!
2
Preface
This report is about the first part of a PhD research which focused on the needs of the driver with
respect to integrated driver assistance. By means of an Internet questionnaire, car drivers could
indicate their needs for support from the car during certain driving tasks and situations. These needs
served as a basis for creating an integrated driver support system, the so-called “Congestion
Assistant”. The impacts of this system on the driver and the traffic flow will be investigated by means
of a driving simulator experiment and a traffic simulation study in the remains of the research. The
PhD project forms a part of the research program of knowledge centre Applications of Integrated
Driver Assistance (AIDA). The AIDA knowledge centre has been realised by TNO and the University
of Twente. This PhD research is also part of the research program Transitions towards Sustainable
Mobility (TRANSUMO). It belongs to the cluster Traffic Management and the project Intelligent
Vehicles (sub project I.8).
We would like to take the opportunity to thank a few people for their contribution to the user needs
survey. We are grateful to Wytse Mensonides for programming the paper version of the questionnaire
into an Internet version. We wish to thank Edouard Buning from market agency RM Interactive for the
nice cooperation with respect to gathering participants for the user needs survey from their Internet
panel. Thanks to Marloes Verhoeven for sending the invitation to the participants of her survey. The
user group of AIDA, and in particular TNO and ANWB, are gratefully thanked for helping with
inviting participants for the survey as well. Dr. Klaas Poortema is thanked for his helpful comments
regarding statistical analysis. Last but not least, thanks to all the respondents who filled in the
questionnaire!
August 2005
Cornelie van Driel
Bart van Arem
3
4
Contents
Summary ................................................................................................................................................. 7
1.
Introduction ..................................................................................................................................... 9
2.
Theoretical framework .................................................................................................................. 11
2.1
Market research.................................................................................................................... 11
2.2
Earlier research .................................................................................................................... 11
3.
User needs survey.......................................................................................................................... 17
3.1
Focus.................................................................................................................................... 17
3.2
Target group and sample size............................................................................................... 17
3.3
Questionnaire design............................................................................................................ 18
3.4
Invitation of participants ...................................................................................................... 20
3.5
Pilot tests.............................................................................................................................. 21
4.
Results ........................................................................................................................................... 23
4.1
Sample ................................................................................................................................. 23
4.2
Driver support functions ...................................................................................................... 24
4.3
Ideal driver support system.................................................................................................. 27
4.4
Influence of sample characteristics ...................................................................................... 31
5.
Discussion ..................................................................................................................................... 39
5.1
Implications for integrated driver assistance........................................................................ 39
5.2
Contribution to conceptual model........................................................................................ 39
5.3
Conducting the survey ......................................................................................................... 41
6.
Conclusions and future work......................................................................................................... 43
6.1
Conclusions.......................................................................................................................... 43
6.2
Future work.......................................................................................................................... 43
References ............................................................................................................................................. 45
Appendix A – The questionnaire........................................................................................................... 49
Appendix B – The AIDA user group .................................................................................................... 67
Appendix C – Driver support functions: frequencies............................................................................ 69
Appendix D – McNemar test................................................................................................................. 99
Appendix E – Ideal driver support system: frequencies...................................................................... 103
Appendix F – Influence of driver characteristics ................................................................................ 107
5
6
Summary
Over the coming years, drivers will have an increasing variety of Intelligent Transport Systems (ITS)
at their disposal, including Advanced Driver Assistance Systems (ADAS). ADAS are in-vehicle
systems that support the driver with the driving task. These systems are expected to lead to a safer,
cleaner and more efficient and comfortable transport system. However, whether ADAS will meet
these high expectations is greatly dependent on the willingness of car drivers to use these systems. An
important question is to what extent drivers are eager to have ‘intelligent vehicles’. When several
ADAS can enter the vehicle, another question arises of which systems should be integrated and how
this should be done. This paper presents the results from a user needs survey that gives an answer to
these questions.
By means of an Internet questionnaire, more than 1000 Dutch car drivers indicated their needs for
driver assistance with certain driving tasks (e.g. congestion driving) and situations (e.g. driver fatigue).
It appeared that warnings for downstream traffic conditions and traffic in blind spots were favoured.
According to McNemar tests, the need for a warning for downstream traffic conditions on motorways
(90%) significantly differed from the need for other driver support functions (<84%). It was concluded
that this function was most popular of all. Moreover, the ideal driver support system should support
the driver with critical situations, such as an imminent crash and reduced visibility. From McNemar
tests, it appeared that drivers preferred the ideal system to help them with these situations and car
following (>51%) to other driving tasks and situations (<43%).
Characteristics of the driver, system and traffic scene affected the needs for driver support. Gender, for
example, appeared to have the biggest influence on the types of driver assistance wanted in the ideal
driver support system. Respondents would mainly like their cars to help them by giving information or
warnings. Automatic actions from the car were unpopular, except when driving in traffic jams for
example. Generally, drivers would like their cars to provide support in critical situations. Besides,
there was a great need for driver assistance on motorways. The needs of the driver indicated
consequences for the integration of driver assistance. Driver support functions should exchange
information to extend their individual fields of activity, for example by inter-vehicle communication
(e.g. warning for downstream traffic conditions) or sensor data fusion (e.g. warning for an imminent
crash). Furthermore, an integrated interface should be used that gives priority to important messages or
actions to prevent the driver from overload or confusion.
7
8
1.
Introduction
Modern societies are increasingly confronted with problems in traffic and transport, such as traffic
accidents, congestion and emissions. An important contribution to the solution of such problems lies in
Intelligent Transport Systems (ITS) (Ministry of Transport, 2004). Over the coming years, motorists
will have an increasing variety of ITS at their disposal, including Advanced Driver Assistance
Systems (ADAS). ADAS are in-vehicle systems that support the driver by partly or entirely taking
over the driving task. When more and more ADAS enter the vehicle, the question of how these
systems should be integrated, raises.
High expectations rest on ADAS (European Commission, 2002). Governments expect them to lead to
a more efficient, safer and cleaner transport system. Car industries and suppliers hope to contribute to
the safety and comfort of their customers by developing and selling distinctive driver support systems.
The driver that will use these systems faces the prospect of travelling in a safer, faster and more
comfortable way. However, whether ADAS will meet the high expectations, is greatly dependent on
the willingness of drivers to purchase and use the systems. An important question is to what extent
motorists are eager to have ‘intelligent vehicles’.
This report presents the results from a user needs survey that gives an answer to both questions. It
focuses on integrated driver assistance from the perspective of the driver. The outline of this report is
as follows. In chapter 2 a theoretical framework for the survey is presented. Chapter 3 discusses the
design of the user needs survey, based on this framework. The results from the survey are presented
and discussed in the chapters 4 and 5. Finally, conclusions are drawn and an outlook for further work
is highlighted.
The user needs survey is part of a research project into the assessment of integrated driver assistance
based on: (a) user needs, (b) impacts on the driver in terms of driver behaviour, workload and
acceptance and (c) impacts on the traffic flow in terms of safety and efficiency (Van Driel & Van
Arem, 2004). The results from the survey serve as a basis for creating an integrated driver support
system. The impacts of this system on the driver will be investigated by means of a driving simulator
experiment. These outcomes will be placed in perspective by assessing the impacts of the integrated
system on the traffic flow using microscopic traffic simulation. The project results will reflect
concepts of integrated driver assistance. It should become clear which driver support functions should
be integrated and which aspects of integration will be of importance, for example issues surrounding
the functional operation. An outline of the project is presented in figure 1.
User needs survey
Driving simulator experiment
Concepts of integrated driver assistance
Traffic simulation study
Figure 1. Project outline
9
10
2.
Theoretical framework
This theoretical framework consists of two parts. First, relevant literature from market research is
presented. This indicates which methods are frequently used to reveal (latent) needs of consumers.
Second, a state-of-the-art on user needs for ADAS is given. It shows which factors are of influence on
the needs for driver assistance.
2.1
Market research
Market research is about consumer behaviour: activities people undertake when obtaining, consuming
and disposing of products and services. The Consumer Decision Process (CDP) model in figure 2
shows this behaviour schematically (Blackwell et al., 2001).
1.
2.
3.
4.
5.
6.
7.
Need recognition
Search for information
Pre-purchase evaluation of alternatives
Purchase
Consumption
Post-consumption evaluation
Divestment
Figure 2. Consumer Decision Process (CDP) model (Blackwell et al., 2001)
The first step of this model – need recognition – plays an important role, when investigating to what
extent motorists would like to have driver support systems. Most of these systems are not yet available
on the market. Therefore, latent needs have to be revealed by reminding the consumer of a (potential)
need. The most frequently used method for this is questioning people. A distinction can be made
between qualitative and quantitative questioning methods. Qualitative methods, such as focus groups,
have a small-scale and exploratory character. Quantitative methods, such as surveys, are large-scale
and aim at presenting a phenomenon in numerical values.
When questioning people to reveal latent needs, Stated Preference (SP) approaches are often used
(Ortúzar, 2000). SP data provide insight into people’s choices in hypothetical situations. In general,
two options are possible: (a) compositional approach and (b) decompositional approach (Marchau,
2000). These approaches are used to determine the importance of attributes of alternatives. A
compositional approach involves that respondents indicate preferences for each attribute, such as price
and colour. In a decompositional approach – also known as conjoint analysis – respondents have to
indicate their overall preference for a hypothetical alternative described in terms of attribute levels, for
example ‘cheap and red’ versus ‘expensive and blue’. Respondents are forced to make trade-offs
among attributes. The overall preference is decomposed into the weights these respondents attach to
the separate attributes.
2.2
Earlier research
A lot of research has been performed to assess the needs for ADAS among potential users. Some
studies have examined this aspect after participants had limited or extended experience with the
system(s) in a driving simulator or in an actual vehicle on the road. Other studies focused on the
opinion of respondents who were not (yet) exposed to the ADAS. This section is about the latter group
of studies.
11
The explored studies differed in design and method. Some studies focused on one ADA system, while
most studies were about more ADAS. Anyhow, respondents had to give their opinions on hypothetical
forms of driver assistance. In this case, the word ‘opinion’ stands for a broad concept. It includes
aspects like attitudes, perceptions, preferences, needs and acceptability. The most frequently used
methods for investigating these opinions were focus groups and questionnaires. Almost all
questionnaires were set up according to the compositional SP approach. Adaptive Cruise Control
(ACC), Intelligent Speed Adaptation (ISA) and similar systems were most often the subject under
discussion.
The studied literature is discussed using a conceptual model that shows an overall picture of earlier
research into user needs for driver support systems; see figure 3. The opinion of the driver on driver
assistance is presented by arrow [1]. This opinion may be influenced by characteristics of the driver
and the system. Besides, it may be influenced by characteristics of the traffic scene; see arrow [2]. The
conceptual model is explained below by referring to the explored studies.
Traffic scene
•
•
•
•
Critical situation
Weather/visibility
Road type
Other traffic
Driver
•
•
•
•
•
•
•
•
•
•
•
•
Gender
Age
Nationality
Target group
Driving style
Driving experience
Familiarity with ADAS
Type of car possession
Psychological factors
Education
Family situation
Income
[2]
Driver assistance
[1]
•
•
•
•
•
System/function
Level of support
Feedback
Overrulability
Price
Vehicle
Figure 3. Conceptual model of user needs for driver assistance
Characteristics of the driver
The opinion of the driver on driver assistance may be influenced by characteristics of the driver. For
example, gender seems to have an effect on the needs for driver assistance. Generally, women were
more positive about help from their cars during driving than men (TRG, 1998; Chalmers, 2001).
However, Blythe & Curtis (2004) found that men had a more positive attitude than women thinking
ADAS should assist or take over instead of simply warn. Also the preference for typical ADAS can
differ between men and women. According to Piao et al. (2004) men liked ACC best and women ISA.
Also Rienstra & Rietveld (1996) found that women were more in favour of speed-regulating devices,
such as a speed limiter. However, in the COMUNICAR survey no differences with respect to gender
were found (Mariani et al., 2000). Also age seems to influence the needs for driver assistance. In
general, older drivers were more positive about driver support systems than younger drivers (CRA,
1998; TRG, 1998; Chalmers, 2001; Piao et al. 2004). There seems to be a relation between nationality
and driver assistance as well (Mariani et al., 2000). Compared to English motorists, Dutch and
Norwegian motorists were less positive about ISA (TRG, 2003). The IN-ARTE survey showed that
12
there was less demand for information and warning functions in Italy compared to Germany (Büsher
& Frese, 1998). Cauzard et al. (1998) found no big differences between the needs for a distance
control system between several European countries. However, the needs for other systems differed, for
example with respect to an alcohol-meter. According to Young et al. (2003) the attitudes of (young
novice) metropolitan and rural drivers were generally very similar. It seems to matter to which target
group drivers belong. Truck drivers and car fleet operators considered driver support systems more
attractive than other driver and fleet operator groups (Marchau et al., 2001). Chalmers (2001) found
that bus and truck drivers generally had a greater need for driver assistance than private drivers. Also
driving style seems to have an effect on the needs for driver assistance. According to Chalmers (2001)
and TRG (1998) cautious and careful drivers were more positive about help from their cars than
drivers with a more ‘sportive’ driving style.
Next to driving style, driving experience seems to be of influence. Less experienced drivers thought
ADAS were more attractive compared to more experienced drivers (Marchau, 2000). Besides the
amount of kilometers driven, personal experience with certain situations may affect the opinion on
driver assistance. A driver monitoring system was rated positive mainly by drivers that had
encountered a (near-)accident due to tiredness (Bekiaris et al., 1997; CRA, 1998). Personal experience
with congestion also motivated positive reactions to a Stop & Go system (TRG, 2003). Opinions on
ADAS seem to be influenced by familiarity with ADAS as well. Piao et al. (2004) stated that the
indicated needs for ACC were connected with the dissemination activities about ACC for the last ten
years. They also found that drivers with experience with cruise control were significantly more willing
to purchase an ACC than those without. However, this was not the case in the study of Turrentine et
al. (1991). Avid users of cruise control appeared not to be early adopters of more automated controls,
like ACC. A possible explanation for these conflicting results may be the increasing publicity of ACC
recently. Type of car possession seems to be another explanatory variable. In the COMUNICAR
survey it was found that owners of a city car had different opinions on driver assistance than owners of
an upper class car (Mariani et al., 2000). A similar result was found by Marchau (2000): especially
business drivers liked support from a navigation system. Psychological factors seem to influence in
particular speed-regulating devices. According to Rienstra & Rietveld (1996) people who frequently
transgress speed limits were more against speed limiters. Also Garvill et al. (2003) found this
relationship: moral and perceived difficulty to keep the speed limits influenced the acceptance of an
Electronic Speed Checker. In a few studies, the effects of education, family situation and income on
the perceived needs for driver assistance were studied. TRG (1998) found that respondents with
children <15 years were more interested in ACC and Collision Warning (CW) systems than
respondents with older children or without children. Highly educated people were more against speed
limiters according to Rienstra & Rietveld (1996). The influence of income was unclear. Dutch and
Norwegian motorists with a high income were negative about ISA, while English motorists with a
high income were positive (TRG, 2003). Possibly, the needs for ISA were more related to nationality
than to income.
Characteristics of the driver assistance
The opinion on driver assistance may be affected by characteristics of the system as well. The general
accepted notion in literature was that systems should be 100% reliable and have no false alarm rates.
Besides, it appeared to be of influence which driving tasks or situations the system supports. In
general, lane keeping systems did not seem to be very popular among drivers. Lane Departure
Warning (LDW) was rated (much) lower than Stop & Go, ACC and ISA (Piao et al., 2004). The
results from the IN-ARTE survey also showed that respondents were indifferent to lateral control
functions (Wevers et al., 1999). However, participants in the study of Regan et al. (2002) felt that
LDW was a good idea. In contrary to lane keeping systems, anti-collision systems seemed to be
popular. CRA (1998) found that all participants viewed Collision Avoidance Systems favourably. The
IN-ARTE survey revealed a strong preference for warnings on front obstacles and warnings to slow
down (Wevers et al., 1999). The results from the COMUNICAR survey showed that driver support
functions related to the primary driving task and enhancing safety were valued as the most preferred
ones (Mariani et al., 2000). These included for example: automatically set cruising speed, and
13
warnings on lateral collisions, front obstacles and car overtaking. Also systems that provide visual
support were rated high. According to Blythe & Curtis (2004), respondents were most positive about
Collision Warning and Prevention, ACC and Driver Alertness Monitoring. Bekiaris et al. (1997) also
found that the acceptance of this last system was high. Chalmers (2001) indicated that generally there
was a major acceptability of systems that control longitudinal distances between vehicles. This was
already concluded by Cauzard et al. (1998), who said that two third among European drivers would
find it very or fairly useful to have a distance control system in their cars. Besides type of system or
function, level of support seems to affect the needs for driver assistance. CRA (1998) found that the
appeal to ACC was more limited, possibly due to the automatic control of car following. Similarly, the
results from the IN-ARTE survey showed that users were generally negative about automatic
intervention, probably because personal control is a crucial issue in driving behaviour (Wevers et al.,
1999). However, acceptance of system intervention was limited to more critical situations and
situations with dense traffic (i.e. convoy driving). Also the COMUNICAR survey revealed a greater
need for information and warnings than automatic actions (Mariani et al., 2000). Garvill et al. (2003)
concluded that informative devices were rated better than intervening ISA devices. 60% of the
respondents in the study of McIntosh (1997) had doubts about automatic vehicle operation (loss of
“hands on” control, safety concerns), while 27% thought it was a good idea (easier to drive, improve
safety). Blythe & Curtis (2004) found that nearly half of the respondents thought ADAS should simply
warn the driver; the other half thought ADAS should assist (38%) or take over (16%). Hoedemaeker
(1999) reported that half of the respondents indicated to dislike it when the car takes over the driving
task.
In some studies the type of feedback was studied. Modalities for giving feedback to the driver can be
visual (e.g. text, icon), auditory (e.g. beep, voice), haptic (e.g. counterforce on gas pedal) and tactile
(e.g. vibration in steering wheel). The COMUNICAR survey revealed that warning signals should be
given visually (41%) or with a sound (35%), rather than with a synthesised voice or with combinations
of these feedback types (Mariani et al., 2000). Garvill et al. (2003) found that light and sound signals
when exceeding the speed limit were preferable to a display presenting the current speed limit and an
active gas pedal making it more difficult to exceed the speed limit. According to TRG (2003), most
respondents preferred warning ISA (visual or audio) to haptic ISA. However, in a follow-up study
with a driving simulator, it was concluded that drivers were more positive about haptic ISA after
having gained experience with it. Frequently, the type of feedback is considered in such (follow-up)
studies, rather than in user needs surveys. Overrulability appears to be very important in the
acceptance of driver assistance. It is about the possibility for the driver to overrule (i.e. ignore) the
system, for example by turning it off. Blythe &Curtis (2004) found that the attitude of the majority of
respondents was favourable if ADAS can be switched on or off by the driver. By contrast, if the driver
cannot switch off ADAS, the attitude of the majority of respondents was negative. Only one
application, Collision Warning and Prevention, still received a favourable attitude. Also Regan et al.
(2002) came to the conclusion that voluntary systems, which drivers can choose to disable, were more
acceptable among the respondents than mandatory systems, which cannot be turned off by the driver.
