ThesisPDF Available

Driver experience and acceptance of driver support systems - a case of speed adaption

Abstract and Figures

Substantial research and development efforts are being made to add driver support systems to the arsenal of traffic safety measures. Obviously, the system cannot reduce fatalities and trauma until it is actually used. Hence, drivers’ experiences and acceptance of the system are of paramount importance. A driver support system (ISA) has been investigated by means of real life trials in Sweden, Hungary and Spain, and the results show that the incentive for drivers to use an ISA system might be the money and embarrassment saved by avoiding speeding tickets, rather than increased traffic safety. Further, to assess the ‘final’, long-term experiences of the system, a longer period than one month of usage is necessary. This thesis conducts a literature review to systematically investigate how acceptance has been defined and how it has been measured within the driver support area. A new definition of acceptance is proposed: “the degree to which an individual intends to use a system and, when available, to incorporate the system in his/her driving”. Additionally, it explores whether the Unified Theory of Acceptance and Use of Technology model (UTAUT), which was originally developed for information technology, may be used as an acceptance model for driver support systems. A pilot test supported to some extent the use of the model. The model constructs ‘performance expectancy’ and ‘social influence’ affect drivers’ intention to use the system.
Content may be subject to copyright.
Driver experience and acceptance
of driver support systems
- a case of speed adaptation
Emeli Adell
Driver experience and acceptance
of driver support systems
- a case of speed adaptation
Emeli Adell
Doctoral Thesis CODEN:LUTVDG/(TVTT-1038)1-171/2009
Bulletin - Lunds Universitet, Tekniska högskolan i Lund, ISBN 978-91-628-7947-1
Institutionen för teknik och samhälle, 251 ISSN 1653-1930
Emeli Adell
Driver experience and acceptance of driver support systems
– a case of speed adaptation
2009
Keywords:
Driver experiences, acceptance, driver support systems, ISA, speed support, UTAUT, ADAS, field
trial
Abstract:
Substantial research and development efforts are being made to add driver support systems to the
arsenal of traffic safety measures. Obviously, the system cannot reduce fatalities and trauma until it
is actually used. Hence, drivers’ experiences and acceptance of the system are of paramount
importance. A driver support system (ISA) has been investigated by means of real life trials in
Sweden, Hungary and Spain, and the results show that the incentive for drivers to use an ISA
system might be the money and embarrassment saved by avoiding speeding tickets, rather than
increased traffic safety. Further, to assess the ‘final’, long-term experiences of the system, a longer
period than one month of usage is necessary. This thesis conducts a literature review to
systematically investigate how acceptance has been defined and how it has been measured within
the driver support area. A new definition of acceptance is proposed: “the degree to which an
individual intends to use a system and, when available, to incorporate the system in his/her
driving”. Additionally, it explores whether the Unified Theory of Acceptance and Use of
Technology model (UTAUT), which was originally developed for information technology, may be
used as an acceptance model for driver support systems. A pilot test supported to some extent the
use of the model. The model constructs ‘performance expectancy’ and ‘social influence’ affect
drivers’ intention to use the system.
Citation:
Adell, Emeli. Driver experience and acceptance of driver support systems – a case of speed ADAPTATION.
Institutionen för Teknik och samhälle, Trafik och väg, 2009. Bulletin - Lunds Universitet, Tekniska
högskolan i Lund, Institutionen för teknik och samhälle, 251
Institutionen för Teknik och samhälle
Lunds Tekniska Högskola
Trafik och väg
Box 118, 221 00 LUND, Sverige
Department of Technology and Society
Lund Institute of Technology
Traffic and Roads
Box 118, SE-221 00 Lund, Sweden
Doctoral Thesis CODEN:LUTVDG/(TVTT-1038)1-171/2009
Bulletin - Lunds Universitet, Tekniska högskolan i Lund, ISBN 978-91-628-7947-1
Institutionen för teknik och samhälle, 251 ISSN 1653-1930
Emeli Adell
Driver experience and acceptance of driver support systems
– a case of speed adaptation
2009
Keywords:
Driver experiences, acceptance, driver support systems, ISA, speed support, UTAUT, ADAS, field
trial
Abstract:
Substantial research and development efforts are being made to add driver support systems to the
arsenal of traffic safety measures. Obviously, the system cannot reduce fatalities and trauma until it
is actually used. Hence, drivers’ experiences and acceptance of the system are of paramount
importance. A driver support system (ISA) has been investigated by means of real life trials in
Sweden, Hungary and Spain, and the results show that the incentive for drivers to use an ISA
system might be the money and embarrassment saved by avoiding speeding tickets, rather than
increased traffic safety. Further, to assess the ‘final’, long-term experiences of the system, a longer
period than one month of usage is necessary. This thesis conducts a literature review to
systematically investigate how acceptance has been defined and how it has been measured within
the driver support area. A new definition of acceptance is proposed: “the degree to which an
individual intends to use a system and, when available, to incorporate the system in his/her
driving”. Additionally, it explores whether the Unified Theory of Acceptance and Use of
Technology model (UTAUT), which was originally developed for information technology, may be
used as an acceptance model for driver support systems. A pilot test supported to some extent the
use of the model. The model constructs ‘performance expectancy’ and ‘social influence’ affect
drivers’ intention to use the system.
Citation:
Adell, Emeli. Driver experience and acceptance of driver support systems – a case of speed ADAPTATION
Institutionen för Teknik och samhälle, Trafik och väg, 2009. Bulletin - Lunds Universitet, Tekniska
högskolan i Lund, Institutionen för teknik och samhälle, 251
Institutionen för Teknik och samhälle
Lunds Tekniska Högskola
Trafik och väg
Box 118, 221 00 LUND, Sverige
Department of Technology and Society
Lund Institute of Technology
Traffic and Roads
Box 118, SE-221 00 Lund, Sweden
ISBN 978-91-628-7947-1
© Emeli Adell, 2009
Printed in Sweden
Media-Tryck, Lund, 2009
ISBN 978-91-628-7947-1
© Emeli Adell, 2009
Printed in Sweden
Media-Tryck, Lund, 2009
Preface
I was fortunate to be able to start my PhD-research within a unique, already
ongoing experiment – the large-scale field trial with ISA (Intelligent Speed
Adaptation) carried out in Lund between 1999 and 2001. My task in this project
was to analyse the driver questionnaires and investigate the drivers’ experience,
opinions and acceptance of the system.
As the work progressed it became clear that it was not possible to determine the
acceptance of the system since there was no clear definition of acceptance and the
measurements used gave partly contradictory results. These experiences led to the
decision to examine the concept of acceptance more thoroughly.
I consider myself lucky to have had the opportunity to be part of several important
and interesting experiments and trials during my PhD-period, both in Sweden and
in Europe. It is very rewarding to work in different environments, with different
people and ideas. I look forward to continuing the research on the questions raised
in this thesis. As my supervisor always says: the doctoral degree is the ‘driving
licence’ for doing research on your own – there is plenty of time after the
dissertation. From that perspective I am grateful for being given the opportunity
to start this research and take pleasure in the thought of being able to continue
with it.
I hope you, the reader, find the research interesting and the thesis enjoyable to
read.
Preface
I was fortunate to be able to start my PhD-research within a unique, already
ongoing experiment – the large-scale field trial with ISA (Intelligent Speed
Adaptation) carried out in Lund between 1999 and 2001. My task in this project
was to analyse the driver questionnaires and investigate the drivers’ experience,
opinions and acceptance of the system.
As the work progressed it became clear that it was not possible to determine the
acceptance of the system since there was no clear definition of acceptance and the
measurements used gave partly contradictory results. These experiences led to the
decision to examine the concept of acceptance more thoroughly.
I consider myself lucky to have had the opportunity to be part of several important
and interesting experiments and trials during my PhD-period, both in Sweden and
in Europe. It is very rewarding to work in different environments, with different
people and ideas. I look forward to continuing the research on the questions raised
in this thesis. As my supervisor always says: the doctoral degree is the ‘driving
licence’ for doing research on your own – there is plenty of time after the
dissertation. From that perspective I am grateful for being given the opportunity
to start this research and take pleasure in the thought of being able to continue
with it.
I hope you, the reader, find the research interesting and the thesis enjoyable to
read.
Contents
List of publications
Abbreviations
1Introduction .............................................................................................. 1
1.1The traffic safety problem ....................................................................... 3
1.2Driver support systems – potential means to improve traffic safety .......... 3
1.3Evaluating traffic safety effects of driver support systems – a case of ISA . 5
1.4Research objectives ................................................................................. 9
1.5The scope of the thesis ............................................................................ 9
2Driver experiences of ISA ........................................................................ 11
2.1ISA systems ............................................................................................ 13
2.2Two field-trials with ISA ........................................................................ 15
2.3Effects on driver behaviour .................................................................... 16
2.4Driver experiences .................................................................................. 17
2.5Implications of the findings of the field trials ......................................... 22
3What is acceptance? ................................................................................. 25
3.1Present definitions of driver acceptance .................................................. 27
3.2Proposal for a new definition of acceptance ............................................ 28
3.3Assessing acceptance .............................................................................. 31
4Acceptance Models .................................................................................. 37
4.1Most used frameworks for acceptance of driver support systems ............. 39
4.2Acceptance models within the area of information technology ............... 40
4.3The Unified Theory of Acceptance and Use of Technology ................... 41
4.4Using the UTAUT model in the context of driver support systems ........ 43
5Discussion ............................................................................................... 51
5.1Thesis contribution ................................................................................ 53
5.2Methodology discussion ......................................................................... 57
5.3Conclusions and final remarks ............................................................... 59
Acknowledgement ........................................................................................ 61
References .................................................................................................... 65
Appended papers
Contents
List of publications
Abbreviations
1Introduction .............................................................................................. 1
1.1The traffic safety problem ....................................................................... 3
1.2Driver support systems – potential means to improve traffic safety .......... 3
1.3Evaluating traffic safety effects of driver support systems – a case of ISA . 5
1.4Research objectives ................................................................................. 9
1.5The scope of the thesis ............................................................................ 9
2Driver experiences of ISA ........................................................................ 11
2.1ISA systems ............................................................................................ 13
2.2Two field-trials with ISA ........................................................................ 15
2.3Effects on driver behaviour .................................................................... 16
2.4Driver experiences .................................................................................. 17
2.5Implications of the findings of the field trials ......................................... 22
3What is acceptance? ................................................................................. 25
3.1Present definitions of driver acceptance .................................................. 27
3.2Proposal for a new definition of acceptance ............................................ 28
3.3Assessing acceptance .............................................................................. 31
4Acceptance Models .................................................................................. 37
4.1Most used frameworks for acceptance of driver support systems ............. 39
4.2Acceptance models within the area of information technology ............... 40
4.3The Unified Theory of Acceptance and Use of Technology ................... 41
4.4Using the UTAUT model in the context of driver support systems ........ 43
5Discussion ............................................................................................... 51
5.1Thesis contribution ................................................................................ 53
5.2Methodology discussion ......................................................................... 57
5.3Conclusions and final remarks ............................................................... 59
Acknowledgement ........................................................................................ 61
References .................................................................................................... 65
Appended papers
List of publications
This thesis is based on the following papers, which will be referred to in the text
by their Roman numerals. The papers are appended at the end of the thesis.
Paper I
Adell, E. (2007) Drivers’ evaluations of the Active Accelerator Pedal
in a real-life trial, IATSS RESEARCH. 31(1) pp. 89-99.
Paper II
Adell, E. and Várhelyi, A. (2008) Driver comprehension and
acceptance of the Active Accelerator Pedal after long-term use.
Transportation Research, Part F: Traffic Psychology and
Behaviour, 11(1) pp. 37-51.
My contribution: Analysis of questionnaire answers and writing of
the larger part of the paper.
Paper III
Adell, E., Várhelyi, A. and Hjälmdahl, M. (2008) Auditory and
haptic systems for in-car speed management – a comparative real life
study, Transportation Research Part F: Traffic Psychology and
Behaviour. 11(6) pp. 445-458.
My contribution: Elaboration of questionnaires, analysis of answers
and writing of the corresponding part of the paper.
Paper IV
Adell, E. On the acceptance of driver support systems. Submitted to
IET Intelligent Transport Systems
List of publications
This thesis is based on the following papers, which will be referred to in the text
by their Roman numerals. The papers are appended at the end of the thesis.
Paper I
Adell, E. (2007) Drivers’ evaluations of the Active Accelerator Pedal
in a real-life trial, IATSS RESEARCH. 31(1) pp. 89-99.
Paper II
Adell, E. and Várhelyi, A. (2008) Driver comprehension and
acceptance of the Active Accelerator Pedal after long-term use.
Transportation Research, Part F: Traffic Psychology and
Behaviour, 11(1) pp. 37-51.
My contribution: Analysis of questionnaire answers and writing of
the larger part of the paper.
Paper III
Adell, E., Várhelyi, A. and Hjälmdahl, M. (2008) Auditory and
haptic systems for in-car speed management – a comparative real life
study, Transportation Research Part F: Traffic Psychology and
Behaviour. 11(6) pp. 445-458.
My contribution: Elaboration of questionnaires, analysis of answers
and writing of the corresponding part of the paper.
Paper IV
Adell, E. On the acceptance of driver support systems. Submitted to
IET Intelligent Transport Systems
Abbreviations
AAP Active Accelerator Pedal, an Intelligent Speed Adaptation system,
mainly providing haptic feedback, for further description see
chapter 2.1.1.
ABS Anti-lock Braking System
ACAS Automotive Collision Avoidance System
ACC Adaptive cruise control
ADAS Advanced Driver Assistance Systems
AVCSS Advanced Vehicle Control and Safety systems
BEEP An Intelligent Speed Adaptation system, mainly providing auditory
feedback, for further description see chapter 2.1.2
BI Behavioural intention to use the system
EE Effort expectancy
ESC Electronic Stability Control systems
FCW Forward Collision Warning
HMI Human Machine Interaction
ISA Intelligent Speed Adaptation
IT Information Technology
ITS Intelligent Transport Systems
IVIS In-Vehicle Information Systems
LCA Lane Change Assist
LDW Lane Departure Warning
NHTSA National Highway Traffic Safety Administration
PE Performance expectancy
PROSPER Project for Research On Speed Adaptation Policies on European
Roads, an EU-project carried out between 2003 and 2005
RDCW Road-Departure Crash Warning System
SASPENCE Safe Speed and Safe Distance, an EU-project, subproject to
PReVENT, carried out between 2004 and 2007
SI Social influence
TAM Technology acceptance model
TPB Theory of Planned Behaviour
UTAUT Unified Theory of Acceptance and Use of Technology
Abbreviations
AAP Active Accelerator Pedal, an Intelligent Speed Adaptation system,
mainly providing haptic feedback, for further description see
chapter 2.1.1.
ABS Anti-lock Braking System
ACAS Automotive Collision Avoidance System
ACC Adaptive cruise control
ADAS Advanced Driver Assistance Systems
AVCSS Advanced Vehicle Control and Safety systems
BEEP An Intelligent Speed Adaptation system, mainly providing auditory
feedback, for further description see chapter 2.1.2
BI Behavioural intention to use the system
EE Effort expectancy
ESC Electronic Stability Control systems
FCW Forward Collision Warning
HMI Human Machine Interaction
ISA Intelligent Speed Adaptation
IT Information Technology
ITS Intelligent Transport Systems
IVIS In-Vehicle Information Systems
LCA Lane Change Assist
LDW Lane Departure Warning
NHTSA National Highway Traffic Safety Administration
PE Performance expectancy
PROSPER Project for Research On Speed Adaptation Policies on European
Roads, an EU-project carried out between 2003 and 2005
RDCW Road-Departure Crash Warning System
SASPENCE Safe Speed and Safe Distance, an EU-project, subproject to
PReVENT, carried out between 2004 and 2007
SI Social influence
TAM Technology acceptance model
TPB Theory of Planned Behaviour
UTAUT Unified Theory of Acceptance and Use of Technology
1
1 Introduction
While in-car safety systems have greatly improved the
chances of surviving an accident, more attention now
needs to be given to systems that can actually prevent
accidents from happening.
Ertico (2009)
1
1 Introduction
While in-car safety systems have greatly improved the
chances of surviving an accident, more attention now
needs to be given to systems that can actually prevent
accidents from happening.
Ertico (2009)
Adell | Driver experiences and acceptance of driver support systems
2
Adell | Driver experiences and acceptance of driver support systems
2
1 Introduction
3
1.1 The traffic safety problem
The estimated number of road traffic fatalities worldwide is about 1.2 million each
year. The number of injured could be as high as 50 million (Peden et al., 2004).
Over 1 270 000 accidents resulted in personal injury and more than 42 000
people were killed (EC, 2009) in the EU in 2007, which is more than 115 lives
lost a day in the EU alone and almost 3 300 lives lost globally every day. The
tragedy behind these figures is unimaginable. Road traffic injuries were the
eleventh leading cause of death worldwide in 2002. Road traffic was the second
leading cause of death among children and young people between 5 and 29 years
of age. If nothing is done, road traffic injuries are predicted to become the third
leading cause of death in 2020 (Peden et al., 2004). These figures make it
unnecessary to further point out the urgent need to continue to address the traffic
safety problem with all available measures – both traditional and new.
The connection between accidents and driver behaviour is well established. Driver
behaviour, such as speeding, driving with too short headway, drinking and driving
and neglecting to use restraint systems or safety devices, has been proven to
increase accident risk and/or injury severity, see e.g. Finch et al. (1994), ETSC,
(2001), Najm et al. (2003), van Kampen (2003), Elvik and Vaa (2004), Nilsson
(2004) and Baldock et al. (2005).
1.2 Driver support systems – potential means to improve traffic safety
Traditionally, the traffic safety problem has been tackled by means such as
physical measures in the road environment, enforcement, education and passive
in-vehicle safety systems. In recent years, substantial research and development has
been carried out in order to add a new type of measure to the arsenal – driver
support systems.
A driver support system may be defined as an in-vehicle system that collects
information from the driving environment, processes it and provides information,
feedback, or vehicle control to support the driver in optimal vehicle operation (van
Driel, 2007). This means that these are active systems, working to prevent
accidents and are not only meant to mitigate the effects when an accident is
1 Introduction
3
1.1 The traffic safety problem
The estimated number of road traffic fatalities worldwide is about 1.2 million each
year. The number of injured could be as high as 50 million (Peden et al., 2004).
Over 1 270 000 accidents resulted in personal injury and more than 42 000
people were killed (EC, 2009) in the EU in 2007, which is more than 115 lives
lost a day in the EU alone and almost 3 300 lives lost globally every day. The
tragedy behind these figures is unimaginable. Road traffic injuries were the
eleventh leading cause of death worldwide in 2002. Road traffic was the second
leading cause of death among children and young people between 5 and 29 years
of age. If nothing is done, road traffic injuries are predicted to become the third
leading cause of death in 2020 (Peden et al., 2004). These figures make it
unnecessary to further point out the urgent need to continue to address the traffic
safety problem with all available measures – both traditional and new.
The connection between accidents and driver behaviour is well established. Driver
behaviour, such as speeding, driving with too short headway, drinking and driving
and neglecting to use restraint systems or safety devices, has been proven to
increase accident risk and/or injury severity, see e.g. Finch et al. (1994), ETSC,
(2001), Najm et al. (2003), van Kampen (2003), Elvik and Vaa (2004), Nilsson
(2004) and Baldock et al. (2005).
1.2 Driver support systems – potential means to improve traffic safety
Traditionally, the traffic safety problem has been tackled by means such as
physical measures in the road environment, enforcement, education and passive
in-vehicle safety systems. In recent years, substantial research and development has
been carried out in order to add a new type of measure to the arsenal – driver
support systems.
A driver support system may be defined as an in-vehicle system that collects
information from the driving environment, processes it and provides information,
feedback, or vehicle control to support the driver in optimal vehicle operation (van
Driel, 2007). This means that these are active systems, working to prevent
accidents and are not only meant to mitigate the effects when an accident is
Adell | Driver experiences and acceptance of driver support systems
4
unavoidable. Anti-lock braking systems (ABS), Electronic Stability Control
systems (ESC)) and similar recovery systems are normally not included in the term
“driver support systems”. Besides, a variety of other terms are found in the
literature, for example Advanced Driver Assistance Systems (ADAS), In-Vehicle
Information Systems (IVIS) and Advanced Vehicle Control and Safety systems
(AVCSS). The term ‘driver support systems’ has been chosen for this thesis
because it refers to the most important aspect of such a system; i.e., supporting the
driver in carrying out the driving task.
