Concept for visualizing concealed objects to improve the driver’s anticipation
Simon Nestler
Fachgebiet Augmented Reality
Technische Universita¨t Mu¨nchen
Fakulta¨t fu¨r Informatik
Boltzmannstraße 3, 85748 Garching bei Mu¨nchen
Germany
Markus Duschl
Fachgebiet Augmented Reality
Technische Universita¨t Mu¨nchen
Fakulta¨t fu¨r Informatik
Boltzmannstraße 3, 85748 Garching bei Mu¨nchen
Germany
Darya Popiv
Lehrstuhl fu¨r Ergonomie
Technische Universita¨t Mu¨nchen
Fakulta¨t fu¨r Maschinenwesen
Boltzmannstraße 15, 85748 Garching bei Mu¨nchen
Germany
Mariana Rakic
BMW Group Forschung und Technik GmbH
Hanauerstraße 46
80992 Mu¨nchen
Germany
Gudrun Klinker
Fachgebiet Augmented Reality
Technische Universita¨t Mu¨nchen
Fakulta¨t fu¨r Informatik
Boltzmannstraße 3, 85748 Garching bei Mu¨nchen
Germany
Current advanced driver assistance systems (e.g. Emergency Brake Assistance, Lane Departure Warning, Lane Keeping
Assistance and Blind Spot Detection) assist the driver in reacting to time-critical and unstable situations in a proper way.
However, the anticipation of situations which are lying in the farer future is currently left primarily to the driver. In this
paper, we present visualization concepts for concealed objects in order to support smart deceleration. Smart deceleration
requires the anticipation of future trac condition and the assistance of the driver in performing deceleration phases
eciently. In addition, safety is increased by reduction of potential criticality through the early deceleration phase. We
have identified and categorized situations in which a broader anticipation is possible: situations with permanent obstacles,
situations with temporarily stopped vehicles and situations with slower driving vehicles.
An important issue when presenting information to the driver is the identification of the most suitable perspective. For
visualizing the trac situation in the surroundings of the driver’s car we established a virtual bird-eye perspective (VBEP),
showing the trac scene from a 3D viewpoint that is slightly raised above the driver and rigidly tethered to the car. This
VBEP is a powerful concept to draw the driver’s attention to situations in the further future. On the basis of this concept
we developed dierent visualizations and integrated them in the digital instrument cluster between the speedometer and
the revolution counter.
MOTIVATION
The goal of this paper is to visualize concealed objects
in order to support anticipative driving behavior. All dierent
reasons for the concealing of objects lead to the same prob-
lem: The driver’s perception is incomplete. Aside from expe-
rience, knowledge and attention, drivers’ anticipation bases
also on their perception. An incomplete perception leads to
wrong anticipation – which might result in critical situations.
Some of such critical situations could be avoided by an im-
provement of driver perception.
INTRODUCTION
Current advanced driver assistance systems (ADAS),
such as Emergency Brake Assistance (Tamura, Inoue, Watan-
abe, & Maruko, 2001), Lane Departure Warning (Labayrade,
Douret, & Aubert, 2006), Lane Keeping Assistance (Ishida
& Gayko, 2004), Blind Spot Detection (Matuszyk, Zelinsky,
Nilsson, & Rilbe, 2004), Night Vision (Bellotti, Bellotti, Glo-
ria, Andreone, & Mariani, 2004), and Overtaking Assistance
(Batavia, Pomerleau, & Thorpe, 1997) support the driver in
reacting to these time-critical and unstable situations in a
proper way. These current ADAS do not, however, focus on
assisting the driver in the very early identification of critical
situations. The anticipation of situations in the farer future is
left primarily to the driver.
Especially in situations in which the driver’s field of vi-
sion is limited, e.g. because of fog, vertical curves or turns
proper anticipation is impossible due to limited cognition. In
order to anticipate properly, the driver requires information on
the upcoming trac situation. When the driver is distracted,
proper anticipation is hindered due to the incomplete situation
assessment.
We designed a visualization concept which overcomes
the cognition limitations by visualizing the upcoming trac
situation from a bird-eye perspective. Additionally our visu-
alization concept provides a situation assessment which leads
the driver to correct anticipation even if he is partially dis-
tracted. We mainly focus on trac situations in which the
visualization of anticipatory information leads to a smart de-
celeration process.
