS also becomes the first cellular
systems allowing to transmit both in unicast and multicast
mode, implementing MBMS. With MBMS, up to 6 time slots
can be assigned in the downlink to a multicast/broadcast
session or up to 4 if feedback is used in the uplink, thus
consuming part of the available cell resources dedicated to
unicast transmissions.
As discussed in [12], even adopting the most robust CS it
is not possible to achieve a sufficient QoS without adopting
some additional technique; the full coverage with a net data
rate of 45 kbps is thus obtained through a single repetition
of all transmitted packets.
In Fig. 6, the refresh time related to MBMS over GPRS
at 45 kbps (denoted as MBMS-2G) is reported. Numerical
1061
results show that adopting TPEG RTM this technology does
not even allow to update the road classes from FRC 0
to FRC2 without exceeding the 5 minutes threshold. On
the contrary, if the short messages solution is adopted, the
“progressive coverage” of road segments can be completed
in less than 1 minute.
Note that the adoption of GPRS has the advantage to exploit
a deeply developed network, which guarantees a full cover-
age in any mobility condition; however, the transmission of
the smart navigator service occupies the resources of 6 to 12
voice calls (depending on the silence suppression efficiency).
2) MBMS over UMTS: UMTS represents the third gener-
ation (3G) of cellular systems. It is based on wideband-code
division multiple access (W-CDMA) and was conceived for
the transmission of both circuit and packet switched services
with QoS provision. From the release 6 UMTS also allows
to transmit data in broadcast and multicast modes adopting
the MBMS technology, with a bit-rate ranging from 64 kbps
up to a maximum of 384 kbps. MBMS consumes in this case
part of the available cell capacity, thus limiting the number of
available dedicated channels that can be established, that is
the number of resources dedicated to voice calls and unicast
data sessions. In particular, the base station is required to
pre-assign a certain amount of power to the MBMS services
depending on the coverage planning and the desired bit rate.
The assessment of the required power level is not an
easy task, since it depends on various aspects, including the
considered scenario, the base stations density, and the vehicle
mobility conditions. As an example, in [13] a minimum of
20.9% of the total power is needed to gain a 95% coverage
with QoS for a 64 kbps transmission in a vehicular environ-
ment where cars move at 3 km/h (higher speed allow better
performance due to higher diversity gain); a lower percentage
(4.8%) is needed if all terminals allow soft combining of up
to three radio links. In general, simulation results shows that
a 64 kbps transmission is practically feasible without exploit-
ing some transmission diversity, although it heavily impacts
on the system capacity, while a 128 kbps transmission is
feasible only if diversity techniques are adopted.
In Fig. 6, the refresh time for the envisioned service is
shown adopting MBMS over UMTS at 64 and 128 kbps (de-
noted as MBMS-3G 64k and MBMS-3G 128k, respectively).
Only the 128 kbps data rate allows a full refresh in less than
five minutes adopting TPEG RTM, while both solutions are
able to guarantee a refresh time lower than one minute if the
short message solution is adopted.
C. DVB-T
DVB-T is the transmission technology developed for the
digital terrestrial television by the DVB Project [14]. It
envisions a single-frequency network (SFN) consisting of
neighboring transmitters synchronized to each other, that
transmit identical moving picture experts group (MPEG)
transport streams with digital audio, video and data. To
guarantee a correct reception also in the presence of fre-
quency selective fading, DVB-T adopts the coded orthogonal
frequency division multiplexing (COFDM) [15], with either
6817 carriers, in the so-called 8K mode, or 1705 carriers in
the 2K mode. Several combinations of modulation schemes
and coding rates are allowed, with data-rates ranging from
4.98 to 31.67 Mbps. One of the mostly adopted DVB-
T configurations allows an useful bit rate of 24.13 Mbps,
which is shared among different television channels whose
bit rate ranges from 2 to 5 Mbps. For this reason, in the
numerical results we considered an useful bit rate of 2 Mbps,
which corresponds to the satisfactory (yet not excellent)
transmission of one television channel.