The extensive development work within the area of driver support systems has
resulted in a range of different systems aimed at improving traffic safety. The
systems may e.g. be categorized as providing longitudinal and lateral support or as
systems monitoring the driver. Examples of longitudinal support systems are:
Intelligent Speed Adaptation (ISA) (e.g. Brookhuis & de Ward, 1999; Regan et al.,
2002; Várhelyi et al., 2004), Adaptive cruise control (ACC) (e.g. Hoedemaeker &
Brookhuis, 1998), Forward Collision Warning (FCW) (e.g. Regan et al., 2002; and
Adell et al., 2009), and Automotive Collision Avoidance System (ACAS) (e.g. Najm
et al., 2006). Examples of lateral support systems: Road-Departure Crash Warning
System (RDCW) (e.g. Wilson et al., 2007), Lane Departure Warning (LDW) (e.g.
Regan, et al., 2002; Yu et al., 2008), Lane Change Assistant (LCA) (e.g. Rüder et
al., 2002) and Blind spot monitoring (e.g. Ehlgen et al., 2008, Kuwana, & Itoh,
2008). Fatigue Monitoring (see e.g. Anund & Hjälmdahl, 2009; Rogado, et al.,
2009) is an example of a system monitoring and warning the driver if his/her state
of alertness is below a suitable level. A description of the systems and a more
thorough categorisation of support systems can be found in van Driel (2007).
Knowledge of the safety effects of these systems is of great importance for
decisions on their development and deployment. The safety evaluation of driver
support systems is commonly classified into three areas: System Safety, Human
Machine Interaction (HMI) and Traffic Safety. System safety covers safety issues
concerning hardware and software design, particularly focusing on reliability, the
tendency to malfunction or to go into a dangerous and/or unanticipated system
mode. The HMI deals with interaction between the user and the system. Key
issues are means of dialogue between the user and the systems, feedback to the
user, design of buttons and controls, location in the car etc. Inappropriate design
can lead to overload, underload or distraction of the driver. Traffic safety refers to
the overall safety effect of system use and the outcomes of system safety and HMI.
It also covers how a system may affect road user behaviour so as to alter the
interaction between the driver, the vehicle, the road infrastructure and other road
users. Evaluations of System Safety and HMI are wide areas and not part of this
thesis. For an overview of these areas see e.g. ETSC (1999).
Adell | Driver experiences and acceptance of driver support systems
4
unavoidable. Anti-lock braking systems (ABS), Electronic Stability Control
systems (ESC)) and similar recovery systems are normally not included in the term
“driver support systems”. Besides, a variety of other terms are found in the
literature, for example Advanced Driver Assistance Systems (ADAS), In-Vehicle
Information Systems (IVIS) and Advanced Vehicle Control and Safety systems
(AVCSS). The term ‘driver support systems’ has been chosen for this thesis
because it refers to the most important aspect of such a system; i.e., supporting the
driver in carrying out the driving task.
The extensive development work within the area of driver support systems has
resulted in a range of different systems aimed at improving traffic safety. The
systems may e.g. be categorized as providing longitudinal and lateral support or as
systems monitoring the driver. Examples of longitudinal support systems are:
Intelligent Speed Adaptation (ISA) (e.g. Brookhuis & de Ward, 1999; Regan et al.,
2002; Várhelyi et al., 2004), Adaptive cruise control (ACC) (e.g. Hoedemaeker &
Brookhuis, 1998), Forward Collision Warning (FCW) (e.g. Regan et al., 2002; and
Adell et al., 2009), and Automotive Collision Avoidance System (ACAS) (e.g. Najm
et al., 2006). Examples of lateral support systems: Road-Departure Crash Warning
System (RDCW) (e.g. Wilson et al., 2007), Lane Departure Warning (LDW) (e.g.
Regan, et al., 2002; Yu et al., 2008), Lane Change Assistant (LCA) (e.g. Rüder et
al., 2002) and Blind spot monitoring (e.g. Ehlgen et al., 2008, Kuwana, & Itoh,
2008). Fatigue Monitoring (see e.g. Anund & Hjälmdahl, 2009; Rogado, et al.,
2009) is an example of a system monitoring and warning the driver if his/her state
of alertness is below a suitable level. A description of the systems and a more
thorough categorisation of support systems can be found in van Driel (2007).
Knowledge of the safety effects of these systems is of great importance for
decisions on their development and deployment. The safety evaluation of driver
support systems is commonly classified into three areas: System Safety, Human
Machine Interaction (HMI) and Traffic Safety. System safety covers safety issues
concerning hardware and software design, particularly focusing on reliability, the
tendency to malfunction or to go into a dangerous and/or unanticipated system
mode. The HMI deals with interaction between the user and the system. Key
issues are means of dialogue between the user and the systems, feedback to the
user, design of buttons and controls, location in the car etc. Inappropriate design
can lead to overload, underload or distraction of the driver. Traffic safety refers to
the overall safety effect of system use and the outcomes of system safety and HMI.
It also covers how a system may affect road user behaviour so as to alter the
interaction between the driver, the vehicle, the road infrastructure and other road
users. Evaluations of System Safety and HMI are wide areas and not part of this
thesis. For an overview of these areas see e.g. ETSC (1999).
1 Introduction
5
1.3 Evaluating traffic safety effects of driver support systems – a case of ISA
The traffic safety outcome of driver support systems may be estimated at three
levels. The maximum safety potential states the highest safety benefit the system can
provide, given the characteristics of the system. The effects when in use provide the
safety effects when drivers use the system in their ordinary driving including
interactions with other road users, behavioural changes, etc. The true effects
provide the safety effects that actually reduce fatalities and trauma because they are
the observable effects of the system when it is implemented. This is dependent on
if (and how much) the drivers employ the system in reality.
1.3.1 The maximum traffic safety potential of driver support systems
The maximum traffic safety potential is the theoretical capacity of a system to
improve traffic safety. It is an estimate of the number of accidents, injuries and
fatalities that the functionality of the system may prevent or mitigate. The
functionality of the system provides information about the positive behavioural
change that may be expected. Knowledge about the connection between this
behaviour and the accident risk provides the basis for estimating the safety effect.
These estimates assume that the drivers use the system as intended by the
designers, and any other behavioural change the usage might lead to is
disregarded.
The maximum traffic safety potential of ISA
The principle behind the Intelligent Speed Adaptation (ISA) system is to support
the driver not to exceed the legal speed limits. The system monitors the current
speed limit (through e.g. GPS and digital maps containing information about
speed limits) and informs/advises the driver not to exceed it or automatically limits
the speed of the vehicle to the speed limit. In addition to this functionality, some
versions of the system integrate further speed restrictions in critical situations (e.g.
sharp curves, slippery road or poor visibility).
In theory, the maximum safety potential of this system is achieved if it prevents all
speeding (and in some systems further limits the speed in critical situations). This
implies, firstly, that the system is impossible to override or that drivers always
choose to follow its recommendations, and, secondly, that all vehicles are
equipped with the system and it is always active. Knowledge about the speeding
behaviour and the relationship between speed and accidents contributes to
ascertaining the maximum safety potential of ISA.
Studies calculating the maximum safety potential of ISA have estimated a
reduction of injury accidents (including fatalities) by between 20 and 70 %
1 Introduction
5
1.3 Evaluating traffic safety effects of driver support systems – a case of ISA
The traffic safety outcome of driver support systems may be estimated at three
levels. The maximum safety potential states the highest safety benefit the system can
provide, given the characteristics of the system. The effects when in use provide the
safety effects when drivers use the system in their ordinary driving including
interactions with other road users, behavioural changes, etc. The true effects
provide the safety effects that actually reduce fatalities and trauma because they are
the observable effects of the system when it is implemented. This is dependent on
if (and how much) the drivers employ the system in reality.
1.3.1 The maximum traffic safety potential of driver support systems
The maximum traffic safety potential is the theoretical capacity of a system to
improve traffic safety. It is an estimate of the number of accidents, injuries and
fatalities that the functionality of the system may prevent or mitigate. The
functionality of the system provides information about the positive behavioural
change that may be expected. Knowledge about the connection between this
behaviour and the accident risk provides the basis for estimating the safety effect.
These estimates assume that the drivers use the system as intended by the
designers, and any other behavioural change the usage might lead to is
disregarded.
The maximum traffic safety potential of ISA
The principle behind the Intelligent Speed Adaptation (ISA) system is to support
the driver not to exceed the legal speed limits. The system monitors the current
speed limit (through e.g. GPS and digital maps containing information about
speed limits) and informs/advises the driver not to exceed it or automatically limits
the speed of the vehicle to the speed limit. In addition to this functionality, some
versions of the system integrate further speed restrictions in critical situations (e.g.
sharp curves, slippery road or poor visibility).
In theory, the maximum safety potential of this system is achieved if it prevents all
speeding (and in some systems further limits the speed in critical situations). This
implies, firstly, that the system is impossible to override or that drivers always
choose to follow its recommendations, and, secondly, that all vehicles are
equipped with the system and it is always active. Knowledge about the speeding
behaviour and the relationship between speed and accidents contributes to
ascertaining the maximum safety potential of ISA.
Studies calculating the maximum safety potential of ISA have estimated a
reduction of injury accidents (including fatalities) by between 20 and 70 %
Adell | Driver experiences and acceptance of driver support systems
6
depending on type of ISA, country and type of road (see e.g. Várhelyi, 2002;
Carsten & Tate, 2005; Carsten et al., 2006).
1.3.2 The traffic safety effects of driver support systems in use
A more realistic estimate of the traffic safety effects requires evaluation of the
effects of the system while it is in use. When calculating the maximum safety
potential, the functionality, as described by the designers/engineers, provides the
basis for the evaluation. At this level, the actual behaviour of drivers constitutes the
basis for providing an estimate of the traffic safety effects of using the system. How
the system is used by drivers may be quite different from how it was intended to
be used. Hence, actual use may also provide a number of unexpected and
sometimes unwanted behavioural changes, which may be due to many different
things, e.g., the driver’s understanding of what the system does, and is capable of
doing, or perception of how the system best fulfils his/her needs and requirements.
Moreover, this can be influenced by external factors like penetration rate,
coexistence with other systems, familiarity with the route, etc.
A number of guidelines, manuals and recommendations are available for the
evaluation of the traffic safety effects when the system is in use, e. g. the
ADVISORS framework (Parkes et al., 2001), the VIKING Guidelines (Kulmala
et al., 2002), the RESPONSE 3 Code of Practice (Schwarz et al., 2006), the
TEMPO handbook (Tarry et al., 2007) and the FESTA handbook (FESTA,
2008).
Investigation of the traffic safety effects of a system in use could either be carried
out in real life (field trials) or in a driving simulator. Both settings have pros and
cons. The imaginary world of the driving simulator offers the possibility of
repeating the same situation, and flexibility when composing the test route and
traffic environment. It also makes it possible to study critical safety situations,
which would be too dangerous to study in real life. Still, the driving simulator
does not offer the complexity of the real driving environment and cannot provide
real interactions with other road users like the real world can do. The real life trials
provide a very realistic quasi experimental setting that limits the validation
problems.
When studying these traffic safety effects, it is important to include both driver
behaviour and state. Driver behaviour should be investigated in terms of
desired/expected behaviour and (unexpected/unwanted) behavioural adaptations.
Which indicators to use are influenced by the hypothesised effects and the
experimental layout, but some important variables regarding traffic safety are
speed, headway, law abidance and interaction with other road users. Driver state
Adell | Driver experiences and acceptance of driver support systems
6
depending on type of ISA, country and type of road (see e.g. Várhelyi, 2002;
Carsten & Tate, 2005; Carsten et al., 2006).
1.3.2 The traffic safety effects of driver support systems in use
A more realistic estimate of the traffic safety effects requires evaluation of the
effects of the system while it is in use. When calculating the maximum safety
potential, the functionality, as described by the designers/engineers, provides the
basis for the evaluation. At this level, the actual behaviour of drivers constitutes the
basis for providing an estimate of the traffic safety effects of using the system. How
the system is used by drivers may be quite different from how it was intended to
be used. Hence, actual use may also provide a number of unexpected and
sometimes unwanted behavioural changes, which may be due to many different
things, e.g., the driver’s understanding of what the system does, and is capable of
doing, or perception of how the system best fulfils his/her needs and requirements.
Moreover, this can be influenced by external factors like penetration rate,
coexistence with other systems, familiarity with the route, etc.
A number of guidelines, manuals and recommendations are available for the
evaluation of the traffic safety effects when the system is in use, e. g. the
ADVISORS framework (Parkes et al., 2001), the VIKING Guidelines (Kulmala
et al., 2002), the RESPONSE 3 Code of Practice (Schwarz et al., 2006), the
TEMPO handbook (Tarry et al., 2007) and the FESTA handbook (FESTA,
2008).
Investigation of the traffic safety effects of a system in use could either be carried
out in real life (field trials) or in a driving simulator. Both settings have pros and
cons. The imaginary world of the driving simulator offers the possibility of
repeating the same situation, and flexibility when composing the test route and
traffic environment. It also makes it possible to study critical safety situations,
which would be too dangerous to study in real life. Still, the driving simulator
does not offer the complexity of the real driving environment and cannot provide
real interactions with other road users like the real world can do. The real life trials
provide a very realistic quasi experimental setting that limits the validation
problems.
When studying these traffic safety effects, it is important to include both driver
behaviour and state. Driver behaviour should be investigated in terms of
desired/expected behaviour and (unexpected/unwanted) behavioural adaptations.
Which indicators to use are influenced by the hypothesised effects and the
experimental layout, but some important variables regarding traffic safety are
speed, headway, law abidance and interaction with other road users. Driver state
1 Introduction
7
might be described by permanent and temporary factors. Permanent factors are
those that stay (relatively) constant in the short term e.g. age, gender and mileage.
Temporary factors such as workload, distraction, and the driver’s emotional state
vary in a shorter time span. Emotional responses like anger and irritation while
driving may interfere with attention, perception, information processing and
motoric reactions and make the driver more disposed to aggressive behaviour (for
a review see Ulleberg, 2004). Further, it is important to consider the layout of the
trial especially regarding selection of participants, choice of test area/route,
experimental design, duration of the trial and instructions/information given to
the participants.
The traffic safety effects of ISA when in use
ISA systems have been tested in various field and simulator tests mostly across
Europe and Australia. These trials have found that the systems reduced mean
speeds and speed variance (see e.g. paper III, Brookhuis & de Waard, 1997;
Carsten & Fowkes, 2000; Lahrmann et al., 2001; Hjälmdahl, et al., 2002; Regan
et al., 2004; Várhelyi et al., 2004; Regan et al., 2005). In addition, improved
behaviour towards other road users and slightly larger headways have been found,
but also some negative behavioural modifications such as forgetting to adjust the
speed to the actual speed limit when not supported by the system (Hjälmdahl &
Várhelyi, 2004).
Armed with knowledge of driver behaviour and driver state, one may make an
estimate of the traffic safety effects when the system is in use. However, to be able
to do so it is also necessary to examine the relationship between driver behaviour
/driver state and traffic safety, but knowledge of this relationship is in many cases
insufficient. Since the most examined relationship is the one between speed and
accident risk/severity (see e.g. Finch et al., 1994; Elvik & Vaa, 2004; Nilsson,
2004), most numeric estimates are based on speed changes. Nevertheless, other
driver behaviour or information about driver state should not be disregarded.
Based on the reduction in mean speeds, the ISA system is estimated to reduce
injury accidents by between 8 % and 25 % and fatal accidents by between 10 %
and 32 %, under the assumption that all cars are equipped with an AAP (Active
Accelerator Pedal) and that the system is permanently active (see e.g. Hjälmdahl et
al., 2002). Regarding the driver state, several studies have shown small increases in
workload and deterioration in the drivers’ emotional response, especially for
drivers who were sceptical about the system before the trial started or for drivers
who experienced system malfunction, see e.g. paper I, paper II and paper III. The
reduction in speed level and speed variance, the tendency of increased headway
and improved behaviour towards other road users speak for improved traffic
safety. On the other hand, forgetting to adjust the speed when the system is off,
1 Introduction
7
might be described by permanent and temporary factors. Permanent factors are
those that stay (relatively) constant in the short term e.g. age, gender and mileage.
Temporary factors such as workload, distraction, and the driver’s emotional state
vary in a shorter time span. Emotional responses like anger and irritation while
driving may interfere with attention, perception, information processing and
motoric reactions and make the driver more disposed to aggressive behaviour (for
a review see Ulleberg, 2004). Further, it is important to consider the layout of the
trial especially regarding selection of participants, choice of test area/route,
experimental design, duration of the trial and instructions/information given to
the participants.
The traffic safety effects of ISA when in use
ISA systems have been tested in various field and simulator tests mostly across
Europe and Australia. These trials have found that the systems reduced mean
speeds and speed variance (see e.g. paper III, Brookhuis & de Waard, 1997;
Carsten & Fowkes, 2000; Lahrmann et al., 2001; Hjälmdahl, et al., 2002; Regan
et al., 2004; Várhelyi et al., 2004; Regan et al., 2005). In addition, improved
behaviour towards other road users and slightly larger headways have been found,
but also some negative behavioural modifications such as forgetting to adjust the
speed to the actual speed limit when not supported by the system (Hjälmdahl &
Várhelyi, 2004).
Armed with knowledge of driver behaviour and driver state, one may make an
estimate of the traffic safety effects when the system is in use. However, to be able
to do so it is also necessary to examine the relationship between driver behaviour
/driver state and traffic safety, but knowledge of this relationship is in many cases
insufficient. Since the most examined relationship is the one between speed and
accident risk/severity (see e.g. Finch et al., 1994; Elvik & Vaa, 2004; Nilsson,
2004), most numeric estimates are based on speed changes. Nevertheless, other
driver behaviour or information about driver state should not be disregarded.
Based on the reduction in mean speeds, the ISA system is estimated to reduce
injury accidents by between 8 % and 25 % and fatal accidents by between 10 %
and 32 %, under the assumption that all cars are equipped with an AAP (Active
Accelerator Pedal) and that the system is permanently active (see e.g. Hjälmdahl et
al., 2002). Regarding the driver state, several studies have shown small increases in
workload and deterioration in the drivers’ emotional response, especially for
drivers who were sceptical about the system before the trial started or for drivers
who experienced system malfunction, see e.g. paper I, paper II and paper III. The
reduction in speed level and speed variance, the tendency of increased headway
and improved behaviour towards other road users speak for improved traffic
safety. On the other hand, forgetting to adjust the speed when the system is off,
Adell | Driver experiences and acceptance of driver support systems
8
increased workload and deterioration in emotional state may have negative effects
on traffic safety.
In addition to the estimate of the traffic safety effects summarized above, it is
important to address the long-term effects of using driver support systems. Most
evaluations of systems are done over a short time period, not allowing the drivers
to really use the system as they would if the system was permanently installed in
their own cars. It has been shown that the duration the system is used has a
significant impact on its effects, where e.g. the speed reduction decreases over time
(Hjälmdahl, 2004) and the emotional state deteriorates further (paper I). This
indicates a danger of overestimating the traffic safety effects when systems are only
evaluated after short-term usage.
1.3.3 The true traffic safety effects of driver support systems
To estimate the true safety effects when a system is implemented, the effects when
the system is in use have to be complemented with an estimate of how much it is
actually going to be used. Only when the system is employed can it reduce
fatalities and trauma. Hence, acceptance of it is vital. As e.g. Najm et al. (2006)
states: “driver acceptance is the precondition that will permit new automotive
technologies to achieve their forecasted benefit levels”.
Acceptance is individual and based on what is known, understood and believed by
the driver. It is the driver’s personal attitudes, expectations, experiences and
subjective evaluation that form the acceptance (or lack thereof) (Schade & Baum,
2007). Driver opinions about and experiences with a certain system are often
collected and presented as information on acceptance. Although this information
is interesting, it is not the same as acceptance of the system. The link between
driver experiences and acceptance is missing, which hinders the evaluation of
acceptance and how it is affected by the drivers’ expectations and experiences. This
in turn influences the extent of usage and consequently the true effects.