The proposed visualization concept assists the driver in
performing deceleration phases eciently. In addition, safety
is increased by reduction of potential criticality due to early
deceleration.
Section 3 gives a short overview of related work in the
field of integrating dierent ADAS functionalities in a con-
sistent HCI concept. In Section 4, dierent trac situations
in which we want to improve the drivers’ perception are pre-
sented. Our virtual bird-eye perspective is presented and dis-

cussed in Section 5. The visualization concept for the situa-
tions of Section 4 is described in Section 6. Options for haptic
assistance are presented in Section 7. Section 8 gives a short
outlook on our future work.
RELATED WORK
According to (Gruyer, Rakotonirainy, & Vrignon, 2005)
the integration of a wide range of functionalities can improve
the reliability of ADAS. Moreover, systems which monitor
car dynamics and environment perception should be com-
bined with systems which assess the driver’s state. They pre-
sented an integrated ADAS which merges these dierent ex-
isting functionalities. The challenge is to avoid the isolation
of human computer interactions from the rest of activities in
which the driver is involved within HCI research approaches
(Rakotonirainy, 2003).
The increasing information access (e.g. phone calls, traf-
fic information, speed limits) leads to new challenges in driver
distraction considerations. The combined impact of multiple
in-vehicle devices on the driver’s distraction has been ana-
lyzed by (Brooks & Rakotonirainy, 2005).
Among others, (Dugarry, 2004) faced the problem of in-
formation overload caused by ADAS, in-vehicle communi-
cation systems (IVCS) and in-vehicle information systems
(IVIS), creating potentially dangerous conditions. Dugarry
focused on the prioritization and presentation of currently rel-
evant information in order to prevent overload in most condi-
tions.
TRAFFIC SITUATIONS
We categorized situations in which smart deceleration
can be useful into three classes: 1) situations with perma-
nently stationary obstacles, 2) situations with temporarily
stopped vehicles and 3) situations with slower driving vehi-
cles. In situations with permanent obstacles, the driver is ap-
proaching an obstacle that does not move, e.g., a construction
site. In situations with temporarily stopped vehicles, an driv-
ing vehicle ahead stops and becomes a temporary obstacle,
e.g., a car in the process of parking or letting passengers get
out. In the situations with slower driving vehicles, the vehicle
ahead is moving, but its speed is considerably lower than the
speed of the own car, e.g., trucks or tractors.
Using these three classes, we developed a visualization
concept that can be adapted to a large range of trac situa-
tions, as described below.
Speed limitation ahead
Speed signs in front of the driver’s car belong to the
group of situations with permanent obstacles. We have con-
sidered two cases. In the first case (see Fig. 1), the driver’s
car is on a rural road which changes into an urban road. In the
second case (see Fig. 2), the driver is on a (German) highway
without speed limitation, followed by a section with a speed
limitation to 100 km/h. In the first case, the deceleration phase
starts at 100 km/h and ends at 50 km/h whereas in the second
case, the deceleration phase starts somewhere between 130
km/h and 200 km/h and ends at 100 km/h.
From the technical point of view, dierent options for
recognizing these speed signs are feasible: camera based
recognition, maps with additional information or speed signs
with wireless communication capabilities.
Figure 1. Speed sign ahead on a rural road
Figure 2. Speed sign ahead on a highway
Construction site ahead
Construction sites in front of the driver’s car also belong
to the group of situations with permanent obstacles. In this
case (see Fig. 3) the driver’s car is, for instance, on a rural
road with a construction site on his side of the road. Because
of the oncoming trac he has to decelerate in front of the con-
struction site. The smarter the driver decelerates, the more
likely it is that he do not have to come to a complete stop;
he can make use of one of the gaps in the oncoming trac.
This deceleration phase starts at 100 km/h and ends at 0 km/h
(in the worst case when the driver performs a hard braking
maneuver in close distance to the obstacle).
