Available online at www.sciencedirect.

com

Procedia - Social and Behavioral Sciences 53 (2012) 732 – 741

SIIV - 5th Internationnal Congress - Sustainabiliity of Road Innfrastructuress

Road Sigh
ht Design and
a Driver Eye
E Heightt: Experimeental Measu
ures

F
Francesco Saverio Capaaldoa*
a
Departmennt of Transportation
n Engineering «Luig
gi Tocchetti», University Federico II, 880125 Naples, Italyy

Abstractt

The roadd design Rules offten refer to information, values anda methodologiees recommendedd in bibliographicc references: the
driver eyye height from ground
g is one of these. In this work
w a series of height
h values waas obtained by crross pictures of
travellingg cars; a second data set comes frrom scale sketchees of passenger cars.
c The results, even though parrtial, have made
possible to study sides of the road safety due
d to real values of the driver eyee height. This is oone of the parameeters to estimate
the minim mum radius in veertical crest and marginally
m be con
nsidered also to check
c the minimuum lateral distance of an obstacle
to visibility in horizontal curves.
c

© 2012
© T Authors.
The
2012 The Publishedby
Authors. Published by Elsevier
ElsevieerLtd.
Ltd. Selection and/or peer-review under responsibility of SIIV2012
Selectionn Committee
Scientific and/or peer-revi
iew
Open under
access respons
under CCsibility of SIIV20
BY-NC-ND 012 Scientific Com
license. mmittee
Keywords: Driver eye height;; two-lane highway
ys; experimental measure; sight distancces; road safety.

1. Introoduction

Without adequate surveys,

s the national
n rules for
f road designn, often refer to information n, values and
methodo ologies suggesteed in well consoolidated biblioggraphic referencces: the driver eeye height from ground is one
of thesee values. Whilee the world perriodically check ks the variation of this parameeter in Italy itss value resists,
unmovin ng, since more than thirty years despite the technical evollution, in time, of the charactteristics of the
nationall vehicular fleet and the changees of the driver population.
p
The road
r design ten nds to avoid siituations that could
c lead to driver
d behavior undecided. Th his principle is
applied in the design standards in mostm Countries: the standards are based, parrtially, on technical data, on
anthropoometric and on behavioral dataa. All these dataa change in spacce (between diffferent Countriess) and over the
years. Without
W adequatte domestic survveys often referrs the driver eyye height with tthe values showwn in Highway
Capacityy Manual [1] annd in AASTHO Green Book [2]. These two exxcellent texts aree built basically y on researches
related to
t U.S. reality: localized surveeys are always needed to knoow which approoximation can be b adopted for

* Corrresponding author. Tel.:

T +39.81.768394
42; fax: +39.81.768
83946.
E-maill address: fcapaldo@unina.it

1877-0428 © 2012 The Authors. Published by Elsevier Ltd. Selection and/or peer-review under responsibility of SIIV2012 Scientific Committee
Open access under CC BY-NC-ND license. doi:10.1016/j.sbspro.2012.09.923
 Francesco Saverio Capaldo / Procedia - Social and Behavioral Sciences 53 (2012) 732 – 741 733

those parameters measured in different environmental situations. Specifically, in Italy, the height of the driver
eye has already been considered by the Standards of 1973 (Consiglio Nazionale delle Ricerche CNR [3]) with a
value of 1,10 m (approximation and rounding up of old U.S. measure of 3,5 ft). Other countries have adopted a
set of measures which differ up to 20%.
In previous papers [4 and 5] has been carried out experimental investigations in order to determine a reliable
value for the driver eye height. The aim of this paper is to submit recent surveys for a significant percentage of
passenger cars and to determine the statistical parameters for measured values. This new analysis showed that
value of driver eye height is different from the value fixed in the Italian Rules and how this difference has
changed over the years 2003 and 2010.
The measure of the driver eye height depends on the model of passenger cars, the average height of the driver
and the driving behavior. Also, and main thing, the driver tends to sit into the vehicle in order to ensure the
maximum comfort [6]. Not all drivers with the same height take the same position in the car and this position
also varies between drivers of different sex. Finally, increasingly accurate and custom controls for driving seats
(longitudinal positioning, vertical positioning, backrest and headrest) can vary the influence of parameter value
for a given type of driver.
The attempt to define the driver eye height through population surveys (heights, anthropometric
measurements, typologies) has not led to concrete results [4 and 5].
Now the probable values of the driver eye height have been studied with three different procedures:
x the first procedure has made it possible to obtain a distribution of values of the driver eye height from the
ground on a sample of vehicular fleet in the Campania region through examination of cross pictures taken
of running vehicles;
x the second procedure has made it possible to get a database of comparable values for a different sample of
vehicles and drivers starting from the measures of scale sketches for a number of passenger cars taken from
a widely magazine (Quattroruote [7]);
x the third procedure, marginal but significant, has allowed to obtain a driver's eye height value with a large
sample of heights of passenger cars produced from 2004 to 2011.
Data analysis showed that the average height of driver eye, as well as the 15th percentile of the data
distribution, is greater than the value indicated by Italian Standards [8].