True traffic safety effects of ISA
The true traffic safety effects of ISA are unfortunately not possible to ascertain,
since the actual use of the system cannot be estimated. If more was known about
the acceptance of ISA, it would be easier to estimate the extent of usage and
thereby also the true effects. This knowledge gap leads us to examine the concept
of acceptance and in what way experiences influence drivers’ acceptance.
Adell | Driver experiences and acceptance of driver support systems
8
increased workload and deterioration in emotional state may have negative effects
on traffic safety.
In addition to the estimate of the traffic safety effects summarized above, it is
important to address the long-term effects of using driver support systems. Most
evaluations of systems are done over a short time period, not allowing the drivers
to really use the system as they would if the system was permanently installed in
their own cars. It has been shown that the duration the system is used has a
significant impact on its effects, where e.g. the speed reduction decreases over time
(Hjälmdahl, 2004) and the emotional state deteriorates further (paper I). This
indicates a danger of overestimating the traffic safety effects when systems are only
evaluated after short-term usage.
1.3.3 The true traffic safety effects of driver support systems
To estimate the true safety effects when a system is implemented, the effects when
the system is in use have to be complemented with an estimate of how much it is
actually going to be used. Only when the system is employed can it reduce
fatalities and trauma. Hence, acceptance of it is vital. As e.g. Najm et al. (2006)
states: “driver acceptance is the precondition that will permit new automotive
technologies to achieve their forecasted benefit levels”.
Acceptance is individual and based on what is known, understood and believed by
the driver. It is the driver’s personal attitudes, expectations, experiences and
subjective evaluation that form the acceptance (or lack thereof) (Schade & Baum,
2007). Driver opinions about and experiences with a certain system are often
collected and presented as information on acceptance. Although this information
is interesting, it is not the same as acceptance of the system. The link between
driver experiences and acceptance is missing, which hinders the evaluation of
acceptance and how it is affected by the drivers’ expectations and experiences. This
in turn influences the extent of usage and consequently the true effects.
True traffic safety effects of ISA
The true traffic safety effects of ISA are unfortunately not possible to ascertain,
since the actual use of the system cannot be estimated. If more was known about
the acceptance of ISA, it would be easier to estimate the extent of usage and
thereby also the true effects. This knowledge gap leads us to examine the concept
of acceptance and in what way experiences influence drivers’ acceptance.
1 Introduction
9
1.4 Research objectives
The original objective of this research was to investigate driver experiences and
acceptance of a driver support system, namely Intelligent Speed Adaptation (ISA).
As the work progressed it became clear that what was meant by ‘acceptance’ was
not clear, making it impossible to achieve the original objective. The extended
objective of the research is hence to investigate driver experiences of ISA and to
examine the concept of acceptance of driver support systems and how drivers’
experiences influence their acceptance.
1.5 The scope of the thesis
The scope of this thesis is to investigate drivers’ experiences of the driver support
system ISA, the concept of acceptance in the context of driver support systems and
how driver experiences influence acceptance. The thesis consists of four papers and
an introductory/summarizing section. Papers I, II and III examine driver
experiences. Papers II and III identify problems with present-day research
regarding driver acceptance. Paper IV considers the concept of acceptance of driver
support systems and a pilot test of an acceptance model. For a schematic diagram
of the scope of the thesis see Figure 1.
Figure 1: The scope of the thesis.
Evaluatingdriversupportsystems
asmeanstoimprovetrafficsafety
Scopeofthethesis
True
effects
Usage
extent
Effects
wheninuse
Howthe
systemis
usedbythe
drivers
Maximum
safetypotential
PaperI:
Drivers’evaluationof
theactiveaccelerator
pedalinareallifetrial
PaperII:
Drivercomprehension
andacceptanceofthe
activeacceleratorpedal
afterlongtermuse
PaperIII:
Auditoryandhaptic
systemsforincarspeed
management–A
comparativereallife
stud
y
PaperIV:
Ontheacceptanceof
driversupportsystems
Acceptancemodel
Driver
experiences
Driver
acceptance
1 Introduction
9
1.4 Research objectives
The original objective of this research was to investigate driver experiences and
acceptance of a driver support system, namely Intelligent Speed Adaptation (ISA).
As the work progressed it became clear that what was meant by ‘acceptance’ was
not clear, making it impossible to achieve the original objective. The extended
objective of the research is hence to investigate driver experiences of ISA and to
examine the concept of acceptance of driver support systems and how drivers’
experiences influence their acceptance.
1.5 The scope of the thesis
The scope of this thesis is to investigate drivers’ experiences of the driver support
system ISA, the concept of acceptance in the context of driver support systems and
how driver experiences influence acceptance. The thesis consists of four papers and
an introductory/summarizing section. Papers I, II and III examine driver
experiences. Papers II and III identify problems with present-day research
regarding driver acceptance. Paper IV considers the concept of acceptance of driver
support systems and a pilot test of an acceptance model. For a schematic diagram
of the scope of the thesis see Figure 1.
Figure 1: The scope of the thesis.
Evaluatingdriversupportsystems
asmeanstoimprovetrafficsafety
Scopeofthethesis
True
effects
Usage
extent
Effects
wheninuse
Howthe
systemis
usedbythe
drivers
Maximum
safetypotential
PaperI:
Drivers’evaluationof
theactiveaccelerator
pedalinareallifetrial
PaperII:
Drivercomprehension
andacceptanceofthe
activeacceleratorpedal
afterlongtermuse
PaperIII:
Auditoryandhaptic
systemsforincarspeed
management–A
comparativereallife
stud
y
PaperIV:
Ontheacceptanceof
driversupportsystems
Acceptancemodel
Driver
experiences
Driver
acceptance
Adell | Driver experiences and acceptance of driver support systems
10
The introductory/summarizing section consists of 5 chapters:
xThis chapter, chapter 1, provides the background and objectives of the
thesis.
xChapter 2 focuses on drivers’ experiences with ISA (Intelligent Speed
Adaptation). The ISA-systems AAP (Active Accelerator Pedal) and BEEP
(an ISA system, mainly providing auditory feedback) are described, as are
the two field trials used to study drivers’ experiences of ISA-systems. The
effects on driver behaviour, found in the two trials, are briefly described
and a more extensive summary of drivers’ experiences with the systems is
provided. The chapter ends with a discussion on the implications of the
findings.
xChapter 3 contains various definitions of acceptance. Considerations when
defining and working on acceptance are discussed and a new definition is
proposed. The chapter also describes the various ways of assessing
acceptance and how these relate to the definitions. It ends with a
discussion on the limitations of the methods used today.
xChapter 4 provides an overview of prevalent frameworks and models for
understanding what affects acceptance. It also presents the results of a first
pilot test to explore the possibilities of using the Unified Theory of
Acceptance and Use of Technology for studying acceptance of driver
support systems.
xChapter 5 discusses the implications and conclusions of the thesis and
proposes ideas for further research.
Adell | Driver experiences and acceptance of driver support systems
10
The introductory/summarizing section consists of 5 chapters:
xThis chapter, chapter 1, provides the background and objectives of the
thesis.
xChapter 2 focuses on drivers’ experiences with ISA (Intelligent Speed
Adaptation). The ISA-systems AAP (Active Accelerator Pedal) and BEEP
(an ISA system, mainly providing auditory feedback) are described, as are
the two field trials used to study drivers’ experiences of ISA-systems. The
effects on driver behaviour, found in the two trials, are briefly described
and a more extensive summary of drivers’ experiences with the systems is
provided. The chapter ends with a discussion on the implications of the
findings.
xChapter 3 contains various definitions of acceptance. Considerations when
defining and working on acceptance are discussed and a new definition is
proposed. The chapter also describes the various ways of assessing
acceptance and how these relate to the definitions. It ends with a
discussion on the limitations of the methods used today.
xChapter 4 provides an overview of prevalent frameworks and models for
understanding what affects acceptance. It also presents the results of a first
pilot test to explore the possibilities of using the Unified Theory of
Acceptance and Use of Technology for studying acceptance of driver
support systems.
xChapter 5 discusses the implications and conclusions of the thesis and
proposes ideas for further research.
11
2 Driver experiences of ISA
The only source of knowledge is experience.”
Albert Einstein
11
2 Driver experiences of ISA
The only source of knowledge is experience.”
Albert Einstein
Adell | Driver experiences and acceptance of driver support systems
12
Adell | Driver experiences and acceptance of driver support systems
12
2 Driver experiences of ISA
13
The driver’s individual decision on the use of a certain driver support system is
based on his/her knowledge, understanding and beliefs regarding issues related to
the system. One very influential source for this is the drivers’ own experiences with
it, which might be different from the effects of the system measured by external
observers.
2.1 ISA systems
Intelligent Speed Adaptation (ISA) is the generic name for a driver support system
that “knows” the actual speed limit and uses that information to
inform/support/limit the driver to comply with it. There are several ways to
categorise these systems. One dimension used is how intervening the systems are:
An Informative system displays the speed limit to the driver; an Intervening system
reminds the driver when the speed limit is exceeded. A Limiting system limits
vehicle speed to the speed limit.
Another dimension relates to the characteristics of the speed recommended by the
system. A Fixed system uses the posted speed limits; a Variable system additionally
lowers the recommended speed due to prevailing conditions e.g. road
characteristics like sharp curves, zebra crossings. A Dynamic system also lowers the
recommended speed due to dynamically changing conditions e.g. fog, slippery
road surface, accident ahead.
A third dimension for differentiating ISA systems is how voluntary the use of the
system is. The level of voluntariness can either refer to whether the driver can turn
on and off the ISA system that is installed in his/her car, or to whether it is
voluntary or mandatory to have the system installed in one’s car.
There are many different types of ISA systems, using different means to interact
with the driver. In different trials, information has e.g. been given to the driver
through a speed limit sign below the speedometer, text messages displayed on the
simulator screen, a colour coded display together with auditory messages, flashing
red light together with beep-sound, flashing red LED display together with a
spoken message, haptic throttle (counter pressure in the accelerator pedal), ‘dead
throttle’ (pressing down the accelerator pedal when the speed limit is reached will
have no effect, and no feedback in the accelerator pedal is given) and seatbelt
vibrations, see e.g. papers I, II and III, Saad and Malaterre (1982), Nilsson and
2 Driver experiences of ISA
13
The driver’s individual decision on the use of a certain driver support system is
based on his/her knowledge, understanding and beliefs regarding issues related to
the system. One very influential source for this is the drivers’ own experiences with
it, which might be different from the effects of the system measured by external
observers.
2.1 ISA systems
Intelligent Speed Adaptation (ISA) is the generic name for a driver support system
that “knows” the actual speed limit and uses that information to
inform/support/limit the driver to comply with it. There are several ways to
categorise these systems. One dimension used is how intervening the systems are:
An Informative system displays the speed limit to the driver; an Intervening system
reminds the driver when the speed limit is exceeded. A Limiting system limits
vehicle speed to the speed limit.
Another dimension relates to the characteristics of the speed recommended by the
system. A Fixed system uses the posted speed limits; a Variable system additionally
lowers the recommended speed due to prevailing conditions e.g. road
characteristics like sharp curves, zebra crossings. A Dynamic system also lowers the
recommended speed due to dynamically changing conditions e.g. fog, slippery
road surface, accident ahead.
A third dimension for differentiating ISA systems is how voluntary the use of the
system is. The level of voluntariness can either refer to whether the driver can turn
on and off the ISA system that is installed in his/her car, or to whether it is
voluntary or mandatory to have the system installed in one’s car.
There are many different types of ISA systems, using different means to interact
with the driver. In different trials, information has e.g. been given to the driver
through a speed limit sign below the speedometer, text messages displayed on the
simulator screen, a colour coded display together with auditory messages, flashing
red light together with beep-sound, flashing red LED display together with a
spoken message, haptic throttle (counter pressure in the accelerator pedal), ‘dead
throttle’ (pressing down the accelerator pedal when the speed limit is reached will
have no effect, and no feedback in the accelerator pedal is given) and seatbelt
vibrations, see e.g. papers I, II and III, Saad and Malaterre (1982), Nilsson and
Adell | Driver experiences and acceptance of driver support systems
14
Berlin (1992), Almquist and Nygård (1997), Brookhuis and de Waard (1997),
Brookhuis and de Waard (1999), Carsten and Fowkes (2000), Lahrman et al.
(2001), Varhelyi and Mäkinen (2001) and Adell et al. (2009).
Two different systems were used in the trials reported in this thesis; the active
accelerator pedal (AAP) and the BEEP system. Both can be classified as fixed,
intervening systems. Within the test area the use of the system was mandatory; the
system could not be switched off. Outside the test area the use of the system was
voluntary.
2.1.1 The Active Accelerator Pedal system
The Active Accelerator Pedal (AAP) system used in the trials gave the driver
continuous visual information about the prevailing speed limit and haptic
support/feedback when the current speed limit was exceeded. The visual
information consisted of a dashboard-mounted display, see Figure 2. The haptic
feedback consisted of an Active Accelerator Pedal (AAP) exerting a counterforce in
the throttle at speeds over the current speed limit. The throttle had to be pressed
approximately three to five times harder than normal in order to override the
counterforce. The actual speed limit was provided by an onboard digital map
combined with GPS.
Figure 2: The display showing the current speed limit. (Published with the kind
permission of SG Utveckling)
2.1.2 The BEEP system
The BEEP system used in the trial gave the driver continuous visual information
about the prevailing speed limit, and auditory and visual warnings when the speed
limit was exceeded. The visual information was identical to the information given
by the AAP system, see Figure 2. When the speed limit was exceeded, a small red
light flashing on the display and a beep sound warned the driver. The sound was a
3500 Hz tone with duration of 0.1 seconds. It was repeated as long as the speed
Adell | Driver experiences and acceptance of driver support systems
14
Berlin (1992), Almquist and Nygård (1997), Brookhuis and de Waard (1997),
Brookhuis and de Waard (1999), Carsten and Fowkes (2000), Lahrman et al.
(2001), Varhelyi and Mäkinen (2001) and Adell et al. (2009).
Two different systems were used in the trials reported in this thesis; the active
accelerator pedal (AAP) and the BEEP system. Both can be classified as fixed,
intervening systems. Within the test area the use of the system was mandatory; the
system could not be switched off. Outside the test area the use of the system was
voluntary.
2.1.1 The Active Accelerator Pedal system
The Active Accelerator Pedal (AAP) system used in the trials gave the driver
continuous visual information about the prevailing speed limit and haptic
support/feedback when the current speed limit was exceeded. The visual
information consisted of a dashboard-mounted display, see Figure 2. The haptic
feedback consisted of an Active Accelerator Pedal (AAP) exerting a counterforce in
the throttle at speeds over the current speed limit. The throttle had to be pressed
approximately three to five times harder than normal in order to override the
counterforce. The actual speed limit was provided by an onboard digital map
combined with GPS.
Figure 2: The display showing the current speed limit. (Published with the kind
permission of SG Utveckling)
2.1.2 The BEEP system
The BEEP system used in the trial gave the driver continuous visual information
about the prevailing speed limit, and auditory and visual warnings when the speed
limit was exceeded. The visual information was identical to the information given
by the AAP system, see Figure 2. When the speed limit was exceeded, a small red
light flashing on the display and a beep sound warned the driver. The sound was a
3500 Hz tone with duration of 0.1 seconds. It was repeated as long as the speed
2 Driver experiences of ISA
15
was higher than the speed limit and the frequency of the repetitions increased the
more the speed limit was exceeded. The longest time interval between the beeps
was 1.5 seconds at the lowest speeding, and when the speed limit was exceeded by
20 km/h or more the beep turned into a continuous tone. The loudness level was
75 dBA when intermittent and 78 dBA when continuous.
2.2 Two field-trials with ISA
Two different field trials with ISA are employed in this thesis to study the drivers’
experiences of ISA-systems, namely a large-scale, long-term field trial with AAP
and a comparative field study on AAP and BEEP. First, short descriptions of the
two trials are given below, followed by a discussion of the effects on driver
behaviour found in those two trials. Thereafter the drivers’ experiences are
presented in terms of experienced effects, workload, emotional state and
acceptance-related issues, followed by findings regarding effects of duration of use,
driver characteristics, region, type of ISA and experiences of a malfunctioning
system. Finally, the chapter ends with a discussion of the implications from the
findings.
2.2.1 Large-scale field trial with AAP
A long-term field trial with the Active Accelerator Pedal system (AAP) was carried
out between 2000 and 2001 in Lund, Sweden, as part of a national large-scale ISA
trial. The system was installed in 281 passenger cars (247 owned by private drivers
and 34 owned by companies) and the owners continued to use their cars with the
system running for between 6 and 12 months. The system was activated
automatically when the vehicle was within the city of Lund (the test area) and
could not be turned off. The evaluation was designed as a short/long-term usage
within-subjects study, using gender, age, initial attitude towards the system, and
driver type (private or company car driver) as between-subject factors. Driver
experiences were elicited by questionnaires after one month of use (short-term use)
and at the end of the trial (long-term use). The short-term response rate was 86%,
the long-term 82%, and 80% of the drivers answered both questionnaires. For
further information about the trial and the analysis of the data see papers I and II.
Two thirds of the drivers reported some level of malfunctioning of the system.
These issues were mainly related to technological problems with the ISA system or
the interaction between the system and the car. Problems experienced by the
drivers were e.g. delayed throttle response and continuous counter pressure in the
throttle. Errors in the digital map providing the speed limits or difficulty with the
navigation unit also caused some problems with the functionality. In the group of
drivers reporting malfunctioning, there was an overrepresentation of company car
2 Driver experiences of ISA
15
was higher than the speed limit and the frequency of the repetitions increased the
more the speed limit was exceeded. The longest time interval between the beeps
was 1.5 seconds at the lowest speeding, and when the speed limit was exceeded by
20 km/h or more the beep turned into a continuous tone. The loudness level was
75 dBA when intermittent and 78 dBA when continuous.
2.2 Two field-trials with ISA
Two different field trials with ISA are employed in this thesis to study the drivers’
experiences of ISA-systems, namely a large-scale, long-term field trial with AAP
and a comparative field study on AAP and BEEP. First, short descriptions of the
two trials are given below, followed by a discussion of the effects on driver
behaviour found in those two trials. Thereafter the drivers’ experiences are
presented in terms of experienced effects, workload, emotional state and
acceptance-related issues, followed by findings regarding effects of duration of use,
driver characteristics, region, type of ISA and experiences of a malfunctioning
system. Finally, the chapter ends with a discussion of the implications from the
findings.
2.2.1 Large-scale field trial with AAP
A long-term field trial with the Active Accelerator Pedal system (AAP) was carried
out between 2000 and 2001 in Lund, Sweden, as part of a national large-scale ISA
trial. The system was installed in 281 passenger cars (247 owned by private drivers
and 34 owned by companies) and the owners continued to use their cars with the
system running for between 6 and 12 months. The system was activated
automatically when the vehicle was within the city of Lund (the test area) and
could not be turned off. The evaluation was designed as a short/long-term usage
within-subjects study, using gender, age, initial attitude towards the system, and
driver type (private or company car driver) as between-subject factors. Driver
experiences were elicited by questionnaires after one month of use (short-term use)
and at the end of the trial (long-term use). The short-term response rate was 86%,
the long-term 82%, and 80% of the drivers answered both questionnaires. For
further information about the trial and the analysis of the data see papers I and II.
Two thirds of the drivers reported some level of malfunctioning of the system.
These issues were mainly related to technological problems with the ISA system or
the interaction between the system and the car. Problems experienced by the
drivers were e.g. delayed throttle response and continuous counter pressure in the
throttle. Errors in the digital map providing the speed limits or difficulty with the
navigation unit also caused some problems with the functionality. In the group of
drivers reporting malfunctioning, there was an overrepresentation of company car
Adell | Driver experiences and acceptance of driver support systems
16
drivers (32 of the 34 were company car drivers). No differences in gender (2,
p=0.29), age ( 2
, p=0.47) or initial attitude ( 2
, p=0.99) could be found among
test drivers with malfunctioning systems compared to drivers not reporting
problems.