Parking vehicle ahead
Parking vehicles in front of the driver’s car belong to the
group of situations with temporarily stopped obstacles. As an
example for this case (see Fig. 4), the driver’s car is on an
urban road with a car parking in his own lane. Because of the
oncoming trac he has to decelerate in front of the parking
vehicle. To some extent, this situation is similar to the sit-
uation with the construction site; however it diers in three
aspects: Firstly, the early identification of a parking vehicle
is more dicult than the identification of a construction site
(especially realizing the fact that the car is standing still in the
lane might be dicult for inexperienced drivers). Secondly,
the obstacle is only temporary and might go away within the
next seconds. Thirdly, the speed on urban roads is signifi-
cantly lower than on rural roads. Here, the deceleration phase
starts at 50 km/h and ends at 0 km/h (in the worst case).
Figure 3. Construction site ahead on a rural road, with trac in the
opposite lane

Figure 4. Parking vehicle ahead on an urban road, with the time
until further movement of the obstacle unknown
Figure 5. Red trac light ahead on an urban road, with the time
until change of the light unknown
Red trac light ahead
Red trac lights with standing vehicles belong to the
group of situations with temporarily stopped vehicles. In
this case (see Fig. 5), the driver’s car might be on an ur-
ban road and approach a crossing with a trac light. The
light is currently red, and some stopped cars are standing in
front of it. Such cars make this situation more complex than
what has recently been covered in designs of trac sign assis-
tants (Moutarde, Bargeton, Herbin, & Chanussot, 2007) and
(Thoma, Klinker, & Lindberg, 2007). The driver waits for the
trac light to switch to green during the deceleration phase.
The smarter his deceleration is the more likely it is that he
does not have to stop completely. This deceleration phase
starts at 50 km/h and ends at 0 km/h (if the red phase of the
trac light is very long).
Jam ahead
Depending on the driving speed of the cars in the jam, the
situation either belongs to the group with temporarily stopped
vehicles or to the group with slower driving vehicles. It even
might be argued that a completely standing trac jam could
be considered a permanent obstacle due to the fact that its
dissolution within the next seconds is practically impossible.
As an example for the first situation (see Fig. 6), the driver’s
car approaches a trac jam on a highway in which the cars
have come to a complete standstill. In the second situation
(see Fig. 7), the trac jam is still moving at about 60 km/h.
In the first case, the deceleration phase starts between 130
km/h and 200 km/h and ends at 0 km/h whereas in the second
case, the deceleration phase starts between 130 km/h and 200
km/h and ends at 60 km/h.
Slower driving vehicle ahead
Slower driving vehicles in front of the driver’s car require
the driver to decelerate. We consider two cases. In the first
case (see Fig. 8), the driver’s car is on a rural road, with
overtaking being prohibited. In the second case (see Fig. 9),
the driver is on a rural road with significant oncoming trac.
The driver is unable to overtake the slower vehicle because of
this oncoming trac. Both deceleration phases start at 100
Figure 6. Trac jam ahead on a highway
Figure 7. Slowly moving trac jam ahead on a highway
km/h and end at 70 km/h, i.e., the speed of the slower driving
vehicle.
VIRTUAL BIRD-EYE
PERSPECTIVE
An important issue when presenting information to the
driver is the identification of the most suitable perspective.
(Milgram & Colquhoun, 1999) and (Milgram & Kishino,
1994) distinguish between four cases: 1) information is pre-
sented in an egocentric view of the user (here: driver), 2) in-
formation is shown from a viewpoint that is raised above and
behind the user, yet rigidly tethered to the user, 3) the view-
point is world-referenced, i.e., independent of the user, yet at
a low height, and 4) the view point is very high above the user,
e.g. on a satellite. The first two cases are called egocentric,
the latter two are called exocentric. All dierent perspectives
are shown in Figure 10.
The options for presenting trac information to the
driver depend on the camera position and orientation, relative
to the egocentric reference frame of the driver. For visual-
izing the trac situation in the surroundings of the driver’s
car a rigidly tethered 3D perspective seems to be most suit-
able. This rigid tethering can be described by the follow-
ing metaphor: A bird carrying a camera is flying behind the
driver’s car; the trac situation is shown from the bird’s per-
spective. Therefore we call this special egocentric perspective
virtual bird-eye perspective (VBEP).