2. Visibility on vertical crest and on horizontal curves

In design of the vertical alignment should be inserted a vertical crest between the different slopes, to ensure
regularity of the gear, safety and comfort. Compliance with the current Italian road [8] requires vertical crests of
a parabolic type (with vertical axis and minimum curvature radius in the vertex point) between the slopes: these
crest types are most suitable compared to those circular previous [9] to ensure uniform operation for vehicles that
running at constant speed (vertical acceleration and uniform gradient per length unit).
The traffic and road safety on vertical crest is related always to have sufficient visibility (D) for the driver, this
visibility is generically called safety sight distance, and is used to make a stopping maneuver or an overtaking
maneuver (on two-lane highways).
For vertical crests the minimum radius of the osculating circle, in the vertex point of the parabola, is linked to
stopping sight distances (Da) or overtaking distance (Ds) according to different next relationships.

'i ˜ D 2
L (1)

2 ˜ h1  h2  2 h1h2 
D2
R (2)

2 ˜ h1  h2  2 h1h2 
734 Francesco Saverio Capaldo / Procedia - Social and Behavioral Sciences 53 (2012) 732 – 741

where: L = vertical crest length; D = safety sight distance (stopping sight distances or overtaking distance); 'i =
value of the gradient difference (percentage) for two slopes; h1 = driver eye height; h2 = fixed obstacle height for
one-way roadways with D=Da; eye height of the driver that running in the opposite direction for two-lane
highways, with D=Ds.
If the length (L) of vertical crest is greater than the safety sight distance (D) (Figure 1) this length should be
calculated with the relationship (1). The minimum radius in the vertex point (given about L = R 'i) can
calculated with the relation (2) and radius is, of course, not dependent from gradient difference percentage for
slopes.

Figure 1 - Vertical crest (L>D) Figure 2 - Vertical crest (L<D)

If the length (L) of the vertical crest is less than the safety sight distance (D) (Figure 2) this should be
calculated with the relation (3). The minimum radius is calculated with the relation (4) and depends from gradient
difference percentage for slopes.
For the Italian Rules (MIT 2002) the values of h1 and h2 are h1=1,10 m and h2=0,10 m in case of Stop (Da) or
h2=1,10 m in case of allowed passing maneuver (Ds) for two-lane highways.

§ h  h  2 h1h2 ·
L 2˜¨D  1 2 ¸ (3)
¨ 'i ¸
© ¹
2 § h  h  2 h1h2 ·
R ˜¨D  1 2 ¸
'i ¨ 'i ¸
© ¹ (4)

To ensure that there exists at least the minimum sight distance in a plan circular curve should be checked that
the side obstacles (Figure 3) are at a distance from the inside lane axis at least equal to:

§ D ·
' U ¨¨1  cos ¸
© 2 U ¸¹
(5)

where: '= lateral obstacles distance from the inside lane axis; D = safety sight distance (Da or Ds); U= circular
radius of the inside lane axis.
The distance (D) must be ensured at height equal to the driver eye height. In case of obstacles due to hedges,
walls, crash barrier or other mainly vertical this (') must be always guaranteed from obstacle to the axis of the
inside lane. If the sideline limits of the obstacle are not vertical, as in road trenches, it can «save» about a meter
of the whole distance, taking into account the angle of the obstacle limit with horizontal: the saving ('1) is a
function of slope angle and of the driver eye height according to the relation (Figure 4):
 Francesco Saverio Capaldo / Procedia - Social and Behavioral Sciences 53 (2012) 732 – 741 735

h1
'1 (6)
tgE

where: '= additional distance from obstacle to inside lane edge; h1 = driver eye height; E= angle of side slope
with horizontal.