According to Comte et al. (2000), participants thought mandatory ISA would be most useful, however,
they preferred the idea of a voluntary system. Opinions on ADAS seem to be influenced by price as
well. In the study of TRG (1998), respondents were first asked to give the total cost of their “reference
car”, including VAT and optional feature. After having introduced ADAS, such as ACC and CW
systems, the question was asked again. On an average, the ratio between the second price and the
initial price was 1.08. The average cost of the “reference car” was €18270; the supplementary price
that people would agree to pay for ADAS would thus be about €1460. Van der Heijden & Molin
(1999) found that respondents were prepared to pay little for an ISA system in their cars. However, the
willingness to purchase such a system seemed to be higher when it was combined with other ADA
systems. Marchau et al. (2001) concluded that, as to be expected, higher prices had a negative
influence on the acceptability of ADAS. By means of a decompositional SP approach, several
hypothetical ADAS alternatives were described with the following attributes: distance keeping, speed
adaptation, navigation and price. By decomposing the overall preference for the alternatives into the
weights attached to these attributes, it appeared that – after navigation – price was the most important
14
attribute of a driver support system. A similar study from Schmeidler (2002) revealed corresponding
results with respect to price. However, in this study price appeared to be the third important attribute,
after navigation and ISA. A majority of respondents were either not prepared to pay anything extra for
nine ADA applications (50%) or would be willing to pay a sum of up to £500 (37%) (Blythe & Curtis,
2004). In general, it can be stated that the willingness to pay for ADAS is rather low. This is contrary
to the ‘noncommitment bias’ stated by Polydoropoulou et al. (1997). They suggested that respondents
in SP studies might overestimate their willingness to pay, because no actual payment was concerned.
Becker et al. (1995) stated that the use of questionnaires with text descriptions of the system
performance for SP studies may not provide a sufficiently accurate or detailed picture of the system
for the subject to give an accurate response. Like type of feedback, price is also frequently considered
in (follow-up) studies that include gaining experience with a system, rather than in user needs surveys.
TRG (1998) referred in her study to earlier research showing that 30% of the participants were willing
to purchase an ACC at an average price of $490, rising to 87% and $616 after prototype rides.
Characteristics of the traffic scene
The opinion of the driver on driver assistance may also be influenced by characteristics of the traffic
scene. In general, there seems to be a greater need for driver assistance in critical situations, such as
near-accidents. Bridger & Patience (1998) concluded that in normal driving situations, the driver
wants to be in control. In more adverse situations, assistance from the car is wanted and in dangerous
situations, the driver even relies on assistance from the car. A higher level of support in more critical
situations was also found by Wevers et al. (1999) in the IN-ARTE survey. According to Várhelyi
(2000), a majority of the respondents accepted ISA that warns the driver or even automatically reduces
the driving speed in imminent crash situations. Weather/visibility conditions also appear to be
connected to the acceptability of ADAS. TRG (1998) showed that drivers would like to use ACC
especially when it is foggy or at night. A majority of the respondents was positive about ISA during
slipperiness or reduced visibility situations (Várhelyi, 2000). The COMUNICAR survey revealed that
respondents were interested in visual support in poor visibility situations (Mariani et al., 2000).
Besides, road type seems to influence the needs for driver assistance. Schmeidler (2002) reported that
respondents thought ACC was especially useful on motorways and rural roads, while ISA should be
used mainly on urban roads. This result on ISA was also found by Van der Heijden & Molin (1999)
and Van Hoorebeeck (2000). Participants in the study of Regan et al. (2002) felt that LDW was a good
idea in rural areas. Results from the IN-ARTE survey revealed that in particular velocity determined
the demand for driver assistance. For example, the higher the velocity, the more important was
information about road geometry. Besides, warnings to change lane were wanted on motorways and
country roads, but not on city roads (Büsher & Frese, 1998). The presence of other road users may
also be of importance when considering the opinions on ADAS. Participants in the study of TRG
(1998) liked to use ACC or CW when the traffic density was low. It was concluded in the IN-ARTE
survey that the needs for certain driver support functions were related to external factors and driving
situations, such as a car cutting in from the right or driving in convoy (Wevers et al., 1999). For
example, the acceptance of autonomous braking was higher in situations with dense traffic (e.g.
convoy driving) as compared to clear running situations. Note that this result is somewhat contrary to
the earlier mentioned result about using ACC (that may brake automatically) when the traffic density
is low.
Summary
Despite the amount of literature on the needs for ADAS, no general theory was found so far to
describe the relations between the driver, the system and the traffic scene. However, a conceptual
model (figure 3) was made that presented an overall picture of earlier research into user needs for
driver support systems. It was found that the opinion of drivers on driver assistance may be influenced
by characteristics of themselves, such as gender and age, and characteristics of the system, such as
type and level of support. Besides, this opinion may be influenced by characteristics of the traffic
scene, such as critical situations and road type. Although the conceptual model provided more insight
into the needs for driver assistance, still some knowledge gaps can be distinguished. Most studies
presented one or more driver support systems to potential users based on technical possibilities and
15
envisaged user requirements. Therefore, little is known about the ‘ideal’ system according to the
driver. Which (combinations of) driving tasks and situations should be supported? Earlier research did
not always show clear relations between the driver, the driver assistance and the traffic scene.
Sometimes (seemingly) conflicting results were found. A more consistent view on these relations is
wanted. Thus, this research aimed at filling knowledge gaps in the conceptual model and moving the
state-of-the-art on driver assistance forward.
16
3.
User needs survey
This chapter describes a user needs survey about integrated driver assistance from the driver’s point of
view. First, the focus of the survey is highlighted. Next, more information is given about the target
group and the required sample size. Then, the design of the questionnaire is explained, followed by the
ways in which participants for the survey were gathered. This chapter ends with some remarks about
pilot tests before the questionnaire was distributed.
3.1
Focus
This research attached importance to the opinion of drivers about ‘intelligent vehicles’. Detailed
information is wanted on the driver, the system and the traffic scene. Therefore, a user needs survey
was set up. Drivers were asked to indicate their (latent) needs for driver assistance by means of a
questionnaire. To explore the broad field of these user needs, the compositional SP approach was used.
Several attributes in terms of driver support were presented separately.
So far, most research into user needs for driver assistance concerned drivers’ preferences for one or
more systems in contrast to drivers’ preferences for certain functions that support the driving task.
Because of this ‘technology driven’ approach, little is known about the extent to which drivers want
assistance from their cars during certain driving tasks and situations. This user needs survey, therefore,
focused on driver support functions related to driving tasks and situations. So instead of presenting a
system such as ISA, that regulates speed according to the speed limit, several driver support functions
which regulate speed were presented.
Because earlier research mostly centred on a specific system, little is known about preferred
combinations of driver support systems and functions. Therefore, the user needs survey included
questions about combinations of driver assistance. The needs of the driver determine implications for
integrated driver assistance. The integration of driver support functions may be related to the
following aspects:
• Technology: functions can make use of the same components, such as sensors;
• Human-Machine-Interface (HMI): functions can make use of one interface to present information
to the driver;
• Functional operation: functions can make use of each other’s information to extend their individual
field of activity.
3.2
Target group and sample size
The target group of the user needs survey consisted of Dutch car drivers of a passenger car.
Participants did not need to possess a car. However, they should have a valid driving licence.
Passenger cars were considered because of:
a. The high number of vehicle kilometres driven by passenger cars: 146.1 billion km (76.5%) by car
drivers and passengers in the Netherlands in 2003 (CBS, 2003);
b. The high number of accidents with passenger cars involved: 607 deaths (64.6%) and 6442 inpatients (70.1%) among car drivers and passengers in the Netherlands in 2002 (SWOV, 2002);
c. A practical reason, namely using the TNO passenger car driving simulator in the next phase of the
project.
The required sample size was estimated by using the guidelines in Buijs (1993) and Cohen (1988). In
Buijs (1993) the formula for defining a sample size is: n z2* 2/a2 with n=sample size, z=standard
score corresponding to given confidence level, =standard deviation and a=proportion of sampling
error. This formula is valid for variables with a normal distribution. Many distributions can be
17
approximated by a normal distribution. Therefore, Buijs (1993) states that this formula may be used
when the population N is large compared to the sample n (n<0.10*N). For proportions the formula can
be rewritten as follows: n z2*p(1-p)/a2 with p=estimated proportion. This formula is valid when n is
sufficiently large (n 200). To estimate a maximal ‘safe’ required sample size, p should be set at 0.50.
In this research, the formula was used with z=1.96, p=0.50 and a=0.05. This gave a required sample
size of at least 385.
Cohen’s power analysis was also used to estimate the required sample size. Cohen (1988) states that
the power of a statistical test of a null hypothesis is the probability that it will lead to the rejection of
the null hypothesis, i.e. the alternative hypothesis is true. The power of a statistical test depends upon
the significance criterion, the reliability of the sample results and the effect size, i.e. the degree to
which the null hypothesis is false. The required sample size depends on the type of statistical test and
is a function of significance criterion (a), effect size (ES) and the amount of desired power (power).
The significance criterion represents the Type I error, i.e. the risk of a false null rejection. Usually, a
significance criterion of 0.05 is chosen. The effect size serves as an index of degree of departure from
the null hypothesis. When the null hypothesis is true, the effect size is zero; when the alternative
hypothesis is true, the effect size is a specific nonzero value. When there is no other basis for setting
the desired power value, it is proposed to use the value 0.80. The Type II error, i.e. the risk of false
null acceptance represented by b=1-power, will then be 0.20. This means that the relative seriousness
of Type I to Type II error (b/a) is 4 when a=0.05. This means that mistakenly rejecting the null
hypothesis is considered four times more serious than mistakenly accepting it. In this research, the
sample size tables in the handbook were used with a=0.05, ES=small (corresponding to values per
statistical test) and power=0.80. Dependent on the statistical test, the required sample size should
consist of at least 400 car drivers.
With these values for the required sample size in mind, it was aimed to gather responses to the survey
from a minimum of 500 Dutch car drivers.
3.3
Questionnaire design
The structure of the survey consisted of the three parts defined below:
• Driver support functions: to what extent do drivers want assistance from the car during driving? On
which type of road, during which driving tasks and situations, and with what level of support?
• Ideal driver support system: what is the ideal assistance according to the driver? Which
combinations of driver support, and which other characteristics?
• General information: background questions about car possession, driving experience, socioeconomic variables, and the like.
Participants were asked to fill in a computer-based questionnaire for distribution on the Internet. This
type of questionnaire was chosen because of the following advantages: (a) personalization of the
questionnaire by interactively showing relevant questions and responses, based on previous answers,
(b) collection and storage of data in an electronic database and (c) quick and cheap way of data
gathering. The questionnaire as it was shown on the Internet can be found in appendix A.
Driver support functions
The survey started with information about the goal and procedure of the survey. Next, some
background questions on car possession and usage were asked. The next part of the questionnaire – on
the needs for driver support functions – started with information about the ways in which the car can
assist the driver with several driving tasks and situations. Table 1 shows the driving tasks and
situations that were included in the questionnaire. This list was established by using the work of
McKnight & Adams (1970) on task analysis of car driving. The driving tasks related to the operational
and tactical level of driving (Michon, 1985). A distinction was made between three road types:
motorways (M), rural roads (R) and urban roads (U).
18
Driving tasks
Regulating speed (M, R, U)
Lane keeping (M, R, U)
Car following (M, R, U)
Lane changing (M, R, U)
Congestion driving (M, R, U)
Negotiating non-signalised intersections (R, U)
Negotiating signalised intersections (R, U)
Situations
Reduced visibility
Driver fatigue
Imminent crash
Table 1. Driving tasks and situations in the survey
For each driving task and situation, several driver support functions were defined. Participants had to
indicate on a five-point scale to what extent they have a need for these functions on each road type
(1=great need, 5=certainly no need). Figure 4 shows an example of speed assistance. The driver
support functions consisted of information, warning or control. In this survey, it was assumed that the
driver could overrule the driver assistance, for example by turning it off. Several driver support
functions included a form of integration. Consider, for example, the function ‘warning for unsafe
speed regarding actual situation, e.g. fog, curve, nearby school’. Communication with other vehicles
or roadside systems is necessary to detect fog, and communication with a digital map may be
necessary to detect a curve or nearby school.
Figure 4. Example of a survey question: speed assistance
Ideal driver support system
After having introduced possibilities for driver assistance, the next part of the survey focused on the
ideal driver support system. A ‘personalized’ table was presented first that showed all driver support
functions for which the participant had a significant need according to his or her previous answers (i.e.
category 1 or 2). It was assumed that drivers prefer a system that consists of less driver support
19
functions than a summing-up of ‘single’ functions. This summing-up might be ‘too much’, for
example. Participants could formulate the ideal driver support system by indicating at most six
favoured types of assistance out of maximal twenty-two. These types of assistance corresponded to the
driving tasks and situations in the first part of the questionnaire; see table 1. In a concept version of the
questionnaire, participants had to indicate their most favoured driver support functions per favoured
type of assistance again. However, in the final version of the questionnaire this step was left out.
Firstly, because the questionnaire should not be too long (15-20 minutes for filling in). Secondly,
because during the pilot test, participants thought this step was (practically) the same as the first part
of the questionnaire. Statements about typical characteristics of the ideal system were also presented in
the questionnaire. Participants had to indicate on a five-point scale (1=totally agree, 5=totally
disagree) to what extent they agree with the statements in table 2.
Statement
1. The ideal system should make car driving easier and thus support in a way that is similar to my ‘normal’
driving behaviour.
2. The ideal system should make car driving better and thus support in a way that optimizes my ‘normal’
driving behaviour, i.e. the system compensates for my errors, mistakes, etc.
3. In the ideal system I can decide on the settings with respect to type of assistance and feedback.
4. The ideal system provides me more support during deviant situations, for example when the weather is
bad or when I am tired.
5. The ideal system provides me more support when the traffic is busy.
6. In case of informing or warning messages at the same time, the ideal system decides which message to
present to me first.
7. In case of automatic actions at the same time, the ideal system decides which action to perform first, for
example a choice between automatic car following and automatic lane changing.
8. When the ideal system takes over driving tasks, I think it would be appealing to me to do other things, for
example reading, putting on music, telephoning, making up (hair, make-up).
Table 2. Statements about the ideal driver support system
When participants indicated in the first part to have (certainly) no need for driver support with respect
to a certain driving task, an extra question was asked. This extra question was about the extent to
which the participant would have a need for help from the car with the concerned driving task during
deviant situations, for example when he or she is tired or when the weather is bad. When participants
were indifferent or negative about the driver support functions presented in the first part of the survey,
they did not have to fill in the second part about the ideal driver support system. They were directly
referred to the background questions about socio-economic variables and the like.
3.4
Invitation of participants
Participants for the user needs survey were invited in two ways, namely via a market agency and via
personal and business contacts. Market agency RM Interactive was called in to invite members of her
Internet panel. This panel consists of more than 25.000 members who are representative for the Dutch
population. Members are invited to participate in a survey once a month on average. Up to date, more
than 200 surveys were conducted with the Internet panel.
A contract was signed that guaranteed 300 completed questionnaires. Members of the Internet panel
received €2.50 for filling in the questionnaire. Besides, they could win a €20 gift voucher (1 per 25
participants). It was tried to represent the target group of Dutch car drivers with respect to gender, age
and education. The main characteristics of Dutch people holding a driving licence were obtained from
Statistics Netherlands (CBS) (2003). The following procedure was used by RM Interactive to carry out
the fieldwork:
- Invitation by e-mail. An e-mail was sent to a gross sample of members of the Internet panel. This email contained a hyperlink to the on-line questionnaire and a personal access code. This code
20
prevents a member to fill in the questionnaire more than once. The e-mail also explained briefly the
topic of the survey and that the survey was conducted by the University of Twente;
- Screening of participants. At the beginning of the questionnaire, the participant was asked for how
long he or she is holding a driving licence for a passenger car. If the answer was ‘do not have a
driving licence’, the participant could not continue with the questionnaire.
From 6-9 September 2004, RM Interactive gathered response from her Internet panel. Table 3 shows
some characteristics about the response rate.
Response justification
Gross sample
Did not fill in screening
Did not satisfy screening
Satisfied screening and started questionnaire
Did not complete questionnaire
Completed questionnaire
3009
2349
40
620
243
377
100%
78%
1%
21%
100%
39%
61%
Table 3. Response rate of RM Interactive panel
A second way of inviting participants for the user needs survey was via personal and business
contacts. An invitation by e-mail containing a hyperlink to the questionnaire was sent to family,
friends, colleagues and the members of the AIDA user group; see appendix B. It was also asked to
forward the e-mail to possibly interested others (snowball method). Besides, an invitation to the
questionnaire was put on the websites of research organisation TNO and consumer organisation
ANWB. Participants could win a €20 gift voucher (1 per 50 participants). The questionnaire was
online from 24 August 2004 until 1 October 2004.
3.5
Pilot tests
A pilot test of the user needs survey was performed to check whether the survey would provide the
necessary information. Twenty-one people participated in the pilot test by completing a paper version
of the questionnaire. Participants included colleagues, friends and family. Based on their answers and
comments, the questionnaire was shortened and clarified. The questionnaire was then programmed
into an Internet version. It was hosted on the webserver of AIDA: www.aida.utwente.nl/enquete. To
check whether the alterations had the desired effect, people from the target group went through the online version again.
21
22
4.
Results
This chapter presents the results from the user needs survey. The needs of car drivers with respect to
help from the car during certain driving tasks and situations revealed starting points for the possible
integration of driver support functions. First, the sample is discussed. Next, the needs for driver
support functions are explained, followed by features of the ideal driver support system. Finally, the
influence of sample characteristics on the results is discussed.
4.1
Sample
Before analysing the results from the survey, the data were verified. It appeared that a few respondents
had sent the data twice. This was possible when one was thumbing through the questionnaire. In this
way, the ‘send button’ could be used more than once. By checking the IP address, the answers and the
date and end time of filling in the survey, double data were filtered out. Without these data, the
questionnaire was completed by 1152 respondents. 103 respondents did this in less than 10 minutes.
These answers were left out of consideration, because it was assumed that one could not seriously fill
in the questionnaire in less than 10 minutes. Therefore, the results are based on the answers from 1049
respondents. Of those, 740 respondents were invited via personal and business contacts (including the
website invitations) and 309 respondents were members of the Internet panel of the market agency.
Table 4 shows some characteristics of the sample and the target group of Dutch car drivers (CBS,
2003). 757 of the respondents were male, 292 were female. The average age of the respondents was 41
years (SD 12, minimum 18, maximum 79). Most of them were highly educated. Chi-square tests were
carried out to check whether the distributions of gender, age and education in the sample differed from
the distributions of these variables in the target group. Based on the results from these tests, it turned
out that the sample significantly differed from the target group (p<0.05).
Characteristic
Gender
male
female
Age
18-24 years
25-44 years
45-64 years
65 years
unknown
Education
primary education
lower secondary education
higher secondary education
higher education/university
other
Sample
Target group
72.2%
27.8%
54%
46%
6.8%
51.9%
38.7%
2.6%
0.1%
16%
32%
32%
20%
-
0.2%
9.2%
22.6%
67.8%
0.2%
10%
27%
30%
33%
-
Table 4. Characteristics of sample and target group
Table 5 shows more background information on the sample. It can be seen that most respondents had a
driving licence for more than 10 years, possessed their own car and drove in it frequently. Almost half
of the respondents were somewhat familiar with ACC. 38 respondents said that they have the system
in their cars (regardless of using it). With respect to the current, low penetration rate of ACC in the
Netherlands, it is thought that 38 seems too high. It is assumed that respondents confused ACC with
regular cruise control. Nearly all respondents indicated to use the Internet regularly.