2.2.2 Comparative study on AAP and BEEP
The comparative field trials with the AAP and the BEEP systems were carried out
in Hungary and Spain in 2003 and 2004. The systems were installed in the
participants’ private cars and each system was used in their regular driving for one
month. All 39 participants (20 Hungarian +19 Spaniards) tested both systems;
half of the drivers used the AAP first followed by the BEEP and the other half vice
versa. The systems were activated automatically when the vehicle was within the
test areas (the cities of Debrecen and Mataró) and could not be turned off. The
evaluation was designed as a before/during/after study using system as a within-
subject factor, and country as a between-subject factor. Data on the use of each
system was collected during (logged data) and after (questionnaire) one month of
usage. All the drivers answered all the questionnaires with the exception of one
Hungarian driver who did not answer the last questionnaire. For further
information about the trial and the analysis of the data see paper III.
The systems were based on an improved version of the AAP-system used in the
large-scale field trial held in Lund, Sweden, between 2000 and 2001. Nevertheless,
there were still some technological problems during the trial, mainly regarding the
accuracy of the map and the fact that the system was also active outside the test
area. In total, 12 (of the 38 drivers answering the question) drivers reported
problems with the AAP and 11 with the BEEP system.
2.3 Effects on driver behaviour
The effects on driver behaviour when using the AAP in the large-scale field trial
were reported by Hjälmdahl et al. (2002), Hjälmdahl and Várhelyi (2004) and
Várhelyi et al. (2004). The results showed that the test drivers drove at lower and
more even speeds when they used the AAP. Compliance with the speed limits
improved. There were no significant changes in time consumption, but the results
indicate a decrease rather than an increase. The average amount of CO and NOx-
emissions per car decreased. The results provided no evidence of compensatory
behaviour, either at intersections (approach speeds and turning speeds) or outside
the test area (speed level). Further, there were no signs of spill-over effects in the
form of lower speeds outside the test area. However, there were some tendencies of
negative behavioural effects where drivers forgot to adapt their speed to the speed
limits or to the prevailing conditions when they were not supported by the system.
Adell | Driver experiences and acceptance of driver support systems
16
drivers (32 of the 34 were company car drivers). No differences in gender (2,
p=0.29), age ( 2
, p=0.47) or initial attitude ( 2
, p=0.99) could be found among
test drivers with malfunctioning systems compared to drivers not reporting
problems.
2.2.2 Comparative study on AAP and BEEP
The comparative field trials with the AAP and the BEEP systems were carried out
in Hungary and Spain in 2003 and 2004. The systems were installed in the
participants’ private cars and each system was used in their regular driving for one
month. All 39 participants (20 Hungarian +19 Spaniards) tested both systems;
half of the drivers used the AAP first followed by the BEEP and the other half vice
versa. The systems were activated automatically when the vehicle was within the
test areas (the cities of Debrecen and Mataró) and could not be turned off. The
evaluation was designed as a before/during/after study using system as a within-
subject factor, and country as a between-subject factor. Data on the use of each
system was collected during (logged data) and after (questionnaire) one month of
usage. All the drivers answered all the questionnaires with the exception of one
Hungarian driver who did not answer the last questionnaire. For further
information about the trial and the analysis of the data see paper III.
The systems were based on an improved version of the AAP-system used in the
large-scale field trial held in Lund, Sweden, between 2000 and 2001. Nevertheless,
there were still some technological problems during the trial, mainly regarding the
accuracy of the map and the fact that the system was also active outside the test
area. In total, 12 (of the 38 drivers answering the question) drivers reported
problems with the AAP and 11 with the BEEP system.
2.3 Effects on driver behaviour
The effects on driver behaviour when using the AAP in the large-scale field trial
were reported by Hjälmdahl et al. (2002), Hjälmdahl and Várhelyi (2004) and
Várhelyi et al. (2004). The results showed that the test drivers drove at lower and
more even speeds when they used the AAP. Compliance with the speed limits
improved. There were no significant changes in time consumption, but the results
indicate a decrease rather than an increase. The average amount of CO and NOx-
emissions per car decreased. The results provided no evidence of compensatory
behaviour, either at intersections (approach speeds and turning speeds) or outside
the test area (speed level). Further, there were no signs of spill-over effects in the
form of lower speeds outside the test area. However, there were some tendencies of
negative behavioural effects where drivers forgot to adapt their speed to the speed
limits or to the prevailing conditions when they were not supported by the system.
2 Driver experiences of ISA
17
The effects on driver behaviour when using the AAP and the BEEP systems in the
comparative study carried out in Hungary and Spain were reported in paper III.
The results are very much in line with the findings of the large-scale field trial in
Lund, Sweden. Both the AAP and the BEEP system decreased the mean speed as
well as the 85 percentile speed. The largest effects of both ISA systems were found
at the highest speeds. The speed variance decreased on all the analysed road types,
except on motorways with a 120 km/h speed limit in Spain. The AAP had larger
speed reducing effects than the BEEP on roads with speed limits of 50 and 80
km/h. The results on 50 km/h streets showed that the effects of both systems were
larger in Spain than in Hungary. After the systems were removed from the
vehicles, the speed levels increased, in both countries, to almost that of the before
situation.
2.4 Driver experiences
The results below give a general picture of the drivers’ experiences of the system
studied. The results from papers I and II showed that technological problems
influenced the drivers’ experiences. To illustrate driver experiences of the system,
separated from technological problems, the results in cases where the drivers
reported malfunctioning system are duly analysed and reported separately below.
More detailed results of the trials are given in papers I, II and III.
2.4.1 Experienced effects
The largest experienced effect when driving with the ISA systems was a
considerable reduction in the risk of being fined for speeding in all three countries.
The drivers also experienced a considerable reduction in speed within the test area.
According to the test drivers, the use of the AAP also made it easier to comply
with the speed limits (paper I) and they expected the speeds to decrease if all cars
were equipped with the system (paper III). The drivers also felt a slight increase in
their own safety and an increase in travel time when driving with the ISA systems.
2.4.2 Workload
There were some differences in the reported workload in the three countries. The
subjective workload was assessed by six factors according to NASA-TLX (Byers et
al., 1989). In Sweden, one workload factor, namely time pressure, showed an
increase in workload, and two workload factors, i.e., performance and the need to
accelerate and brake (long-term experiences), showed a decrease. The results from
Spain showed no indications of increased workload when driving with either of
the two ISA systems, whereas the Hungarian drivers noted an increased workload
2 Driver experiences of ISA
17
The effects on driver behaviour when using the AAP and the BEEP systems in the
comparative study carried out in Hungary and Spain were reported in paper III.
The results are very much in line with the findings of the large-scale field trial in
Lund, Sweden. Both the AAP and the BEEP system decreased the mean speed as
well as the 85 percentile speed. The largest effects of both ISA systems were found
at the highest speeds. The speed variance decreased on all the analysed road types,
except on motorways with a 120 km/h speed limit in Spain. The AAP had larger
speed reducing effects than the BEEP on roads with speed limits of 50 and 80
km/h. The results on 50 km/h streets showed that the effects of both systems were
larger in Spain than in Hungary. After the systems were removed from the
vehicles, the speed levels increased, in both countries, to almost that of the before
situation.
2.4 Driver experiences
The results below give a general picture of the drivers’ experiences of the system
studied. The results from papers I and II showed that technological problems
influenced the drivers’ experiences. To illustrate driver experiences of the system,
separated from technological problems, the results in cases where the drivers
reported malfunctioning system are duly analysed and reported separately below.
More detailed results of the trials are given in papers I, II and III.
2.4.1 Experienced effects
The largest experienced effect when driving with the ISA systems was a
considerable reduction in the risk of being fined for speeding in all three countries.
The drivers also experienced a considerable reduction in speed within the test area.
According to the test drivers, the use of the AAP also made it easier to comply
with the speed limits (paper I) and they expected the speeds to decrease if all cars
were equipped with the system (paper III). The drivers also felt a slight increase in
their own safety and an increase in travel time when driving with the ISA systems.
2.4.2 Workload
There were some differences in the reported workload in the three countries. The
subjective workload was assessed by six factors according to NASA-TLX (Byers et
al., 1989). In Sweden, one workload factor, namely time pressure, showed an
increase in workload, and two workload factors, i.e., performance and the need to
accelerate and brake (long-term experiences), showed a decrease. The results from
Spain showed no indications of increased workload when driving with either of
the two ISA systems, whereas the Hungarian drivers noted an increased workload
Adell | Driver experiences and acceptance of driver support systems
18
in four of the six factors, namely physical demand, time pressure, effort and
performance.
2.4.3 Emotional state
The Swedish drivers noted a reduction in driving enjoyment and an increase in the
feeling of obstructing the traffic. The Hungarian and Spanish drivers reported
increased irritation, increased feeling of obstructing the traffic and increased
feeling of being controlled as well as a tendency of reduced enjoyment when
driving with the AAP.
2.4.4 Acceptance-related findings
The results below cover some of the indicators of driver acceptance used today.
The concept of the ISA was rated positively by a majority of the Swedish,
Hungarian and Spanish drivers on a five-grade scale from “very bad” to “very
good”.
The system was considered to be useful, while it was considered neutral on the
satisfaction scale (van der Laan et al. 1997). In Sweden (where a somewhat
modified version was used based on a previous translation of the items into
Swedish), the system was considered ‘good’, ‘important’, ‘effective’, ‘clear’ and
‘informing’ rather than their opposites, which is an indication of relatively high
usefulness. Further, the system was reported to be slightly ‘pleasant’ and ‘ugly’ and
neither ‘irritating’ nor ‘soothing’ and neither ‘uncomfortable’ nor ‘comfortable’
which indicated neither positive nor negative satisfaction. In Hungary and Spain
both systems were considered ‘good’, ‘effective’, ‘useful’, ‘assisting’ and ‘raising
alertness’ as well as neither ‘unpleasant’ nor ‘pleasant’ and neither ‘undesirable’ nor
‘desirable’. The BEEP system was considered to be ‘annoying’ and ‘irritating’
while the AAP was not.
The Swedish drivers generally state that they ‘never’ or ‘seldom’ overrode the
system. Whether this should be interpreted as an indication of acceptance or not
depends, like all other measurements, on the definition of acceptance as well as on
whether the possibility of overriding the system is seen as utilizing a functionality
of the system or as turning off the system.
After the trials were finished, about 28 % of the Swedish drivers wanted to keep
the system in their car. In Hungary and Spain about 50 % wanted to keep the
AAP while about 65 % wanted to keep the BEEP system.
Adell | Driver experiences and acceptance of driver support systems
18
in four of the six factors, namely physical demand, time pressure, effort and
performance.
2.4.3 Emotional state
The Swedish drivers noted a reduction in driving enjoyment and an increase in the
feeling of obstructing the traffic. The Hungarian and Spanish drivers reported
increased irritation, increased feeling of obstructing the traffic and increased
feeling of being controlled as well as a tendency of reduced enjoyment when
driving with the AAP.
2.4.4 Acceptance-related findings
The results below cover some of the indicators of driver acceptance used today.
The concept of the ISA was rated positively by a majority of the Swedish,
Hungarian and Spanish drivers on a five-grade scale from “very bad” to “very
good”.
The system was considered to be useful, while it was considered neutral on the
satisfaction scale (van der Laan et al. 1997). In Sweden (where a somewhat
modified version was used based on a previous translation of the items into
Swedish), the system was considered ‘good’, ‘important’, ‘effective’, ‘clear’ and
‘informing’ rather than their opposites, which is an indication of relatively high
usefulness. Further, the system was reported to be slightly ‘pleasant’ and ‘ugly’ and
neither ‘irritating’ nor ‘soothing’ and neither ‘uncomfortable’ nor ‘comfortable’
which indicated neither positive nor negative satisfaction. In Hungary and Spain
both systems were considered ‘good’, ‘effective’, ‘useful’, ‘assisting’ and ‘raising
alertness’ as well as neither ‘unpleasant’ nor ‘pleasant’ and neither ‘undesirable’ nor
‘desirable’. The BEEP system was considered to be ‘annoying’ and ‘irritating’
while the AAP was not.
The Swedish drivers generally state that they ‘never’ or ‘seldom’ overrode the
system. Whether this should be interpreted as an indication of acceptance or not
depends, like all other measurements, on the definition of acceptance as well as on
whether the possibility of overriding the system is seen as utilizing a functionality
of the system or as turning off the system.
After the trials were finished, about 28 % of the Swedish drivers wanted to keep
the system in their car. In Hungary and Spain about 50 % wanted to keep the
AAP while about 65 % wanted to keep the BEEP system.
2 Driver experiences of ISA
19
2.4.5 Effects of duration of use on driver experiences of AAP
Data from two time periods were considered in the Swedish trial, short-term (after
one month of usage) and long-term (after an additional 5 to 11 months of usage).
When comparing these periods, differences were found in the drivers’ assessment
of speed, their performance, driving enjoyment, stress and awareness of speed
limits outside the test area.
The experienced speed change differed in magnitude between the two
measurement periods. In the short-term, the drivers reported changes of the same
magnitude regardless of speed limits, as opposed to larger reductions in streets
with lower speed limits in the long term. The driving performance was rated
higher in the long-term compared to the short-term, whereas the driving
enjoyment decreased further and the stress increased. The awareness of speed
limits outside the test area decreased over time.
2.4.6 Effects of driver characteristics on their experiences of AAP
The large-scale field trial with AAP in Lund, Sweden, where 281 drivers
participated, made it possible to study the experiences of different driver groups.
The drivers were categorised according to their age, gender, initial attitude towards
the system and whether they used the system in their private car or in a company
car.
Main effects were found for initial attitude, age, gender and driver type. Most of
the effects were found for initial attitude which mainly influenced workload,
emotional state and usage. Overall, the initially negative drivers distinguished
themselves from the initially positive drivers in that they experienced more
difficulty with the system. Four of the six workload factors were negatively affected
(attention, time pressure, effort and frustration), as were two of the four factors for
emotional state (stress and driving enjoyment). The initially negative drivers also
found the system less attractive and pleasant and stated that they overrode the
system more frequently than the positive drivers did.
Effects of age were found for usefulness/satisfaction, willingness to keep the system
after the trial ended, workload and usage, indicating that younger drivers found
the usefulness of and satisfaction with the system to be worse compared to older
drivers, and were willing to keep the system after the trial ended to a lower extent
compared to middle-aged drivers. Older drivers rated their driving performance
with the system more highly and reported lower frustration than the younger
drivers did. Further, main effects of age groups were found in the usage of the
system, which indicated that the older drivers perceived the counter force as more
of a command to lower their speed than as a support to keep the speed limit.
2 Driver experiences of ISA
19
2.4.5 Effects of duration of use on driver experiences of AAP
Data from two time periods were considered in the Swedish trial, short-term (after
one month of usage) and long-term (after an additional 5 to 11 months of usage).
When comparing these periods, differences were found in the drivers’ assessment
of speed, their performance, driving enjoyment, stress and awareness of speed
limits outside the test area.
The experienced speed change differed in magnitude between the two
measurement periods. In the short-term, the drivers reported changes of the same
magnitude regardless of speed limits, as opposed to larger reductions in streets
with lower speed limits in the long term. The driving performance was rated
higher in the long-term compared to the short-term, whereas the driving
enjoyment decreased further and the stress increased. The awareness of speed
limits outside the test area decreased over time.
2.4.6 Effects of driver characteristics on their experiences of AAP
The large-scale field trial with AAP in Lund, Sweden, where 281 drivers
participated, made it possible to study the experiences of different driver groups.
The drivers were categorised according to their age, gender, initial attitude towards
the system and whether they used the system in their private car or in a company
car.
Main effects were found for initial attitude, age, gender and driver type. Most of
the effects were found for initial attitude which mainly influenced workload,
emotional state and usage. Overall, the initially negative drivers distinguished
themselves from the initially positive drivers in that they experienced more
difficulty with the system. Four of the six workload factors were negatively affected
(attention, time pressure, effort and frustration), as were two of the four factors for
emotional state (stress and driving enjoyment). The initially negative drivers also
found the system less attractive and pleasant and stated that they overrode the
system more frequently than the positive drivers did.
Effects of age were found for usefulness/satisfaction, willingness to keep the system
after the trial ended, workload and usage, indicating that younger drivers found
the usefulness of and satisfaction with the system to be worse compared to older
drivers, and were willing to keep the system after the trial ended to a lower extent
compared to middle-aged drivers. Older drivers rated their driving performance
with the system more highly and reported lower frustration than the younger
drivers did. Further, main effects of age groups were found in the usage of the
system, which indicated that the older drivers perceived the counter force as more
of a command to lower their speed than as a support to keep the speed limit.
Adell | Driver experiences and acceptance of driver support systems
20
Younger drivers stated that they overrode the system more frequently compared to
older drivers. No systematic correlations of driver experiences due to gender or
driver type were found.
The interaction effects indicated that women were more influenced by their initial
attitude and their driver type (private/company car) compared to men. Similarly,
company car drivers and older drivers were more influenced by their initial
attitude than private car drivers and younger drivers, respectively.
Interaction effects with time were mainly found between time and driver type and
time and attitude. Generally, the effects showed larger changes over time for
company car drivers and for the initially negative drivers. Interaction effects were
mainly found on experienced speed changes and emotional state.
2.4.7 Differences in experiences of Swedish, Hungarian and Spanish drivers
Most of the short-term results from Sweden concur with the results of AAP use in
Hungary and Spain, see Table 1. The differences among the countries are to be
found in the workload assessment, the effect on driving enjoyment and the
willingness to keep the system after the trials ended. The Hungarian drivers stated
that the workload increased, whereas no change was found in Spain, while the
Swedish drivers reported both an increase and a decrease in various workload
factors. The decrease in driving enjoyment communicated by the drivers in
Sweden was not confirmed by the drivers in Hungary and Spain, although some
indications in that direction were found, especially in Hungary. The difference in
willingness to keep the system after the trial ended between Swedish drivers on the
one hand and Hungarian and Spanish drivers on the other, was most likely due to
the relatively larger number of system failures in the Swedish trial.
Adell | Driver experiences and acceptance of driver support systems
20
Younger drivers stated that they overrode the system more frequently compared to
older drivers. No systematic correlations of driver experiences due to gender or
driver type were found.
The interaction effects indicated that women were more influenced by their initial
attitude and their driver type (private/company car) compared to men. Similarly,
company car drivers and older drivers were more influenced by their initial
attitude than private car drivers and younger drivers, respectively.
Interaction effects with time were mainly found between time and driver type and
time and attitude. Generally, the effects showed larger changes over time for
company car drivers and for the initially negative drivers. Interaction effects were
mainly found on experienced speed changes and emotional state.
2.4.7 Differences in experiences of Swedish, Hungarian and Spanish drivers
Most of the short-term results from Sweden concur with the results of AAP use in
Hungary and Spain, see Table 1. The differences among the countries are to be
found in the workload assessment, the effect on driving enjoyment and the
willingness to keep the system after the trials ended. The Hungarian drivers stated
that the workload increased, whereas no change was found in Spain, while the
Swedish drivers reported both an increase and a decrease in various workload
factors. The decrease in driving enjoyment communicated by the drivers in
Sweden was not confirmed by the drivers in Hungary and Spain, although some
indications in that direction were found, especially in Hungary. The difference in
willingness to keep the system after the trial ended between Swedish drivers on the
one hand and Hungarian and Spanish drivers on the other, was most likely due to
the relatively larger number of system failures in the Swedish trial.