In Fig. 6, the refresh time for DVB-T is shown: it can be
observed that a refresh time lower than one minute can be
achieved adopting the “progressive coverage” strategy also
with the resource eager TPEG RTM. If all roads from FRC
1 to FRC 6 in the whole nation are considered, more than
two minutes are necessary.
Although the DVB-T technology has not been conceived
for an in-car usage, recent advances in the receiver design
[16] showed that it can work with a good performance also
at a vehicular speed of 180 Km/h. Nonetheless, in a dense
urban environment it can suffer a poor received power at the
street-level, especially in the presence of urban canyons. For
this reason in the following paragraph we also investigate
the DVB-H technology, which is, in some way, a version of
DVB-T designed for mobile usage.
D. DVB-H
The DVB-H standard has been developed in order to
broadcast MPEG streams also to mobile terminals. The
physical layer transmission is performed according to the
DVB-T standard, with the introduction of the 4K mode
with 4096 OFDM subcarriers; this allows for a doubled
transmitter distance in SFNs compared to the 2K mode, and
makes DVB-H less susceptible to Doppler frequencies in
case of mobile reception compared to the 8K mode. Also
for DVB-H, depending on the modulation and coding rate,
different data rates can be achieved, ranging from 4.98 to
31.67 Mbps. In the numerical results we considered an useful
bit rate of 384 Kbps, which corresponds to a H.264 video
stream used for deploying the mobile TV service to DVB-H
terminals.
The refresh time that can be achieved with DVB-H is
shown in Fig. 6: a refresh time of 80 seconds is obtained
adopting the “progressive coverage” strategy and the TPEG
RTM packet structure. A reduction of the number of updated
roads or a reduction of the packet length (8 bytes, for
instance) is needed if one minute maximum is targeted.
E. Broadcast systems comparison
So far we have considered “pure” broadcasting systems
(RDS-TMC, DVB-T, DVB-H) and “hybrid” broadcasting
systems (MBMS over cellular systems). Numerical results
show that, due to its scarce resource and consequent high re-
fresh time, RDS-TMC does not represent a suitable solution
for the envisioned service. In contrast, DVB-T and DVB-H
could be considered as candidate for the broadcasting of a
certain percentage of roads condition updates.
1062
MBMS over cellular systems has performance in between
RDS-TMC and DVB. It represents a valid solution for
the real-time broadcasting of information considering one
minute as the target refresh time only if short messages are
transmitted.
Let us observe, however, that the envisioned service is
based on the real-time collection of information from ve-
hicles, allowed by on-board cellular systems. This reason,
together with their widespread diffusion and their good per-
formance in high mobility scenarios, makes cellular systems
a valid solutions to support smart navigation.
VI. UNICAST TRANSMISSION
With unicast transmission technologies each driver could
receive personalized information, related to the average-
speeds of all road segments (except FRC 7) interested by
slowdowns along his way to the destination. In this case
the on-board navigator should previously inform the control
center about its route or at least its destination.
Although this solution allows to update only the interested
users, it also entails some drawbacks that would increase with
a higher penetration of the service: firstly, a dedicated unicast
communication is required also in the uplink direction;
secondly, the same information will be often sent to a not
negligible number of vehicles, wasting useful resources.
The strategy adopted to identify the roads of interest for
each vehicle must thus be carefully designed, since it sets
the trade-off between a sufficient picture of the road traffic
conditions and an oversized amount of transferred data both
in the downlink and the uplink directions.
In this paper we suppose to update the road segments
encompassed by an ellipse (as shown in Fig. 7) whose
focuses are the actual vehicle position and either the next
intermediate point or the final destination. In particular,
the final destination will be directly targeted only in the
case of a short travel, for instance shorter than 30 Km.
When, on the contrary, a long travel has to be planned,
the on-board navigator evaluates, first of all, the best route
according to the static information stored in its memory, then
it detects a number of intermediate points along the way
and, finally, it informs the control center about the next-to-
come intermediate point. According to the user preferences,
periodic updates can also be sent during the trip.
