You are on page 1of 303

DESIGN AND DEVELOPMENT OF AN IN-VEHICLE HUMAN

MACHINE INTERFACE FOR ECO-SAFE DRIVING

Atiyeh Vaezipour
BSc (Computer Engineering)
MSc (Information Engineering and Management)

Submitted in fulfilment of the requirements for the degree of


Doctor of Philosophy (PhD)

Centre for Accident Research and Road Safety – Queensland (CARRS-Q)

School of Psychology and Counselling

Institute of Health and Biomedical Innovation (IHBI)

Faculty of Health

Queensland University of Technology

2018
Keywords

Driver acceptance; Driver Behaviour; Eco-driving; Fuel Consumption; Human

Computer Interaction; Human Machine Interface; In-vehicle system; Intelligent

Transport Systems; Road safety; Safe driving; Technology Acceptance Model; User-

centred design; Usability; Driver Workload

Design and Development of an In-vehicle Human Machine Interface for Eco-Safe Driving i
Abstract

Road transport plays a significant role in most industries and mobility services

around the globe and has a vital impact on our daily lives. However, reliance on

motorised vehicles has serious impacts on both human health and the environment. In

recent years, rapid developments in technology, such as in-vehicle human-machine

interfaces (HMI) and smartphone applications, have provided opportunities to

encourage drivers to drive in a more fuel-efficient manner, referred to as an eco-driving

style. In-vehicle HMIs have also been used to promote safe driving. However, limited

research has targeted both eco-driving and safe driving with a single integrated system.

This can be referred to as eco-safe driving, a driving style that reduces fuel

consumption while also improving safe driving behaviour.

This thesis by publication adopts a user-centred approach to the design,

development and evaluation of an in-vehicle HMI to provide real-time advice and

feedback to drivers. A total of five papers highlighting different aspects and phases of

this user-centred design approach are included in this thesis. The HMI developed as

part of this research aims to encourage eco-safe driving behaviour, while attaining a

high level of driver acceptance and having minimal impact on driver workload. In

order to achieve this research objective, three studies were conducted using a mixed-

methods design consisting of both qualitative and quantitative approaches.

Study 1 addressed Research Question 1, which asked what is the relevant

information that should be conveyed to drivers in an in-vehicle HMI to improve eco-

safe driving? The study involved a critical review of the literature to identify the

research gaps and gain a comprehensive understanding of the relationship between

eco-driving and safety. In addition, the review sought to identify best-practice

ii Design and Development of an In-vehicle Human Machine Interface for Eco-Safe Driving
principles associated with the design and development of in-vehicle HMI systems.

These findings informed the subsequent studies.

Study 2 also addressed Research Question 1, but did so from a user perspective

and involved a series of focus groups and an expert workshop. The research explored

novel ways to provide real-time eco-safe driving advice and feedback to drivers via an

in-vehicle HMI. Ultimately, the user-centred design approach guided the development

of a prototype in-vehicle HMI system and the Technology Acceptance Model (TAM)

was used as a theoretical framework to guide the system evaluation. Taken together,

the findings of the critical review and qualitative studies guided the design and

development of three innovative and novel eco-safe in-vehicle HMI prototypes. They

varied in the types of advice and feedback messages provided to drivers, specifically:

(i) visual advice only; (ii) visual feedback only; and (iii) visual advice and feedback.

Finally, Study 3 addressed Research Question 2, which asked what impact does

advice and/or feedback characteristics of an in-vehicle HMI have on (a) driver

acceptance of technology, (b) driver workload, and (c) eco-safe driving performance?

The study involved a driving simulator experiment and a driver survey, and sought to

evaluate the three HMI prototypes based on the type of advice and feedback provided.

Objective measures including fuel consumption and a range of eco-safe driving

behaviours were analysed, as well as subjective measures including driver acceptance,

attitudes and demographics. In addition, Study 3 also addressed Research Question 3,

which asked what impact do incentives associated with using an in-vehicle eco-safe

HMI have on eco-safe driving performance?