23
Characteristic
Frequency of car driving
>3 times a week
1-3 times a week
1-3 times a month
<1 time a month
Years of driving experience
<5 years
5-10 years
>10 years
Average annual mileage
<10.000 km
10.000-20.000 km
>20.000 km
Type of car possession
no car
private: <50% business
private: >50% business
business: under one’s own trust
business: lease
other
Familiarity with ACC
not familiar with
somewhat familiar with (media, etc.)
familiar with (in car, but not in use)
very familiar with (in car and in use)
(very) familiar with (but not in possession)
Frequency of Internet usage
rarely
sometimes
regularly
#
%
753
197
66
33
71.8
18.8
6.3
3.1
93
184
772
8.9
17.5
73.6
251
408
390
23.9
38.9
37.2
69
726
105
34
104
11
6.6
69.2
10.0
3.2
9.9
1.0
266
525
11
27
220
25.4
50.0
1.0
2.6
21.0
28
1021
2.7%
97.3%
Table 5. More background information on the sample
4.2
Driver support functions
More insight is gained into the needs for certain driver support functions by analyzing the answers to
the first part of the questionnaire. In this part the respondents indicated to what extent they want
assistance from the car during driving: during which driving tasks and situations, on which type of
road and with what level of support.
Driving tasks and situations
The results from the survey showed that the most popular driver support functions related to the
following driving tasks and situations: regulating speed, lane changing, negotiating intersections,
imminent crash and reduced visibility. Table 6 shows these most popular functions based on the
answers ‘great need’ and ‘need’ (categories 1 and 2 respectively). Especially warnings for downstream
traffic conditions and warnings for traffic in blind spots were favoured. Apparently, drivers appreciate
being well informed when driving a car. Knowledge about what is happening further down the road
(e.g. an accident or road works) can help the driver to accordingly regulate his speed or allows the
driver to turn off the road and seek an alternative route to his destination. Being aware of what is
happening in the direct vicinity of the car (e.g. another vehicle in the left blind spot or a bicycle in the
right blind spot at an intersection) provides the driver with a more accurate idea of the situation, so
that dangerous situations can be prevented or at least anticipated. Respondents also expressed a need
for ‘general’ warnings during imminent crash situations (e.g. Forward Collision Warning) and
information during reduced visibility situations (e.g. Night Vision).
24
Driver support function
1. Warning for downstream traffic condition – motorway
2. Warning for downstream traffic condition – rural road
3. Blind spot warning during lane changing – motorway
4. Blind spot warning at non-signalised intersection – urban road
5. Blind spot warning at non-signalised intersection – rural road
6. Blind spot warning during lane changing – rural road
7. Blind spot warning at signalised intersection – urban road
8. Blind spot warning at signalised intersection – rural road
9. Warning for imminent crash
10. Presentation of badly visible objects on windscreen
(Great) need
946
90.2%
884
84.3%
861
82.1%
860
82.0%
833
79.4%
787
75.0%
769
73.3%
71.4%
749
736
70.2%
734
70.0%
Table 6. Top 10 of most popular driver support functions
The top 10 looked somewhat differently when we only considered the most wanted driver support
functions based on the answer ‘great need’ (category 1). Eight out of ten functions were the same. The
two ‘new’ functions concerned warnings on an urban road with respect to downstream traffic
condition and blind spot during lane changing. The order of the functions also differed, except for the
first four functions. 517 respondents (49.3%) indicated to have a great need for the most favoured
function ‘warning for downstream traffic condition – motorway’.
It was investigated to what extent there seemed to be a significantly greater need for certain driver
support functions than others. Therefore, McNemar tests were performed with data from the top 10 of
most popular functions. The McNemar test is a nonparametric test for two related dichotomous
variables (Rice, 1994). It focuses on changes in responses from one sample response (e.g. function 1)
to another (e.g. function 2) using the chi-square distribution; see also appendix D. It appeared that the
first function ‘warning for downstream traffic condition – motorway’ significantly differed from the
other functions (p 0.001). This also applied to the functions 2 to 5, which significantly differed from
the first function and from the functions 6 to 10. The last five functions differed significantly from the
first five functions. Based on this, the top 10 of functions could be divided into three groups; see table
7. Within the groups the perceived (great) need for the specific functions did not differ significantly,
except for function 2 versus 5.
Driver support function
1. Warning for downstream traffic condition – motorway
2. Warning for downstream traffic condition – rural road
3. Blind spot warning during lane changing – motorway
4. Blind spot warning at non-signalised intersection – urban road
5. Blind spot warning at non-signalised intersection – rural road
6. Blind spot warning during lane changing – rural road
7. Blind spot warning at signalised intersection – urban road
8. Blind spot warning at signalised intersection – rural road
9. Warning for imminent crash
10. Presentation of badly visible objects on windscreen
(Great) need
946
90.2%
884
84.3%
861
82.1%
860
82.0%
833
79.4%
787
75.0%
769
73.3%
71.4%
749
736
70.2%
734
70.0%
Table 7. Three groups of most popular driver support functions
McNemar tests were also performed with data from the top 10 of most popular functions based on the
answer ‘great need’. It again appeared that the function ‘warning for downstream traffic condition –
motorway’ significantly differed from the other functions (p 0.001). It can be concluded that based
on frequency and statistics, this driver support function was most popular of all.
When respondents indicated to have (certainly) no need for driver support with respect to the
concerned driving task, an extra question was asked. This extra question was about the extent to which
25
they would have a need for help from the car with the driving task during deviant situations, for
example when they are tired on when the weather is bad. Table 8 shows the results with respect to this
extra question.
Task or situation
Regulating speed
Lane keeping
Car following
Lane changing
Congestion driving
Negotiating non-signalised intersections
Negotiating signalised intersections
#
1
107
40
92
19
15
35
no help
1
52
25
53
14
10
25
maybe
0
49
12
30
4
4
10
yes help
0
6
3
9
1
1
0
Table 8. Results with respect to the extra ‘(certainly) no need’ question
The results showed that only three respondents were indifferent or negative (i.e. categories 3 to 5)
about all driver support functions presented in the first part of the survey. 107 respondents out of 1049
(10.2%) indicated to have (certainly) no need for lane keeping support. From these, only 6 respondents
said that they would like to have support with this driving task in deviant situations. Half of them said
they still did not want the car to help and half of them were indifferent. Also support with lane
changing was not very wanted. Almost 9% of the respondents did not need any help from the car with
this driving task. However, it should be kept in mind that more than 82% of the respondents indicated
to have a (great) need for a blind spot warning during lane changing on motorways. Only one
respondent had to fill in the extra question with respect to support with regulating speed. This showed
that nearly all respondents had a need for at least one of the driver support functions related to
regulating speed.
Road type
In general, the results showed that there was a great need for driver assistance during driving on
motorways. Among other things, the most popular driver support function for each driving task
(except negotiating intersections) related to driving on a motorway. Respondents wanted less help
from their cars with driving on rural roads and even less on urban roads. However, this tendency did
not hold up for all driver support functions.
Per road type, respondents indicated their perceived (great) need for 26 driver support functions that
related to: regulating speed, lane keeping, car following, congestion driving and lane changing.
McNemar tests were performed to check whether the needs for certain functions were significantly
different on the three road types. It turned out that 18 out of 26 driver support functions scored highest
on motorways, two scored highest on rural roads, none on urban roads and the rest scored indifferent
(p 0.001). The two driver support functions that were more wanted on rural roads than on motorways
and urban roads all give support with regulating speed:
• Information on speed limit;
• Warning for exceeding speed limit.
Six functions were wanted on rural roads as much as on motorways, among which warnings for unsafe
speed regarding the actual situation (e.g. fog, curve) and automatic lane keeping on winding roads.
McNemar tests were also performed to check whether the needs for intersection support were
significantly different on rural and urban roads. Also the function ‘warning for opposing traffic’ (lane
changing) was taken into account, because this function was not considered on motorways. It appeared
that 6 out of 14 functions scored higher on rural roads than on urban roads, the rest scored indifferent
(p 0.001). Among these six functions were warnings for opposing traffic when changing lanes and
information on approaching a (non-)signalised intersection. There was no greater need for specific
functions on urban roads compared to rural roads.
26
Level of support
According to the results, respondents would mainly like their cars to help them by giving information
or warnings. They indicated hardly any need for driver support functions that consisted of control. The
most unpopular driver support functions were related to automatic actions with respect to lane
changing, negotiating intersections and lane keeping. However, in some cases respondents do want the
car to take over control.
It turned out from McNemar tests that respondents preferred automatic actions from the car when they
want to maintain a self-chosen speed on motorways or rural roads (p 0.001). They also would like the
car to take over the longitudinal driving task or even the whole driving task when they are driving in
traffic jams, irrespective of the road type. Besides, there were no significant differences between the
needs for lane keeping support on all three road types with respect to warnings and automatic actions
in case of road works.
4.3
Ideal driver support system
More insight is gained into the ideal driver support system by analyzing the answers to the second part
of the questionnaire. In this part the respondents indicated at most six driving tasks and situations that
the ideal system should provide help with. Three respondents were indifferent or negative about the
driver support functions presented in the first part of the survey. Therefore, they did not have to fill in
the second part of the survey about the ideal driver support system. Answers from the remaining 1046
respondents were used for the analyses below.
Types of driver assistance
The respondents chose various types of driver assistance in their ideal driver support system, so the
ideal system seemed to be very personal. Table 9 shows the most popular driving tasks and situations
that the ideal system should provide help with. The ideal system should, according to the respondents,
not support lane keeping on urban roads and negotiating signalised intersections on rural and urban
roads. Less than 2.8% of the respondents indicated to want assistance from the ideal system with these
driving tasks. A simple data check revealed that help with reduced visibility situations and help with
congestion driving were not related to respectively the weather and the (partial) holiday season during
the (on-line) questionnaire.
Type of assistance
1. Reduced visibility
2. Imminent crash
3. Car following – motorway
4. Regulating speed – motorway
5. Congestion driving – motorway
6. Driver fatigue
7. Regulating speed – rural road
8. Car following – rural road
9. Negotiating non-signalised intersection – rural road
10. Negotiating non-signalised intersection – urban road
Yes in system
633
60.5%
588
56.2%
535
51.1%
458
43.8%
442
42.3%
369
35.3%
354
33.8%
298
28.5%
234
22.4%
208
19.9%
Table 9. Top 10 of most popular types of assistance in the ideal system
Comparable to the single driver support functions, it was investigated to what extent there seemed to
be a significantly greater need for certain types of driver assistance in the ideal system than others.
McNemar tests were performed with data from the top 10 of most popular driver assistance in the
ideal system. It appeared that the first three types of assistance did not significantly differ from each
other (p 0.001). However, the difference between these three and the rest of the top 10 was
significant. Based on frequency and statistics, help from the ideal system with reduced visibility,
imminent crash and car following on motorways was most popular of all; see table 9. Within the group
27
of type 4 to 10 the perceived need for the specific types of assistance differed significantly, except for
consecutive numbers. This means, for example, that respondents were indifferent about help from the
ideal system with regulating speed on motorways (type 4) and congestion driving on motorways (type
5), but that they preferred help with regulating speed on motorways (type 4) to help from the system
when they are fatigued (type 6).
Number of types of assistance
Respondents could indicate at most six out of twenty-two types of driver assistance in their ideal
driver support system. Table 10 shows how many driving tasks and situations the system should
provide help with according to the respondents. It can be seen that most respondents (56.2%) thought
the ideal system should provide support with six driving tasks and situations.
# types of assistance
1
2
3
4
5
6
# respondents
29
55
115
142
117
588
Table 10. Number of types of assistance in the ideal system
Combinations of driver assistance
Respondents could formulate their ideal driver support system. By doing this, it became clear which
combinations of driver assistance were most popular. Each respondent indicated at most six driving
tasks and situations for which he or she would like to have help from the ideal system. Based on these
combinations, the most wanted pairs of driver assistance were determined. Table 11 shows the most
frequently mentioned pairs. It can be seen that the three most popular pairs consisted of combinations
of the three most popular ‘single’ driving tasks and situations in the ideal system (table 8).
Combination
1. Reduced visibility & imminent crash
2. Car following (motorway) & reduced visibility
3. Car following (motorway) & imminent crash
4. Regulating speed (motorway) & regulating speed (rural road)
5. Regulating speed (motorway) & car following (motorway)
6. Car following (motorway) & car following (rural road)
7. Reduced visibility & driver fatigue
8. Driver fatigue & imminent crash
9. Congestion driving (motorway) & reduced visibility
10. Regulating speed (motorway) & reduced visibility
#
384
304
299
272
262
261
259
250
247
238
%
36.7
29.1
28.6
26.0
25.1
25.0
24.8
23.9
23.6
22.8
Table 11. Most popular combinations of driver assistance in the ideal system
Other characteristics of ideal system
Respondents indicated on a five-point scale to what extent they agree with statements about typical
characteristics of the ideal driver support system (table 2). Table 12 shows what characteristics the
ideal driver support system should have besides giving support with the above-mentioned driving
tasks and situations.
28
Totally agree
Somewhat agree
Neutral
Somewhat disagree
Totally disagree
Making car
driving easier
#
%
467
44.6
418
40.0
108
10.3
30
2.9
23
2.2
Making car
driving better
#
%
373
35.7
410
39.2
156
14.9
78
7.5
29
2.8
Adjustable
settings
#
%
539
51.5
328
31.4
121
11.6
38
3.6
20
1.9
More support
deviant situations
#
%
530
50.7
385
36.8
89
8.5
27
2.6
15
1.4
Management of
messages
#
%
318
30.4
413
39.5
219
20.9
68
6.5
28
2.7
Management of
actions
#
%
151
14.4
254
24.3
312
29.8
197
18.8
132
12.6
Doing other things
Totally agree
Somewhat agree
Neutral
Somewhat disagree
Totally disagree
More support
busy traffic
#
%
323
30.9
460
44.0
172
16.4
66
6.3
25
2.4
#
102
118
116
170
540
%
9.8
11.3
11.1
16.3
51.6
Table 12. Answers to statements about the ideal system
Respondents thought the ideal driver support system should make car driving easier. It should provide
assistance similar to the ‘normal’ driving behaviour of the driver concerned. The system should be
easily adjustable, so that the driver can choose the type of assistance and feedback. Respondents
would like to have more help from the system during deviant situations, for example when they are
tired or when the weather is bad. Ideally, the system should include information management to
prevent the driver from overload and confusion in case of simultaneous information and warnings.
This outcome was less clear for automatic actions instead of messages. In case the ideal system takes
over driving tasks, most respondents did not want to do other things that are not related to driving.
McNemar tests were performed to check whether the answers to the statements about the ideal system
significantly differed from each other. These tests were especially interesting for the following pairs of
statements:
- Making car driving easier versus making car driving better;
- Giving more support in deviant situations versus giving more support in busy traffic;
- Prioritizing simultaneous messages versus prioritizing simultaneous actions.
So, for example, it was tested whether the number of respondents answering ‘totally agree’ to the
statement about making car driving easier (467) significantly differed from the number of respondents
answering ‘totally agree’ to the statement about making car driving better (373). This was also tested
for the combined category ‘(totally) agree’ (i.e. 885 versus 783). It appeared that the answers to all
pairs significantly differed from each other (p 0.001). This means that respondents preferred the ideal
system to make car driving easier instead of better. It also means that respondents would like to have
more support from the ideal system in deviant situations compared to busy traffic. Finally, prioritizing
simultaneous messages by the ideal system was preferred to prioritizing simultaneous automatic
actions. This last result may be related to the earlier observed preference for warnings to control
actions.
Ideal system in relation to single functions
It was assumed that drivers prefer a system that consists of less driver support functions than a
summing-up of ‘single’ functions. This summing-up might be ‘too much’, for example. In this
questionnaire, the ideal driver support system was not “directly” linked to the single driver support
functions and vice versa. Respondents had to formulate their ideal system on a higher level of
abstraction by indicating driving tasks and situations (not functions) that the system should support.
Although there was not a direct relationship, it was investigated to what extent the composition of the
ideal system was related to the indicated needs for the driver support functions. Driving tasks and
situations “corresponding” with the driver support functions for which respondents indicated a (great)
29
need, could or could not be chosen in the ideal system. Table 13 shows the types of driver assistance
in the ideal system that “correspond” to the ten most popular driver support functions. The last column
presents the % of respondents that had a (great) need for the concerned function, that also wanted the
ideal system to support with the corresponding driving task or situation.
Driver support function
1. Warning for downstream traffic condition – motorway
2. Warning for downstream traffic condition – rural road
3. Blind spot warning during lane changing – motorway
4. Blind spot warning at non-signalised intersection – urban road
5. Blind spot warning at non-signalised intersection – rural road
6. Blind spot warning during lane changing – rural road
7. Blind spot warning at signalised intersection – urban road
8. Blind spot warning at signalised intersection – rural road
9. Warning for imminent crash
10. Presentation of badly visible objects on windscreen
“Corresponding” task or situation in
ideal system
Regulating speed – motorway
Regulating speed – rural road
Lane changing – motorway
Non-signalised intersection – urban road
Non-signalised intersection – rural road
Lane changing – rural road
Signalised intersection – urban road
Signalised intersection – rural road
Imminent crash
Reduced visibility
%
44.4
33.8
11.8
21.9
24.4
8.5
2.6
2.3
63.9
64.9
Table 13. Types of assistance “corresponding” with most popular driver support functions
It can be seen that, for example, 64.9% of the respondents that indicated a (great) need for the
presentation of badly visible objects on the windscreen wanted the ideal system to help with reduced
visibility situations. In this case, most respondents that had a need for the single driver support system
related to reduced visibility also indicated a need for help from the ideal system with this situation.
However, this did not apply to, for example, help with negotiating signalised intersections on rural
roads. Only 2.3% of the respondents that indicated a (great) need for blind spot warnings at signalised
intersections on rural roads wanted the ideal system to help them with this driving task. It can be
concluded that when having to establish priorities, it seems that respondents thought it was more
important to receive other types of assistance from the ideal system, for example support during
reduced visibility and imminent crash situations, than support with negotiating signalised intersections
or lane changing.
Respondents did not have to indicate (again) what their most wanted driver support function per
favoured type of assistance in the ideal system would be. It was thought that this was too labourintensive. Besides, during the pilot test which did include this step, participants stated that it was much
of a repetition of what they did in the first part of the questionnaire. Based on this, we can carefully
say that table 14 shows the most wanted driver support functions that “correspond” to each favoured
type of assistance in the ideal system.
Type of assistance
1. Reduced visibility
2. Imminent crash
3. Car following – motorway
4. Regulating speed – motorway
5. Congestion driving – motorway
6. Driver fatigue
7. Regulating speed – rural road
8. Car following – rural road
9. Negotiating non-sign. intersection – rural road
10. Negotiating non-sign. intersection – urban road
“Corresponding” function
Presentation of badly visible objects on windscreen
Warning for imminent crash
Warning for unsafe following distance regarding actual situation
Warning for downstream traffic condition
The car automatically drives in congestion, incl. lane keeping
Drowsy driver warning
Warning for downstream traffic condition
Warning for unsafe following distance regarding actual situation
Blind spot warning
Blind spot warning
Table 14. Functions “corresponding” with most popular types of assistance in the ideal system
Remarks from respondents
Respondents had the possibility to make comments at the end of the survey. Irrespective of
information, warnings or automatic actions, respondents indicated to want to stay in full command of
30
the car. Generally, it should be up to the driver to decide whether or not the car should provide help
during car driving. At the same time, people indicated that when traffic safety comes into play, the car
could perform actions on its own, for example execute an emergency stop. Respondents thought it was
important to preserve the pleasure of driving. Therefore, automatic actions were considered less
attractive than information and warnings. However, there were also respondents that would enjoy an
“autopilot”, in particular on motorways, so that they could rest or do something else. Besides,
respondents indicated that they had a need for traffic and travel information, for example a navigation
system. It was remarkable that driver assistance was often linked to the behaviour of other road users.