2 Driver experiences of ISA
21
Table 1: Comparison of results of experienced effects of AAP after one month’s use in
three different “regional typical” countries in Europe
Hungary Spain Sweden
Experienced effects
Safety Moderate increase Small increase Small increase
Speed change Moderate decrease
Small to si
g
nificant
decrease (varying with
speed limit)
Moderate decrease
Gettin
g
fined for
speeding Moderate decrease Significant decrease Significant decrease
Travel time Moderate increase Moderate increase Moderate increase
Fuel consumption No change No change No change
Workload
Workload Moderate increase No change No overall change
Emotional state
Irritation Small increase Small increase
Small increase
(only those with mal-
functioning system)
Stress No change No change No change
Driving enjoyment No change No change Small decrease
Acceptance-related issues
The concept of AAP 17 of 19 drivers
positive (89 %)
16 of 18 drivers
positive (89 %)
126 of 160 drivers
positive (79 %)
System features
Usefulness: moderate
positive
Satisfactory: neutral
Usefulness: moderate
positive
Satisfactory: neutral
Usefulness: moderate
positive
Satisfactory: neutral
Keeping the system 10 of 20 drivers
positive (50 %)
9 of 17 drivers
positive (53 %)
44 of 155 drivers
positive (28%)
2.4.8 Differences in experiences of using AAP and BEEP
Comparing the Hungarian and Spanish drivers’ experiences when using the AAP
versus the BEEP system, workload factors, aspects of the emotional state and
system features stand out. According to the drivers, the AAP increased the
‘physical demand’ and the ‘effort’ of driving. The BEEP increased the ‘mental
demand’ according to the Spanish drivers. When using the AAP, there were some
indications of reduced ‘driving enjoyment’. The BEEP system was considered to
be more ‘annoying’ and ‘irritating’ and ‘raising alertness’ compared to the AAP.
Nevertheless, the drivers were more positive to having the BEEP system in their
cars compared to having the AAP. When choosing between the systems, more
drivers preferred the BEEP system compared to the AAP.
2 Driver experiences of ISA
21
Table 1: Comparison of results of experienced effects of AAP after one month’s use in
three different “regional typical” countries in Europe
Hungary Spain Sweden
Experienced effects
Safety Moderate increase Small increase Small increase
Speed change Moderate decrease
Small to si
g
nificant
decrease (varying with
speed limit)
Moderate decrease
Gettin
fined for
speeding Moderate decrease Significant decrease Significant decrease
Travel time Moderate increase Moderate increase Moderate increase
Fuel consumption No change No change No change
Workload
Workload Moderate increase No change No overall change
Emotional state
Irritation Small increase Small increase
Small increase
(only those with mal-
functioning system)
Stress No change No change No change
Driving enjoyment No change No change Small decrease
Acceptance-related issues
The concept of AAP 17 of 19 drivers
positive (89 %)
16 of 18 drivers
positive (89 %)
126 of 160 drivers
positive (79 %)
System features
Usefulness: moderate
positive
Satisfactory: neutral
Usefulness: moderate
positive
Satisfactory: neutral
Usefulness: moderate
positive
Satisfactory: neutral
Keeping the system 10 of 20 drivers
positive (50 %)
9 of 17 drivers
positive (53 %)
44 of 155 drivers
positive (28%)
2.4.8 Differences in experiences of using AAP and BEEP
Comparing the Hungarian and Spanish drivers’ experiences when using the AAP
versus the BEEP system, workload factors, aspects of the emotional state and
system features stand out. According to the drivers, the AAP increased the
‘physical demand’ and the ‘effort’ of driving. The BEEP increased the ‘mental
demand’ according to the Spanish drivers. When using the AAP, there were some
indications of reduced ‘driving enjoyment’. The BEEP system was considered to
be more ‘annoying’ and ‘irritating’ and ‘raising alertness’ compared to the AAP.
Nevertheless, the drivers were more positive to having the BEEP system in their
cars compared to having the AAP. When choosing between the systems, more
drivers preferred the BEEP system compared to the AAP.
Adell | Driver experiences and acceptance of driver support systems
22
2.4.9 Effects on driver experiences of malfunctioning system
Technological problems caused two thirds of the Swedish test drivers to report a
malfunctioning system. These drivers’ experiences were similar to those of the
initially negative drivers (no difference in initial attitude could be found among
the drivers reporting technological problems). They reported higher workload
(effort and frustration) as well as deterioration in their emotional state (irritation
and stress) and higher override rate compared to drivers who did not report
technological problems. However, they also noted smaller effects of the system,
such as a smaller reduction in the risk of being fined for speeding and smaller
speed reductions.
The interaction effects indicated that women were more influenced by system
failure compared to men. No interaction effects with time were found, i.e., time
did not influence the experiences with malfunctioning systems.
2.5 Implications of the findings of the field trials
The studies in papers I, II and III show that the drivers found the system to be
effective in decreasing their speed and believed that their risk of being fined for
speeding decreased significantly. The emotional state deteriorated somewhat and
the subjective workload was also affected for some of the drivers. The drivers
found the concept of the system good and rated the usefulness of the system
moderately high. However, satisfaction with the system and the willingness to
keep the system were lower.
Generally, the results suggest that the most typically sceptical driver was a young,
male, company car driver with an initially negative attitude to the system. The
most enthusiastic driver was an older, female, private driver with an initially
positive attitude.
The results also showed that the experienced effects (speed, travel time, risk of
being fined for speeding etc.) were relatively uncorrelated with any of the driver
variables, namely initial attitude, age, gender and driver type. The drivers’ initial
attitude correlated with their emotional experiences, their assessment of the system
features and their override rate. The age correlated with the drivers’ assessment of
the system’s features and usage of the system (including override rate) as well as
the willingness to keep it. No systematic correlations due to gender or driver type
were found.
The results from paper I as well as results regarding observed speed behaviour
(Várhelyi et al., 2004) point to the fact that adaptation to the system was still
Adell | Driver experiences and acceptance of driver support systems
22
2.4.9 Effects on driver experiences of malfunctioning system
Technological problems caused two thirds of the Swedish test drivers to report a
malfunctioning system. These drivers’ experiences were similar to those of the
initially negative drivers (no difference in initial attitude could be found among
the drivers reporting technological problems). They reported higher workload
(effort and frustration) as well as deterioration in their emotional state (irritation
and stress) and higher override rate compared to drivers who did not report
technological problems. However, they also noted smaller effects of the system,
such as a smaller reduction in the risk of being fined for speeding and smaller
speed reductions.
The interaction effects indicated that women were more influenced by system
failure compared to men. No interaction effects with time were found, i.e., time
did not influence the experiences with malfunctioning systems.
2.5 Implications of the findings of the field trials
The studies in papers I, II and III show that the drivers found the system to be
effective in decreasing their speed and believed that their risk of being fined for
speeding decreased significantly. The emotional state deteriorated somewhat and
the subjective workload was also affected for some of the drivers. The drivers
found the concept of the system good and rated the usefulness of the system
moderately high. However, satisfaction with the system and the willingness to
keep the system were lower.
Generally, the results suggest that the most typically sceptical driver was a young,
male, company car driver with an initially negative attitude to the system. The
most enthusiastic driver was an older, female, private driver with an initially
positive attitude.
The results also showed that the experienced effects (speed, travel time, risk of
being fined for speeding etc.) were relatively uncorrelated with any of the driver
variables, namely initial attitude, age, gender and driver type. The drivers’ initial
attitude correlated with their emotional experiences, their assessment of the system
features and their override rate. The age correlated with the drivers’ assessment of
the system’s features and usage of the system (including override rate) as well as
the willingness to keep it. No systematic correlations due to gender or driver type
were found.
The results from paper I as well as results regarding observed speed behaviour
(Várhelyi et al., 2004) point to the fact that adaptation to the system was still
2 Driver experiences of ISA
23
ongoing after one month of use. It takes more than one month of driving to get
used to and incorporate a new driver support system in the driving task. The
reduction in driving enjoyment and increase in stress suggest that negative
emotional experiences may increase rather than decrease over time. The reduced
awareness of speed limits outside the test area points to an increased delegation of
responsibility over time. These time effects have to be acknowledged when
evaluating driver support systems.
In a real life trial, there are always external variables included in the drivers’
evaluation of the system, apart from the system itself. In the case of ISA, the
functionality of the system is based on speed limits determined by the road
authority. The correspondence between the speed limits and the driver’s opinion
about an appropriate speed will, for example, have an impact on how the system is
experienced by the drivers. In that sense, it is not only the system itself that is
evaluated by the drivers, but the “package” of using the system in real life.
Depending on the “settings” in real life, the experiences of using it may vary. It is
often difficult to separate the system effects from other external effects.
The similar results from the trials in the different regions in Europe indicate,
however, that the different “settings” within Europe have no major influence on
the experiences of using the systems; hence the findings might be generalized from
one European country to another. The results, presented in paper III, show that
many overall trends are the same in Hungary, Spain and Sweden although the
exact magnitudes cannot be compared. However, the results also show that the
subjective workload should not be generalized.
Comparing the AAP and the BEEP systems in Hungary and Spain, the drivers
generally preferred the BEEP system to the AAP, Despite assessing the BEEP
system as being more ‘irritating’ and ‘annoying’ compared to the AAP. However,
there were indications of higher subjective workload when using the AAP.
Technological problems are difficult to avoid in field trials with newly developed
driver support systems. These are likely to occur especially when prototypes are
used. The ambition would of course be to avoid any technological problems by
having a perfect system. Unfortunately, this is not achievable even in the best of
worlds and in the real world one has to settle for “good enough”. In the large-scale
field trial in Lund, carried out between 2000 and 2001, there were significant
technological problems. It is sometimes hard to know the extent of technological
problems in advance and one may of course question whether the system used in
that trial was “good enough”. In an attempt to separate the effects of the
technological problems, the results concerning drivers who had malfunctioning
systems have been shown separately. This may also be debated, since more
2 Driver experiences of ISA
23
ongoing after one month of use. It takes more than one month of driving to get
used to and incorporate a new driver support system in the driving task. The
reduction in driving enjoyment and increase in stress suggest that negative
emotional experiences may increase rather than decrease over time. The reduced
awareness of speed limits outside the test area points to an increased delegation of
responsibility over time. These time effects have to be acknowledged when
evaluating driver support systems.
In a real life trial, there are always external variables included in the drivers’
evaluation of the system, apart from the system itself. In the case of ISA, the
functionality of the system is based on speed limits determined by the road
authority. The correspondence between the speed limits and the driver’s opinion
about an appropriate speed will, for example, have an impact on how the system is
experienced by the drivers. In that sense, it is not only the system itself that is
evaluated by the drivers, but the “package” of using the system in real life.
Depending on the “settings” in real life, the experiences of using it may vary. It is
often difficult to separate the system effects from other external effects.
The similar results from the trials in the different regions in Europe indicate,
however, that the different “settings” within Europe have no major influence on
the experiences of using the systems; hence the findings might be generalized from
one European country to another. The results, presented in paper III, show that
many overall trends are the same in Hungary, Spain and Sweden although the
exact magnitudes cannot be compared. However, the results also show that the
subjective workload should not be generalized.
Comparing the AAP and the BEEP systems in Hungary and Spain, the drivers
generally preferred the BEEP system to the AAP, Despite assessing the BEEP
system as being more ‘irritating’ and ‘annoying’ compared to the AAP. However,
there were indications of higher subjective workload when using the AAP.
Technological problems are difficult to avoid in field trials with newly developed
driver support systems. These are likely to occur especially when prototypes are
used. The ambition would of course be to avoid any technological problems by
having a perfect system. Unfortunately, this is not achievable even in the best of
worlds and in the real world one has to settle for “good enough”. In the large-scale
field trial in Lund, carried out between 2000 and 2001, there were significant
technological problems. It is sometimes hard to know the extent of technological
problems in advance and one may of course question whether the system used in
that trial was “good enough”. In an attempt to separate the effects of the
technological problems, the results concerning drivers who had malfunctioning
systems have been shown separately. This may also be debated, since more
Adell | Driver experiences and acceptance of driver support systems
24
sceptical drivers may be more prone to report technological problems. However,
there was no statistically significant correlation between the initial attitude of the
driver and whether the driver reported problems. The technological problems had
an impact on the drivers’ experiences with the system. In fact those with
technological problems were the most negative driver group, also reporting less
positive effects of the system (smaller decrease in the risk of being fined for
speeding). This illustrates how fundamental the reliability of the system is to the
drivers. The lack of interaction effects between time and malfunctioning systems
suggests that technological problems influence the driver quite early in the
adaptation process. These effects are also unlikely to change over time.
Attempts to assess the acceptance of the systems are made in Papers II and III. The
drivers’ opinion about the concept of ISA, their assessment of the usefulness of
and satisfaction with the system as well as their willingness to have/keep and pay
for the system are used as indicators of acceptance. The findings in paper II show
that the concept of the AAP was rated positively by almost 80 % of the Swedish
drivers while barely 30 % of these drivers were willing to keep the system. If these
two measurements had been used in two different set-ups (different systems, driver
groups, experimental settings etc), two different conclusions on the level of
acceptance would have been made, going in two opposing directions.
A similar problem was found in the study reported in paper III, where more
drivers wanted to keep the BEEP system than the AAP even though the BEEP
system was perceived as more ‘irritating’ and ‘annoying’, both of which are
satisfaction items on the usefulness/satisfaction acceptance scale (van der Laan et
al., 1997).
The results reported in papers II and III do not give a clear picture of the
acceptance of the systems. Instead, they highlight the problematic situation in
which today’s acceptance research on driver support systems finds itself. The weak
common ground in terms of the definition of acceptance has produced numerous
different measurements, generating problems when e.g. interpreting and
comparing results from different studies regarding acceptance of driver support
systems.
Adell | Driver experiences and acceptance of driver support systems
24
sceptical drivers may be more prone to report technological problems. However,
there was no statistically significant correlation between the initial attitude of the
driver and whether the driver reported problems. The technological problems had
an impact on the drivers’ experiences with the system. In fact those with
technological problems were the most negative driver group, also reporting less
positive effects of the system (smaller decrease in the risk of being fined for
speeding). This illustrates how fundamental the reliability of the system is to the
drivers. The lack of interaction effects between time and malfunctioning systems
suggests that technological problems influence the driver quite early in the
adaptation process. These effects are also unlikely to change over time.
Attempts to assess the acceptance of the systems are made in Papers II and III. The
drivers’ opinion about the concept of ISA, their assessment of the usefulness of
and satisfaction with the system as well as their willingness to have/keep and pay
for the system are used as indicators of acceptance. The findings in paper II show
that the concept of the AAP was rated positively by almost 80 % of the Swedish
drivers while barely 30 % of these drivers were willing to keep the system. If these
two measurements had been used in two different set-ups (different systems, driver
groups, experimental settings etc), two different conclusions on the level of
acceptance would have been made, going in two opposing directions.
A similar problem was found in the study reported in paper III, where more
drivers wanted to keep the BEEP system than the AAP even though the BEEP
system was perceived as more ‘irritating’ and ‘annoying’, both of which are
satisfaction items on the usefulness/satisfaction acceptance scale (van der Laan et
al., 1997).
The results reported in papers II and III do not give a clear picture of the
acceptance of the systems. Instead, they highlight the problematic situation in
which today’s acceptance research on driver support systems finds itself. The weak
common ground in terms of the definition of acceptance has produced numerous
different measurements, generating problems when e.g. interpreting and
comparing results from different studies regarding acceptance of driver support
systems.
25
3 What is acceptance?
“The contrast between the drivers’ response to the concept
of the AAP and the willingness to keep the system puts a
clear focus on the importance of defining acceptance and
developing a tool to ensure reliable assessment of
acceptance and make inter-study comparisons possible.”
Paper II
25
3 What is acceptance?
“The contrast between the drivers’ response to the concept
of the AAP and the willingness to keep the system puts a
clear focus on the importance of defining acceptance and
developing a tool to ensure reliable assessment of
acceptance and make inter-study comparisons possible.”
Paper II
Adell | Driver experiences and acceptance of driver support systems
26
Adell | Driver experiences and acceptance of driver support systems
26
3 What is acceptance?
27
Despite the recognized importance of acceptance, the understanding of acceptance
of driver support systems is often taken for granted. While many studies claim to
have measured acceptance, few have defined what it is. As Regan et al. (2002) put
it: “While everyone seems to know what acceptability is, and all agree that
acceptability is important, there is no consistency across studies as to what
‘acceptability’ is and how to measure it”. Although there is no common definition
of acceptance, some definitions of acceptance as well as descriptions of different
types of acceptances can be found in the literature.
3.1 Present definitions of driver acceptance
3.1.1 Five different ways of defining acceptance
As the analysis of the literature review presented in paper IV shows, the acceptance
definitions identified in the literature can be classified into five categories. The
first category uses the word “accept” to define acceptance. The second category is
concerned with the needs and requirements of users (and other stakeholders). This
may be interpreted as the usefulness of the system. The third category of definition
sees acceptance as the sum of all attitudes, implying that other, for example, more
emotionally formed attitudes are added to the more “rational” evaluation of the
usefulness of the system (as in category 2). The fourth category focuses on the will
to use the system. This definition of acceptance aims for a behavioural change and
may be seen as being based on the earlier categories, in that the will to use a system
is based on drivers’ assessment of the usefulness of the system (as in category 2) as
well as all other attitudes to the system and its effects (as in category 3). This
fourth category stresses the will to act as a consequence. The fifth category of
acceptance emphasizes the actual use of the system, which presumably is
influenced by the will to use it (as in category 4).
Viewing the categories like this, they may to some extent be seen as a progression
from assessing the usefulness of a system towards the actual use of that system, the
later categories including the earlier ones, see Figure 3. This progression
perspective, however, cannot include category 1, which uses the word “accept” to
define acceptance, but does not provide any information about what is implied by
acceptance or accept.
3 What is acceptance?
27
Despite the recognized importance of acceptance, the understanding of acceptance
of driver support systems is often taken for granted. While many studies claim to
have measured acceptance, few have defined what it is. As Regan et al. (2002) put
it: “While everyone seems to know what acceptability is, and all agree that
acceptability is important, there is no consistency across studies as to what
‘acceptability’ is and how to measure it”. Although there is no common definition
of acceptance, some definitions of acceptance as well as descriptions of different
types of acceptances can be found in the literature.
3.1 Present definitions of driver acceptance
3.1.1 Five different ways of defining acceptance
As the analysis of the literature review presented in paper IV shows, the acceptance
definitions identified in the literature can be classified into five categories. The
first category uses the word “accept” to define acceptance. The second category is
concerned with the needs and requirements of users (and other stakeholders). This
may be interpreted as the usefulness of the system. The third category of definition
sees acceptance as the sum of all attitudes, implying that other, for example, more
emotionally formed attitudes are added to the more “rational” evaluation of the
usefulness of the system (as in category 2). The fourth category focuses on the will
to use the system. This definition of acceptance aims for a behavioural change and
may be seen as being based on the earlier categories, in that the will to use a system
is based on drivers’ assessment of the usefulness of the system (as in category 2) as
well as all other attitudes to the system and its effects (as in category 3). This
fourth category stresses the will to act as a consequence. The fifth category of
acceptance emphasizes the actual use of the system, which presumably is
influenced by the will to use it (as in category 4).
Viewing the categories like this, they may to some extent be seen as a progression
from assessing the usefulness of a system towards the actual use of that system, the
later categories including the earlier ones, see Figure 3. This progression
perspective, however, cannot include category 1, which uses the word “accept” to
define acceptance, but does not provide any information about what is implied by
acceptance or accept.
Adell | Driver experiences and acceptance of driver support systems
28
Definition categories
1 2 3 4 5
Using the word
“accept”
Satisfying needs
and requirements
Sum of
attitudes
Willingness to
use
Actual use
Figure 3: The five categories of acceptance, based on definitions found in the literature
review.
3.1.2 Different dimensions of acceptance
There are also different types of acceptances described in the literature. Authors
have made distinctions between attitudinal and behavioural acceptance (Kollman,
2000; Franken, 2007), between social and practical acceptance (Nielsen, 1993)
and between different levels of problem awareness of the individual (Katteler,
2005).
Attitudinal acceptance is, according to Franken (2007), based on emotion and
experience and provides a basis for accepting a system. Behavioural acceptance is
displayed in the form of observable behaviour. Comparing to the definition
categories described above, attitudinal acceptance is comparable to ‘sum of
attitudes’ (category 3) and behavioural acceptance to the ‘actual use’ (category 5).
Similar to this, Kollmann (2000) describes acceptance as consisting of three levels:
the general connection of inner assessment and expectation (the attitude level), the
acquisition or purchase of the product (the action level) and its voluntary use with
a frequency greater than that of other traffic participants (the utilization level).
Social acceptability is described by an example (Nielsen, 1993): “Consider a
system to investigate whether people applying for unemployment benefits are
currently gainfully employed and thus have submitted fraudulent applications.