Figure 8. Slower driving vehicle ahead on a rural road, with over-
taking being prohibited

Figure 9. Slower driving vehicle ahead on a rural road with trac
in the opposite lane
Figure 10. Dierent perspectives, from egocentric to exocentric
(according to Milgram & Colquhoun, 1999 and Milgram & Kishino,
1994)
The complete semi sphere on which the camera can the-
oretically be placed for a VBEP is shown in Figure 11. The
angle  describes the rotation around the x-axis, whereas the
angle  describes the rotation around the y-axis.1 A rotation
around the z-axis is not feasible. The distance d is the distance
between camera and car; d is the length of the translation vec-
tor.
To experiment with dierent perspectives, we built an
Augmented Reality-based design tool, consisting of a web-
cam and an ARToolKit marker (Kato & Billinghurst, 1999).
This Toolkit is able to determine the position and orientation
of a camera relative to live video images of markers. It can
then augment the images with virtual objects in real-time. For
the VBEP design tool, we use a piece of paper showing a
street and a marker at the position where the driver’s car is
supposed to be. The car is shown as a virtual object that is
overlaid on the image stream (see Fig. 12a). By moving the
webcam, the designer can explore dierent viewing perspec-
tives onto the car.
Additionally we explore various trac situations by plac-
Figure 11. Semi sphere on which the camera can be placed
(a) Augmented driver’s car (b) Augmented trac jam
Figure 12. VBEPs with webcam and ARToolKit marker
ing obstacles in relation to the driver’s vehicle (see Fig. 12b).
The camera based tracking requires the designer to rotate the
camera such that the camera sees the marker. Consequently
the marker is visible in all perspectives which are selected
by the engineer. Thus, our approach prohibits the designer
from choosing perspectives in which the driver’s car is not
visible and the driver might become disoriented, loosing the
overview.
The VBEP concept and the design tool enable the de-
signer of a trac visualization concepts to freely adapt the
camera position as well as the region of interest to the current
situation. The tool enables the designer to intentionally ex-
plore turning the driver’s spotlight to objects, which currently
might not be visible in reality. The advantage of the VBEP
design tool setup is that the designer can adapt all parameters
(, , d) of the bird-eye perspective in real time while the car
goes through one or more of the envisaged trac situations.
Setting the region of interest on objects in the farer future is
done by decreasing distance d and decreasing angle . Setting
the region of interest on the direct surroundings of the driver’s
car is done by increasing angle  and increasing distance d.
When using this camera based VBEP design tool, the engi-
neer does not have to think about angles and distances, but
just translates and rotates the camera such that all important
objects are visible in the VBEP.
For visualizing the trac situations of Section 4, we
found that translating the camera position in the negative z
direction as well as in the positive y direction leads to the
best overview on the current trac situation. Additionally the
camera has to be rotated clockwise around the x-axis. These
three modifications enable us to present information in the
farer future best. Consequently in our presented visualization
concept, all information which is required by the driver is vi-
sualized from this special virtual bird-eye perspective.
Summarizing our line of thought thus far: It is currently
much speculated that an egocentric presentation of informa-
tion within a driver’s field of view (e.g. contact-analog pre-
sentations in a head-up display) may be of great benefit for
ADAS. We here object that a purely egocentric presentation
is not suitable for supporting driver anticipation of events in
the further future since such presentation needs to show ob-
jects that cannot yet be seen. In our visualization concept for
concealed objects, all information for driver anticipation of
deceleration maneuvers is therefore visualized from a virtual
bird-eye perspective (VBEP). It still adheres to an egocentric
reference frame. Yet, it is capable of adding a well-dosed
amount of overview support, thereby strengthening driver an-
ticipation.
1 We use a right handed coordinate system, the y-axis is oriented
upwards, the z-axis goes into the projection plane and the x-axis goes
to the left.

VISUALIZATION CONCEPT
The VBEP is a powerful concept to draw the driver’s at-
tention to situations in the further future. On the basis of this
concept we developed visualizations of trac situations from
Section 4, and integrated them in the digital instrument cluster
between the speedometer and the revolution counter.
We integrated an abstract version of the VBEP. The visu-
alization diers from the one used for the VBEP design tool
(Fig. 12). It is more schematic and less realistic to account for
the fact that the visualizations will be presented to the driver
during car motion. In the visualization concepts we tried to
focus on the essential; nevertheless we tried to visualize all
information which is necessary for the driver’s correct antici-
pation.