Figure 3 - Sight lane axis in circular curve Figure 4 - Sight distance in road trench

3. Surveys

Three different procedures have been used for the data acquisition of the driver eye height. It must be said also
that were made two kinds of measures: the first is the driver eye height and the second is the height of the
headlight centroids. In this paper consider only the first kind of measures.
A first series of experimental samples of driver eye heights was made by cross pictures of running passenger
cars (Figure 5). On these pictures was measured, the driver eye height from ground (h1), and known the vehicle
height (hV) declared by the producer, was calculated ratio of driver eye height and the car height (Rh= h1/hV).
Finally was measured the headlight heights from the ground (h3). The class of vehicle displacement has been
defined properly but roughly (many cars have same body but different engines). The ratio between the two
heights (Rh) was utilized to obtain a sufficiently probable values of driver eye heights when are known the
passenger car heights. The total sample was approximately 200 passenger cars of which 35% driven by women
and 65% driven by men.
The significance of picture number for surveys sample was valued by the confidence interval of the average of
the Gaussian random variable at 95% with a probable error of 1,25 cm.
Scale sketchs of the all cars tested in the year 2007 by Quattroruote magazine [7] were utilized for a second
data set. Even in these sketchs were recorded the heights measures of similar features (see Figure 6). In this case,
the class of vehicle displacement was sure known. All sample scale sketch for the year 2007 was about 70
passenger cars.
The heights of the cars produced between 2004 and 2011 (for a total of more than 500 values, as producer
declared) were collected in a representative sample. Even the car heights were split according to class of engine
displacement. The 0-800 cc class was not considered: the few vehicles with this displacement were joined to the
next class (801-1200).
736 Francesco Saverio Capaldo / Procedia - Social and Behavioral Sciences 53 (2012) 732 – 741

Figure 5 - Sample picture Figure 6 - Sample scale sketch

4. Data Analysis

Table 1 shows the characteristic values of the distributions of driver eye height (cm) obtained from real drivers
(men, women and total) and from those scale sketches taken by widely magazine. The values taken by the
magazine are obtained from the drawing position of dummies in cars, and for this reason have not been counted
with the other recorded values but kept separately as control. In the following figures (Figures 7, 8 and 9) shows
the trends of the distributions for driver men, women and the total of surveys combined with the normal
distributions with equal average and Standard Deviation. In the Figure 10 shows the distributions of eye height
from scale sketches.

Table 1. Characteristic distribution values (height in cm)

Characteristic values Men Women Total Sketches

Average 126,54 122,53 125,19 132,11
Standard Dev. Pop. 9,39 6,81 8,81 10,71
Standard Dev. Sample 9,43 6,86 8,84 10,81
Coefficient of Variation 0,07 0,06 0,07 0,08
th
10 Percentile 117,09 116,01 116,44 120,35
15th Percentile 118,40 116,85 117,81 123,62

Table 2. Percentages of passenger car fleet (year 2010) and survey

Engine Displ. Class (c.c.) Campania Italy Survey

0 - 800 6,66 4,26 3,68
801 - 1200 26,82 22,49 26,84
1201 - 1600 37,95 41,63 40,00
1601 - 2000 23,27 24,23 18,42
2001 - 2500 3,67 4,28 4,74
> 2500 1,59 3,08 6,32
not identified 0,04 0,03 0,00
 Francesco Saverio Capaldo / Procedia - Social and Behavioral Sciences 53 (2012) 732 – 741 737

The very high cars (like Sport Utility Vehicles - SUVs -, or large van) have excluded from total results. These
cars were recorded for approximately 20% of total sample. An approximate evaluation of these few values,
indicates an average of driver eye height of 144 cm, a Standard Deviation of about 6,3 and a Coefficient of
Variation of 0,04. This type of passenger cars has let record an average of car height of 172 cm and a Standard
Deviation of 6,9.
All samples were split in percentages according to the engine displacement sample classes and these were
compared to the percentages of the passenger car fleet in the Region Campania and in Italy (year 2010, [10]) and
are shown in Table 2 and Figure 11.
The analysis of the ratio between the passenger car height and the driver eye height (Rh) for all the vehicle
surveys has provided an Average value of 0,83 with a Standard Deviation of 0,02 and Coefficient of Variation of
0,03. These results allow to consider the passenger car height a good parameter to obtain the driver eye height
from ground (Table 3).