For example, people thought that such systems were very useful to get rid of tailgaters or notorious
racers. Furthermore, it was said that driver assistance should be installed in each car, otherwise it
would be useless or even counterproductive (e.g. dangerous). It appeared that respondents were
worried about the alertness of the driver when supported by driver assistance, especially in case the car
takes over (parts of) the driving task. Other worries with respect to the driver, concerned information
overload, too much trust in technology, loss of driving ability and showing unwanted behaviour.
Furthermore, respondents expressed concerns with technology. For example, the system could fail.
Also, the costs should allow everyone to afford driver assistance. Some respondents had doubts about
the working of the system. That is, they thought technology could not replace the human driver. In
their view, driver assistance should only consist of information and warnings.
4.4
Influence of sample characteristics
It is interesting to investigate whether the results from the user needs survey were dependent on
characteristics of the sample. First, the effects of corrected proportions between the sample and the
target group on the results were studied by weighting cases. Second, the influence of driver
characteristics, such as age and driving experience, on the needs for driver assistance was examined.
Finally, it was tested whether the way participants were gathered for the survey was of influence.
Weighting cases
The sample of the user needs survey significantly differed from the target group with respect to
gender, age and education. Therefore, the results are only valid for this subset of the target group.
However, by weighting cases the proportions between the sample and the target group were corrected.
For example, little weight was attached to the answers of men and great weight to the answers of
women. Again, the needs for driver support functions and the composition of the ideal driver support
system were analysed. Table 15 shows the used weights for gender, age and education. No weights
were calculated for the age category ‘unknown’ and the education category ‘other’. These values were
considered as missing values (MV) in the analysis.
Characteristic
Gender
male
female
Age
18-24 years
25-44 years
45-64 years
65 years
unknown
Education
primary education
lower secondary education
higher secondary education
higher education/university
other
Sample
Target group
Weight
72.2%
27.8%
54%
46%
0.75
1.65
6.8%
51.9%
38.7%
2.6%
0.1%
16%
32%
32%
20%
-
2.35
0.62
0.83
7.69
MV
0.2%
9.2%
22.6%
67.8%
0.2%
10%
27%
30%
33%
-
50
2.93
1.33
0.49
MV
Table 15. Weights for gender, age and education
31
After weighting cases, it appeared that the top 10 of most popular driver support functions was
(almost) identical to the top 10 without weighting cases. This applied to the top 10 based on the
answer ‘great need’ as well as to the top 10 based on the answers ‘great need’ and ‘need’. Dependent
on the weighted variable (gender, age or education), 8 to 10 functions were the same. The most
favoured function was in all cases ‘warning for downstream traffic condition – motorway’. The order
of the rest of the functions sometimes differed.
With respect to the ideal driver support system, the top 10 of most favoured types of driver assistance
was identical. The order of only two types was turned (number 7 and 8) by weighting the variable age.
Weighting cases did not have a big influence on the most popular pairs of driver assistance in the ideal
system either. Dependent on the weighted variable, 9 to 10 pairs were the same. The three most
popular pairs were unchanged. The order of the rest of the pairs sometimes differed. In conclusion,
although the sample did not statistically resemble the target group, the results from the sample gave a
clear indication of those of the target group.
Driver characteristics
It was studied to what extent eight driver characteristics were of influence on the results from the user
needs survey. Table 16 shows these characteristics and their categories.
Characteristic
Gender (GEN)
Age (AGE)
Education (EDU)
Frequency of car driving (FRQ)
Years of driving experience (YRS)
Average annual mileage (KM)
Type of car possession (CAR)
Familiarity with ACC (ACC)
Categories
male, female
18-24 years, 25-44 years, 45-64 years, 65 years
primary & lower secondary, higher secondary, higher & university
>3 times a week, 1-3 times a week, 1-3 times a month, <1 time a month
<5 years, 5-10 years, >10 years
<10.000 km, 10.000-20.000 km, >20.000 km
private, business
not familiar with, somewhat familiar with, (very) familiar with
Table 16. Driver characteristics and categories on behalf of statistical tests
Driver support functions
First, it was explored to what extent driver characteristics affected the needs for the most popular
driver support functions (table 6). Therefore, chi-square tests were carried out. Two types of chisquare tests were performed. On the one hand, the tests focused on the group of respondents that
indicated to have a (great) need for the function in question. It was stated that driver characteristics
were not of influence on the perceived needs, when the distributions of the eight variables in this subsample were equal to the distributions of these variables in the total sample. On the other hand, the
tests focused on all respondents and their answers with respect to the needs for the functions in the top
10. In this way, it was studied to what extent driver characteristics were related to the combined
answer ‘(great) need’ (i.e. categories 1 and 2) versus the combined answer ‘other’ (i.e. categories 3, 4
or 5).
The first type of chi-square tests (only focussing on the respondents that indicated to have a (great)
need for the concerned functions) showed one significant result concerning the distribution of type of
car possession with respect to the group of respondents that indicated to have a (great) need for the
presentation of badly visible objects on the windscreen (p<0.05). Based on the distribution of this
variable in the total sample, it was expected that more private and less business car drivers would have
a need for this function; see table 17. The results indicated the opposite, so it can be said that this
driver support function was more popular among business drivers than private drivers.
32
Private
Business
Total
# observed
568
117
685
% observed
82.9
17.1
100.0
% sample
85.8
14.2
100.0
Table 17. Car possession and ‘presentation of badly visible objects on windscreen’ (sub-sample)
The second type of chi-square tests (focussing on all respondents) showed more significant results
next to the one mentioned above (p<0.05). Table 18 presents an example of the cross tabulation table
for gender and the perceived needs for warnings for downstream traffic conditions. As can be seen,
90.8% of the male respondents and 88.7% of the female respondents indicated to have a (great) need
for this driver support function. Chi-square tests showed that these percentages (90.8% vs. 88.7%) did
not significantly differ from each other. In other words, men and women appeared to have an equal
need for warnings for downstream traffic conditions.
Gender
Male
Female
Total
#
%
#
%
#
%
(Great) need
687
90.8
259
88.7
946
90.2
Other
70
9.2
33
11.3
103
9.8
Total
757
100.0
292
100.0
1049
100.0
Table 18. Gender and ‘warning for downstream traffic condition – motorway’ (total sample)
Table 19 shows the driver characteristics (see above for abbreviations) that had a significant influence
on the perceived needs for the most popular driver support functions (marked with X). It can be seen
that most types of blind spot warnings and warnings for imminent crash situations did not appear to be
related to the driver characteristics. So, respondents had a similar need for these functions, regardless
of their background.
Driver support function
Warning for downstream traffic condition – motorway
Warning for downstream traffic condition – rural road
Blind spot warning during lane changing – motorway
Blind spot warning at non-sign. intersection – urban road
Blind spot warning at non-sign. intersection – rural road
Blind spot warning during lane changing – rural road
Blind spot warning at signalised intersection – urban road
Blind spot warning at signalised intersection – rural road
Warning for imminent crash
Presentation of badly visible objects on windscreen
GEN
AGE
EDU
FRQ
YRS
KM
CAR
X
X
X
X
X
X
X
X
X
Table 19. Influences of driver characteristics on the needs for driver support functions (total sample)
Table 19 can be further explained as follows. Men appeared to have a greater need for the presentation
of badly visible objects on the windscreen during reduced visibility situations than women (71.9% vs.
65.1%). Respondents aged 45-64 years seemed to have a greater need for blind spot warnings than
respondents at the age of 25-44 years (76.1% vs. 67.8%). The other results suggested that older
respondents liked warnings for traffic in blind spots better than younger respondents. The level of
education did not seem to have an influence on the survey results. Respondents that drive more than 3
times a week had a stronger preference for the function ‘presentation of badly visible objects on
windscreen’ than respondents that drive 1-3 times a week (73.2% vs. 60.9%). The other results
suggested that frequently driving respondents favoured this function more than less frequently driving
33
ACC
X
X
respondents. This also applied to respondents that drive a lot based on their average annual mileage.
Respondents that drive more than 20.000 km a year appeared to have a greater need for this function
than respondents that drive less than 10.000 km a year (75.9% vs. 62.5%). Besides, respondents that
drive a lot based on their average annual mileage seemed to have a greater need for warnings for
downstream traffic conditions on rural roads. Respondents that have a driving licence for more than 10
years seemed to like blind spot warnings on signalised intersections on rural roads better than
respondents that have a driving licence for less than 5 years (73.4% vs. 64.5%). The other results
suggested that the longer one had his driving licence, the more one indicated a need for this driver
support function. The type of car possession seemed to be related to the perceived needs for three
functions. The results indicated that respondents that own a business car liked warnings for
downstream traffic conditions on rural roads better than respondents that own a private car (90.6% vs.
83.3%). Business car drivers also liked blind spot warnings when changing lanes on rural roads better
than private car drivers (81.9% vs. 73.9%). Besides, business car drivers indicated to have a greater
need for the presentation of badly visible objects on the windscreen (84.8% vs. 68.4%). This last result
was already found in the first type of chi-square tests. The perceived needs for warnings on
downstream traffic conditions were influenced by the familiarity with ACC. Respondents that are
(very) familiar with ACC seemed to have a greater need for this driver assistance on motorways and
rural roads compared to respondents that are not familiar with ACC (94.6% vs. 85.3% and 87.6% vs.
79.7%). The other results suggested that the more one is familiar with ACC, the more one had a need
for these driver support functions.
In section 4.1 it was found that 38 respondents said that they have an ACC system in their cars
(regardless of using in it). With respect to the current, low penetration rate of ACC in the Netherlands,
it was assumed that respondents confused ACC with regular cruise control. To study the possible
implications of this, chi-square tests without these 38 respondents were performed with respect to the
familiarity with ACC and the needs for driver assistance. The results indicated no differences
compared to the results that included the 38 respondents.
Ideal driver support system
It was studied to what extent driver characteristics affected the needs for the most popular types of
driver assistance in the ideal driver support system (table 9). Again, two types of chi-square tests were
carried out. On the one hand, the tests focused on the group of respondents that indicated to want the
ideal system to include the types of driver assistance in the top 10. On the other hand, the tests focused
on all respondents and their answers with respect to the types of driver assistance in the top 10.
The first type of chi-square tests (only focussing on the respondents that indicated to have a need for
the types of driver assistance in the ideal system) showed several significant results (p<0.05). For
some popular types of driver assistance, the distribution of several variables in the sub-sample differed
from the distribution of these variables in the total sample. Consider, for example, the distribution of
gender in the sub-sample that appeared to have a need for support with reduced visibility in the ideal
system versus the distribution of gender in the total sample; see table 20. It can be seen that 67.5% of
the respondents that indicated to have a need for this type of assistance was male and 32.5% was
female. This distribution appeared to significantly differ from the distribution of gender in the total
sample, namely 72.2% and 27.8% respectively. Thus, this type of driver assistance in the ideal driver
support system was more popular among women than men.
Male
Female
Total
# observed
427
206
633
% observed
67.5
32.5
100.0
% sample
72.2
27.8
100.0
Table 20. Gender and ‘support with reduced visibility in the ideal system’ (sub-sample)
34
Table 21 shows the types of driver assistance and the driver characteristics for which the distributions
did not resemble the distributions in the total sample.
Type of driver assistance
Reduced visibility
Imminent crash
Car following – motorway
Regulating speed – motorway
Congestion driving – motorway
Driver fatigue
Regulating speed – rural road
Car following – rural road
Negotiating non-sign. intersection – rural road
Negotiating non-sign. intersection – urban road
GEN
X
X
X
X
X
X
AGE
EDU
X
X
FRQ
YRS
KM
CAR
ACC
X
X
X
X
Table 21. Influences of driver characteristics on the composition of the ideal system (sub-sample)
Table 21 can be clarified as follows. Gender appeared to have the biggest influence on the types of
driver assistance wanted in the ideal driver support system. Men thought the ideal system should
mainly support with congestion driving on motorways and with regulating speed on motorways and
rural roads. However, women thought the ideal system should mainly support with reduced visibility
situations and negotiating non-signalised intersections on rural and urban roads. The results indicated
differences between certain age groups, but no ‘trend’ between young versus old could be found.
Especially respondents aged 25-44 years appeared to have a greater need for help with congestion
driving on motorways. However, this group seemed to have less need for help with car following on
rural roads compared to the other age groups. The results revealed that the higher one’s education, the
more one indicated a need for support with congestion driving on motorways. Also respondents that
drive more than 20.000 km a year had a great need for this type of support. The results further
suggested that the more one was familiar with ACC, the more one had a need for this type of driver
assistance in the ideal system as well. However, familiarity with ACC did not show a clear
relationship with the need for help with car following on rural roads. Respondents that were not
familiar or very familiar with ACC indicated to have a stronger preference for this kind of support in
the ideal system than respondents that were somewhat familiar with ACC.
The second type of chi-square tests (focussing on all respondents) showed more significant results
next to the ones mentioned above (p<0.05). Table 22 shows the driver characteristics (see above for
abbreviations) that had a significant influence on the perceived needs for types of driver assistance in
the ideal driver support system (marked with X). It can be seen that support with imminent crash
situations did not appear to be related to the driver characteristics. So, respondents had a similar need
for this type of support from the ideal driver support system, regardless of their background.
Especially needs for help with congestion driving on motorways was influenced by several driver
characteristics.
Type of driver assistance
Reduced visibility
Imminent crash
Car following – motorway
Regulating speed – motorway
Congestion driving – motorway
Driver fatigue
Regulating speed – rural road
Car following – rural road
Negotiating non-sign. intersection – rural road
Negotiating non-sign. intersection – urban road
GEN
X
X
X
X
X
X
X
X
AGE
EDU
X
X
X
X
FRQ
YRS
KM
CAR
X
ACC
X
X
X
X
X
Table 22. Influences of driver characteristics on the composition of the ideal system (total sample)
35
Table 22 can be further explained as follows. Gender appeared to have the biggest influence on the
types of driver assistance that are wanted in the ideal driver support system. Men seemed to have a
greater need for some of the assistance types than women. They thought the ideal system should
mainly support with congestion driving on motorways (45.0% vs. 35.1%) and with car following and
regulating speed on motorways (53.8% vs. 44.3% and 46.6% vs. 36.4%) and car following and
regulating speeds on rural roads (36.8% vs. 26.1% and 30.5% vs. 23.4%). However, women thought
the ideal system should mainly support with reduced visibility situations (70.8% vs. 56.6%) and
negotiating non-signalised intersections on rural and urban roads (29.9% vs. 19.5% and 26.8% vs.
17.2%). The relation between age and the composition of the ideal system was unclear. The results
from four types of driver assistance indicated significant differences between certain age groups, but
no ‘trend’ between young versus old, as above with the needs for driver support functions, could be
found. Especially respondents aged 45-64 years seemed to have a greater need for support from the
ideal system with regulating speed on rural roads. Respondents at the age of 65+ years indicated the
ideal system to mainly support with negotiating non-signalised intersections on urban roads. This
group of respondents also seemed to have a greater need for support with negotiating non-signalised
intersections on rural roads compared to respondents at the age of 25-44 years (48.1% vs. 24.4%). The
needs for support with congestion driving on motorways between these two age groups were opposite.
Respondents aged 25-44 years appeared to have a greater need for help with this driving task
compared to respondents aged 65+ years (46.0% vs. 14.8%). The level of education also seemed to
have an influence on the need for support with congestion driving on motorways. It appeared that
highly-educated respondents would like the ideal system to help them with this driving task better than
the least educated respondents (46.2% vs. 27.3%). The other results suggested that the higher one’s
education, the more one indicated a need for this type of driver assistance in the ideal system. The
frequency of car driving and driving experience in years did not appear to have an influence on the
composition of the ideal system; driving experience in kilometres did. Respondents that drive more
than 20.000 km a year had a greater need for support with congestion driving on motorways compared
to respondents that drive 10.000-20.000 km a year (50.6% vs. 36.7%). Similar results applied to
support from the ideal system with driver fatigue situations between respondents that drive more than
20.000 km a year compared to respondents that drive less than 10.000 km a year (40.1% vs. 31.9%).
The other results suggested that the higher one’s average annual mileage, the more one would like the
ideal system to help with congestion driving on motorways and driver fatigue situations. The type of
car possession had a significant influence on the need for support with reduced visibility situations.
Business car drivers appeared to have a greater need for this type of assistance than private car drivers
(71.7% vs. 58.7%). Help with congestion driving was more wanted by respondents that are (very)
familiar with ACC compared to respondents that are not familiar with this system (48.4% vs. 34.2%).
The other results suggested that the more one is familiar with ACC, the more one has a need for this
type of driver assistance in the ideal system. However, familiarity with ACC did not show a clear
relationship with the need for help with car following on rural roads, although this relationship
appeared to be significant. Respondents that are not familiar with ACC indicated to have a stronger
preference for this kind of support in the ideal system than respondents that are somewhat familiar
with ACC (33.5% vs. 23.9%). Respondents that are very familiar with ACC indicated a similar
preference for this kind of support than respondents that are not familiar with ACC.
As above, chi-square tests without the 38 respondents, (presumably) indicating unjustly to have an
ACC system in their cars, were performed with respect to the familiarity with ACC and the needs for
types of driver assistance in the ideal driver support system. The results indicated no differences
compared to the results which included the 38 respondents.
Summary
It may be stated that driver characteristics appeared to influence the perceived needs for certain driver
support functions and types of assistance in the ideal driver support system. However, these influences
were sometimes ambiguous. The following relationships were most clear (p 0.001) (see also
appendix F):
36
• Average annual mileage & presentation of badly visible objects on the windscreen: respondents
that drive many kilometres had a greater need for this driver support function than respondents that
drive fewer kilometres
• Type of car possession & presentation of badly visible objects on the windscreen: business car
drivers had a greater need for this driver support function than private car drivers
• Gender & help with reduced visibility: women had a greater need for this kind of support in the
ideal driver support system than men
• Gender & help with regulating speed on rural roads: men had a greater need for this kind of support
in the ideal driver support system than women
• Gender & help with negotiating non-signalised intersections on rural roads: women had a greater
need for this kind of support in the ideal driver support system than men
• Gender & help with negotiating non-signalised intersections on urban roads: women had a greater
need for this kind of support in the ideal driver support system than men
• Education & help with congestion driving: higher educated respondents had a greater need for this
kind of support in the ideal driver support system than lower educated respondents
• Average annual mileage & help with congestion driving: respondents that drive many kilometres
had a greater need for this kind of support in the ideal driver support system than respondents that
drive fewer kilometres
Driver characteristics did not appear to have an effect on the perceived needs for support with
imminent crash situations. Evidently, all respondents regardless of their background thought this kind
of support was important. Also most ‘types’ of blind spot warnings were not influenced by
characteristics of the driver. Education did not have an influence on the needs for the ten most popular
driver support functions, nor did frequency of car driving and years of driving experience on the ten
most popular types of driver assistance in the ideal driver support system. It was surprising that the
‘single’ function with respect to reduced visibility was more wanted by men than by women, while
help from the ideal system with this situation was more wanted by women than by men. Apparently, it
is of importance whether respondents had to indicate their needs for ‘single’ functions compared to
choosing their most favoured types of driver assistance in the ideal driver support system.
Response group
Participants for the survey were invited in two ways, namely via a market agency and via personal and
business contacts (see 3.4). Both response groups belonged to the target group (i.e. Dutch people
holding a driving licence) and had to fill in the same questions about their needs for driver support.