We use only a very schematic representation of the road.
The update of the trac situation can either be done contin-
uously or in discrete steps (e.g. 1 second). Although the real
road has some horizontal or vertical curves, the road in the
visualization remains straight. Due to the fact that a more de-
tailed course of the road would make the visualization more
dynamic, we decided to exclude the course of the road in our
first visualizations. Below the road the directions of both
lanes are shown. Whereas on a highway the driver’s lane
and the lane left of the driver’s lane are most important for
a proper anticipation, on urban and rural roads the driver’s
lane and the lane of the opposite direction are important for a
proper anticipation. Consequently, on highways, the left lane
is in the driver’s direction and on all other roads the left lane
is in the opposite direction.
The driver’s car is visualized in white and has a static
position on the road; it is always at the lower end of the right
lane. This eect is just a consequence of the VBEP approach
- the scene is shown from a camera which is tethered rigidly
to the driver’s car. Other cars are visualized in orange, with
dierent symbols for oncoming and driving ahead cars. Ad-
ditionally we included a symbol for smart deceleration in our
VBEP which indicates the point in time when the smart de-
celeration has to be started. The deceleration phase itself is
marked by the orange bar at the right side of the road. In most
cases at end of the deceleration phase an orange situation-
specific symbol is included in the visualization. The start
and end of the deceleration phase is discretized in steps of
1 second. Consequently the visualization enables the driver
to get information for the next n seconds, the distance which
is shown in this visualization depends on the driver’s current
speed.
As a consequence of this visualization, symbols automat-
ically become larger the more important (closer) they become
for the driver. The fading out of the symbols can be delayed
as long as the driver has not reacted on the symbol.
Speed limitation ahead
The speed limitations to 50 km/h and 100 km/h are shown
in the visualization by an additional speed sign. In addition
to the monochrome speed signs we developed a similar visu-
alization with the original colors as well. The two situations
(rural, see Fig. 13 and highway, see Fig. 14) dier in the
start speed and the end speed. As a consequence the start
point of the deceleration phase and the length of the decel-
eration phase are dierent. Furthermore the direction of the
left lane is dierent because the first situation is on a rural
road whereas the second situation is on a highway. In this
situation, however, the direction of the left lane is not that
important.
Figure 13. VBEP of speed sign ahead on a rural road
Figure 14. VBEP of speed sign ahead on a highway
Construction site ahead
The construction site is indicated by three street posts.
All other situations are described by the combination of other
cars and trac signs. For this situation we additionally need
an abstract visualization for an obstacle on the road which is
not a car. Therefore we decided to use these street posts as
a metaphor for these kinds of obstacles. Moreover the com-
plete oncoming trac is shown in the visualization in Figure
15. The deceleration symbol indicates the best start of the
deceleration phase according to the oncoming trac and the
construction site.
Parking vehicle ahead
The parking vehicle as well as the oncoming trac is
visualized in orange, whereas the driver’s car is visualized in
white. The speed of the parking car (which is zero of course)
has not been included in the visualization, as shown in Fig-
ure 16. We decided to show the fact that the car is standing
still by its orange warning lights instead. Again, this is just a
metaphor and even if the parking car has not warning lights
in reality they are included in our visualization. The length of
the deceleration phase, however, indicates that the dierence
between the driver’s speed and the speed of the car ahead is
rather large.
Figure 15. VBEP of construction site ahead on a rural road, with
trac in the opposite lane

Figure 16. VBEP of parking vehicle ahead on an urban road, with
the time until further movement of the obstacle unknown
Figure 17. VBEP of red trac light ahead on an urban road, with
the time until change of the light unknown
Red trac light ahead
The major information which has to be communicated to
the driver is the fact that a car is standing at the trac light,
as shown in Figure 17. The information that the car is rather
slow or is even standing can be deduced from the length of
the deceleration phase. Additionally it is important for the
driver to know that overtaking does not make any sense in
this situation because the reason for the stand still of the car
ahead is the red trac light. Therefore we just included a
simple monochrome symbol for the red trac light. Again,
we do not show the crossway itself to the driver. As an alter-
native we developed a visualization with the original colors
of a trac sign.