The system might do this by asking applicants a number of questions and
searching their answers for inconsistencies or profiles that are often indicative of
cheaters. Some people may consider such a fraud-preventing system highly socially
desirable, but others may find it offensive to subject applicants to this kind of
quizzing and socially undesirable to delay benefits for people fitting certain
profiles.” Comparably, a driver might find it socially unacceptable for a
government to impose a driver support system on a user, even if it results in
reduction in road trauma. Practical acceptability includes dimensions like cost,
compatibility, reliability, usefulness, etc.
Katteler (2005) defines different types of acceptances depending on the subject’s
awareness of the problem the support system is aimed at tackling. The well-
Adell | Driver experiences and acceptance of driver support systems
28
Definition categories
1 2 3 4 5
Using the word
“accept”
Satisfying needs
and requirements
Sum of
attitudes
Willingness to
use
Actual use
Figure 3: The five categories of acceptance, based on definitions found in the literature
review.
3.1.2 Different dimensions of acceptance
There are also different types of acceptances described in the literature. Authors
have made distinctions between attitudinal and behavioural acceptance (Kollman,
2000; Franken, 2007), between social and practical acceptance (Nielsen, 1993)
and between different levels of problem awareness of the individual (Katteler,
2005).
Attitudinal acceptance is, according to Franken (2007), based on emotion and
experience and provides a basis for accepting a system. Behavioural acceptance is
displayed in the form of observable behaviour. Comparing to the definition
categories described above, attitudinal acceptance is comparable to ‘sum of
attitudes’ (category 3) and behavioural acceptance to the ‘actual use’ (category 5).
Similar to this, Kollmann (2000) describes acceptance as consisting of three levels:
the general connection of inner assessment and expectation (the attitude level), the
acquisition or purchase of the product (the action level) and its voluntary use with
a frequency greater than that of other traffic participants (the utilization level).
Social acceptability is described by an example (Nielsen, 1993): “Consider a
system to investigate whether people applying for unemployment benefits are
currently gainfully employed and thus have submitted fraudulent applications.
The system might do this by asking applicants a number of questions and
searching their answers for inconsistencies or profiles that are often indicative of
cheaters. Some people may consider such a fraud-preventing system highly socially
desirable, but others may find it offensive to subject applicants to this kind of
quizzing and socially undesirable to delay benefits for people fitting certain
profiles.” Comparably, a driver might find it socially unacceptable for a
government to impose a driver support system on a user, even if it results in
reduction in road trauma. Practical acceptability includes dimensions like cost,
compatibility, reliability, usefulness, etc.
Katteler (2005) defines different types of acceptances depending on the subject’s
awareness of the problem the support system is aimed at tackling. The well-
3 What is acceptance?
29
founded, firm acceptance indicates, apart from a positive attitude towards the
system, that the individual is aware of the problem the system is designed to
tackle. The opportunistic acceptance indicates low problem awareness and is
likely, according to Katteler (2005), to be less stable and more sensitive to changes.
There is also discussion about ‘conditional’ and ‘contextual’ acceptance in the
literature. Conditional acceptance indicates that acceptance is dependent on
certain preconditions (Saad & Dionisio, 2007), e.g. “I will use the system if I am
free to turn it off when I want to” or “I will use the system if everybody else does”.
Similarly, contextual acceptance indicates that acceptance depends on the
situational context (Saad, 2004), e.g. “I will use the system on roads with speed
cameras” or “I won’t use the system in rush hour”.
Goldenbeld (2003) makes a distinction between acceptance and support, where
acceptance is the willingness to be subjected to something (e.g. pay taxes) while
support is the liking for doing so. Vlassenroot et al. (2006) further claims that
(public) support is a precondition for acceptance since it “defines the degree of
acceptance or intentions people have to adapt or not to adapt to the desired
behaviour”. The sum of the individuals’ acceptance indicates whether there is
public support, according to Vlassenroot and de Mol (2007). By the reasoning of
Vlassenroot et al. (2006) the willingness to do something has to be preceded by a
liking for doing it.
Some stress the importance of making a distinction between acceptability and
acceptance. Schade and Schlag (2003) define acceptability as the “prospective
judgement of measures to be introduced in the future”. Acceptability is measured
when the subject has no experiences of the system, and is therefore an attitude
construct. Acceptance, on the other hand, consists of attitudes including
behavioural reactions after the introduction of a measure. Pianelli et al. (2007)
differentiate between two types of acceptability: priori and posteriori acceptability.
Priori acceptability is acceptability without experience of the system while
posteriori acceptability is the acceptability after having tried the system. The
posteriori acceptability includes experiences of the system, but does not necessarily
include behavioural reactions, making it different from the acceptance described
by Schade and Schlag (2003).
The literature also contains some statements on the purpose of investigating
acceptance. Najm et al. (2006) claim that “driver acceptance is the precondition
that will permit new automotive technologies to achieve their forecasted benefit
levels” and that there is a need to determine whether drivers will accept and use
the new technologies as intended. Further, Najm et al. (2006) state that driver
acceptance provides a means to estimate drivers’ interest in purchasing and using
3 What is acceptance?
29
founded, firm acceptance indicates, apart from a positive attitude towards the
system, that the individual is aware of the problem the system is designed to
tackle. The opportunistic acceptance indicates low problem awareness and is
likely, according to Katteler (2005), to be less stable and more sensitive to changes.
There is also discussion about ‘conditional’ and ‘contextual’ acceptance in the
literature. Conditional acceptance indicates that acceptance is dependent on
certain preconditions (Saad & Dionisio, 2007), e.g. “I will use the system if I am
free to turn it off when I want to” or “I will use the system if everybody else does”.
Similarly, contextual acceptance indicates that acceptance depends on the
situational context (Saad, 2004), e.g. “I will use the system on roads with speed
cameras” or “I won’t use the system in rush hour”.
Goldenbeld (2003) makes a distinction between acceptance and support, where
acceptance is the willingness to be subjected to something (e.g. pay taxes) while
support is the liking for doing so. Vlassenroot et al. (2006) further claims that
(public) support is a precondition for acceptance since it “defines the degree of
acceptance or intentions people have to adapt or not to adapt to the desired
behaviour”. The sum of the individuals’ acceptance indicates whether there is
public support, according to Vlassenroot and de Mol (2007). By the reasoning of
Vlassenroot et al. (2006) the willingness to do something has to be preceded by a
liking for doing it.
Some stress the importance of making a distinction between acceptability and
acceptance. Schade and Schlag (2003) define acceptability as the “prospective
judgement of measures to be introduced in the future”. Acceptability is measured
when the subject has no experiences of the system, and is therefore an attitude
construct. Acceptance, on the other hand, consists of attitudes including
behavioural reactions after the introduction of a measure. Pianelli et al. (2007)
differentiate between two types of acceptability: priori and posteriori acceptability.
Priori acceptability is acceptability without experience of the system while
posteriori acceptability is the acceptability after having tried the system. The
posteriori acceptability includes experiences of the system, but does not necessarily
include behavioural reactions, making it different from the acceptance described
by Schade and Schlag (2003).
The literature also contains some statements on the purpose of investigating
acceptance. Najm et al. (2006) claim that “driver acceptance is the precondition
that will permit new automotive technologies to achieve their forecasted benefit
levels” and that there is a need to determine whether drivers will accept and use
the new technologies as intended. Further, Najm et al. (2006) state that driver
acceptance provides a means to estimate drivers’ interest in purchasing and using
Adell | Driver experiences and acceptance of driver support systems
30
new technology. “Driver acceptance measurement also provides a means to
estimate drivers’ interest in purchasing and using new technologies as a basis for
estimating the safety benefit associated with its use.” (Najm et al., 2006). Van der
Laan et al. (1997) also see acceptance as the link to usage, thereby materializing
the potential safety effects, whereas van Driel (2007) sees acceptance as a predictor
of the willingness to buy a system.
3.2 Proposal for a new definition of acceptance
Working on driver acceptance makes it essential to understand the importance of
a driver-centred view, as it is the driver who makes the decision to use or not use a
system. Since acceptance is individual, it can only be based on an individual’s
personal attitudes, expectations, experiences and subjective evaluation of the
system and the effects of using it (Schade & Baum, 2007). The effects of the
system (e.g. reduction in accident risk) can only influence acceptance if they are
known, understood and believed in by the driver. A misunderstanding of the
system will influence acceptance as much as a correct conception.
It is also important to remember that focus should not be on the innovation per se
or on political/policy goals. It has been recognized that, to achieve the acceptance
and use of new systems, personal importance to the users has to be valued more
highly than degree of innovation (see e.g. Ausserer & Risser, 2005). However,
policies and political goals are often confused with the driver’s personal goals.
Societal goals and individuals goals do not necessarily coincide. For example, the
policy goal behind ISA could be to increase traffic safety or to increase speed limit
compliance. These goals might not be relevant to some drivers e.g. due to their
feeling that safety measures are redundant because of their own personal driving
skills (Brookhuis & Brown, 1992), or that speeding is not seen as a “real crime”
(Corbett, 2001). Nevertheless, they might find that the system helps them to
avoid speeding tickets or that they have an interest in innovative systems.
The multidimensional definition of acceptance proposed by Katteler (2005) is
interesting and offers new dimensions to the level of acceptance. Katteler has
studied ISA and uses awareness of the speed problem as the “problem awareness”
dimension. However, this might not be the “problem” for which drivers wish to
use the system. Similar systems are marketed as a problem-solver for speeding
tickets.
The same argument can be raised against the approach chosen by Vlassenroot et
al. (2006). The authors examine ‘public support’ for ISA since it “defines the
degree of acceptance or intentions people have to adapt or not to adapt to the
desired behaviour”. In this approach drivers have to agree that high speeds are a
Adell | Driver experiences and acceptance of driver support systems
30
new technology. “Driver acceptance measurement also provides a means to
estimate drivers’ interest in purchasing and using new technologies as a basis for
estimating the safety benefit associated with its use.” (Najm et al., 2006). Van der
Laan et al. (1997) also see acceptance as the link to usage, thereby materializing
the potential safety effects, whereas van Driel (2007) sees acceptance as a predictor
of the willingness to buy a system.
3.2 Proposal for a new definition of acceptance
Working on driver acceptance makes it essential to understand the importance of
a driver-centred view, as it is the driver who makes the decision to use or not use a
system. Since acceptance is individual, it can only be based on an individual’s
personal attitudes, expectations, experiences and subjective evaluation of the
system and the effects of using it (Schade & Baum, 2007). The effects of the
system (e.g. reduction in accident risk) can only influence acceptance if they are
known, understood and believed in by the driver. A misunderstanding of the
system will influence acceptance as much as a correct conception.
It is also important to remember that focus should not be on the innovation per se
or on political/policy goals. It has been recognized that, to achieve the acceptance
and use of new systems, personal importance to the users has to be valued more
highly than degree of innovation (see e.g. Ausserer & Risser, 2005). However,
policies and political goals are often confused with the driver’s personal goals.
Societal goals and individuals goals do not necessarily coincide. For example, the
policy goal behind ISA could be to increase traffic safety or to increase speed limit
compliance. These goals might not be relevant to some drivers e.g. due to their
feeling that safety measures are redundant because of their own personal driving
skills (Brookhuis & Brown, 1992), or that speeding is not seen as a “real crime”
(Corbett, 2001). Nevertheless, they might find that the system helps them to
avoid speeding tickets or that they have an interest in innovative systems.
The multidimensional definition of acceptance proposed by Katteler (2005) is
interesting and offers new dimensions to the level of acceptance. Katteler has
studied ISA and uses awareness of the speed problem as the “problem awareness”
dimension. However, this might not be the “problem” for which drivers wish to
use the system. Similar systems are marketed as a problem-solver for speeding
tickets.
The same argument can be raised against the approach chosen by Vlassenroot et
al. (2006). The authors examine ‘public support’ for ISA since it “defines the
degree of acceptance or intentions people have to adapt or not to adapt to the
desired behaviour”. In this approach drivers have to agree that high speeds are a
3 What is acceptance?
31
problem and that ISA is a good way of reducing them to demonstrate public
support (Vlassenroot et al., 2006). It is true that these drivers will most likely
demonstrate a high acceptance of ISA, but it is also possible that drivers who do
not show support for the system on these terms may accept the system. If this were
not the case, it would also raise the question of what traffic safety benefits a system
like ISA would have if only used by drivers who already agree that high speeds are
a problem and that ISA is a good way of dealing with them, see e.g. Hjälmdahl
(2004) and Jamson (2006).
The use of the system is vital in striving to improve traffic safety by deploying
driver support systems. It is the use of the system that will materialise its potential
and hopefully produce benefits for the driver and the society. Neither attitudinal
acceptance (Franken, 2007) nor support (Goldenbeld, 2003) requires any impact
on the actual use of a system. Hence, the main aim and focus should be on
behavioural acceptance (Franken, 2007), the utilization level as described by
Kollmann (2000) or the acceptance definition category 5 – actual use (described
above in Chapter 3.1.1), which emphasizes the use of the system. From this
perspective, the second and third categories of acceptance definitions (usefulness
and all attitudes), attitudinal acceptance (Franken, 2007) and the attitude level
described by Kollmann (2000) influence the will to use and the actual usage, but
are not to be seen as acceptance.
Obviously, it is not possible to assess the use of systems that are in the design
phase. It is desirable, though, to accurately predict user acceptance as early as
possible in the design process to be able to evaluate different alternatives and
identify obstacles to overcome. Further, also for systems that are available to
drivers, the use of the system has to be seen as a part of a process including the will
to use as a step towards usage.
The following definition of acceptance is proposed:
Acceptance is the degree to which an individual intends to use a system
and, when available, to incorporate the system in his/her driving
.
This definition has the advantages of considering the intention to use (which can
be used early in the development phase) and of emphasising the importance of
manifesting the intention in actual behaviour. Further, it also stresses the
importance of focusing on the individual who makes the decision to use or not use
a system and his/her understanding of the system. This definition clearly states
that the level of acceptance is not believed to be limited to acceptance/non
acceptance (nominal scale), but to be of a more continuous nature.
3 What is acceptance?
31
problem and that ISA is a good way of reducing them to demonstrate public
support (Vlassenroot et al., 2006). It is true that these drivers will most likely
demonstrate a high acceptance of ISA, but it is also possible that drivers who do
not show support for the system on these terms may accept the system. If this were
not the case, it would also raise the question of what traffic safety benefits a system
like ISA would have if only used by drivers who already agree that high speeds are
a problem and that ISA is a good way of dealing with them, see e.g. Hjälmdahl
(2004) and Jamson (2006).
The use of the system is vital in striving to improve traffic safety by deploying
driver support systems. It is the use of the system that will materialise its potential
and hopefully produce benefits for the driver and the society. Neither attitudinal
acceptance (Franken, 2007) nor support (Goldenbeld, 2003) requires any impact
on the actual use of a system. Hence, the main aim and focus should be on
behavioural acceptance (Franken, 2007), the utilization level as described by
Kollmann (2000) or the acceptance definition category 5 – actual use (described
above in Chapter 3.1.1), which emphasizes the use of the system. From this
perspective, the second and third categories of acceptance definitions (usefulness
and all attitudes), attitudinal acceptance (Franken, 2007) and the attitude level
described by Kollmann (2000) influence the will to use and the actual usage, but
are not to be seen as acceptance.
Obviously, it is not possible to assess the use of systems that are in the design
phase. It is desirable, though, to accurately predict user acceptance as early as
possible in the design process to be able to evaluate different alternatives and
identify obstacles to overcome. Further, also for systems that are available to
drivers, the use of the system has to be seen as a part of a process including the will
to use as a step towards usage.
The following definition of acceptance is proposed:
Acceptance is the degree to which an individual intends to use a system
and, when available, to incorporate the system in his/her driving
.
This definition has the advantages of considering the intention to use (which can
be used early in the development phase) and of emphasising the importance of
manifesting the intention in actual behaviour. Further, it also stresses the
importance of focusing on the individual who makes the decision to use or not use
a system and his/her understanding of the system. This definition clearly states
that the level of acceptance is not believed to be limited to acceptance/non
acceptance (nominal scale), but to be of a more continuous nature.
Adell | Driver experiences and acceptance of driver support systems
32
By this definition it follows that the driver does not necessarily have to like to use
the system to demonstrate acceptance. To show high acceptance it is enough that
the driver decides to use the system, which, under the given circumstances, is seen
as the best option. In this way, tolerating the use of the system can be seen as part
of acceptance. For example: a driver who normally would not choose to use an
ISA system decides to use the system due to the amount of his speeding fines, or a
driver agrees to use the system since it is required by law. Of course, the degree of
acceptance could also be zero when the driver does not use the system and has no
intention to do so.
3.3 Assessing acceptance
Considering the different definitions of acceptance, it is hardly surprising that
there are almost as many ways to measure acceptance as there are researchers trying
to measure it. Besides, the definition and meaning of ‘acceptance’ are usually taken
for granted in ITS research, and most researchers assess acceptance without
defining it. Below is a summary of the different ways to assess acceptance found in
the literature review in paper IV.
3.3.1 Different ways to assess acceptance
The numerous ways of assessing acceptance found in the literature review are
summarized into 9 groups (with 22 subgroups). Most authors use more than one
measurement to assess acceptance, either from the same or from different
measurement groups. Usually, the measurements used in the various studies are
based on questionnaires, but there are also measurements based on interviews,
focus groups, logged data and physiological measures. There are also acceptance
studies with results concerning acceptance, but without describing how the results
are obtained. For more information about the subgroups and references see paper
IV.
xUsing the word “accept/acceptable
This seems to be a quite common way to assess acceptance. Usually
measurements in this category use phrases like “would you accept…?” or
“how acceptable is… ?”. Another way of measuring acceptance included in
this group is the assessment of the “willingness to accept” something.
xUsefulness and satisfaction
The most used method is the usefulness/satisfaction scale developed by van
der Laan et al. (1997). This is a standardised procedure to estimate the
usefulness of and satisfaction with a system. The driver assesses the system in
question by means of nine five-point rating-scale items (useful – useless,
Adell | Driver experiences and acceptance of driver support systems
32
By this definition it follows that the driver does not necessarily have to like to use
the system to demonstrate acceptance. To show high acceptance it is enough that
the driver decides to use the system, which, under the given circumstances, is seen
as the best option. In this way, tolerating the use of the system can be seen as part
of acceptance. For example: a driver who normally would not choose to use an
ISA system decides to use the system due to the amount of his speeding fines, or a
driver agrees to use the system since it is required by law. Of course, the degree of
acceptance could also be zero when the driver does not use the system and has no
intention to do so.
3.3 Assessing acceptance
Considering the different definitions of acceptance, it is hardly surprising that
there are almost as many ways to measure acceptance as there are researchers trying
to measure it. Besides, the definition and meaning of ‘acceptance’ are usually taken
for granted in ITS research, and most researchers assess acceptance without
defining it. Below is a summary of the different ways to assess acceptance found in
the literature review in paper IV.
3.3.1 Different ways to assess acceptance
The numerous ways of assessing acceptance found in the literature review are
summarized into 9 groups (with 22 subgroups). Most authors use more than one
measurement to assess acceptance, either from the same or from different
measurement groups. Usually, the measurements used in the various studies are
based on questionnaires, but there are also measurements based on interviews,
focus groups, logged data and physiological measures. There are also acceptance
studies with results concerning acceptance, but without describing how the results
are obtained. For more information about the subgroups and references see paper
IV.
xUsing the word “accept/acceptable
This seems to be a quite common way to assess acceptance. Usually
measurements in this category use phrases like “would you accept…?” or
“how acceptable is… ?”. Another way of measuring acceptance included in
this group is the assessment of the “willingness to accept” something.
xUsefulness and satisfaction
The most used method is the usefulness/satisfaction scale developed by van
der Laan et al. (1997). This is a standardised procedure to estimate the
usefulness of and satisfaction with a system. The driver assesses the system in
question by means of nine five-point rating-scale items (useful – useless,
3 What is acceptance?
33
pleasant – unpleasant, bad – good, nice – annoying, effective – superfluous,
irritating – likeable, assisting – worthless, undesirable – desirable and raising
alertness – sleep-inducing). These bipolar scales are then combined into one
usefulness score and one satisfaction score for the system. This method has
been used in papers II and III.