Therefore, it was decided to analyze both groups together. To check whether this had consequences
for the results from the survey, statistical analyses were performed.
First, the two groups were compared to each other with respect to the eight driver characteristics (e.g.
gender, age) by means of chi-square tests. Results from these tests showed that the composition of the
groups was not the same. The groups significantly differed from each other with respect to all of the
characteristics, except for average annual mileage and type of car possession (p<0.05). For example,
the ‘Internet panel’ group consisted of (1) more female respondents, (2) more older respondents, (3)
less highly-educated respondents, (4) more respondents that drove over three times a week, (5) more
respondents with over 10 years of driving experience and (6) less respondents that were (very) familiar
with ACC.
Second, chi-square tests were performed to see whether the response group had an influence on the top
10 of most popular driver support functions (table 6). As above, two types of analyses were
performed: (1) only focussing on the sub-sample indicating to have a (great) need for the concerned
driver support function and (2) focussing on all respondents. The first type of chi-square tests showed
no significant results (p<0.05). This means that for each popular function, the distribution of the
variable ‘response group’ in the sub-sample did not differ from the distribution of this variable in the
total sample. Consider, for example, response group and the perceived needs for warnings for
37
downstream traffic conditions on motorways in table 23. It can be seen that 71.0% of the respondents
that indicated to have a (great) need for this function was from the ‘self-gathered’ group and 29.0%
was from the ‘Internet panel’ group. This distribution appeared not to significantly differ from the
distribution of response group in the total sample, namely 70.5% and 29.5% respectively.
Self-gathered
Internet panel
Total
# observed
672
274
946
% observed
71.0
29.0
100.0
% sample
70.5
29.5
100.0
Table 23. Response group and ‘warning for downstream traffic condition – motorway’ (sub-sample)
The second type of chi-square tests showed significant results for blind spot warnings at (non-)
signalised intersections on rural roads (p<0.05). Thus, it appeared that response group had an influence
on the perceived needs for these two driver support functions. Respondents from the Internet panel
group showed a greater need for blind spot warnings at non-signalised intersections on rural roads than
respondents from the self-gathered group (84.5% vs. 77.3%). This also applied to blind spot warnings
at signalised intersections on rural roads (76.1% vs. 69.5%). Besides response group, age and years of
driving experience appeared to have an influence on the needs of this latter driver support function
(see above). However, no other driver characteristics than response group appeared to influence the
needs for the former driver support function. In this case, the way respondents were invited to the
survey probably affected the perceived needs for driver assistance.
Third, chi-square tests were also performed to see whether the response group had an influence on the
top 10 of most popular types of driver assistance in the ideal driver support system (table 9). Again,
two types of analyses were performed: (1) only focussing on the sub-sample indicating to want the
concerned type of driver assistance in the ideal system and (2) focussing on all respondents. Both
types of analyses showed significant results with respect to support from the ideal system with
congestion driving on motorways. It appeared that the distribution of the variable ‘response group’ in
the sub-sample significantly differed from the distribution of this variable in the total sample (p=
.007). Based on the distribution of response group in the total sample, it was expected that more
respondents from the Internet panel group and less respondents from the self-gathered group would
have a need for this type of driver assistance. The results indicated the opposite, so it can be said that
this type of driver assistance was more popular among respondents from the self-gathered group. This
result corresponds to the result from the second type of chi-square tests. Namely, it was found that
respondents from the self-gathered group appeared to have a greater need for the ideal system to
support with congestion driving on motorways than respondents from the Internet panel group (44.9%
vs. 35.9%). This can possibly be explained by the different composition of the two groups with respect
to gender, age, education and familiarity with ACC. These variables already appeared to have an
influence on the perceived needs for types of driver assistance in the ideal driver support system.
Summarizing, it may be stated that response group appeared to have an influence on the perceived
needs for certain driver support functions and types of assistance in the ideal driver support system.
Probably, this influence can be attributed to the different composition of the two groups (self-gathered
and Internet panel) with respect to background characteristics, such as gender, age and education.
38
5.
Discussion
This section discusses the results from the user needs survey. First, implications for the possible
integration of driver support functions are considered. Next, the results are compared to what is known
from the conceptual model of user needs for driver assistance. Finally, advantages and disadvantages
of the used method (i.e. survey) will be dealt with as well.
5.1
Implications for integrated driver assistance
The aim of the survey was to reflect the needs of the driver with respect to driver assistance.
Generally, it seems that respondents want to be well informed when driving. It was noticed that
respondents had a need for a reasonable number of driver support functions. Besides, the ideal system
should certainly provide help with more than one driving task or situation. These needs of the driver
will have consequences for the possible integration of driver support functions. It was recognised that
a shift has to be made from ADAS that only include one kind of support to ADAS that consist of
integrated driver assistance. Technical integration should be taken into account, so that different
functions can use the same components, such as sensors. Aspects surrounding the Human-MachineInterface (HMI) are important to prevent the driver from overload and confusion. Different functions
should use one interface with information management to give priority to certain information. The
HMI should also be easily adjustable, because respondents would like to have the possibility to choose
the type of assistance and feedback. Attention should be given to the functional operation of integrated
driver assistance as well. ADAS should make use of each others information, for example by intervehicle communication. This applies to functions such as warnings for downstream traffic conditions.
Communication between functions within the vehicle needs also consideration. Imagine a driver
support system that includes assistance with car following and assistance during reduced visibility.
The functions in question should exchange information to extend their individual field of activity, for
example, in terms of maintaining an increased headway in reduced visibility situations.
5.2
Contribution to conceptual model
Based on earlier research into user needs for driver assistance, a conceptual model was formulated (see
2.2). This model related these user needs to characteristics of the driver, the driver assistance and the
traffic scene. The user needs survey aimed at filling knowledge gaps in the conceptual model and
moving the state-of-the-art on (integrated) driver assistance forward. Figure 4 shows the conceptual
model. The opinion of the driver on driver assistance is presented by arrow [1]. This opinion may be
influenced by characteristics of the driver and the system. Besides, it may be influenced by
characteristics of the traffic scene; see arrow [2]. The bold characteristics appeared to influence the
perceived needs for driver assistance according to the results from the survey. Characteristics in
parentheses were not taken into account in this survey.
In this research, it was studied to what extent driver characteristics affected the perceived needs for the
ten most popular (1) driver support functions and (2) types of assistance in the ideal driver support
system. The influences of driver characteristics were sometimes ambiguous. However, gender
revealed to have the most connections with the perceived needs. Men and women had equal needs for
driver assistance. However, these needs differed per driving task or situation. For example, women
had a greater need for help from the ideal system with reduced visibility situations and negotiating
non-signalised intersections on rural and urban roads, while men had a greater need for help with
regulating speeds on rural roads. The average annual mileage appeared to be related to a need for
support with congestion driving or driver fatigue. The more kilometres one drove, the more one
indicated to have a need for these kinds of support. Driver characteristics did not appear to have an
effect on the perceived needs for support with imminent crash situations. Evidently, all respondents
39
regardless of their background thought this kind of support was important. It must be noted, that the
results only focussed on the most favoured driver assistance. It might be possible, that more relations
between driver characteristics and perceived needs for driver support appear, when all results are
considered. However, due to time constraint, these analyses were not conducted.
Traffic scene
•
•
•
•
Critical situation
Weather/visibility
Road type
Other traffic
Driver
•
•
•
•
•
•
•
•
•
•
•
•
Gender
Age
Education
Driving experience
Type of car possession
Familiarity with ADAS
(Nationality)
(Target group)
(Driving style)
(Psychological factors)
(Family situation)
(Income)
[2]
Driver assistance
[1]
•
•
•
•
•
System/function
Level of support
(Feedback)
(Overrulability)
(Price)
Vehicle
Figure 4. Influence of several characteristics on the needs for driver assistance
Compared to driver characteristics, aspects of the driver assistance and the traffic scene were more
likely to affect the perceived needs for driver support. Respondents indicated to have the greatest need
for driver support functions related to regulating speed, lane changing, negotiating intersections,
imminent crash and reduced visibility. Especially warnings for downstream traffic conditions and
warnings for traffic in blind spots were favoured. The ideal driver support system should, according to
the respondents, provide help with (combinations of) reduced visibility, imminent crash and car
following on motorways. Respondents indicated in particular to have (certainly) no need for any of the
driver support functions concerning lane keeping. This type of assistance was also very unpopular in
the ideal system. These findings are in accordance with what is already known from earlier literature.
Respondents would mainly like their cars to help them by giving information or warnings. Automatic
actions from the car were unpopular. However, in some cases respondents did want the car to take
over: when they want to maintain a self-chosen speed on motorways or rural roads and when they are
driving in traffic jams, irrespective of the road type. Generally, control functions were less popular,
because drivers indicated to want to stay in full command of the car. This fits in with results from
literature.
According to the results from the survey, respondents thought that the ideal driver support system
should generally make car driving easier. The system should provide assistance similar to the ‘normal’
driving behaviour of the driver concerned. However, it also seems that drivers would like their cars to
help them in critical situations. For example, warnings for an imminent crash and assistance during
reduced visibility supported this view. This is consistent with earlier research. Besides, most
respondents indicated that the ideal system should give more support in busy traffic and deviant
situations, for example when they are tired or when the weather is bad. In general, the results showed
that there was a great need for driver assistance during driving on motorways. However, driver support
40
functions that were more wanted on rural roads than on motorways and urban roads were information
on speed limit and warning for exceeding speed limit. There appeared to be no greater need for
specific functions on urban roads compared to rural roads.
5.3
Conducting the survey
Drivers can be regarded as ‘hands-on’ experts in car driving. Therefore, it seems relevant to ask them
to indicate their needs for driver assistance. However, this methodology can bring along some
limitations. A limitation of this user needs survey is the fact that the driver support functions were
presented hypothetically. A few respondents said at the end of the survey, that it was rather hard to
indicate whether or not they had needs for the presented driver assistance, because they did not have
any experience with it. Instead of needs, one could also speak of interests. To some extent, it is
uncertain whether respondents understood everything correctly. Pilot testing was done to diminish the
chance of wrong comprehension. The survey included one picture to clarify the ‘vision enhancement’
function. This function turned out to be the tenth most popular driver support function. It is unclear
whether presenting the picture had an influence on the perceived popularity of the function. A few
respondents wrote down at the end of the survey that the 5-point rating scales seemed contra-intuitive
to them, because 1 concerned a positive answer and 5 a negative. However, it is thought that
respondents did not mix up the answering categories, because a mouse tip was included that showed
the meaning of the number next to the pointer of the mouse (see appendix A). It was expected that
giving a socially desirable answer played a role when answering the statement on the ideal driver
support system about doing other things while the car takes over driving tasks. Most respondents did
not want to do other things that are not related to driving. However, this conflicts with what often can
be seen in the ‘real’ world. Furthermore, respondents may be influenced by their familiarity with
forms of driver assistance. In this study, it appeared that familiarity with ACC affected the perceived
needs for warnings for downstream traffic conditions and help with car following and congestion
driving on motorways. This possibly also explains the great perceived need for support on motorways,
because systems already available on the market, such as regular cruise control and ACC, are merely
designed to operate on motorways. Besides, research shows that in most cases drivers were more
positive about a driver support system after having gained experience with it (TRG, 2003). This user
needs survey will be continued by means of a driving simulator experiment (see 6.2). In this way the
predictive power of the survey will be investigated.
In this questionnaire, respondents had to express the extent to which they would like to have driver
assistance by indicating their needs for driver support functions and by formulating their ideal driver
support system. The ideal system was not “directly” linked to the single driver support functions,
because respondents had to indicate driving tasks and situations (not functions) that the system should
provide support with. For future research, it is recommended to formulate the ideal system in terms of
driver support functions. For example, respondents to this survey were presented a ‘personalized’ table
that showed all driver support functions for which the respondent had a significant need according to
his or her previous answers. Instead of choosing driving tasks and situations to be supported by the
ideal system, respondents could also have indicated which driver support functions from the
‘personalized’ table should be included in the ideal system.
Several ways exist to distribute a questionnaire. In this study, the choice was made to design a
computer-based questionnaire for distribution on the Internet. In spite of the many benefits of this
approach, there is one major drawback. Access via the Internet introduces bias, because only Internet
users can fill in the questionnaire. This could also be seen from the frequency of Internet usage: more
than 97% of the respondents indicated to regularly make use of the Internet. However, it should be
kept in mind that the number of current Internet users is high and still growing (TNS NIPO, 2002).
Another limitation may be the selection of participants. Some people may be more willing to
participate in the survey than others, for example, because they like the subject of the survey
(McKinsey, 2005). In particular, the invitation of participants via personal and business contacts can
41
lead to inherent bias. Statistical tests revealed that response group (i.e. self-gathered or Internet panel)
appeared to have an influence on the perceived needs for certain driver support functions and types of
assistance in the ideal driver support system. Probably, this influence can be attributed to the different
composition of the two groups with respect to background characteristics, such as gender and age. The
total sample appeared not to significantly resemble the target group of Dutch car drivers. However,
weighting cases compensated for this bias. It was stated that the results from the sample could be
regarded as strongly directional for those of the target group.
42
6.
Conclusions and future work
In this chapter conclusions are drawn and an outlook for further work is highlighted.
6.1
Conclusions
The car driver of tomorrow will enjoy an increasing variety of in-vehicle systems that assist him or her
in the driving task. This report discussed the perceived needs of the driver for driver assistance. By
means of a user needs survey, car drivers were asked to indicate their needs for support from their cars
during certain driving tasks and situations. It appeared that warnings for downstream traffic conditions
and warnings for traffic in blind spots were favoured. Besides, the respondents preferred the ideal
system to give support in critical situations, such as an imminent crash and reduced visibility. Based
on these needs, it was recognised that a shift has to be made from ADAS that only include one kind of
support to ADAS that consist of integrated driver assistance. Because most preferred driver support
functions consist of information and warnings, issues surrounding the HMI are important. Different
functions should use one interface with information management to prevent the driver from overload
and confusion during car driving. Furthermore, functional integration should be emphasized.
Functions should make use of each other’s information to extend their individual field of activity.
Consider, for example, warnings for imminent crash situations based on sensor data fusion. Or think of
inter-vehicle communication that applies to functions such as warnings for downstream traffic
conditions.
6.2
Future work
As explained in the introduction, the results from this user needs survey serve as a basis for creating an
integrated driver support system. Based on the survey results, a “Congestion Assistant” was
developed. This system fits with the following needs that scored high in the user needs survey:
• Warnings for downstream traffic conditions
• Help with congestion driving
The Congestion Assistant provides the driver with the following types of help dependent on the state
of the traffic flow; see table 24.
Phase
Before traffic jam
Just before traffic jam
In traffic jam
Driver support function
Warning for downstream congestion (display)
Support with slowing down by haptic feedback (active gas pedal)
Take-over of longitudinal driving task (automatic Stop & Go)
Information on duration of congestion (display)
Table 24. Characteristics of the “Congestion Assistant”
It is assumed that the system can decide on the state of the traffic flow, and thus the type of help to
provide, based on data from the surrounding traffic by means of inter-vehicle communication. The
transitions between the phases are of particular interest. The impacts of the Congestion Assistant on
the driver were investigated by means of a driving simulator experiment. This experiment took place
at TNO in Soesterberg in May 2005. 37 subjects drove four trips in different conditions. These
conditions consisted of the combinations of two factors: system (on, off) and visibility (normal, fog).
The fog condition was added, because help from the car during reduced visibility situations was often
mentioned in the user needs survey. Data analysis will focus on the effects of the system on driving
behaviour, workload and acceptance. Driving behaviour was investigated by measuring the lateral and
longitudinal control performance. Workload was assessed by means of the Peripheral Detection Task
(PDT), heart rate and the subjective mental effort scale (RSME). Furthermore, questionnaires were
43
used to examine the acceptance of and the willingness to buy the Congestion Assistant (e.g. Van der
Laan scale and Juster scale).
The outcomes of the driver simulator experiment will be placed in perspective by assessing the
impacts of the Congestion Assistant on the traffic flow, in terms of traffic safety and efficiency, using
microscopic traffic simulation.
44
References
Becker, S. Brockmann, M., Mertens, A., Niu, R. and Sonntag, A. (1995) ‘User Acceptance and
Willingness to Pay for Advanced Driver Support Systems’. Paper presented at the TRAFFIC
Technology Europe Conference, Berlin
Bekiaris, E., Petica, S. and Brookhuis, K.A. (1997) ‘Driver needs and public acceptance regarding
telematic in-vehicle emergency control aids’. Paper presented at the 4th World Congress on ITS,
Berlin
Blackwell, R.D., Miniard, P.W. and Engel, J.F. (2001). Consumer behavior, Harcourt College
Publishers, Fort Worth
Blythe, P.T. and Curtis, A.M. (2004) ‘Advanced Driver Assistance Systems: gimmick or reality?’.