The critical review (Study 1) demonstrated that improved fuel efficiency can

be achieved as a result of positive interactions between an in-vehicle HMI and drivers.

It highlighted a variety of eco-safe driving parameters, as well as best-practice design

Design and Development of an In-vehicle Human Machine Interface for Eco-Safe Driving iii
and feedback characteristics, which were used to inform the design and development

of the prototype system. In addition, the review enhanced our understanding of factors

such as driver acceptance, which plays a significant role in the voluntary and sustained

use of in-vehicle HMIs, and thus the overall effectiveness of any system.

The driver focus groups and expert workshop (Study 2), demonstrated that an

eco-safe in-vehicle HMI can experience relatively high levels of driver acceptance if

carefully designed with the needs and requirements of users in mind. Drivers expressed

a range of motivations to use such a system, including both extrinsic (e.g., monetary

savings associated with fuel and vehicle maintenance costs) and intrinsic (e.g.,

environmental benefits and road safety). A consideration of user requirements and

needs appeared to be critical and users expressed the desire to customise device

characteristics to suit their specific preferences. In addition, several potential barriers

to driver acceptance were revealed, including lack of perceived trust and privacy of

the system, distraction and the inadvertent promotion of unsafe and illegal driving

behaviours.

The findings on driver acceptance in the driving simulator experiment and the

survey suggested that all three prototypes of the eco-safe in-vehicle HMI examined

(i.e., advice only, feedback only, advice/feedback) were associated with relatively high

levels of acceptance. Nonetheless, the advice only system was associated with

significantly higher levels of acceptance, including usability, ease of use and overall

system satisfaction. While all system types produced relatively low levels of workload

for drivers, the two systems involving feedback messages significantly increased the

workload associated with using the interface.

The findings regarding the impact of the system on driving behaviours

suggested that the advice/feedback system led to the greatest improvement in both fuel

iv Design and Development of an In-vehicle Human Machine Interface for Eco-Safe Driving
consumption and eco-safe driving behaviour, including acceleration, deceleration,

speed, and headway. That is, overall, the advice/feedback system had a significantly

greater impact on eco-safe driving behaviour than either the advice only or feedback

only systems. These findings suggest that both message types working in a

complimentary manner are important for system effectiveness, rather than either

message type in isolation.

There was limited evidence of any additional impact of incentives and

competition on eco-safe driving behaviour. However, there was some evidence to

suggest that a range of extrinsic and intrinsic incentives may be beneficial for

increasing intentions to use the system, which has strong practical significance.

Specifically, drivers typically reported low levels of voluntary intentions to use the

system, regardless of how positively they perceived the system to be in terms of

usefulness and usability. There was limited difference regarding system preferences,

however older drivers who perceived themselves to be safer, did reveal a preference

for the advice only system.

Overall, this program of research highlighted the effectiveness of an eco-safe

in-vehicle HMI to not only provide real-time advice and feedback to drivers regarding

their eco-safe driving performance, but positively influence and improve this

behaviour. Furthermore, this research demonstrates the utility and benefits of

technology that allows for the communication between vehicles and infrastructure in

the traffic environment, known as V2X. As this technology becomes more

commonplace, such systems are likely to experience more widespread development

and implementation.

This research also contributes new and valuable knowledge towards enhancing

current understanding of the nature and content of design features of eco-safe in-

Design and Development of an In-vehicle Human Machine Interface for Eco-Safe Driving v
vehicle HMIs. Such information provides an evidence-based foundation for designers

and developers to increase the effectiveness, as well as the driver acceptance, of such

systems. This research explores some of the practical implications associated with the

development and implementation of such systems, including effects on driver

workload, the influence of competition and incentives on intentions to use the system,

as well as how various feedback and device characteristics impact upon perceived

system usefulness and intentions to use. Furthermore, by examining both driver

acceptance and system effectiveness, this research acknowledges that system

effectiveness may not necessarily ensure high levels of driver acceptance, and vice

versa.