Many studies use a variety of usefulness measurements as measures of
acceptance, e.g. whether the system facilitates the driving task, affects the
driving performance of oneself and/or affects the driving performance of
others. This group also includes opinions about the effectiveness of the system
and information about what kind of instructions/corrections one wants to
receive from the system.
Some studies also use other measurements to assess satisfaction as a measure
of acceptance. Examples from this group are a general assessment of
satisfaction with the system, opinions on whether having the system is an
advantage or disadvantage, the attractiveness/unattractiveness of the system,
whether the system is disturbing or annoying and whether it is supportive or
constructive.
xWillingness to submit to something
Quite a few studies use the willingness to subject to something as a measure of
acceptance, and investigate the willingness to pay, either by posing an open
question or a closed one with different price intervals. Further, willingness to
buy, accept, have, keep, use and install, as well as the wish to shut down the
system are indicators assigned to this group. This type of measurement is used
in papers II and III where the drivers are asked whether they want to have,
keep and pay for the system.
xUse
One group of measurements, which focuses on the use of the system, includes
measurements of voluntary use, frequency of use and the action of shutting
down the system. This group is sometimes measured by observed behaviour.
In paper I, the drivers state how often they overrode the system. This may be
seen as a (indirect) measure of use. However, if it is depends on whether one
sees overriding the system as not using the system or as utilizing a certain
feature of the system.
3 What is acceptance?
33
pleasant – unpleasant, bad – good, nice – annoying, effective – superfluous,
irritating – likeable, assisting – worthless, undesirable – desirable and raising
alertness – sleep-inducing). These bipolar scales are then combined into one
usefulness score and one satisfaction score for the system. This method has
been used in papers II and III.
Many studies use a variety of usefulness measurements as measures of
acceptance, e.g. whether the system facilitates the driving task, affects the
driving performance of oneself and/or affects the driving performance of
others. This group also includes opinions about the effectiveness of the system
and information about what kind of instructions/corrections one wants to
receive from the system.
Some studies also use other measurements to assess satisfaction as a measure
of acceptance. Examples from this group are a general assessment of
satisfaction with the system, opinions on whether having the system is an
advantage or disadvantage, the attractiveness/unattractiveness of the system,
whether the system is disturbing or annoying and whether it is supportive or
constructive.
xWillingness to submit to something
Quite a few studies use the willingness to subject to something as a measure of
acceptance, and investigate the willingness to pay, either by posing an open
question or a closed one with different price intervals. Further, willingness to
buy, accept, have, keep, use and install, as well as the wish to shut down the
system are indicators assigned to this group. This type of measurement is used
in papers II and III where the drivers are asked whether they want to have,
keep and pay for the system.
xUse
One group of measurements, which focuses on the use of the system, includes
measurements of voluntary use, frequency of use and the action of shutting
down the system. This group is sometimes measured by observed behaviour.
In paper I, the drivers state how often they overrode the system. This may be
seen as a (indirect) measure of use. However, if it is depends on whether one
sees overriding the system as not using the system or as utilizing a certain
feature of the system.
Adell | Driver experiences and acceptance of driver support systems
34
xGeneral assessments
Many researchers also use a general assessment to measure acceptance.
Examples in this group would be to judge the concept/idea of a system, assess
the popularity of the system, whether the driver is in favour of the system and
whether the user would recommend or appreciate it if loved ones used the
system. This type of measurement is used in papers II and III where the drivers
are asked to rate the concept of the system.
xImportance of the system
Some researchers assess the importance of the system as an indicator of
acceptance, e.g. ranking the importance of the system compared to other
systems/measurements or a judgement of its necessity. Whether the driver
supports an implementation is also included in this group.
xReliability of the system
A few studies measure drivers’ level of trust in the system or the credibility of
the system (unclear whether the authors refer to their own judgement of the
system or to the drivers’ opinion about the system).
xHMI assessments
Some studies use assessment of the HMI (Human-Machine-Interaction) as a
measurement for acceptance e.g. about the timing of the information given to
the drivers, if they were startled by the information, if they could identify the
source of the alert and if they wanted to modify the intensity of the feedback.
xPhysiological reactions
One study also uses physiological reactions interpreted as stress to measure
acceptance. This is done by measuring the heart rate of the driver.
3.3.2 Acceptance measures in relation to the definitions of acceptance
Most of the measurements used can be related to the definition categories of
acceptance based on findings in the literature, see above in chapter 3.1. In this
respect, the assessment using the word “accept” clearly relates to the first definition
category but does not provide any further information about the meaning of
acceptance. The usefulness/satisfaction scale proposed by van der Laan et al.
(1997) reflects the third category of definition – the sum of attitudes. The
usefulness measurement may be associated with the second definition (needs and
requirements), and is a subset of the third definition category, as is the satisfaction
assessment. The willingness to subject to something partly reflects the fourth
category of definition – the willingness to use. However, the willingness to pay,
Adell | Driver experiences and acceptance of driver support systems
34
xGeneral assessments
Many researchers also use a general assessment to measure acceptance.
Examples in this group would be to judge the concept/idea of a system, assess
the popularity of the system, whether the driver is in favour of the system and
whether the user would recommend or appreciate it if loved ones used the
system. This type of measurement is used in papers II and III where the drivers
are asked to rate the concept of the system.
xImportance of the system
Some researchers assess the importance of the system as an indicator of
acceptance, e.g. ranking the importance of the system compared to other
systems/measurements or a judgement of its necessity. Whether the driver
supports an implementation is also included in this group.
xReliability of the system
A few studies measure drivers’ level of trust in the system or the credibility of
the system (unclear whether the authors refer to their own judgement of the
system or to the drivers’ opinion about the system).
xHMI assessments
Some studies use assessment of the HMI (Human-Machine-Interaction) as a
measurement for acceptance e.g. about the timing of the information given to
the drivers, if they were startled by the information, if they could identify the
source of the alert and if they wanted to modify the intensity of the feedback.
xPhysiological reactions
One study also uses physiological reactions interpreted as stress to measure
acceptance. This is done by measuring the heart rate of the driver.
3.3.2 Acceptance measures in relation to the definitions of acceptance
Most of the measurements used can be related to the definition categories of
acceptance based on findings in the literature, see above in chapter 3.1. In this
respect, the assessment using the word “accept” clearly relates to the first definition
category but does not provide any further information about the meaning of
acceptance. The usefulness/satisfaction scale proposed by van der Laan et al.
(1997) reflects the third category of definition – the sum of attitudes. The
usefulness measurement may be associated with the second definition (needs and
requirements), and is a subset of the third definition category, as is the satisfaction
assessment. The willingness to subject to something partly reflects the fourth
category of definition – the willingness to use. However, the willingness to pay,
3 What is acceptance?
35
buy, accept, have, keep and install are also used as assessments of acceptance
within this group. The other measurements do not clearly belong to any of the
definition types, but one can assume that “have”, “keep” and “install” aim at
estimating the willingness to use. Some measurements aim at assessing the use of
the system, which is the fifth category. The general assessment, importance and
reliability of the system do not clearly reflect any of the definitions. However,
these measurements indicate, to some degree, the attitude towards the system
(category 3) and the usefulness of the system (category 2). Opinions about HMI
and physiological reactions of stress do not reflect any of the definition categories.
3.3.3 Limitations of current ways of measuring acceptance
How to measure acceptance in a valid way does depend on how acceptance is
defined. It is not surprising therefore that the weak common ground regarding
acceptance definition has resulted in a large number of different attempts to
measure acceptance. The large differences in the measures used indicate quite a
large discrepancy in the understanding of acceptance as well as in what are
believed to be important and valid indicators of acceptance.
The many different ways of assessing acceptance may cause confusion and lead to
incorrect conclusions or interpretations. Illustrations of problems with the current
ways of measuring acceptance are found in paper II and paper III.
One interpretation of the different results is that the measurements used do not
measure the same kind of acceptance. The concept of the system and the
usefulness/satisfaction scale relate to definition category 3 (sum of all attitudes)
while the willingness to keep relates to category 4 (willingness to use). However,
even if more measurements assessing the same kind of acceptance were used, there
would be no guarantee that the results would concur, since validations of the
measurements used are virtually non-existent. Most researchers define acceptance
implicitly by the measurements they use to assess it, making validation
superfluous.
The present situation is troublesome. If acceptance has not been defined, then we
cannot be sure that the tool we use to measure it will give valid results. Without
knowing how acceptance is defined, it is impossible to understand how drivers’
experiences influence it. The inconstancy of acceptance definitions (implicitly
defined or not), and of measurements and thereby the diversity of results, even
though collected in the same experiment, present a breeding ground for
misinterpretations and misuse of the results. What is more, it makes comparisons
between systems and settings almost impossible.
3 What is acceptance?
35
buy, accept, have, keep and install are also used as assessments of acceptance
within this group. The other measurements do not clearly belong to any of the
definition types, but one can assume that “have”, “keep” and “install” aim at
estimating the willingness to use. Some measurements aim at assessing the use of
the system, which is the fifth category. The general assessment, importance and
reliability of the system do not clearly reflect any of the definitions. However,
these measurements indicate, to some degree, the attitude towards the system
(category 3) and the usefulness of the system (category 2). Opinions about HMI
and physiological reactions of stress do not reflect any of the definition categories.
3.3.3 Limitations of current ways of measuring acceptance
How to measure acceptance in a valid way does depend on how acceptance is
defined. It is not surprising therefore that the weak common ground regarding
acceptance definition has resulted in a large number of different attempts to
measure acceptance. The large differences in the measures used indicate quite a
large discrepancy in the understanding of acceptance as well as in what are
believed to be important and valid indicators of acceptance.
The many different ways of assessing acceptance may cause confusion and lead to
incorrect conclusions or interpretations. Illustrations of problems with the current
ways of measuring acceptance are found in paper II and paper III.
One interpretation of the different results is that the measurements used do not
measure the same kind of acceptance. The concept of the system and the
usefulness/satisfaction scale relate to definition category 3 (sum of all attitudes)
while the willingness to keep relates to category 4 (willingness to use). However,
even if more measurements assessing the same kind of acceptance were used, there
would be no guarantee that the results would concur, since validations of the
measurements used are virtually non-existent. Most researchers define acceptance
implicitly by the measurements they use to assess it, making validation
superfluous.
The present situation is troublesome. If acceptance has not been defined, then we
cannot be sure that the tool we use to measure it will give valid results. Without
knowing how acceptance is defined, it is impossible to understand how drivers’
experiences influence it. The inconstancy of acceptance definitions (implicitly
defined or not), and of measurements and thereby the diversity of results, even
though collected in the same experiment, present a breeding ground for
misinterpretations and misuse of the results. What is more, it makes comparisons
between systems and settings almost impossible.
Adell | Driver experiences and acceptance of driver support systems
36
Adell | Driver experiences and acceptance of driver support systems
36
37
4 Acceptance Models
“Researchers are confronted with a choice among a multitude
of models and find that they must “pick and choose” constructs
across the models, or choose a “favoured model” and largely
ignore the contributions from alternative models. Thus, there is
a need for a review and synthesis in order to progress toward a
unified view of user acceptance.”
Venkatesh et al. (2003)
37
4 Acceptance Models
“Researchers are confronted with a choice among a multitude
of models and find that they must “pick and choose” constructs
across the models, or choose a “favoured model” and largely
ignore the contributions from alternative models. Thus, there is
a need for a review and synthesis in order to progress toward a
unified view of user acceptance.”
Venkatesh et al. (2003)
Adell | Driver experiences and acceptance of driver support systems
38
Adell | Driver experiences and acceptance of driver support systems
38
4 Acceptance Models
39
A definition of acceptance has now been proposed. However, for an
understanding of how acceptance is formed and what factors influence it, a model
is a useful tool.
An important component in this definition of acceptance is behaviour (use). The
most frequently used model to explain behaviour in the area of traffic safety is the
Theory of Planned Behaviour (Ajzen, 1991). According to this theory, behaviour
is determined by people’s attitude towards the behaviour, their subjective norm
and their perceived behavioural control via their intention to perform the
behaviour. In the context of driver support systems, the Theory of Planned
Behaviour has been used to study e.g. violations of speed limits (Wallén Warner &
Åberg, 2006 and Wallén Warner & Åberg, 2008), but the model has not been
used to study the use of a system per se.
4.1 Most used frameworks for acceptance of driver support systems
There are a few frameworks within which one may understand acceptance in the
driver support system literature. The National Highway Traffic Safety
Administration (NHTSA) strategic plan, 1997-2002, stated that driver acceptance
should be understood in terms of ease of use, ease of learning, adaptation and
perception (Najm et al. 2006). These aspects of driver support systems should
show whether the system satisfies the needs and requirements of the drivers (our
second definition category, chapter 3.1). In 2001, the framework was revised to
include ease of use, ease of learning, perceived value, driving performance and
advocacy of the system or willingness to endorse it (Sterns et al., 2002 and Najm
et al., 2006). Regan et al. (2002) state that acceptability is dependent on
usefulness, ease of use, effectiveness, affordability and social acceptability (this is
also how the authors define acceptance). When studying ISA, Molin and
Brookhuis (2007) showed, by means of a Structural Equation Model (SEM), that
acceptability of the system was related to “belief that speed causes accidents”,
whether the system can “contribute to personal or societal goals” and “if one
prefers an ever limiting ISA”. The authors did not define acceptance; nevertheless,
it was measured by the indicators “intention to buy ISA if it is for free”, “wants to
possess ISA” and “support for policy to impose ISA on all cars”. These indicators
do not clearly fit into any of the five definition categories described in chapter 3.1.
However, the first two indicators suggest an acceptance definition in category 4
4 Acceptance Models
39
A definition of acceptance has now been proposed. However, for an
understanding of how acceptance is formed and what factors influence it, a model
is a useful tool.
An important component in this definition of acceptance is behaviour (use). The
most frequently used model to explain behaviour in the area of traffic safety is the
Theory of Planned Behaviour (Ajzen, 1991). According to this theory, behaviour
is determined by people’s attitude towards the behaviour, their subjective norm
and their perceived behavioural control via their intention to perform the
behaviour. In the context of driver support systems, the Theory of Planned
Behaviour has been used to study e.g. violations of speed limits (Wallén Warner &
Åberg, 2006 and Wallén Warner & Åberg, 2008), but the model has not been
used to study the use of a system per se.
4.1 Most used frameworks for acceptance of driver support systems
There are a few frameworks within which one may understand acceptance in the
driver support system literature. The National Highway Traffic Safety
Administration (NHTSA) strategic plan, 1997-2002, stated that driver acceptance
should be understood in terms of ease of use, ease of learning, adaptation and
perception (Najm et al. 2006). These aspects of driver support systems should
show whether the system satisfies the needs and requirements of the drivers (our
second definition category, chapter 3.1). In 2001, the framework was revised to
include ease of use, ease of learning, perceived value, driving performance and
advocacy of the system or willingness to endorse it (Sterns et al., 2002 and Najm
et al., 2006). Regan et al. (2002) state that acceptability is dependent on
usefulness, ease of use, effectiveness, affordability and social acceptability (this is
also how the authors define acceptance). When studying ISA, Molin and
Brookhuis (2007) showed, by means of a Structural Equation Model (SEM), that
acceptability of the system was related to “belief that speed causes accidents”,
whether the system can “contribute to personal or societal goals” and “if one
prefers an ever limiting ISA”. The authors did not define acceptance; nevertheless,
it was measured by the indicators “intention to buy ISA if it is for free”, “wants to
possess ISA” and “support for policy to impose ISA on all cars”. These indicators
do not clearly fit into any of the five definition categories described in chapter 3.1.
However, the first two indicators suggest an acceptance definition in category 4
Adell | Driver experiences and acceptance of driver support systems
40
(willingness to use) and the third indicator might be connected to categories 2 or 3
(needs and requirements or sum of attitudes) (compare the discussion regarding
the connection between measurements and definitions in chapter 3.3.2).
Neither Najm et al. (2006) nor Regan et al. (2002) have shown if and how these
aspects influence the acceptance of a system. This limits the use of these
frameworks for understanding how acceptance is formed and how to affect it. The
causal model regarding acceptance of ISA described by Molin and Brookhuis
(2007) is interesting, and points to the importance of the perceived usefulness of
the system, but it is too specialized to describe what stimulates acceptance in a
wider perspective. In conclusion, there is a need for a model to satisfactorily
describe what influences acceptance vis-à-vis driver support systems.
4.2 Acceptance models within the area of information technology
Following the rapid development of new technologies and software in computer
science, interest in the acceptance and use of these technologies has increased
significantly. A number of different models are used in the information technology
area, which today includes one of the most comprehensive research bodies on
acceptance and use of new technology. However, the scientist is faced with a
choice of numerous models, for example:
ǦThe Pleasure, Arousal and Dominance paradigm (Mehrabian & Russell, 1974)
ǦTheory of Reasoned Action (Ajzen & Fishbein, 1980)
ǦExpectation Disconfirmation theory (Oliver 1980)
ǦSocial Exchange Theory (Kelley, 1979, Emersson, 1987)
ǦTechnology Acceptance Model (TAM) (Davis, 1989)
ǦTheory of Planned Behaviour (TPB) (Ajzen, 1991)
Ǧthe Model of PC Utilization (Thompson et al., 1991)
ǦSocial Influence Model (Fulk, et al., 1990 and Fulk, 1993)
ǦMotivational Model (Davis, et al., 1992)
ǦA combined model of TAM and TPB (Taylor and Todd, 1995)
ǦSocial Cognitive Theory (Compeau & Higgins, 1995)
ǦInnovation Diffusion Theory (Rogers, 1995)
ǦTask technology fit (Goodhue & Thompson, 1995)
ǦSystem Implementation (Clegg, 2000)
ǦTechnology Readiness (Parasuraman, 2000)
ǦIS Continuance (Bhattachrjee, 2001)
ǦThree-Tier Use Model (Liaw et al., 2006)
ǦMotivation variable of LGO (Saadé, 2007)
ǦSocial Identity Theory (e.g. Yang et al., 2007)
Adell | Driver experiences and acceptance of driver support systems
40
(willingness to use) and the third indicator might be connected to categories 2 or 3
(needs and requirements or sum of attitudes) (compare the discussion regarding
the connection between measurements and definitions in chapter 3.3.2).
Neither Najm et al. (2006) nor Regan et al. (2002) have shown if and how these
aspects influence the acceptance of a system. This limits the use of these
frameworks for understanding how acceptance is formed and how to affect it. The
causal model regarding acceptance of ISA described by Molin and Brookhuis
(2007) is interesting, and points to the importance of the perceived usefulness of
the system, but it is too specialized to describe what stimulates acceptance in a
wider perspective. In conclusion, there is a need for a model to satisfactorily
describe what influences acceptance vis-à-vis driver support systems.
4.2 Acceptance models within the area of information technology
Following the rapid development of new technologies and software in computer
science, interest in the acceptance and use of these technologies has increased
significantly. A number of different models are used in the information technology
area, which today includes one of the most comprehensive research bodies on
acceptance and use of new technology. However, the scientist is faced with a
choice of numerous models, for example:
ǦThe Pleasure, Arousal and Dominance paradigm (Mehrabian & Russell, 1974)
ǦTheory of Reasoned Action (Ajzen & Fishbein, 1980)
ǦExpectation Disconfirmation theory (Oliver 1980)
ǦSocial Exchange Theory (Kelley, 1979, Emersson, 1987)
ǦTechnology Acceptance Model (TAM) (Davis, 1989)
ǦTheory of Planned Behaviour (TPB) (Ajzen, 1991)
Ǧthe Model of PC Utilization (Thompson et al., 1991)
ǦSocial Influence Model (Fulk, et al., 1990 and Fulk, 1993)
ǦMotivational Model (Davis, et al., 1992)
ǦA combined model of TAM and TPB (Taylor and Todd, 1995)
ǦSocial Cognitive Theory (Compeau & Higgins, 1995)
ǦInnovation Diffusion Theory (Rogers, 1995)
ǦTask technology fit (Goodhue & Thompson, 1995)
ǦSystem Implementation (Clegg, 2000)
ǦTechnology Readiness (Parasuraman, 2000)
ǦIS Continuance (Bhattachrjee, 2001)
ǦThree-Tier Use Model (Liaw et al., 2006)
ǦMotivation variable of LGO (Saadé, 2007)
ǦSocial Identity Theory (e.g. Yang et al., 2007)
4 Acceptance Models
41
In 2003, Venkatesh et al. integrated eight of the most significant models of
individual acceptance (in bold in the list above) into one comprehensive model.