Driver eye height Men Driver eye height Women

0,50 0,500
0,45 0,450
0,40 0,400
0,35 0,350
0,30 0,300
0,25 0,250
0,20 0,200
0,15 0,150
0,10 0,100
0,05 0,050
0,00 0,000
100 105 110 115 120 125 130 135 140 145 150 155 100 105 110 115 120 125 130 135 140 145 150 155

Figure 7 - Driver eye height men (cm) Figure 8 - Driver eye height women (cm)

Driver eye height Total Driver eye height from Scale Sketches

0,50 0,50
0,45 0,45
0,40 0,40
0,35 0,35
0,30 0,30
0,25 0,25
0,20 0,20
0,15 0,15
0,10 0,10
0,05 0,05
0,00 0,00
100 105 110 115 120 125 130 135 140 145 150 155 100 105 110 115 120 125 130 135 140 145 150 155

Figure 9 - Driver eye height total (cm) Figure 10 - Driver eye height from Scale Sketches

In Figure 12 shows a proportional sketch with the limits of Rh (average and 10/15th percentile, lower) for the
men eye height distribution.
The percentages of vehicles split for engine displacement classes and utilized in the analysis were quite
738 Francesco Saverio Capaldo / Procedia - Social and Behavioral Sciences 53 (2012) 732 – 741

similar to those classes of vehicles running in Campania and very close to the percentages of vehicles running in
Italy. The main difference in the test samples regards the class 1601 - 2000 c.c.: not greater than about 5% in
Campania Region and not greater than about 6% in . The sample survey percentage is always greater than
running car fleet (4,7% and 3,2%) for vehicles with a displacement class over 2500 c.c. But vehicles with this
displacements class were those with small height compared to that of the numerically dominant classes and
considering their weight in the final evaluation, the error is in behalf of safety.

Table 3. Characteristic distribution values for Rh

Characteristic values Men Women Sketches Total

Average 0,82 0,81 0,85 0,83
Standard Dev. Pop. 0,02 0,02 0,01 0,02
Standard Dev. Sample 0,02 0,02 0,01 0,02
Coefficient of Variation 0,03 0,02 0,02 0,03
10thPercentile 0,80 0,78 0,83 0,80
15th Percentile 0,80 0,79 0,83 0,80

Table 4. Characteristic distribution values for sample of passenger car heights from 2004 to 2011 (in cm)

Engine Displ. Class (c.c.) 0-1200 1200-1600 1600-2000 2000-2500 > 2500 Total
Average 150,13 159,42 154,45 158,43 151,50 155,57
Standard Dev. Sample 5,75 97,10 12,58 15,96 19,81 57,92
Coefficient of Variation 0,04 0,61 0,08 0,10 0,13 0,37
10thPercentile 144,00 144,60 142,10 141,70 130,00 140,37
15th Percentile 146,75 146,10 142,40 142,50 131,52 142,01
Total Values 26 178 146 50 128 528

The Table 4 shows the characteristic distribution values for the all sample of passenger car heights from the
years 2004 to 2011. The maximum value of average height is for the class of 1200-1600 cc. In this class are
many family cars with more than four seats. Also the dispersion values are the highest of the whole sample set
(Standard Dev. Sample 97,10 and Coefficient of Variation 0,61). At last for the total of 528 passenger cars the
characteristic values record an average of about 156 cm and a 15th percentile of about 142 cm. This value
multiplied for the average value of Rh (Total, fifth column of Table 3) gives a result of about 118 cm.

5. Results and discussions

It is correct to choose the lowest value in Table 1 as the 15th percentile of the height distributions: 117 cm and
round it down to 115 cm in order to ensure at least 85% of drivers adequate safety sight distance.
Other Countries have Rules with different values of driver eye height (see Table 5) and the localized surveys
realized in time have shown sometimes contrasting trends for the reference value. Also the same correction of the
values found [11, 12 and 13] was different from Country to Country. The United States have agreed to correct
their Rules for a single centimeter of variation of the parameter value; Australia has done so for an larger order of
magnitude (10 cm) rounding down the reference value.
 Francesco Saverio Capaldo / Procedia - Social and Behavioral Sciences 53 (2012) 732 – 741 739

The values take by the magazine sketches are again slightly increased than those carried out by the pictures of
running vehicles also when compared to previous works [4 and 5]. The reference values (15th percentile) is
increased from 118 cm to 124 cm about for the scale sketches and from 116 cm to about 118 cm by measures
taken from pictures.