This multidisciplinary thesis draws upon human factors, human computer

interactions and transport safety disciplines to enhance our understanding of the needs

and requirements of drivers with regards to an eco-safe in-vehicle HMI and the barriers

to driver acceptance. This understanding provides a foundation to inform the future

design and development of innovative, persuasive and acceptable in-vehicle HMIs. It

also contributes to improved eco-safe driving, while having minimal adverse impact

on the driver.

vi Design and Development of an In-vehicle Human Machine Interface for Eco-Safe Driving
Table of Contents

Keywords .................................................................................................................................. i
Abstract .................................................................................................................................... ii
Table of Contents ................................................................................................................... vii
List of Figures ......................................................................................................................... xi
List of Tables ........................................................................................................................ xiii
List of Abbreviations ..............................................................................................................xv
List of Definitions ................................................................................................................. xvi
List of Publications .............................................................................................................. xvii
Statement of Original Authorship ....................................................................................... xviii
Acknowledgment .................................................................................................................. xix
Chapter 1: Introduction ...................................................................................... 1
1.1 Introductory Comments ..................................................................................................1
1.2 Research Problem and Background ................................................................................1
1.2.1 Environmental Concerns ......................................................................................2
1.2.2 Traffic Crashes .....................................................................................................3
1.2.3 Eco-Driving and Eco-Safe Driving ......................................................................5
1.2.4 In-Vehicle Technologies ......................................................................................8
1.2.5 Driver Acceptance of Technology........................................................................9
1.2.6 Incentives, Gamification and Competition .........................................................11
1.3 Research Gap ................................................................................................................13
1.4 Theoretical Framework .................................................................................................14
1.4.1 Technology Acceptance Model (TAM) .............................................................14
1.4.2 User-Centred Approach to Design .....................................................................17
1.5 Research Objectives......................................................................................................19
1.5.1 Research Approach.............................................................................................20
1.6 Research Contribution and Significance.......................................................................21
1.7 Demarcation of Scope...................................................................................................23
1.8 Outline of Dissertation..................................................................................................24
Chapter 2: Research Design .............................................................................. 27
2.1 Introductory Comments ................................................................................................27
2.2 Research Design and Methodology ..............................................................................27
2.3 Study 1: Literature Review ...........................................................................................29
2.4 Study 2a: User Focus Groups .......................................................................................30
2.5 Study 2b: Expert Workshop..........................................................................................33
2.6 Eco-Safe In-vehicle HMI Description ..........................................................................37
2.7 Study 3: Driving Simulator Experiment & Driver Survey ...........................................40