Paper presented at the 11th World Congress on ITS, Nagoya
Bridger, R. and Patience, C. (1998) Consumer Attitudes to and Acceptance of Driver Assistance
Technologies,
http://www.aapolicy.com/aamotoringtrust/pdf/consumer_attitudes.pdf,
consulted:
March 2003
Buijs, A. (1993) Statistiek om mee te werken, Stenfert Kroese, Leiden [in Dutch]
Büsher and Frese (1998) IN-ARTE; User needs survey, Deliverable 3.2
Cauzard, J.-P. (ed.) (1998) The attitude and behaviour of European car drivers to road safety: Project
on Social Attitudes to Road Traffic Risk in Europe SARTRE 2; Part 2: report on in-depth analyses,
http://sartre.inrets.fr/documents-pdf/repS2V2E.pdf, consulted: January 2004
CBS Statistics Netherlands (2003) Statline, http://statline.cbs.nl, consulted: January 2004 [in Dutch]
Chalmers, I.J. (2001) ‘User attitudes to automated highway systems’. Paper presented at the IEE
International Conference on Advanced Driver Assistance Systems, Birmingham
Cohen, J. (1988) Statistical power analysis for the behavioral sciences, Lawrence Erlbaum Associates,
Hillsdale
Comte, S.L., Wardman, M. and Whelan, G. (2000) ‘Drivers’ acceptance of automatic speed limiters:
implications for policy and implementation’, Transport Policy, Vol. 7, no. 4, pp. 259-267
CRA (1998) Consumer Acceptance of Automotive Avoidance Devices; A report of qualitative
research, Charles River Associates Inc., Boston
European Commission (2002) Final Report of the eSafety Working Group on Road Safety, Brussels
Garvill, J., Marell, A. and Westin, K. (2003) ‘Factors influencing drivers’ decision to install an
electronic speed checker in the car’, Transportation Research Part F, Vol. 6, no. 1, pp. 37-43
Marchau, V.A.W.J. (2000) Technology assessment of Automated Vehicle Guidance; Prospects for
automated driving implementation, Delft University Press, Delft
Marchau, V.A.W.J., Penttinen, M., Wiethoff, M. and E.J.E. Molin (2001) ‘Stated preferences
regarding Advanced Driver Assistance Systems (ADAS) of European drivers’, European Journal of
Transport and Infrastructure Research, Vol. 1, no. 3, pp. 291-308
Mariani, M., Veltri, G., Montanari, R., Peterson, D., Karlsson, B. and Amditis, A. (2000)
COMUNICAR; User needs, Deliverable 2.2
McIntosh, J.L. (1997) ‘Acceptance of '
ITS'by Australian motorists’. Paper presented at 4th World
Congress on ITS, Berlin
McKinsey (2005) 21minuten.nl, http://www.21minuten.nl, consulted: May 2005 [in Dutch]
45
McKnight, A.J, and Adams, B.B. (1970) Driver Education Task Analysis: Volume 1. Task
Descriptions, National Highway Traffic Safety Administration, Washington DC
Michon, J.A. (1985) ‘A critical view of driver behavior models: what do we know, what should we
do?’, in Evans, L. and Schwing, R.C. (eds.) Human behavior and traffic safety, Plenum Press, New
York, pp. 485-524
Ministry of Transport, Public Works and Water Management (2004) Nota Mobiliteit; Naar een
betrouwbare en voorspelbare bereikbaarheid, The Hague [in Dutch]
Ortúzar, J. de D. (2000) Stated preference modelling techniques, PTRC Education & Research
Services, London
Piao, J., McDonald, M. and Vöge, T. (2004) ‘An assessment of user acceptance of ADAS/AVG
systems from a questionnaire in Southampton’. Paper presented at the 11th World Congress on ITS,
Nagoya
Polydoropoulou, A., Gopinath, D.A. and Ben-Akiva, M. (1997) ‘Willingness to pay for Advanced
Traveler Information Systems; SmarTraveler case study’, Transportation Research Record 1588
Regan, M.A., Mitsopoulos, E., Haworth, N., and Young, K.L. (2002) Acceptability of in-vehicle
intelligent transport systems to Victorian drivers, Monash University Accident Research Centre,
Victoria
Rice, J.A. (1994) Mathematical statistics and data analysis, Duxbury Press, Belmont
Rienstra, S. and Rietveld, P. (1996) ‘Speed behavior of car drivers’, Transportation Research Part D,
Vol. 1, no. 2, pp. 97-110
Schmeidler, K. (2002) ‘Acceptance of Advanced Assistance Systems by Czech Drivers’. Paper
presented at the Safety on Roads International Conference (SORIC), Bahrain
SWOV Institute for Road Safety Research (2002) BIS-V, http://www.swov.nl/cognos/cgibin/ppdscgi.exe, consulted: January 2004
TNS NIPO (2002) E-monitor, http://www.tns-nipo.com, consulted: June 2004
TRG (1998) DIATS; User preference and behavioural changes from ATT systems, Deliverable 13-14
TRG (2003) STARDUST; Assessment of behavioural acceptance of intelligent infrastructure and
ADAS/AVG systems, Deliverable 4
Turrentine, T., Sperling, D. and Hungerford, D. (1991) ‘Consumer acceptance of adaptive cruise
control and collision avoidance systems’, Transportation Research Record 1318
Van der Heijden, R.E.C.M. and Molin, E.J.E. (1999) ‘The societal support for electronic driver
support systems; The case of the intelligent speed adapter’, in Van der Heijden, R.E.C.M. and
Wiethoff, M. (eds.) Automation of Car Driving; Exploring societal impacts and conditions, Delft
University Press, Delft, pp. 193-208
Van Driel, C.J.G. & Van Arem, B. (2004) ‘What about the integration of driver support functions?’.
Paper presented at the 11th World Congress on ITS, Nagoya
Van Hoorebeeck, B. (2000) Meting van het maatschappelijk draagvlak voor intelligente
snelheidsbegrenzers, BIVV Belgisch Instituut voor de Verkeersveiligheid, Brussels [in Dutch]
Várhelyi, A. (2000) ‘A large scale trial with Intelligent Speed Adaptation in Lund, Sweden’. Paper
presented at the 7th World Congress on ITS, Turin
Wevers, K., Bekiaris, A., Boverie, S., Burns, P., Frese, T., Harms, L., Martens, M., Saroldi, A., Spigai,
M. and Widlroither, H. (1999) ‘Integration of driver support systems; The IN-ARTE Project’. Paper
presented at the 6th World Congress on ITS, Toronto
46
Young, K.L., Regan, M.A., Mitsopoulos, E. and Haworth, N. (2003) Acceptability of in-vehicle
intelligent transport systems to young novice drivers in NSW, Monash University Accident Research
Centre, Victoria
47
48
Appendix A – The questionnaire
This appendix shows the computer screens of the user needs survey when it was online on the Internet.
The survey was in Dutch.
Introduction
49
Background questions
50
Information about help from the car during driving
51
Questions about the needs for driver support functions
52
53
54
When respondents indicated in the first part to have (certainly) no need for driver support with respect
to the concerned driving task, for example ‘afstand houden’ (car following), the following extra
question was asked. This extra question was about the extent to which the participant would have a
need for help from the car with the driving task during deviant situations, for example when the
weather is bad.
55
56
57
58
59
‘Personalized’ table with indicated needs for driver support functions based on the respondent’s
previous answers
60
Questions about the ideal driver support system
61
When respondents forgot to give an answer, they got the following message to go back and please
answer the question.
62
63
Background questions, including a question about the willingness to participate in the follow-up
of this user needs survey: a driving simulator study
64
End of questionnaire with possibility for comments
65
66
Appendix B – The AIDA user group
The AIDA knowledge centre has a user group in which interested parties from the government and
business world participate. The purpose of the user group is to maximise the usefulness of the research
results by having the group give feedback before and during the research on the plans and intermediate
results.
The following organisations are represented in the user group:
- ANWB
- BOVAG
- DaimlerChrysler, Research and Technology, Information and Communication / Assisting Systems
- Groeneveld
- KNV
- KPN
- Ministry of Transport, Public Works and Water Management, Directorate-General for Passenger
Transport (DGP)
- TNO Science and Industry – Automotive
- Vialis Traffic & Mobility
67
68
Appendix C – Driver support functions: frequencies
This appendix shows the results from the frequency statistics with respect to the 113 presented driver
support functions. Note that this appendix is for the most part in Dutch.
a = auto(snel)weg – motorway
p = provinciale weg – rural road
s = weg in stad/dorp – urban road
Snelheid regelen – Regulating speed
Informatie over snelheidslimiet - a
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
211
402
209
153
74
1049
Percent
20.1
38.3
19.9
14.6
7.1
100.0
Valid Percent
20.1
38.3
19.9
14.6
7.1
100.0
Cumulative
Percent
20.1
58.4
78.4
92.9
100.0
Informatie over snelheidslimiet - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
272
399
243
96
39
1049
Percent
25.9
38.0
23.2
9.2
3.7
100.0
Valid Percent
25.9
38.0
23.2
9.2
3.7
100.0
Cumulative
Percent
25.9
64.0
87.1
96.3
100.0
Informatie over snelheidslimiet - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
218
318
224
193
96
1049
Percent
20.8
30.3
21.4
18.4
9.2
100.0
Valid Percent
20.8
30.3
21.4
18.4
9.2
100.0
Cumulative
Percent
20.8
51.1
72.4
90.8
100.0
69
Waarschuwing bij overschrijden van geldende snelheidslimiet - a
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
176
365
245
182
81
1049
Percent
16.8
34.8
23.4
17.3
7.7
100.0
Valid Percent
16.8
34.8
23.4
17.3
7.7
100.0
Cumulative
Percent
16.8
51.6
74.9
92.3
100.0
Waarschuwing bij overschrijden van geldende snelheidslimiet - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
226
391
252
122
58
1049
Percent
21.5
37.3
24.0
11.6
5.5
100.0
Valid Percent
21.5
37.3
24.0
11.6
5.5
100.0
Cumulative
Percent
21.5
58.8
82.8
94.5
100.0
Waarschuwing bij overschrijden van geldende snelheidslimiet - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
214
357
235
158
85
1049
Percent
20.4
34.0
22.4
15.1
8.1
100.0
Valid Percent
20.4
34.0
22.4
15.1
8.1
100.0
Cumulative
Percent
20.4
54.4
76.8
91.9
100.0
De auto houdt automatisch snelheid constant volgens geldende snelheidslimiet - a
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
132
273
235
224
185
1049
Percent
12.6
26.0
22.4
21.4
17.6
100.0
Valid Percent
12.6
26.0
22.4
21.4
17.6
100.0
Cumulative
Percent
12.6
38.6
61.0
82.4
100.0
70
De auto houdt automatisch snelheid constant volgens geldende snelheidslimiet - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
131
268
240
237
173
1049
Percent
12.5
25.5
22.9
22.6
16.5
100.0
Valid Percent
12.5
25.5
22.9
22.6
16.5
100.0
Cumulative
Percent
12.5
38.0
60.9
83.5
100.0
De auto houdt automatisch snelheid constant volgens geldende snelheidslimiet - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
110
191
204
312
232
1049
Percent
10.5
18.2
19.4
29.7
22.1
100.0
Valid Percent
10.5
18.2
19.4
29.7
22.1
100.0
Cumulative
Percent
10.5
28.7
48.1
77.9
100.0
Waarschuwing bij overschrijden van zelfgekozen snelheid - a
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
121
317
262
226
123
1049
Percent
11.5
30.2
25.0
21.5
11.7
100.0
Valid Percent
11.5
30.2
25.0
21.5
11.7
100.0
Cumulative
Percent
11.5
41.8
66.7
88.3
100.0
Waarschuwing bij overschrijden van zelfgekozen snelheid - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
125
317
263
228
116
1049
Percent
11.9
30.2
25.1
21.7
11.1
100.0
Valid Percent
11.9
30.2
25.1
21.7
11.1
100.0
Cumulative
Percent
11.9
42.1
67.2
88.9
100.0
71
Waarschuwing bij overschrijden van zelfgekozen snelheid - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
118
262
252
264
153
1049
Percent
11.2
25.0
24.0
25.2
14.6
100.0
Valid Percent
11.2
25.0
24.0
25.2
14.6
100.0
Cumulative
Percent
11.2
36.2
60.2
85.4
100.0
De auto houdt automatisch zelfgekozen snelheid constant - a
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
232
401
208
121
87
1049
Percent
22.1
38.2
19.8
11.5
8.3
100.0
Valid Percent
22.1
38.2
19.8
11.5
8.3
100.0
Cumulative
Percent
22.1
60.3
80.2
91.7
100.0
De auto houdt automatisch zelfgekozen snelheid constant - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
171
384
250
147
97
1049
Percent
16.3
36.6
23.8
14.0
9.2
100.0
Valid Percent
16.3
36.6
23.8
14.0
9.2
100.0
Cumulative
Percent
16.3
52.9
76.7
90.8
100.0
De auto houdt automatisch zelfgekozen snelheid constant - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
106
215
211
317
200
1049
Percent
10.1
20.5
20.1
30.2
19.1
100.0
Valid Percent
10.1
20.5
20.1
30.2
19.1
100.0
Cumulative
Percent
10.1
30.6
50.7
80.9
100.0
72
Waarschuwing bij onveilige snelheid gezien actuele omstandigheden, bv. mist, bocht,
gladheid, nabijheid van kruispunt of school - a
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
257
392
254
106
40
1049
Percent
24.5
37.4
24.2
10.1
3.8
100.0
Valid Percent
24.5
37.4
24.2
10.1
3.8
100.0
Cumulative
Percent
24.5
61.9
86.1
96.2
100.0
Waarschuwing bij onveilige snelheid gezien actuele omstandigheden, bv. mist, bocht,
gladheid, nabijheid van kruispunt of school - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
261
410
233
107
38
1049
Percent
24.9
39.1
22.2
10.2
3.6
100.0
Valid Percent
24.9
39.1
22.2
10.2
3.6
100.0
Cumulative
Percent
24.9
64.0
86.2
96.4
100.0
Waarschuwing bij onveilige snelheid gezien actuele omstandigheden, bv. mist, bocht,
gladheid, nabijheid van kruispunt of school - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
221
350
247
162
69
1049
Percent
21.1
33.4
23.5
15.4
6.6
100.0
Valid Percent
21.1
33.4
23.5
15.4
6.6
100.0
Cumulative
Percent
21.1
54.4
78.0
93.4
100.0
De auto past automatisch snelheid aan aan actuele omstandigheden - a
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
94
197
317
268
173
1049
Percent
9.0
18.8
30.2
25.5
16.5
100.0
Valid Percent
9.0
18.8
30.2
25.5
16.5
100.0
Cumulative
Percent
9.0
27.7
58.0
83.5
100.0
73
De auto past automatisch snelheid aan aan actuele omstandigheden - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
91
203
311
264
180
1049
Percent
8.7
19.4
29.6
25.2
17.2
100.0
Valid Percent
8.7
19.4
29.6
25.2
17.2
100.0
Cumulative
Percent
8.7
28.0
57.7
82.8
100.0
De auto past automatisch snelheid aan aan actuele omstandigheden - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
75
176
279
301
218
1049
Percent
7.1
16.8
26.6
28.7
20.8
100.0
Valid Percent
7.1
16.8
26.6
28.7
20.8
100.0
Cumulative
Percent
7.1
23.9
50.5
79.2
100.0
Waarschuwing over verkeerscondities verderop, bv. file, werkzaamheden, ongeluk - a
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
517
429
81
13
9
1049
Percent
49.3
40.9
7.7
1.2
.9
100.0
Valid Percent
49.3
40.9
7.7
1.2
.9
100.0
Cumulative
Percent
49.3
90.2
97.9
99.1
100.0
Waarschuwing over verkeerscondities verderop, bv. file, werkzaamheden, ongeluk - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
420
464
131
21
13
1049
Percent
40.0
44.2
12.5
2.0
1.2
100.0
Valid Percent
40.0
44.2
12.5
2.0
1.2
100.0
Cumulative
Percent
40.0
84.3
96.8
98.8
100.0
74
Waarschuwing over verkeerscondities verderop, bv. file, werkzaamheden, ongeluk - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
297
376
219
110
47
1049
Percent
28.3
35.8
20.9
10.5
4.5
100.0
Valid Percent
28.3
35.8
20.9
10.5
4.5
100.0
Cumulative
Percent
28.3
64.2
85.0
95.5
100.0
De auto past automatisch snelheid aan aan verkeerscondities verderop - a
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
104
232
339
247
127
1049
Percent
9.9
22.1
32.3
23.5
12.1
100.0
Valid Percent
9.9
22.1
32.3
23.5
12.1
100.0
Cumulative
Percent
9.9
32.0
64.3
87.9
100.0
De auto past automatisch snelheid aan aan verkeerscondities verderop - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
86
221
336
269
137
1049
Percent
8.2
21.1
32.0
25.6
13.1
100.0
Valid Percent
8.2
21.1
32.0
25.6
13.1
100.0
Cumulative
Percent
8.2
29.3
61.3
86.9
100.0
De auto past automatisch snelheid aan aan verkeerscondities verderop - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
69
168
293
326
193
1049
Percent
6.6
16.0
27.9
31.1
18.4
100.0
Valid Percent
6.6
16.0
27.9
31.1
18.4
100.0
Cumulative
Percent
6.6
22.6
50.5
81.6
100.0
75
Koershouden – Lane keeping
Waarschuwing bij dreigen van onbedoeld verlaten van rijstrook op grotendeels rechte
stukken - a
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
126
325
308
207
83
1049
Percent
12.0
31.0
29.4
19.7
7.9
100.0
Valid Percent
12.0
31.0
29.4
19.7
7.9
100.0
Cumulative
Percent
12.0
43.0
72.4
92.1
100.0
Waarschuwing bij dreigen van onbedoeld verlaten van rijstrook op grotendeels rechte
stukken - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
102
301
283
260
103
1049
Percent
9.7
28.7
27.0
24.8
9.8
100.0
Valid Percent
9.7
28.7
27.0
24.8
9.8
100.0
Cumulative
Percent
9.7
38.4
65.4
90.2
100.0
Waarschuwing bij dreigen van onbedoeld verlaten van rijstrook op grotendeels rechte
stukken - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
36
115
243
419
236
1049
Percent
3.4
11.0
23.2
39.9
22.5
100.0
Valid Percent
3.4
11.0
23.2
39.9
22.5
100.0
Cumulative
Percent
3.4
14.4
37.6
77.5
100.0
De auto blijft automatisch binnen rijstrook op grotendeels rechte stukken - a
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
102
268
313
229
137
1049
Percent
9.7
25.5
29.8
21.8
13.1
100.0
Valid Percent
9.7
25.5
29.8
21.8
13.1
100.0
Cumulative
Percent
9.7
35.3
65.1
86.9
100.0
76
De auto blijft automatisch binnen rijstrook op grotendeels rechte stukken - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
83
240
308
266
152
1049
Percent
7.9
22.9
29.4
25.4
14.5
100.0
Valid Percent
7.9
22.9
29.4
25.4
14.5
100.0
Cumulative
Percent
7.9
30.8
60.2
85.5
100.0
De auto blijft automatisch binnen rijstrook op grotendeels rechte stukken - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
34
99
223
422
271
1049
Percent
3.2
9.4
21.3
40.2
25.8
100.0
Valid Percent
3.2
9.4
21.3
40.2
25.8
100.0
Cumulative
Percent
3.2
12.7
33.9
74.2
100.0
Waarschuwing bij dreigen van onbedoeld verlaten van rijstrook op bochtige stukken - a
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
114
279
317
240
99
1049
Percent
10.9
26.6
30.2
22.9
9.4
100.0
Valid Percent
10.9
26.6
30.2
22.9
9.4
100.0
Cumulative
Percent
10.9
37.5
67.7
90.6
100.0
Waarschuwing bij dreigen van onbedoeld verlaten van rijstrook op bochtige stukken - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
94
263
320
262
110
1049
Percent
9.0
25.1
30.5
25.0
10.5
100.0
Valid Percent
9.0
25.1
30.5
25.0
10.5
100.0
Cumulative
Percent
9.0
34.0
64.5
89.5
100.0
77
Waarschuwing bij dreigen van onbedoeld verlaten van rijstrook op bochtige stukken - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
49
125
241
402
232
1049
Percent
4.7
11.9
23.0
38.3
22.1
100.0
Valid Percent
4.7
11.9
23.0
38.3
22.1
100.0
Cumulative
Percent
4.7
16.6
39.6
77.9
100.0
De auto blijft automatisch binnen de rijstrook op bochtige stukken - a
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
87
235
303
274
150
1049
Percent
8.3
22.4
28.9
26.1
14.3
100.0
Valid Percent
8.3
22.4
28.9
26.1
14.3
100.0
Cumulative
Percent
8.3
30.7
59.6
85.7
100.0
De auto blijft automatisch binnen de rijstrook op bochtige stukken - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
86
220
289
297
157
1049
Percent
8.2
21.0
27.6
28.3
15.0
100.0
Valid Percent
8.2
21.0
27.6
28.3
15.0
100.0
Cumulative
Percent
8.2
29.2
56.7
85.0
100.0
De auto blijft automatisch binnen de rijstrook op bochtige stukken - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
42
109
238
397
263
1049
Percent
4.0
10.4
22.7
37.8
25.1
100.0
Valid Percent
4.0
10.4
22.7
37.8
25.1
100.0
Cumulative
Percent
4.0
14.4
37.1
74.9
100.0
78
Waarschuwing bij dreigen van onbedoeld verlaten van rijstrook bij versmalde rijstroken
door werkzaamheden - a
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
131
308
319
204
87
1049
Percent
12.5
29.4
30.4
19.4
8.3
100.0
Valid Percent
12.5
29.4
30.4
19.4
8.3
100.0
Cumulative
Percent
12.5
41.8
72.3
91.7
100.0
Waarschuwing bij dreigen van onbedoeld verlaten van rijstrook bij versmalde rijstroken
door werkzaamheden - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
109
284
319
234
103
1049
Percent
10.4
27.1
30.4
22.3
9.8
100.0
Valid Percent
10.4
27.1
30.4
22.3
9.8
100.0
Cumulative
Percent
10.4
37.5
67.9
90.2
100.0
Waarschuwing bij dreigen van onbedoeld verlaten van rijstrook bij versmalde rijstroken
door werkzaamheden - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
62
157
277
352
201
1049
Percent
5.9
15.0
26.4
33.6
19.2
100.0
Valid Percent
5.9
15.0
26.4
33.6
19.2
100.0
Cumulative
Percent
5.9
20.9
47.3
80.8
100.0
De auto blijft automatisch binnen rijstrook bij versmalde rijstroken door werkzaamheden a
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
141
266
294
223
125
1049
Percent
13.4
25.4
28.0
21.3
11.9
100.0
Valid Percent
13.4
25.4
28.0
21.3
11.9
100.0
Cumulative
Percent
13.4
38.8
66.8
88.1
100.0