Car Fleet and Engine Displacements

50
45 Campania
40 Italy
Percentages

35 Survey
30
25
20
15
10
5
-

not identified
801 - 1200

> 2500
1201 - 1600

1601 - 2000

2001 - 2500
0 - 800

Figure 11 - Car fleet and engine displacements (cc) Figure 12 - Men Average and 10/15th Percentile (lower)

For the results extracted from the pictures of the running vehicles the changes of the value measured during
five years (less than 1%) is sure not very significant and may be due to variability of the random sample.
The percentiles selected from the height distributions differ from the value of the Italian rules [8] from 13% to
6% respectively.
Using the relation (2), for example, it is possible to calculate the percentage changes of the minimum radii of
the osculating circles for a vertical crest due to different values of the driver eye height. For a height of the fixed
obstacle (h2) of 0,10 m and a eye height (h1) of 1,24 m and 1,18 m the minimum radii are, respectively, lower
than 9% and 5% of those calculated with the value of Rules. These percentages are increases in order to
calculation of the minimum radius required to allow the passing maneuver.
On the other hand keeping the minimum radii calculated with the standard driver eye height can have
available a safety sight distance greater than 4,7% and 2,4% (for h1=1,24 m and 1,18 m).
For the increase of the safety sight on circular curve can be simply to point out that the term '1 is directly
related to h1 according to 1/tgE (relation 6).

6. Critical considerations and further developments

This paper has shown the results of a recent survey on the driver eye height from the ground. The survey was
conducted in the Campania Region (Italy). These results are still the only ones available in Italy where the rules,
[8] continue on keeping unchanged the value since the years seventies [3] despite changes in the passenger car
fleet and driver behavior.
The minimum actual value compared to the current standards is greater than about 6%: this value has shown a
trend to increase in the year, (trend to check and study even more). Using the result values obtained instead that
the standard values can lead to sizing vertical crests with smaller minimum radii. On the other hand the radii
calculated according to the current rules allow the driver to have safety distances slightly greater.
740 Francesco Saverio Capaldo / Procedia - Social and Behavioral Sciences 53 (2012) 732 – 741

The designer has to control always the coordination between the horizontal and vertical alignments [14, 8 and
15] and so the radii of vertical connection curves are usually higher than the minimum value.
The value of the ratio between the eye height and the passenger car height (0,83) must be considered among
the obtained results. This result (and parameters show in Table 3) is a just more accurate than that obtained in a
previous work [16] and shows that drivers, within certain limits and regardless of their stature, tend to sit in car
quite similar (height of the head compared to the car roof), using the seat controls available. They, then, seek to
maintain, during driving, the eyes to about 1/3 of the vertical projection of the front windscreen from the car roof.
As first approximation this result allows to obtain a reliable value of the driver eye height just by knowing the
declared car height. For example multiplying the value of 15th percentile of the passenger car height detected by
the pictures (142 cm) for the coefficient found it is possible to get a driver eye height value of 117,9 cm: this
value is comparable to the corresponding percentile in the third column of Table 1 (Total).
The survey was also conducted with car engine displacement classes. The results obtained could have been
weighed according to the vehicular fleet in the region or in Italy. But considering the «similarity» between the
diagrams shown in Figure 11, the results will be not very different from those reported in the third column of
Table 1.
Has used sketches drawn from Quattroruote magazine. These are complete sketches that sometimes had
needed of scale corrections for distortions due to editing reasons. Otherwise it may also use the scale vector
sketches of over 5000 cars that CCVision [17] produces for different purposes. These, however, do not have the
drawings of the dummies on board but it is the most complete collection of vector drawing for commercial
vehicles in production (front, rear, side and top view).
These results suggest to accept, at least for now and without further studies, the standard current value of
driver eye height without special reservations for two reasons:
x the little savings that could be obtained, in the road construction, with adjusting the height value to the
reality of running vehicles;
x the increase driving safety due to a sight distance available greater than that strictly necessary (increase real
but not much significant in percentage).
Also the increase of sight distance in circular curve for road trenches is directly proportional to h1 according to
constant 1/tgE. If the trench has a typical slope gradient of 1/1 (45°) the difference of additional side visibility
compared to that due to the road section adopted is slightly less than 3%.
At the end it is important underline that the suggested percentile value is a «safe value» because all higher cars
(SUVs and large van) were shut out from calculation.
The found upward trend of the parameter «height of eye», then, advises to monitor the value currently
measured through surveys of more large samples. Also with the correction factors used in the search, the
Campania is not the world, and it is not even Italy. It would be more desirable an exchange of information and
data among Italian researchers and above all with researchers of different Countries.