This strategy avoids the transmission of information re-
lated to road segments too far from the actual vehicle
position, which would be out of date when the vehicle
will get there. Moreover, since only the transmission of the
coordinates of two points is needed from the navigation
system to the control center, the amount of data transmitted
in the uplink is very small, thus limiting costs and resource
occupation.
The suitability of unicast technologies to support the above
described service is investigated hereafter, with particular
reference to GPRS and UMTS, which are actually the
most widespread technologies to guarantee full coverage in
mobility.
Fig. 7. The area considered for unicast updating varying the distancefrom the actual vehicle position and the targeted point (destination or nextintermediate point).
Numerical results will be provided in terms of the time
needed to complete the updates transmission, varying the
distance between the ellipse’s focuses (thus increasing the
number of interested road segments), with one focus always
taken in the city center of Florence (refer to Fig. 7). A ratio
of 1/2 is considered between the minor and the major radii
of the ellipse.
Similarly to the broadcast case, TPEG RTM is adopted for
the average-speeds update and 10% of the road segments un-
der investigation (those from FRC 1 to FRC 6 encompassed
by the ellipses) are supposed to suffer of some slowdowns.
A. GPRS
As recalled in Sec. V-B, GPRS represents the evolution
of GSM toward packet switched networks. As far as unicast
transmissions are concerned, GPRS allows a mobile station
to transmit on one to eight time slots per TDMA frame.
Channels are allocated only when data packets are sent
or received, then released after the transmission. For bursty
traffic this results in an efficient usage of radio resources
as multiple users can share one physical channel. GPRS
theoretically allows data rate up to 170 Kbit/s, depending
on the adopted code rate and the number of used time slots.
In Fig. 8 the time needed for an update of the 10% of road
segments from FRC 1 to FRC 6 inside the discussed ellipse
is given as a function of the distance between the focuses,
considering a data rate of 32 kbps. Such speed follows the
use of 4 time slots (i.e., of half the capacity of a carrier)
with the most robust coding scheme.
Numerical results show that the GPRS performance is
borderline with respect to the acceptable one: on one hand,
it provides sufficiently low latencies after the connection
setup (50 seconds are needed to complete the update for
a distance of 30 km between focuses); on the other hand,
however, a maximum of 2.3 or 19.8 requests can be served
per minute adopting an entire carrier (2 groups of 4 time
slots each) if the distance between the focuses is 30 or 5
km, respectively. It follows that depending on the amount
1063
5 10 15 20 25 3010
−2
10−1
100
101
102
Focuses distance [km]
Update
dura
tion [s]
GPRS 32 kbps
UMTS 64 kbps
HSDPA 7.2 Mbps
Fig. 8. Time needed to complete the update of 10% of roads adoptingeither GPRS at 32 kbps, UMTS at 64 kbps or UMTS-HSDPA at 7.2 Mbps,as a function of the focuses distance.
of users and the ellipse dimension the connection-setup time
could be annoying.
B. UMTS
UMTS represents the most performing cellular system
actually deployed, allowing unicast bit-rates beyond 10 Mbps
with the introduction of high speed downlink packet access
(HSDPA) in 3GPP Releases 5 [17]. Due to the adoption
of CDMA, the number of active channels in UMTS is a
consequence of the trade-off between coverage and capacity,
and the amount of resources occupied by each transmission
is given in terms of used power: On one hand, a higher
data rate as well as a higher distance from the base require
a higher power for a sufficient QoS; on the other hand, a
higher power reduces the cell capacity. The power is, in
fact, a limited resource at the base station (in downlink) and
each transmission turns into an interference to all other active
communications (in both directions).
In Fig. 8, the time required for an update of the 10%
of the road links inside the ellipse is shown, considering
UMTS data rates of 64 kbps and 7.2 Mbps. The former value
corresponds to the minimum data rate for an unicast data
transmission allowing full coverage and minimum resource
occupation, while the latter corresponds to the maximum
achievable data rate with HSDPA for the majority of ter-
minals actually available (it corresponds to users equipment
belonging to category 8 [18]).