79
De auto blijft automatisch binnen rijstrook bij versmalde rijstroken door werkzaamheden p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
110
258
285
256
140
1049
Percent
10.5
24.6
27.2
24.4
13.3
100.0
Valid Percent
10.5
24.6
27.2
24.4
13.3
100.0
Cumulative
Percent
10.5
35.1
62.2
86.7
100.0
De auto blijft automatisch binnen rijstrook bij versmalde rijstroken door werkzaamheden s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
57
139
249
372
232
1049
Percent
5.4
13.3
23.7
35.5
22.1
100.0
Valid Percent
5.4
13.3
23.7
35.5
22.1
100.0
Cumulative
Percent
5.4
18.7
42.4
77.9
100.0
Afstand houden – Car following
Waarschuwing bij onvelige afstand tot voorligger: overschrijding van zelfgekozen
minimum volgafstand - a
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
166
407
301
137
38
1049
Percent
15.8
38.8
28.7
13.1
3.6
100.0
Valid Percent
15.8
38.8
28.7
13.1
3.6
100.0
Cumulative
Percent
15.8
54.6
83.3
96.4
100.0
Waarschuwing bij onvelige afstand tot voorligger: overschrijding van zelfgekozen
minimum volgafstand - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
133
383
317
170
46
1049
Percent
12.7
36.5
30.2
16.2
4.4
100.0
Valid Percent
12.7
36.5
30.2
16.2
4.4
100.0
Cumulative
Percent
12.7
49.2
79.4
95.6
100.0
80
Waarschuwing bij onvelige afstand tot voorligger: overschrijding van zelfgekozen
minimum volgafstand - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
75
168
284
362
160
1049
Percent
7.1
16.0
27.1
34.5
15.3
100.0
Valid Percent
7.1
16.0
27.1
34.5
15.3
100.0
Cumulative
Percent
7.1
23.2
50.2
84.7
100.0
De auto houdt automatisch (zelfgekozen) veilige volgafstand tot voorligger - a
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
156
348
319
144
82
1049
Percent
14.9
33.2
30.4
13.7
7.8
100.0
Valid Percent
14.9
33.2
30.4
13.7
7.8
100.0
Cumulative
Percent
14.9
48.0
78.5
92.2
100.0
De auto houdt automatisch (zelfgekozen) veilige volgafstand tot voorligger - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
132
309
335
172
101
1049
Percent
12.6
29.5
31.9
16.4
9.6
100.0
Valid Percent
12.6
29.5
31.9
16.4
9.6
100.0
Cumulative
Percent
12.6
42.0
74.0
90.4
100.0
De auto houdt automatisch (zelfgekozen) veilige volgafstand tot voorligger - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
76
148
258
363
204
1049
Percent
7.2
14.1
24.6
34.6
19.4
100.0
Valid Percent
7.2
14.1
24.6
34.6
19.4
100.0
Cumulative
Percent
7.2
21.4
45.9
80.6
100.0
81
Waarschuwing bij onveilige volgafstand gezien actuele omstandigheden, bv. mist, bocht,
gladheid, nabijheid van kruispunt of school - a
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
201
464
265
92
27
1049
Percent
19.2
44.2
25.3
8.8
2.6
100.0
Valid Percent
19.2
44.2
25.3
8.8
2.6
100.0
Cumulative
Percent
19.2
63.4
88.7
97.4
100.0
Waarschuwing bij onveilige volgafstand gezien actuele omstandigheden, bv. mist, bocht,
gladheid, nabijheid van kruispunt of school - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
180
434
294
110
31
1049
Percent
17.2
41.4
28.0
10.5
3.0
100.0
Valid Percent
17.2
41.4
28.0
10.5
3.0
100.0
Cumulative
Percent
17.2
58.5
86.6
97.0
100.0
Waarschuwing bij onveilige volgafstand gezien actuele omstandigheden, bv. mist, bocht,
gladheid, nabijheid van kruispunt of school - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
105
257
310
266
111
1049
Percent
10.0
24.5
29.6
25.4
10.6
100.0
Valid Percent
10.0
24.5
29.6
25.4
10.6
100.0
Cumulative
Percent
10.0
34.5
64.1
89.4
100.0
De auto past automatisch volgafstand aan aan actuele omstandigheden - a
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
129
282
307
224
107
1049
Percent
12.3
26.9
29.3
21.4
10.2
100.0
Valid Percent
12.3
26.9
29.3
21.4
10.2
100.0
Cumulative
Percent
12.3
39.2
68.4
89.8
100.0
82
De auto past automatisch volgafstand aan aan actuele omstandigheden - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
117
252
331
229
120
1049
Percent
11.2
24.0
31.6
21.8
11.4
100.0
Valid Percent
11.2
24.0
31.6
21.8
11.4
100.0
Cumulative
Percent
11.2
35.2
66.7
88.6
100.0
De auto past automatisch volgafstand aan aan actuele omstandigheden - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
66
146
278
360
199
1049
Percent
6.3
13.9
26.5
34.3
19.0
100.0
Valid Percent
6.3
13.9
26.5
34.3
19.0
100.0
Cumulative
Percent
6.3
20.2
46.7
81.0
100.0
De auto past automatisch volgafstand aan aan verkeerscondities verderop - a
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
136
297
325
194
97
1049
Percent
13.0
28.3
31.0
18.5
9.2
100.0
Valid Percent
13.0
28.3
31.0
18.5
9.2
100.0
Cumulative
Percent
13.0
41.3
72.3
90.8
100.0
De auto past automatisch volgafstand aan aan verkeerscondities verderop - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
123
271
336
212
107
1049
Percent
11.7
25.8
32.0
20.2
10.2
100.0
Valid Percent
11.7
25.8
32.0
20.2
10.2
100.0
Cumulative
Percent
11.7
37.6
69.6
89.8
100.0
83
De auto past automatisch volgafstand aan aan verkeerscondities verderop - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
71
149
297
349
183
1049
Percent
6.8
14.2
28.3
33.3
17.4
100.0
Valid Percent
6.8
14.2
28.3
33.3
17.4
100.0
Cumulative
Percent
6.8
21.0
49.3
82.6
100.0
Filerijden – Congestion driving
Waarschuwing bij onvelige afstand tot voorligger: overschrijding van zelfgekozen
minimum volgafstand - a
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
139
347
269
206
88
1049
Percent
13.3
33.1
25.6
19.6
8.4
100.0
Valid Percent
13.3
33.1
25.6
19.6
8.4
100.0
Cumulative
Percent
13.3
46.3
72.0
91.6
100.0
Waarschuwing bij onvelige afstand tot voorligger: overschrijding van zelfgekozen
minimum volgafstand - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
112
318
290
229
100
1049
Percent
10.7
30.3
27.6
21.8
9.5
100.0
Valid Percent
10.7
30.3
27.6
21.8
9.5
100.0
Cumulative
Percent
10.7
41.0
68.6
90.5
100.0
Waarschuwing bij onvelige afstand tot voorligger: overschrijding van zelfgekozen
minimum volgafstand - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
56
177
246
361
209
1049
Percent
5.3
16.9
23.5
34.4
19.9
100.0
Valid Percent
5.3
16.9
23.5
34.4
19.9
100.0
Cumulative
Percent
5.3
22.2
45.7
80.1
100.0
84
De auto houdt automatisch veilige volgafstand tot voorligger, incl. optrekken, afremmen - a
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
219
343
256
143
88
1049
Percent
20.9
32.7
24.4
13.6
8.4
100.0
Valid Percent
20.9
32.7
24.4
13.6
8.4
100.0
Cumulative
Percent
20.9
53.6
78.0
91.6
100.0
De auto houdt automatisch veilige volgafstand tot voorligger, incl. optrekken, afremmen - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
163
338
281
169
98
1049
Percent
15.5
32.2
26.8
16.1
9.3
100.0
Valid Percent
15.5
32.2
26.8
16.1
9.3
100.0
Cumulative
Percent
15.5
47.8
74.5
90.7
100.0
De auto houdt automatisch veilige volgafstand tot voorligger, incl. optrekken, afremmen - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
94
196
252
312
195
1049
Percent
9.0
18.7
24.0
29.7
18.6
100.0
Valid Percent
9.0
18.7
24.0
29.7
18.6
100.0
Cumulative
Percent
9.0
27.6
51.7
81.4
100.0
De auto rijdt automatisch file, incl. binnen rijstrook blijven - a
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
292
322
244
115
76
1049
Percent
27.8
30.7
23.3
11.0
7.2
100.0
Valid Percent
27.8
30.7
23.3
11.0
7.2
100.0
Cumulative
Percent
27.8
58.5
81.8
92.8
100.0
85
De auto rijdt automatisch file, incl. binnen rijstrook blijven - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
219
318
271
153
88
1049
Percent
20.9
30.3
25.8
14.6
8.4
100.0
Valid Percent
20.9
30.3
25.8
14.6
8.4
100.0
Cumulative
Percent
20.9
51.2
77.0
91.6
100.0
De auto rijdt automatisch file, incl. binnen rijstrook blijven - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
130
179
256
296
188
1049
Percent
12.4
17.1
24.4
28.2
17.9
100.0
Valid Percent
12.4
17.1
24.4
28.2
17.9
100.0
Cumulative
Percent
12.4
29.5
53.9
82.1
100.0
Rijstrookwisselen – Lane changing
Waarschuwing bij verkeer in dode hoek - a
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
403
458
143
29
16
1049
Percent
38.4
43.7
13.6
2.8
1.5
100.0
Valid Percent
38.4
43.7
13.6
2.8
1.5
100.0
Cumulative
Percent
38.4
82.1
95.7
98.5
100.0
Waarschuwing bij verkeer in dode hoek - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
352
435
189
57
16
1049
Percent
33.6
41.5
18.0
5.4
1.5
100.0
Valid Percent
33.6
41.5
18.0
5.4
1.5
100.0
Cumulative
Percent
33.6
75.0
93.0
98.5
100.0
86
Waarschuwing bij verkeer in dode hoek - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
352
371
186
95
45
1049
Percent
33.6
35.4
17.7
9.1
4.3
100.0
Valid Percent
33.6
35.4
17.7
9.1
4.3
100.0
Cumulative
Percent
33.6
68.9
86.7
95.7
100.0
Waarschuwing bij tegenliggers, bv. op bochtige of heuvelachtige stukken of wanneer uw
voorligger het zicht wegneemt - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
237
404
267
102
39
1049
Percent
22.6
38.5
25.5
9.7
3.7
100.0
Valid Percent
22.6
38.5
25.5
9.7
3.7
100.0
Cumulative
Percent
22.6
61.1
86.6
96.3
100.0
Waarschuwing bij tegenliggers, bv. op bochtige of heuvelachtige stukken of wanneer uw
voorligger het zicht wegneemt - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
166
280
295
221
87
1049
Percent
15.8
26.7
28.1
21.1
8.3
100.0
Valid Percent
15.8
26.7
28.1
21.1
8.3
100.0
Cumulative
Percent
15.8
42.5
70.6
91.7
100.0
Aanwijzing dat het veilig is om van rijstrook te wisselen - a
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
154
306
350
161
78
1049
Percent
14.7
29.2
33.4
15.3
7.4
100.0
Valid Percent
14.7
29.2
33.4
15.3
7.4
100.0
Cumulative
Percent
14.7
43.9
77.2
92.6
100.0
87
Aanwijzing dat het veilig is om van rijstrook te wisselen - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
150
301
341
174
83
1049
Percent
14.3
28.7
32.5
16.6
7.9
100.0
Valid Percent
14.3
28.7
32.5
16.6
7.9
100.0
Cumulative
Percent
14.3
43.0
75.5
92.1
100.0
Aanwijzing dat het veilig is om van rijstrook te wisselen - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
104
185
311
294
155
1049
Percent
9.9
17.6
29.6
28.0
14.8
100.0
Valid Percent
9.9
17.6
29.6
28.0
14.8
100.0
Cumulative
Percent
9.9
27.6
57.2
85.2
100.0
De auto wisselt automatisch van rijstrook - a
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
38
56
212
364
379
1049
Percent
3.6
5.3
20.2
34.7
36.1
100.0
Valid Percent
3.6
5.3
20.2
34.7
36.1
100.0
Cumulative
Percent
3.6
9.0
29.2
63.9
100.0
De auto wisselt automatisch van rijstrook - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
30
48
204
376
391
1049
Percent
2.9
4.6
19.4
35.8
37.3
100.0
Valid Percent
2.9
4.6
19.4
35.8
37.3
100.0
Cumulative
Percent
2.9
7.4
26.9
62.7
100.0
88
De auto wisselt automatisch van rijstrook - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
22
28
154
397
448
1049
Percent
2.1
2.7
14.7
37.8
42.7
100.0
Valid Percent
2.1
2.7
14.7
37.8
42.7
100.0
Cumulative
Percent
2.1
4.8
19.4
57.3
100.0
Ongeregeld kruispunt – Non-signalised intersection
Informatie over naderen van (gevaarlijk) ongeregeld kruispunt - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
122
389
377
120
41
1049
Percent
11.6
37.1
35.9
11.4
3.9
100.0
Valid Percent
11.6
37.1
35.9
11.4
3.9
100.0
Cumulative
Percent
11.6
48.7
84.7
96.1
100.0
Informatie over naderen van (gevaarlijk) ongeregeld kruispunt - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
109
322
361
200
57
1049
Percent
10.4
30.7
34.4
19.1
5.4
100.0
Valid Percent
10.4
30.7
34.4
19.1
5.4
100.0
Cumulative
Percent
10.4
41.1
75.5
94.6
100.0
Informatie over voorrangssituatie - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
144
432
318
120
35
1049
Percent
13.7
41.2
30.3
11.4
3.3
100.0
Valid Percent
13.7
41.2
30.3
11.4
3.3
100.0
Cumulative
Percent
13.7
54.9
85.2
96.7
100.0
89
Informatie over voorrangssituatie - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
148
418
301
139
43
1049
Percent
14.1
39.8
28.7
13.3
4.1
100.0
Valid Percent
14.1
39.8
28.7
13.3
4.1
100.0
Cumulative
Percent
14.1
54.0
82.7
95.9
100.0
Waarschuwing bij naderend verkeer, bv. bij slecht zicht op het kruispunt - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
166
452
327
73
31
1049
Percent
15.8
43.1
31.2
7.0
3.0
100.0
Valid Percent
15.8
43.1
31.2
7.0
3.0
100.0
Cumulative
Percent
15.8
58.9
90.1
97.0
100.0
Waarschuwing bij naderend verkeer, bv. bij slecht zicht op het kruispunt - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
150
431
320
107
41
1049
Percent
14.3
41.1
30.5
10.2
3.9
100.0
Valid Percent
14.3
41.1
30.5
10.2
3.9
100.0
Cumulative
Percent
14.3
55.4
85.9
96.1
100.0
Waarschuwing bij verkeer in dode hoek bij ongeregeld kruispunt, bv. fietsers - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
344
489
154
50
12
1049
Percent
32.8
46.6
14.7
4.8
1.1
100.0
Valid Percent
32.8
46.6
14.7
4.8
1.1
100.0
Cumulative
Percent
32.8
79.4
94.1
98.9
100.0
90
Waarschuwing bij verkeer in dode hoek bij ongeregeld kruispunt, bv. fietsers - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
388
472
141
33
15
1049
Percent
37.0
45.0
13.4
3.1
1.4
100.0
Valid Percent
37.0
45.0
13.4
3.1
1.4
100.0
Cumulative
Percent
37.0
82.0
95.4
98.6
100.0
Aanwijzing dat het veilig is om het kruispunt te passeren - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
102
301
376
198
72
1049
Percent
9.7
28.7
35.8
18.9
6.9
100.0
Valid Percent
9.7
28.7
35.8
18.9
6.9
100.0
Cumulative
Percent
9.7
38.4
74.3
93.1
100.0
Aanwijzing dat het veilig is om het kruispunt te passeren - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
97
283
374
219
76
1049
Percent
9.2
27.0
35.7
20.9
7.2
100.0
Valid Percent
9.2
27.0
35.7
20.9
7.2
100.0
Cumulative
Percent
9.2
36.2
71.9
92.8
100.0
De auto remt automatisch, stopt en (indien voorligger aanwezig) houdt afstand tot
voorligger wanneer ongeregeld kruispunt niet vrij is - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
73
165
308
333
170
1049
Percent
7.0
15.7
29.4
31.7
16.2
100.0
Valid Percent
7.0
15.7
29.4
31.7
16.2
100.0
Cumulative
Percent
7.0
22.7
52.0
83.8
100.0
91
De auto remt automatisch, stopt en (indien voorligger aanwezig) houdt afstand tot
voorligger wanneer ongeregeld kruispunt niet vrij is - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
63
161
309
334
182
1049
Percent
6.0
15.3
29.5
31.8
17.3
100.0
Valid Percent
6.0
15.3
29.5
31.8
17.3
100.0
Cumulative
Percent
6.0
21.4
50.8
82.7
100.0
De auto passeert automatisch en veilig een ongeregeld kruispunt - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
74
127
282
334
232
1049
Percent
7.1
12.1
26.9
31.8
22.1
100.0
Valid Percent
7.1
12.1
26.9
31.8
22.1
100.0
Cumulative
Percent
7.1
19.2
46.0
77.9
100.0
De auto passeert automatisch en veilig een ongeregeld kruispunt - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
65
126
269
349
240
1049
Percent
6.2
12.0
25.6
33.3
22.9
100.0
Valid Percent
6.2
12.0
25.6
33.3
22.9
100.0
Cumulative
Percent
6.2
18.2
43.9
77.1
100.0
Geregeld kruispunt – Signalised intersection
Informatie over naderen van (gevaarlijk) geregeld kruispunt - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
94
313
342
229
71
1049
Percent
9.0
29.8
32.6
21.8
6.8
100.0
Valid Percent
9.0
29.8
32.6
21.8
6.8
100.0
Cumulative
Percent
9.0
38.8
71.4
93.2
100.0
92
Informatie over naderen van (gevaarlijk) geregeld kruispunt - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
89
284
344
248
84
1049
Percent
8.5
27.1
32.8
23.6
8.0
100.0
Valid Percent
8.5
27.1
32.8
23.6
8.0
100.0
Cumulative
Percent
8.5
35.6
68.4
92.0
100.0
Informatie over kleur van verkeerslicht - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
56
214
236
354
189
1049
Percent
5.3
20.4
22.5
33.7
18.0
100.0
Valid Percent
5.3
20.4
22.5
33.7
18.0
100.0
Cumulative
Percent
5.3
25.7
48.2
82.0
100.0
Informatie over kleur van verkeerslicht - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
56
193
247
356
197
1049
Percent
5.3
18.4
23.5
33.9
18.8
100.0
Valid Percent
5.3
18.4
23.5
33.9
18.8
100.0
Cumulative
Percent
5.3
23.7
47.3
81.2
100.0
Waarschuwing bij verkeer in dode hoek bij geregeld kruispunt, bv. fietsers - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
294
455
214
61
25
1049
Percent
28.0
43.4
20.4
5.8
2.4
100.0
Valid Percent
28.0
43.4
20.4
5.8
2.4
100.0
Cumulative
Percent
28.0
71.4
91.8
97.6
100.0
93
Waarschuwing bij verkeer in dode hoek bij geregeld kruispunt, bv. fietsers - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
331
438
201
51
28
1049
Percent
31.6
41.8
19.2
4.9
2.7
100.0
Valid Percent
31.6
41.8
19.2
4.9
2.7
100.0
Cumulative
Percent
31.6
73.3
92.5
97.3
100.0
Waarschuwing bij onveilige omstandigheden om linksaf te slaan bij tegemoetkomend
rechtdoorgaand verkeer - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
115
313
383
180
58
1049
Percent
11.0
29.8
36.5
17.2
5.5
100.0
Valid Percent
11.0
29.8
36.5
17.2
5.5
100.0
Cumulative
Percent
11.0
40.8
77.3
94.5
100.0
Waarschuwing bij onveilige omstandigheden om linksaf te slaan bij tegemoetkomend
rechtdoorgaand verkeer - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
114
292
383
196
64
1049
Percent
10.9
27.8
36.5
18.7
6.1
100.0
Valid Percent
10.9
27.8
36.5
18.7
6.1
100.0
Cumulative
Percent
10.9
38.7
75.2
93.9
100.0
De auto remt automatisch, stopt en (indien voorligger aanwezig) houdt afstand tot
voorligger wanneer geregeld kruispunt niet vrij is - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
64
166
291
320
208
1049
Percent
6.1
15.8
27.7
30.5
19.8
100.0
Valid Percent
6.1
15.8
27.7
30.5
19.8
100.0
Cumulative
Percent
6.1
21.9
49.7
80.2
100.0
94
De auto remt automatisch, stopt en (indien voorligger aanwezig) houdt afstand tot
voorligger wanneer geregeld kruispunt niet vrij is - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
63
152
281
339
214
1049
Percent
6.0
14.5
26.8
32.3
20.4
100.0
Valid Percent
6.0
14.5
26.8
32.3
20.4
100.0
Cumulative
Percent
6.0
20.5
47.3
79.6
100.0
De auto passeert automatisch en veilig een geregeld kruispunt - p
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
63
118
243
353
272
1049
Percent
6.0
11.2
23.2
33.7
25.9
100.0
Valid Percent
6.0
11.2
23.2
33.7
25.9
100.0
Cumulative
Percent
6.0
17.3
40.4
74.1
100.0
De auto passeert automatisch en veilig een geregeld kruispunt - s
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
65
104
234
367
279
1049
Percent
6.2
9.9
22.3
35.0
26.6
100.0
Valid Percent
6.2
9.9
22.3
35.0
26.6
100.0
Cumulative
Percent
6.2
16.1
38.4
73.4
100.0
Slecht zicht – Reduced visibility
Meedraaiende koplampen in bochten
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
221
434
315
63
16
1049
Percent
21.1
41.4
30.0
6.0
1.5
100.0
Valid Percent
21.1
41.4
30.0
6.0
1.5
100.0
Cumulative
Percent
21.1
62.4
92.5
98.5
100.0
95
Op voorruit weergeven van slecht zichtbare obstakels op de weg voor u, zoals wandelaars
of overstekende dieren
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
290
444
270
32
13
1049
Percent
27.6
42.3
25.7
3.1
1.2
100.0
Valid Percent
27.6
42.3
25.7
3.1
1.2
100.0
Cumulative
Percent
27.6
70.0
95.7
98.8
100.0
Minder alert – Driver fatigue
Waarschuwing bij verminderde alertheid
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
180
454
325
72
18
1049
Percent
17.2
43.3
31.0
6.9
1.7
100.0
Valid Percent
17.2
43.3
31.0
6.9
1.7
100.0
Cumulative
Percent
17.2
60.4
91.4
98.3
100.0
Geen reactie op waarschuwing: de auto parkeert automatisch
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
84
165
358
264
178
1049
Percent
8.0
15.7
34.1
25.2
17.0
100.0
Valid Percent
8.0
15.7
34.1
25.2
17.0
100.0
Cumulative
Percent
8.0
23.7
57.9
83.0
100.0
Dreigende botsing – Imminent crash
Waarschuwing bij dreigende botsing
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
273
463
243
50
20
1049
Percent
26.0
44.1
23.2
4.8
1.9
100.0
Valid Percent
26.0
44.1
23.2
4.8
1.9
100.0
Cumulative
Percent
26.0
70.2
93.3
98.1
100.0
96
De auto remt automatisch ten behoeve van noodstop
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
227
377
307
94
44
1049
Percent
21.6
35.9
29.3
9.0
4.2
100.0
Valid Percent
21.6
35.9
29.3
9.0
4.2
100.0
Cumulative
Percent
21.6
57.6
86.8
95.8
100.0
De auto wijkt automatisch uit indien mogelijk/veilig
Valid
veel behoefte aan
behoefte aan
misschien behoefte aan
geen behoefte aan
zeker geen behoefte aan
Total
Frequency
213
315
322
118
81
1049
Percent
20.3
30.0
30.7
11.2
7.7
100.0
Valid Percent
20.3
30.0
30.7
11.2
7.7
100.0
Cumulative
Percent
20.3
50.3
81.0
92.3
100.0
97
98
Appendix D – McNemar test
By means of frequency statistics an overview of the perceived needs for driver assistance could be
given. However, from these statistics it was not clear to what extent there exist significantly greater
needs for one thing compared to the other. Therefore, a matched-pairs design was used (Rice, 1994).