Jam ahead
The jam ahead of the driver’s car is indicated by the visu-
alization of all cars in this jam. There are two ways to distin-
guish trac which has come to a complete halt (see Fig. 18)
from slow-moving trac (see Fig. 19): Again, the length of
the deceleration phase is larger if the cars are standing than if
the cars are still moving. Additionally the two situations can
be distinguished by the signs: When the cars are standing the
trac jam sign is shown whereas when the cars are moving
their speed is shown. In both trac jams the low speed of the
cars ahead is indicated by the warning lights.
Slower driving vehicle ahead
There are two situations which lead to problems when a
slower driving vehicle is in front of the driver’s car: A slower
driving vehicle in a zone where overtaking is prohibited, and
a slower driving vehicle while encountering oncoming trac.
When overtaking is prohibited this information is presented
to the driver by a monochrome symbol (see Fig. 20). Again,
another alternative with the original colored symbol has been
developed as well. When the oncoming trac makes the over-
taking impossible, this oncoming trac is shown in the visu-
alization (see Fig. 21). The speed of the slower driving vehi-
cle can be estimated by the length of the deceleration phase.
Figure 18. VBEP of trac jam ahead on a highway
Figure 19. VBEP of slowly moving trac jam ahead on a highway
We want to mention that this visualization does not give any
overtaking suggestions even if overtaking is not prohibited
and no oncoming cars were detected.
HAPTIC ASSISTANCE
The proposed concept of visual assistance can be com-
bined with a haptic assistance, i.e. active accelerator pedal
(AAP) as proposed by (Reichart et al., 1998). In combination
with haptic assistance, the visualization explains the reason
for the proposed deceleration phase, whereas the AAP lets
the driver feel how deceleration process can be performed in
an ecient way. Without the visualization the driver is not
able to identify the reason for the deceleration recommenda-
tion given by the AAP, due to the fact that the reasons for the
smart deceleration are in the further future and might be not
visible from the driver’s egocentric perspective.
This active accelerator pedal applies resistance against
the driver’s foot, indicating that the driver should decelerate
in order to reach the upcoming lower speed which is being
anticipated by the system as an appropriate for considered
driving situations. If the driver decides to obey the haptic
suggestion, he will feel the resistance on the accelerator pedal
until he positions his foot such that exactly needed amount
of acceleration is produced for reaching and keeping the tar-
get speed. It should be mentioned that if the driver decides
Figure 20. VBEP oflower driving vehicle ahead on a rural road,
with overtaking being prohibited

Figure 21. VBEP of slower driving vehicle ahead on a rural road
with trac on the opposite lane
Figure 22. Force-path diagram for the active accelerator pedal (ac-
cording to Reichart et al., 1998)
to decelerate more strongly, i.e. gives less gas as the system
suggests, he also does not experience the resistance (see Fig.
22).
The driver is able in any case to press over the resis-
tance on the gas pedal, with that dismissing the advice of the
system. Activation of the haptic feedback suggesting to the
driver taking his foot o the accelerator pedal is done for all
described situations. Apart from these situations, the acceler-
ator pedal acts in the convenient manner.
CONCLUSIONS AND
FUTURE WORK
This paper presented investigations which perspective is
suitable for visualizing anticipatory driving information in the
farer future. Our current concept considers future to be a mat-
ter of time rather than a matter of space: The visualization
shows the part of the road that the car will drive across in the
next n seconds, depending on the current speed. Thus, the
road in our visualization concept is rather a time line than a
spatial map of the further course of the road. Of course there
is a close connection between what will happen in the near
future (matter of time) and what will happen in front of the
driver’s car (matter of space). Nevertheless some situations
are related to spatial aspects (e.g. a trac jam), some situa-
tions are related to temporal aspects (e.g. oncoming trac)
and some situations are related to time and space aspects (e.g.
overtaking a car with oncoming trac). In our further work
we want to take a closer look at the importance of space in the
support of the driver’s anticipation. It might help the driver
not just to know that a trac jam is 10 seconds away but that
it is behind the next curve in front of him. In order to keep
the visualization simple in the first step we did not include the
course of the road yet, as described in Chapter 6. We expect,
however, that in some situations the course of the road might
improve the driver’s cognition and anticipation.