Design and Development of an In-vehicle Human Machine Interface for Eco-Safe Driving vii
2.8 Ethics Approval............................................................................................................ 51
2.9 Concluding Remarks .................................................................................................... 51
Chapter 3: Reviewing In-vehicle Systems to Improve Fuel Efficiency and
Road Safety (Paper 1) .............................................................................................. 53
3.1 Introductory Comments ............................................................................................... 53
3.2 Abstract ........................................................................................................................ 56
3.3 Introduction .................................................................................................................. 57
3.4 Search Methodology .................................................................................................... 59
3.5 Results .......................................................................................................................... 60
3.5.1 Relationship Between Eco-Driving and Safety ................................................. 60
3.5.2 Effective Use of In-Vehicle Systems on Fuel Consumption and Safety ........... 63
3.5.3 Driving Parameters Influencing Fuel Consumption and Safety ........................ 64
3.5.4 Driver Acceptance of In-Vehicle Systems ......................................................... 67
3.5.5 Social Concepts to Motivate Eco-Safe Driving ................................................. 69
3.6 Conclusions .................................................................................................................. 70
Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human
Machine Interface: A Qualitative Study (Paper 2) ............................................... 73
4.1 Introductory Comments ............................................................................................... 73
4.2 Abstract ........................................................................................................................ 76
4.3 Introduction .................................................................................................................. 76
4.3.1 Eco-Driving as a Climate Change Initiative ...................................................... 77
4.3.2 User-Centred Design and Driver Acceptance .................................................... 79
4.3.3 Study Aims ........................................................................................................ 81
4.4 Method ......................................................................................................................... 82
4.4.1 Participants ........................................................................................................ 82
4.4.2 Procedure & Materials ....................................................................................... 83
4.4.3 Data Analysis ..................................................................................................... 84
4.5 Findings and Discussion .............................................................................................. 85
4.5.1 Number of Statements, Themes and Sub-Themes ............................................. 85
4.5.2 Perceived Importance of Monitoring Eco-Safe Driving Behaviours ................. 86
4.5.3 Perceived Usefulness of an Eco-Safe In-Vehicle System .................................. 88
4.5.4 Intentions to Use an Eco-Safe In-Vehicle HMI ................................................. 91
4.5.5 Eco-Safe In-Vehicle HMI Design Characteristics ............................................. 98
4.5.6 Potential Problems Associated with Using an Eco-Safe In-Vehicle HMI ....... 103
4.6 Conclusion and Implications ...................................................................................... 107
4.6.1 Implications for Practice .................................................................................. 107
4.6.2 Limitation & Future Research ......................................................................... 111
Chapter 5: Design of a Gamified Interface to Improve Fuel Efficiency and
Safe Driving (Paper 3)............................................................................................ 113
5.1 Introductory Comments ............................................................................................. 113
5.2 Abstract ...................................................................................................................... 116
5.3 Introduction ................................................................................................................ 116
5.4 Conceptual Gamified Design ..................................................................................... 119
5.5 Method ....................................................................................................................... 121

viii Design and Development of an In-vehicle Human Machine Interface for Eco-Safe Driving
5.5.1 Participants .......................................................................................................121
5.5.2 Procedure and Materials ...................................................................................122
5.5.3 Data Analysis ...................................................................................................122
5.6 Findings ......................................................................................................................122
5.6.1 User Perceptions of Gamified Concepts...........................................................123
5.6.2 Incentives and Motivations to Use the System .................................................125
5.6.3 Social Persuasion Feedback .............................................................................128
5.7 Discussion and Design Implications ...........................................................................129
5.8 Conclusion ..................................................................................................................131
Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface
for Eco-Safe Driving (Paper 4) ............................................................................. 133
6.1 Introductory Comments ..............................................................................................133
6.2 Abstract.......................................................................................................................136
6.3 Introduction ................................................................................................................136
6.4 Eco-safe in-vehicle HMI design process and description ...........................................139
6.5 Research hypotheses ...................................................................................................142
6.6 Method ........................................................................................................................143
6.6.1 Participants .......................................................................................................143
6.6.2 Driving simulator and scenario ........................................................................144
6.6.3 Experimental Design and Procedure ................................................................146
6.6.4 Questionnaire measures ....................................................................................147
6.6.5 Driving performance measures.........................................................................148
6.6.6 Data Analysis ...................................................................................................150
6.7 Results ........................................................................................................................150
6.7.1 H1: Driver acceptance of the technology .........................................................150
6.7.2 H2: Workload ...................................................................................................153
6.7.3 H3: Eco-safe driving behaviour........................................................................154
6.7.4 Realism of the simulated driving experience ...................................................160
6.8 Discussion ...................................................................................................................161
6.9 Conclusion ..................................................................................................................167
6.10 Limitations ..................................................................................................................168
Chapter 7: A Simulator Study of the Effect of Incentives on Adoption and
Effectiveness of an In-Vehicle Driving HMI (Paper 5) ....................................... 169
7.1 Introductory Comments ..............................................................................................169
7.2 Abstract.......................................................................................................................172
7.3 Introduction ................................................................................................................172
7.4 Study Aims .................................................................................................................177
7.5 Method ........................................................................................................................177
7.5.1 Participants .......................................................................................................177
7.5.2 Driving Simulator and Scenario .......................................................................178
7.5.3 Experimental Design ........................................................................................179
7.5.4 Procedure ..........................................................................................................181
7.5.5 Driving Behaviors Measures ............................................................................183
7.5.6 Subjective Measures .........................................................................................184
7.5.7 Data Analysis ...................................................................................................184