With this, it was possible to compare, for example, the perceived needs for driver support function X
to the perceived needs for driver support function Y. Among others, comparisons between the needs
for the ten most popular driver support functions were made. The five answering categories were split
into: (1) (great) need, i.e. categories 1 + 2 and (2) other, i.e. categories 3 to 5. Table 25 gives an
example of how the data were represented. In this example, the needs for warnings on downstream
traffic conditions on motorways were compared to the needs for these warnings on rural roads (i.e. the
two most popular driver support functions according to the frequency statistics).
Warning for downstream traffic condition - motorway &
Warning for downstream traffic condition - rural road
Warning for
downstream traffic
condition - motorway
1
2
Warning for
downstream traffic
condition - rural road
1
2
881
65
3
100
Table 25. Representation of data for the purpose of matched pair design and McNemar test
The probabilities in the table can be represented as follows:
11
12
21
22
.1
.2
1.
1
2.
The null hypothesis states that the probabilities of the categories ‘(great) need’ and ‘other’ are the
same for warnings for downstream traffic conditions on motorways and these warnings on rural roads.
This means that 1. = .1 and 2. = .2, or H0: 12 = 21. The test of this hypothesis is called the
McNemar test: a nonparametric test for two related dichotomous variables. It focuses on changes in
responses from one sample response (e.g. function 1) to another (e.g. function 2) using the chi-square
distribution (SPSS 11.5, Help function). The maximum likelihood estimates of the cell probabilities
and the McNemar chi-square are computed using the two cells that correspond to a (great) need for
only one of the two driver support functions. Continuity is corrected because the chi-square statistic is
used to approximate a discrete distribution. Table 26 shows the test statistics for the above-mentioned
example. As can be seen, 2 = 54.721, with a corresponding p-value of 0.000. So, the null hypothesis
was rejected in favor of the hypothesis that there was a significantly greater need for warnings for
downstream traffic conditions on motorways than on rural roads.
With this approach, it is assumed that all possible comparisons will be made. In this study, this means,
for example, that the perceived needs for all driver support functions (113 in total) should be
compared to each other. However, this was regarded too much work. Therefore, only the top 10s of
driver support functions and types of driver assistance in the ideal system were taken into account.
Besides, McNemar tests were performed to check whether the needs for driver assistance differed per
road type and level of support, and whether the answers to the statements about the ideal system
differed from each other. Because in the analyses not all possible comparisons were made, a
“cautious” significance criterion of 0.001 was used.
99
Test Statisticsb
N
Chi-Square a
Asymp. Sig.
Warning for
downstream
traffic
condition motorway &
Warning for
downstream
traffic
condition rural road
1049
54,721
,000
a. Continuity Corrected
b. McNemar Test
Table 26. Test statistics of McNemar chi-square
Table 27 shows, as an example, the results from the McNemar tests with data from the top 10 of most
popular functions (table 6).
Pair 12 = comparison between function 1 and 2
Pair 910 = comparison between function 9 en 10
Pair
12
13
14
15
16
17
18
19
110
23
24
25
26
27
28
p-value
.000
.000
.000
.000
.000
.000
.000
.000
.000
.153
.126
.001
.000
.000
.000
Pair
29
210
34
35
36
37
38
39
310
45
46
47
48
49
410
p-value
.000
.000
1.000
.032
.000
.000
.000
.000
.000
.004
.000
.000
.000
.000
.000
Pair
56
57
58
59
510
67
68
69
610
78
79
710
89
810
910
p-value
.001
.000
.000
.000
.000
.256
.012
.004
.005
.010
.069
.066
.500
.457
.956
Table 27. Results from the McNemar tests for pairs of the ten most popular functions
It can be seen that with p 0.001 all pairs significantly differ from each other except for 15 pairs,
namely 23, 24, 34, 35, 45, 67, 68, 69, 610, 78, 79, 710, 89, 810 and 910 (see italic p-values). This
means, for example, that there is a greater need for function 1 compared to function 10, but that there
is no difference between the perceived needs for function 9 compared to function 10. It appeared that
function 1 (i.e. warning for downstream traffic condition – motorway) significantly differed from the
other functions. This also applied to the functions 2 to 5, which significantly differed from the first
100
function and from the functions 6 to 10. The last five functions differed significantly from the first five
functions. Based on this, the top 10 of functions could be divided into three groups; see table 28.
Within the groups the perceived (great) need for the specific functions did not differ significantly,
except for function 2 versus 5.
Driver support function
1. Warning for downstream traffic condition – motorway
2. Warning for downstream traffic condition – rural road
3. Blind spot warning during lane changing – motorway
4. Blind spot warning at non-signalised intersection – urban road
5. Blind spot warning at non-signalised intersection – rural road
6. Blind spot warning during lane changing – rural road
7. Blind spot warning at signalised intersection – urban road
8. Blind spot warning at signalised intersection – rural road
9. Warning for imminent crash
10. Presentation of badly visible objects on windscreen
(Great) need
946
90.2%
884
84.3%
861
82.1%
860
82.0%
833
79.4%
787
75.0%
769
73.3%
71.4%
749
736
70.2%
734
70.0%
Table 28. Three groups of most popular driver support functions
101
102
Appendix E – Ideal driver support system: frequencies
This appendix shows the results from the frequency statistics with respect to the 22 presented types of
driver assistance in the ideal driver support system.
Regulating speed - motorway
Valid
no, not in system
yes, in system
Total
Frequency
591
458
1049
Percent
56.3
43.7
100.0
Valid Percent
56.3
43.7
100.0
Cumulative
Percent
56.3
100.0
Lane keeping - motorway
Valid
no, not in system
yes, in system
Total
Frequency
876
173
1049
Percent
83.5
16.5
100.0
Valid Percent
83.5
16.5
100.0
Cumulative
Percent
83.5
100.0
Car following - motorway
Valid
no, not in system
yes, in system
Total
Frequency
514
535
1049
Percent
49.0
51.0
100.0
Valid Percent
49.0
51.0
100.0
Cumulative
Percent
49.0
100.0
Congestion driving - motorway
Valid
no, not in system
yes, in system
Total
Frequency
607
442
1049
Percent
57.9
42.1
100.0
Valid Percent
57.9
42.1
100.0
Cumulative
Percent
57.9
100.0
Lane changing - motorway
Valid
no, not in system
yes, in system
Total
Frequency
940
109
1049
Percent
89.6
10.4
100.0
Valid Percent
89.6
10.4
100.0
Cumulative
Percent
89.6
100.0
103
Regulating speed - rural road
Valid
no, not in system
yes, in system
Total
Frequency
695
354
1049
Percent
66.3
33.7
100.0
Valid Percent
66.3
33.7
100.0
Cumulative
Percent
66.3
100.0
Lane keeping - rural road
Valid
no, not in system
yes, in system
Total
Frequency
951
98
1049
Percent
90.7
9.3
100.0
Valid Percent
90.7
9.3
100.0
Cumulative
Percent
90.7
100.0
Car following - rural road
Valid
no, not in system
yes, in system
Total
Frequency
751
298
1049
Percent
71.6
28.4
100.0
Valid Percent
71.6
28.4
100.0
Cumulative
Percent
71.6
100.0
Congestion driving - rural road
Valid
no, not in system
yes, in system
Total
Frequency
867
182
1049
Percent
82.7
17.3
100.0
Valid Percent
82.7
17.3
100.0
Cumulative
Percent
82.7
100.0
Lane changing - rural road
Valid
no, not in system
yes, in system
Total
Frequency
972
77
1049
Percent
92.7
7.3
100.0
Valid Percent
92.7
7.3
100.0
Cumulative
Percent
92.7
100.0
Non-signalised intersection - rural road
Valid
no, not in system
yes, in system
Total
Frequency
815
234
1049
Percent
77.7
22.3
100.0
Valid Percent
77.7
22.3
100.0
Cumulative
Percent
77.7
100.0
104
Signalised intersection - rural road
Valid
no, not in system
yes, in system
Total
Frequency
1026
23
1049
Percent
97.8
2.2
100.0
Valid Percent
97.8
2.2
100.0
Cumulative
Percent
97.8
100.0
Regulating speed - urban road
Valid
no, not in system
yes, in system
Total
Frequency
910
139
1049
Percent
86.7
13.3
100.0
Valid Percent
86.7
13.3
100.0
Cumulative
Percent
86.7
100.0
Lane keeping - urban road
Valid
no, not in system
yes, in system
Total
Frequency
1032
17
1049
Percent
98.4
1.6
100.0
Valid Percent
98.4
1.6
100.0
Cumulative
Percent
98.4
100.0
Car following - urban road
Valid
no, not in system
yes, in system
Total
Frequency
945
104
1049
Percent
90.1
9.9
100.0
Valid Percent
90.1
9.9
100.0
Cumulative
Percent
90.1
100.0
Congestion driving - urban road
Valid
no, not in system
yes, in system
Total
Frequency
988
61
1049
Percent
94.2
5.8
100.0
Valid Percent
94.2
5.8
100.0
Cumulative
Percent
94.2
100.0
Lane changing - urban road
Valid
no, not in system
yes, in system
Total
Frequency
1015
34
1049
Percent
96.8
3.2
100.0
Valid Percent
96.8
3.2
100.0
Cumulative
Percent
96.8
100.0
105
Non-signalised intersection - urban road
Valid
no, not in system
yes, in system
Total
Frequency
841
208
1049
Percent
80.2
19.8
100.0
Valid Percent
80.2
19.8
100.0
Cumulative
Percent
80.2
100.0
Signalised intersection - urban road
Valid
no, not in system
yes, in system
Total
Frequency
1020
29
1049
Percent
97.2
2.8
100.0
Valid Percent
97.2
2.8
100.0
Cumulative
Percent
97.2
100.0
Reduced visibility
Valid
no, not in system
yes, in system
Total
Frequency
416
633
1049
Percent
39.7
60.3
100.0
Valid Percent
39.7
60.3
100.0
Cumulative
Percent
39.7
100.0
Driver fatigue
Valid
no, not in system
yes, in system
Total
Frequency
680
369
1049
Percent
64.8
35.2
100.0
Valid Percent
64.8
35.2
100.0
Cumulative
Percent
64.8
100.0
Imminent crash
Valid
no, not in system
yes, in system
Total
Frequency
461
588
1049
Percent
43.9
56.1
100.0
Valid Percent
43.9
56.1
100.0
Cumulative
Percent
43.9
100.0
106
Appendix F – Influence of driver characteristics
It was studied to what extent the following driver characteristics were of influence on the results from
the user needs survey.
Characteristic
Gender (GEN)
Age (AGE)
Education (EDU)
Frequency of car driving (FRQ)
Years of driving experience (YRS)
Average annual mileage (KM)
Type of car possession (CAR)
Familiarity with ACC (ACC)
Categories
male, female
18-24 years, 25-44 years, 45-64 years, 65 years
primary & lower secondary, higher secondary, higher & university
>3 times a week, 1-3 times a week, 1-3 times a month, <1 time a month
<5 years, 5-10 years, >10 years
<10.000 km, 10.000-20.000 km, >20.000 km
private, business
not familiar with, somewhat familiar with, (very) familiar with
Driver support functions
Results from chi-square tests only focussing on the sub-sample of respondents that indicated to have a
(great) need for the ten most popular driver support functions (p<0.05):
Driver support function
Warning for downstream traffic condition – motorway
Warning for downstream traffic condition – rural road
Blind spot warning during lane changing – motorway
Blind spot warning at non-sign. intersection – urban road
Blind spot warning at non-sign. intersection – rural road
Blind spot warning during lane changing – rural road
Blind spot warning at signalised intersection – urban road
Blind spot warning at signalised intersection – rural road
Warning for imminent crash
Presentation of badly visible objects on windscreen
GEN
AGE
EDU
FRQ
YRS
KM
CAR
ACC
0.031
Results from chi-square tests focussing on the perceived needs of all respondents for the ten most
popular driver support functions (p<0.05):
Driver support function
Warning for downstream traffic condition – motorway
Warning for downstream traffic condition – rural road
Blind spot warning during lane changing – motorway
Blind spot warning at non-sign. intersection – urban road
Blind spot warning at non-sign. intersection – rural road
Blind spot warning during lane changing – rural road
Blind spot warning at signalised intersection – urban road
Blind spot warning at signalised intersection – rural road
Warning for imminent crash
Presentation of badly visible objects on windscreen
GEN
AGE
EDU
FRQ
YRS
KM
CAR
0.048
0.029
0.044
0.048
0.031
0.048
0.004
0.001
0.000
107
ACC
0.002
0.038
Ideal driver support system
Results from chi-square tests only focussing on the sub-sample of respondents that indicated to have a
need for the ten most popular types of driver assistance in the ideal driver support system (p<0.05):
Type of driver assistance
Reduced visibility
Imminent crash
Car following – motorway
Regulating speed – motorway
Congestion driving – motorway
Driver fatigue
Regulating speed – rural road
Car following – rural road
Negotiating non-sign. intersection – rural road
Negotiating non-sign. intersection – urban road
GEN
0.008
0.026
0.027
0.008
0.001
0.002
AGE
EDU
0.039
0.009
FRQ
YRS
KM
CAR
0.006
ACC
0.037
0.018
0.021
Results from chi-square tests focussing on the perceived needs of all respondents for the ten most
popular types of driver assistance in the ideal driver support system (p<0.05):
Type of driver assistance
Reduced visibility
Imminent crash
Car following – motorway
Regulating speed – motorway
Congestion driving – motorway
Driver fatigue
Regulating speed – rural road
Car following – rural road
Negotiating non-sign. intersection – rural road
Negotiating non-sign. intersection – urban road
GEN
0.000
0.006
0.003
0.003
0.001
0.023
0.000
0.001
AGE
EDU
0.002
0.000
0.019
0.003
FRQ
YRS
KM
0.000
0.041
CAR
0.004
ACC
0.003
0.005
0.044
The second type of chi-square tests appeared to include more significant results than the first type.
This is because more data (namely from all respondents instead of just a sub-sample) were taken into
account.
108