References

[1] HCM (2010). Highway Capacity Manual, TRB Special Report 209. Transportation Research Board, Washington, D.C., 2011.
[2] AASHTO (2004). A Policy on Geometric Design of Highways and Street. AASHTO, Washington, D.C., 2004.
[3] CNR (1973). Norme sulle caratteristiche geometriche delle strade. Consiglio Nazionale delle Ricerche CNR, Bollettino Ufficiale n. 31,
Roma, 1973.
[4] Capaldo F. S. (2003). Determinazione sperimentale dell'altezza dell'occhio del conducente ed implicazioni sulla sicurezza della
circolazione. In S.I.I.V. 2003, XIII Convegno Nazionale, Padova 30-31 ottobre 2003.
[5] Capaldo F. S. (2004). Driver's Eye Height: an Experimental Localized Analysis. In II PhD CivilExpo - 29-30 January 2004, Budapest.
[6] Parkin S., Mackay G. M., Cooper A. (1995). How Drivers Sit in Cars. Accid. Anal. and Prev., Vol. 27, No 6, pp. 777-783, Pergamon,
1995.
[7] Quattroruote (2007). Ed. Domus, Milano, 2007.
 Francesco Saverio Capaldo / Procedia - Social and Behavioral Sciences 53 (2012) 732 – 741 741

[8] MIT (2002). Norme funzionali e geometriche per la costruzione delle strade. Ministero delle Infrastrutture e dei Trasporti, Ispettorato
generale per la circolazione e la sicurezza stradale, Decreto Ministeriale 5 novembre 2001, S.O. n. 5, G.U.R.I. n. 3 del 4.01.02.
[9] CNR (1980). Norme sulle caratteristiche geometriche delle strade extraurbane. CNR, Bollettino Ufficiale n. 78, Roma, 1980.
[10] ACI (2011). Annuario Statistico 2011. ACI - Automobile Club d'Italia - Sistema Statistico Nazionale, ACI, Roma, 2011
<http://www.aci.it/sezione-istituzionale/studi-e-ricerche/dati-e-statistiche/annuario-statistico-2011.html> (last Mar. 2012).
[11] Sivak M., Flannagan M. J., Budnik E. A., Flannagan C. C., Kojima S. (1996). The Locations of Headlamps and Driver Eye Positions in
Vehicles Sold in the U.S.A.. Report No. UMTRI-96-36, The University of Michigan Transportation Research Institute, November 1996.
[12] Fambro D. B., Fitzpatrick K., Koppa R. J. (2000). New Stopping Sight Distance Model for Use in Highway Geometric Design, Design
Speed, Operating Speed, and Sight Distance Issues. TRR No 1701, Transportation Research Board, Washington, D. C., 2000.
[13] Lamm R., Psarianos B., Mailander T. (1999). Highway Design and Traffic Safety Engineering Handbook. McGraw Hill ed., 1999.
[14] Ferrari P., Giannini F. (1996). Ingegneria stradale Vol. 1, Geometria e progetto di strade, ISEDI ed., 1996.
[15] Capaldo F. S., Guzzo F. (2011). Vademecum dell’Esperto in Infortunistica del Traffico e della Circolazione Stradale. EPC Libri ed.,
Roma, 2011.
[16] Capaldo F. S. (2012). Driver Eye Height: Experimental Determination and Implications on Sight Distances. In Procedia - Social and
Behavioral Sciences, Volume 43, 2012, Pages 375-383, 8th International Conference on Traffic and Transportation Studies (ICTTS 2012).
[17] Creativ Collection Vision, (2011) CAR-SPECIAL V.18, <www.ccvision.de/en/car-special/index.html> (last Mar. 2012).