With reference to the 64 kbps connection, we can observe
that differently from GPRS the delay never exceeds 30
seconds; the maximum number of users that can be served
per physical channel per minute in each cell is 2.3 or
19.8 if the distance between the focuses is 30 or 5 km,
respectively; the number of channels simultaneously active is
highly variable, depending on the network planning, the users
position, the resources exploited by the other common and
dedicated channels. In any case the adoption of the 64 kbps
channel would become critical with an increasing penetration
of the smart navigation service.
In contrast, the full usage of a 7.2 Mbps channel allows
negligible delays and up to 258.4 and 2227.3 requests per
minute for distances between the focuses of 30 and 5 km,
respectively. Note that these values represent upper bounds,
since such a high data-rate is possible only near the base
station, where the channel is highly reliable.
VII. CONCLUSIONS
In this paper we investigated the feasibility of frequent
(say real-time) traffic condition updates through broadcast
and unicast transmission technologies. Attention has been
focused on the amount of data to be transmitted and their re-
fresh time in order to allow smart navigation. Results showed
that only some technologies can guarantee the envisioned
service depending on the target refresh time. Moreover,
broadcast has to be preferred for high densities of vehicles,
whereas unicast UMTS transmission can be considered for
dedicated information at the expense of a capacity reduction
for other UMTS services.
ACKNOWLEDGMENTS
The authors would like to thank Prof. O. Andrisano for
motivating this work, G. Leonardi and T. Pavani for their
valuable support with the Tele Atlas data base.
REFERENCES
[1] “Cvis project,” Website, http://www.cvisproject.org/.[2] Website, www.car-to-car.org/.[3] Website, http://pegasus.octotelematics.com.[4] Website, www.tomtom.com/.[5] Website, http://www.tmcforum.com/.[6] Traffic and Traveller Information (TTI) TTI messages via traffic mes-
sage coding Part 1: Coding protocol for Radio Data System Traffic
Message Channel (RDS-TMC) using ALERT-C, ISO Std. 14 819-1.[7] Traffic and Travel Information (TTI) via Transport Protocol Expert
Group (TPEG) data-streams: Parts 1 to 6, ETSI Std. ISO 18 234.[8] Website, http://www.tpeg.org/.[9] V. Hiestermann, “Map-independent location matching certi-
fied by the AGORA-C standard,” Transportation Research
Part C: Emerging Technologies, vol. 16, no. 3, pp.307 – 319, 2008, emerging Commercial Technologies.[Online]. Available: http://www.sciencedirect.com/science/article/B6VGJ-4S0280N-1/2/b64ec2e00c364e0488866db160182047
[10] S. Cho, K. Geon, Y. Jeong, C.-H. Ahn, S. I. Lee, and H. Lee, “Realtime traffic information service using terrestrial digital multimediabroadcasting system,” Broadcasting, IEEE Transactions on, vol. 52,no. 4, pp. 550 –556, dec. 2006.
[11] Introduction of the Multimedia Broadcast/Multicast Service (MBMS)
in the Radio Access Network (RAN), ETSI Std. 3GPP TS 25.346.[12] L. Provvedi, C. Rattray, J. Hofmann, and S. Parolari, “Provision
of MBMS over the GERAN: technical solutions and performance,”in 3G Mobile Communication Technologies, Fifth IEE International
Conference on, 2004, pp. 494 – 498.[13] S-CCPCH performance for Multimedia Broadcast/Multicast Service
(MBMS), ETSI Std. 3GPP TR 25.803 v.60.0, 2005.[14] Website, http://www.dvb.org/.[15] ETSI EN 300 744 V1.6.1 (2009-01), ETSI Std., 2009.[16] C.-C. Wang, T.-J. Lee, H. K. Lo, S.-P. Lin, and R. Hu, “High-
sensitivity and high-mobility compact DVB-T receiver for in-carentertainment,” Consumer Electronics, IEEE Transactions on, vol. 52,no. 1, pp. 21 – 25, feb. 2006.
[17] H. Holma and A. Toskala, WCDMA for UMTS, 3rd Ed. Wiley, 2005.[18] UE Radio Access capabilities, ETSI Std. 3GPP TS 25.306 V9.1.0.
1064