The next step will be to conduct an evaluation of the pro-
posed visualization concepts. After a pre-evaluation for the
clarification of design, arrangement and structuring the eval-
uation will focus on the eect of anticipation support. We
expect that our visualization concept helps the driver to drive
more anticipatory than without any information on the further
future. Therefore within this evaluation the concept for visu-
alizing concealed objects to improve the driver’s anticipation
will be compared with driving without assistance. Addition-
ally the visual assistance will be compared to the combination
of visual and haptic assistance.
REFERENCES
Batavia, P., Pomerleau, D., & Thorpe, C. (1997, November). Over-
taking vehicle detection using implicit optical flow. In Intelligent
transportation system, itsc ’97 (pp. 729–734).
Bellotti, C., Bellotti, F., Gloria, A. D., Andreone, L., & Mariani,
M. (2004, June). Developing a near infrared based night vision
system. In Intelligent vehicles symposium (pp. 686–691).
Brooks, C., & Rakotonirainy, A. (2005, June). In-vehicle technolo-
gies, advanced driver assistance systems and driver distraction:
Research challenges. In International conference on driver dis-
traction. Sydney, Australia.
Dugarry, A. (2004). Advanced driver assistance systems informa-
tion management and presentation. Unpublished doctoral dis-
sertation, Cranfield University School of Engineering, Applied
Mathematics and Computing Group.
Gruyer, D., Rakotonirainy, A., & Vrignon, J. (2005, February). Ad-
vancement in advanced driving assistance systems tools: Inte-
grating vehicle dynamics, environmental perception and drivers’
behaviours to assess vigilance. In Intelligent vehicles and road
infrastructure conference. Melbourne, Victoria.
Ishida, S., & Gayko, J. (2004, June). Development, evaluation and
introduction of a lane keeping assistance system. In Intelligent
vehicles symposium (pp. 943–944).
Kato, H., & Billinghurst, M. (1999, October). Marker tracking and
hmd calibration for a video-based augmented reality conferenc-
ing system. In Proceedings of the 2nd international workshop on
augmented reality (iwar 99). San Francisco, USA.
Labayrade, R., Douret, J., & Aubert, D. (2006, September). A multi-
model lane detector that handles road singularities. In roceedings
of the ieee itsc 2006, 2006 ieee intelligent transportation systems
conference. Toronto, Canada.
Matuszyk, L., Zelinsky, A., Nilsson, L., & Rilbe, M. (2004, June).
Stereo panoramic vision for monitoring vehicle blind-spots. In
Intelligent vehicles symposium (pp. 31–36).
Milgram, P., & Colquhoun, H. (1999). A taxonomy of real and
virtual worlds display integration. In Mixed reality-merging real
and virtual worlds (pp. 1–16). Berlin.
Milgram, P., & Kishino, F. (1994). A taxonomy of mixed reality
visual display. In Ieice trans, on information and sstems (special
issue on networked reality) (pp. 1321–1329).
Moutarde, F., Bargeton, A., Herbin, A., & Chanussot, L. (2007,
June). Robust on-vehicle real-time visual detection of ameri-
can and european speed limit signs, with a modular trac signs
recognition system. In Intelligent vehicles symposium (pp. 1122–
1126).
Rakotonirainy, A. (2003). Human-computer interactions: Research
challenges for in-vehicle technology. In Proceedings 2003 road
safety research, policing and education conference (pp. 214–
219). Sydney.
Reichart, G., Friedmann, S., Dorrer, C., Rieker, H., Drechsel, E.,
& Wermuth, G. (1998). Potentials of bmw driver assistance to
improve fuel economy. In 27th fisita congress. Paris.
Tamura, M., Inoue, H., Watanabe, T., & Maruko, N. (2001). Re-
search on a brake assist system with a preview function. In
Sae 2001 world congress, session: Human factors in automotive
telematics.
Thoma, S., Klinker, G., & Lindberg, T. (2007, October). Speed
Recommendations During Trac Light Approach: A Compari-
son of Dierent Display Concepts. In Proceedings of the human
factors and ergonomics society europe chapter annual meeting.
Braunschweig, Germany.