The Unified Theory of Acceptance and Use of Technology (UTAUT) is based on
an extensive literature review and empirical comparison of the Theory of Reasoned
Action, the Technology Acceptance Model, the Theory of Planned Behaviour, a model
combining the Technology Acceptance Model and the Theory of Planned Behaviour,
the Model of PC Utilization, the Motivational Model, the Social Cognitive Theory
and the Innovation Diffusion Theory, including their extensions (for references see
Venkatesh et al., 2003). The key element in all these models is the behaviour, i.e.
use of the new technology. The model was validated for acceptance and use of
computer software by workers in the USA. The UTAUT model outperformed the
eight individual models, accounting for 70 percent of the variance (adjusted R2) in
use (Ventkatesh et al., 2003). The authors concluded that the promising results
indicate that the UTAUT is a useful tool for assessing the likelihood of success for
new technology introduction and provides knowledge of what stimulates
acceptance, which may be used to proactively design interventions (including
training, marketing, etc) targeted at populations of users that may be less inclined
to adopt and use new systems (Venkatesh et al., 2003).
4.3 The Unified Theory of Acceptance and Use of Technology
Venkatesh et al. (2003) postulate two direct determinants of use: ‘intention to use’
and ‘facilitating conditions’. ‘Intention to use’ is in turn influenced by
‘performance expectancy’, ‘effort expectancy’ and ‘social influence’. Gender, age,
experience and voluntariness of use act as moderators, see Figure 4.
Figure 4: The Unified Theory of Acceptance and Use of Technology and definitions of
the constructs (Venkatesh et al., 2003).
Behavioural
Intention
Use
behaviour
Performance expectancy
"thedegreetowhichanindividualbelievesthat
usingthesystemwillhelphim/hertoattaingains"
Effort expectancy
"thedegreeofeaseassociatedwiththeuseofthe
system"
Social influence
"thedegreetowhichanindividualperceivesthat
importantothersbelieveheorsheshouldusethe
newsystem”
Facilitating conditions
“thedegreetowhichanindividualbelievesthatan
organizationalandtechnicalinfrastructureexiststo
supportuseofthesystem”
Voluntariness
of use
Age Gender Experience
4 Acceptance Models
41
In 2003, Venkatesh et al. integrated eight of the most significant models of
individual acceptance (in bold in the list above) into one comprehensive model.
The Unified Theory of Acceptance and Use of Technology (UTAUT) is based on
an extensive literature review and empirical comparison of the Theory of Reasoned
Action, the Technology Acceptance Model, the Theory of Planned Behaviour, a model
combining the Technology Acceptance Model and the Theory of Planned Behaviour,
the Model of PC Utilization, the Motivational Model, the Social Cognitive Theory
and the Innovation Diffusion Theory, including their extensions (for references see
Venkatesh et al., 2003). The key element in all these models is the behaviour, i.e.
use of the new technology. The model was validated for acceptance and use of
computer software by workers in the USA. The UTAUT model outperformed the
eight individual models, accounting for 70 percent of the variance (adjusted R2) in
use (Ventkatesh et al., 2003). The authors concluded that the promising results
indicate that the UTAUT is a useful tool for assessing the likelihood of success for
new technology introduction and provides knowledge of what stimulates
acceptance, which may be used to proactively design interventions (including
training, marketing, etc) targeted at populations of users that may be less inclined
to adopt and use new systems (Venkatesh et al., 2003).
4.3 The Unified Theory of Acceptance and Use of Technology
Venkatesh et al. (2003) postulate two direct determinants of use: ‘intention to use’
and ‘facilitating conditions’. ‘Intention to use’ is in turn influenced by
‘performance expectancy’, ‘effort expectancy’ and ‘social influence’. Gender, age,
experience and voluntariness of use act as moderators, see Figure 4.
Figure 4: The Unified Theory of Acceptance and Use of Technology and definitions of
the constructs (Venkatesh et al., 2003).
Behavioural
Intention
Use
behaviour
Performance expectancy
"thedegreetowhichanindividualbelievesthat
usingthesystemwillhelphim/hertoattaingains"
Effort expectancy
"thedegreeofeaseassociatedwiththeuseofthe
system"
Social influence
"thedegreetowhichanindividualperceivesthat
importantothersbelieveheorsheshouldusethe
newsystem”
Facilitating conditions
“thedegreetowhichanindividualbelievesthatan
organizationalandtechnicalinfrastructureexiststo
supportuseofthesystem”
Voluntariness
of use
Age Gender Experience
Adell | Driver experiences and acceptance of driver support systems
42
The items used in assessing the constructs were also selected from the eight
investigated models. Through empirical evaluation, the four most significant items
for each construct were chosen as indicators for the constructs in the UTAUT
model, see Table 2. The intention to use was assessed through three items and use
was measured as duration of use via system logs (Venkatesh et al., 2003).
Table 2: The original UTAUT items used to assess the constructs (Answers were given
on a seven-graded scale from “strongly disagree” (1) to “strongly agree” (7)) (Venkatesh
et al., 2003).
Performance expectancy
I would find the system useful in my job
Using the system enables me to accomplish tasks more quickly
Using the system increases my productivity
If I use the system, I will increase my chances of getting a raise
Effort expectancy
My interaction with the system would be clear and understandable
It would be easy for me to become skilful at using the system
I would find the system easy to use
Learning to operate the system is easy for me
Social influence
People who influence my behaviour would think that I should use the system
People who are important to me would think that I should use the system
The senior management of this business has been helpful in the use of the system
In general, the organization has supported the use of the system
Facilitating conditions
I have the resources necessary to use the system
I have the knowledge necessary to use the system
The system is not compatible with other systems I use
A specific person (or group) is available for assistance with system difficulties
Intention to use
I intend to use the system in the next <n> months
I predict I would use the system in the next <n> months
I plan to use the system in the next <n> months
Venkatesh et al. (2003) found that ‘performance expectancy’ is a determinant of
‘intention to use’ in most situations. The strength of the relationship, however, is
moderated by age and gender, being more significant for men and younger
workers. The effect of ‘effort expectancy’ is also moderated by gender and age,
being more significant for women and older workers. This effect decreases with
experience. The effect of ‘social influence’ on intention to use is conditioned by
Adell | Driver experiences and acceptance of driver support systems
42
The items used in assessing the constructs were also selected from the eight
investigated models. Through empirical evaluation, the four most significant items
for each construct were chosen as indicators for the constructs in the UTAUT
model, see Table 2. The intention to use was assessed through three items and use
was measured as duration of use via system logs (Venkatesh et al., 2003).
Table 2: The original UTAUT items used to assess the constructs (Answers were given
on a seven-graded scale from “strongly disagree” (1) to “strongly agree” (7)) (Venkatesh
et al., 2003).
Performance expectancy
I would find the system useful in my job
Using the system enables me to accomplish tasks more quickly
Using the system increases my productivity
If I use the system, I will increase my chances of getting a raise
Effort expectancy
My interaction with the system would be clear and understandable
It would be easy for me to become skilful at using the system
I would find the system easy to use
Learning to operate the system is easy for me
Social influence
People who influence my behaviour would think that I should use the system
People who are important to me would think that I should use the system
The senior management of this business has been helpful in the use of the system
In general, the organization has supported the use of the system
Facilitating conditions
I have the resources necessary to use the system
I have the knowledge necessary to use the system
The system is not compatible with other systems I use
A specific person (or group) is available for assistance with system difficulties
Intention to use
I intend to use the system in the next <n> months
I predict I would use the system in the next <n> months
I plan to use the system in the next <n> months
Venkatesh et al. (2003) found that ‘performance expectancy’ is a determinant of
‘intention to use’ in most situations. The strength of the relationship, however, is
moderated by age and gender, being more significant for men and younger
workers. The effect of ‘effort expectancy’ is also moderated by gender and age,
being more significant for women and older workers. This effect decreases with
experience. The effect of ‘social influence’ on intention to use is conditioned by
4 Acceptance Models
43
age, gender, experience and voluntariness such that the authors found it to be non-
significant when the data were analyzed without the inclusion of moderators. The
effect of ‘facilitating conditions’ is only significant when examined in combination
with the moderating effects of age and experience – i.e. they only matter for older
workers in later stages of experience (Venkatesh et al., 2003).
4.3.1 The use of the UTAUT model in other areas
The UTAUT model has also been utilized in other areas such as adoption of
mobile services among consumers (Carlsson et al., 2006) and in the health sector
to examine e.g. the viability of motes (tiny, wireless sensor devices) as health
monitoring tools, health professionals’ reluctance to accept and utilise information
and communication technologies, physicians’ acceptance of a pharmacokinetics-
based clinical decision support system and physician adoption of electronic
medical records technology (e.g. Lubrin et al., 2006; Chang et al., 2007;
Hennington & Janz, 2007 and Schaper & Pervan, 2007).
The studies largely support the appropriateness of the UTAUT model in these
areas. However, the social influence was not found to be as strong a predictor as
suggested by the model when investigating information/communication
technologies and decision support in the health sector (Chang et al., 2007 and
Schaper & Pervan, 2007). Extensions/modifications of the model were
recommended both in the adoption of mobile services and within the health sector
(Carlsson et al., 2006; and Lubrin et al., 2006).
4.4 Using the UTAUT model in the context of driver support systems
The Unified Theory of Acceptance and Use of Technology has provided assistance
in understanding what factors either enable or hinder technology acceptance and
use. It is based on a number of significant behavioural models, of which some are
frequently used in traffic safety research as well. Although the model was
developed for the information technology area, it has fruitfully been used in other
areas. As such, it is possible that the UTAUT model could be useful for
understanding and analysing acceptance in the driver support area. A first proposal
to use the UTAUT model for acceptance of driver support systems was made by
Adell in 2007 (Adell, 2007), but the model has not yet been applied to driver
support systems. Although there are many similarities between information
technology (IT) and driver support systems, there are also several important
differences.
4 Acceptance Models
43
age, gender, experience and voluntariness such that the authors found it to be non-
significant when the data were analyzed without the inclusion of moderators. The
effect of ‘facilitating conditions’ is only significant when examined in combination
with the moderating effects of age and experience – i.e. they only matter for older
workers in later stages of experience (Venkatesh et al., 2003).
4.3.1 The use of the UTAUT model in other areas
The UTAUT model has also been utilized in other areas such as adoption of
mobile services among consumers (Carlsson et al., 2006) and in the health sector
to examine e.g. the viability of motes (tiny, wireless sensor devices) as health
monitoring tools, health professionals’ reluctance to accept and utilise information
and communication technologies, physicians’ acceptance of a pharmacokinetics-
based clinical decision support system and physician adoption of electronic
medical records technology (e.g. Lubrin et al., 2006; Chang et al., 2007;
Hennington & Janz, 2007 and Schaper & Pervan, 2007).
The studies largely support the appropriateness of the UTAUT model in these
areas. However, the social influence was not found to be as strong a predictor as
suggested by the model when investigating information/communication
technologies and decision support in the health sector (Chang et al., 2007 and
Schaper & Pervan, 2007). Extensions/modifications of the model were
recommended both in the adoption of mobile services and within the health sector
(Carlsson et al., 2006; and Lubrin et al., 2006).
4.4 Using the UTAUT model in the context of driver support systems
The Unified Theory of Acceptance and Use of Technology has provided assistance
in understanding what factors either enable or hinder technology acceptance and
use. It is based on a number of significant behavioural models, of which some are
frequently used in traffic safety research as well. Although the model was
developed for the information technology area, it has fruitfully been used in other
areas. As such, it is possible that the UTAUT model could be useful for
understanding and analysing acceptance in the driver support area. A first proposal
to use the UTAUT model for acceptance of driver support systems was made by
Adell in 2007 (Adell, 2007), but the model has not yet been applied to driver
support systems. Although there are many similarities between information
technology (IT) and driver support systems, there are also several important
differences.
Adell | Driver experiences and acceptance of driver support systems
44
4.4.1 Differences between IT and driver support systems
Applications of information technology and driver support systems share many
important features: the user interacts with a technology that is often too complex
to fully understand, new applications are incorporated in an existing interaction
between the user and the technology, the information conveyed to the user seeks
to facilitate an ongoing task, etc.
Despite the similarities, there are important differences between the settings in
which information technology applications and the driver support systems are
used, particularly at the operational level. One important difference between
computer use and car driving is the time aspect. When using a computer, the user
normally has the possibility of pausing and pondering and even asking for help
with a process or decision. A continuous decision making or execution is not
usually required. It is different when driving a car. The car driver normally has a
short time span in which a decision (and action) has to be made and does not
normally have the possibility of acquiring assistance with a process or decision.
Car driving also demands continuous decision making and execution of these.
When using a computer the user does not normally have to interact with other
humans, while a car driver must interact with other road users, making the social
dimension of the two settings very different. When a computer user makes a
mistake it is often reparable; the consequence is usually irritating and sometimes
time consuming, but seldom dangerous. When a car driver makes a mistake it
could end in severe physical damage or fatality both for the user and others. The
working environment when using a computer is imaginary, while the use of a car
takes place in the real world.
These differences are important to recognize and address – but they should not
stand in the way of transferring methodology from analysing acceptance and use
of information technology to studies of driver support systems.
4.4.2 Using the UTAUT for a driver support system – a pilot test
Based on the acceptance research of information technology, a pilot test, using the
Unified Theory of Acceptance and Use of Technology (UTAUT), was carried out
on a driver support system in 2008. Data for this pilot test was collected in 2006
and 2007 during field trials to evaluate a prototype driver support system
(SASPENCE). The purpose was to explore the potential of using the model in the
context of driver support systems; thus, the original model was applied as far as
possible. However, the experimental design of the field trials could not be
modified for the evaluation of UTAUT. Nevertheless, additional questions to the
already planned questionnaires allowed data collection for examination of the
inter-relationships of ‘performance expectancy’, ‘effort expectancy’, ‘social influence
Adell | Driver experiences and acceptance of driver support systems
44
4.4.1 Differences between IT and driver support systems
Applications of information technology and driver support systems share many
important features: the user interacts with a technology that is often too complex
to fully understand, new applications are incorporated in an existing interaction
between the user and the technology, the information conveyed to the user seeks
to facilitate an ongoing task, etc.
Despite the similarities, there are important differences between the settings in
which information technology applications and the driver support systems are
used, particularly at the operational level. One important difference between
computer use and car driving is the time aspect. When using a computer, the user
normally has the possibility of pausing and pondering and even asking for help
with a process or decision. A continuous decision making or execution is not
usually required. It is different when driving a car. The car driver normally has a
short time span in which a decision (and action) has to be made and does not
normally have the possibility of acquiring assistance with a process or decision.
Car driving also demands continuous decision making and execution of these.
When using a computer the user does not normally have to interact with other
humans, while a car driver must interact with other road users, making the social
dimension of the two settings very different. When a computer user makes a
mistake it is often reparable; the consequence is usually irritating and sometimes
time consuming, but seldom dangerous. When a car driver makes a mistake it
could end in severe physical damage or fatality both for the user and others. The
working environment when using a computer is imaginary, while the use of a car
takes place in the real world.
These differences are important to recognize and address – but they should not
stand in the way of transferring methodology from analysing acceptance and use
of information technology to studies of driver support systems.
4.4.2 Using the UTAUT for a driver support system – a pilot test
Based on the acceptance research of information technology, a pilot test, using the
Unified Theory of Acceptance and Use of Technology (UTAUT), was carried out
on a driver support system in 2008. Data for this pilot test was collected in 2006
and 2007 during field trials to evaluate a prototype driver support system
(SASPENCE). The purpose was to explore the potential of using the model in the
context of driver support systems; thus, the original model was applied as far as
possible. However, the experimental design of the field trials could not be
modified for the evaluation of UTAUT. Nevertheless, additional questions to the
already planned questionnaires allowed data collection for examination of the
inter-relationships of ‘performance expectancy’, ‘effort expectancy’, ‘social influence
4 Acceptance Models
45
and ‘intention to use’, including gender and age as moderators. A summary of the
trial is given below; more details about the trial are reported in paper IV and Adell
et al. (2009).
The SASPENCE system
The SASPENCE system is a driver support system which assists the driver to keep a
safe speed (according to road and traffic conditions) and a safe distance to the
vehicle ahead. The “Safe Speed and Safe Distance” function informs/warns the
driver when a) the car is too close to the vehicle in front, b) a collision is likely due
to a positive relative speed, c) the speed is too high considering the road layout
and d) the car is exceeding the speed limit.
The driver receives information and feedback from the system by means of an
external speedometer display located on the instrument panel, haptic feedback in
the accelerator pedal or in the seat belt and an auditory message when a too short
headway could lead to imminent danger. For further information about the
system see Adell et al. (2009).
Method
Two different test routes were used, one in Turin, Italy, and one in Valladolid,
Spain. Both routes were approximately 50 km long and contained both urban and
rural road stretches and a motorway section. The test drivers drove the test route
twice, once with the system on and once with the system off, thus serving as their
own controls. The order of driving was altered to minimize bias due to learning
effects.
At each site, 20 randomly selected inhabitants, balanced according to age and
gender, participated in the trial. Unfortunately, the data for one test driver was lost
due to system failure in Italy, and one of the test drives in Spain was cancelled for
safety reasons.
Before the drivers used the SASPENCE system, they were given a brief explanation
of the system. The questions regarding the UTAUT assessment were given to the
drivers as part of the questionnaire after the second drive.
The items for assessing ‘behavioural intention’, ‘performance expectancy’, ‘effort
expectancy’ and ‘social influence’ were adopted from Venkatesh et al. (2003).
Some of the items had to be adapted to fit the context of driver assistance systems,
see Table 3. Each item was measured using a seven-point scale, ranging from
“strongly disagree” (1) to “strongly agree” (7) (identical to Venkatesh et al., 2003).
4 Acceptance Models
45
and ‘intention to use’, including gender and age as moderators. A summary of the
trial is given below; more details about the trial are reported in paper IV and Adell
et al. (2009).
The SASPENCE system
The SASPENCE system is a driver support system which assists the driver to keep a
safe speed (according to road and traffic conditions) and a safe distance to the
vehicle ahead. The “Safe Speed and Safe Distance” function informs/warns the
driver when a) the car is too close to the vehicle in front, b) a collision is likely due
to a positive relative speed, c) the speed is too high considering the road layout
and d) the car is exceeding the speed limit.
The driver receives information and feedback from the system by means of an
external speedometer display located on the instrument panel, haptic feedback in
the accelerator pedal or in the seat belt and an auditory message when a too short
headway could lead to imminent danger. For further information about the
system see Adell et al. (2009).
Method
Two different test routes were used, one in Turin, Italy, and one in Valladolid,
Spain. Both routes were approximately 50 km long and contained both urban and
rural road stretches and a motorway section. The test drivers drove the test route
twice, once with the system on and once with the system off, thus serving as their
own controls. The order of driving was altered to minimize bias due to learning
effects.
At each site, 20 randomly selected inhabitants, balanced according to age and
gender, participated in the trial. Unfortunately, the data for one test driver was lost
due to system failure in Italy, and one of the test drives in Spain was cancelled for
safety reasons.
Before the drivers used the SASPENCE system, they were given a brief explanation
of the system. The questions regarding the UTAUT assessment were given to the
drivers as part of the questionnaire after the second drive.
The items for assessing ‘behavioural intention’, ‘performance expectancy’, ‘effort
expectancy’ and ‘social influence’ were adopted from Venkatesh et al. (2003).
Some of the items had to be adapted to fit the context of driver assistance systems,
see Table 3. Each item was measured using a seven-point scale, ranging from
“strongly disagree” (1) to “strongly agree” (7) (identical to Venkatesh et al., 2003).
Adell | Driver experiences and acceptance of driver support systems
46
Table 3: The original UTAUT items and the modified items used in this study to assess
acceptance of driver support systems.
Original items (Venkatesh et al., 2003) Modified items
Behavioural intention to use the system (BI):
Imagine that the system was on the market and
you could get the system in you own car.
BI1 I intend to use the system in the next <n>
months