Design and Development of an In-vehicle Human Machine Interface for Eco-Safe Driving ix
7.6 Results ........................................................................................................................ 185
7.6.1 System Choice and Intention to Use ................................................................ 185
7.6.2 Factors Associated with System Choice .......................................................... 186
7.6.3 Impact of Incentives and Motivations on Intentions to Use ............................ 188
7.6.4 Impact of Incentives and Competition on Eco-Safe Driving Behaviours ........ 189
7.6.5 Willingness to Purchase the System ................................................................ 191
7.7 Discussion and Conclusion ........................................................................................ 192
Chapter 8: Discussion and Conclusion........................................................... 197
8.1 Introductory Comments ............................................................................................. 197
8.2 Discussion of Findings ............................................................................................... 198
8.3 General Recommendations ........................................................................................ 208
8.4 Research Implications ................................................................................................ 210
8.5 Theoretical Implications ............................................................................................ 212
8.6 Strengths and Limitations .......................................................................................... 213
8.7 Directions for Future Research .................................................................................. 216
8.8 Concluding Remarks .................................................................................................. 218
Bibliography ........................................................................................................... 221
Appendices .............................................................................................................. 233
A. Focus Groups Material ............................................................................................... 233
B. Expert Workshop Materials ....................................................................................... 245
C. Pilot HMI Icons .......................................................................................................... 255
D. Progression of Device Design .................................................................................... 256
E. Simulator Study Survey ............................................................................................. 257
F. Additional Data Analysis (Study 3) ........................................................................... 274
G. HMI Explantion from Simulator Study (Study 3) ...................................................... 278

x Design and Development of an In-vehicle Human Machine Interface for Eco-Safe Driving
List of Figures

Figure 1.1. Transport Emission by Mode in Australia, 2012 ...................................... 3


Figure 1.2. Annual road fatalities in Australia, 2007-2016......................................... 4
Figure 1.3. Annual serious injury crashes in Australia, 2002-2014 ............................ 4
Figure 1.4. Diagram of research gap addressed by this research .............................. 14
Figure 1.5. Technology Acceptance Model Framework ........................................... 16
Figure 1.6. The User-Centred Design Approach....................................................... 18
Figure 1.7. Mixed-Method Design Approach ........................................................... 21
Figure 2.1. Research design as it relates to the user-centred design approach ......... 28
Figure 2.2. First iteration of eco-safe in-vehicle HMI prototype .............................. 34
Figure 2.3. Examples of iterations of the eco-safe in-vehicle HMI prototype
screens. ......................................................................................................... 37
Figure 2.4. Examples of the eco-safe in-vehicle HMI prototype housing. ............... 37
Figure 2.5. Diagram of the circuit ............................................................................. 39
Figure 2.6. Examples advice and feedback messages ............................................... 39
Figure 2.7. Examples advice and feedback messages in context .............................. 40
Figure 2.8. Experimental Design and Procedure....................................................... 42
Figure 2.9. Examples of digital prototypes screens presented on iPad ..................... 44
Figure 2.10. SCANeR™ studio engine specification ............................................... 46
Figure 2.11. SCANeR™studio fuel consumption specification ............................... 46
Figure 3.1 User-Centred Design Approach - Critical Review .................................. 54
Figure 4.1 User-Centred Design Approach- Driver Focus Groups ........................... 74
Figure 4.2. The User-Centred Design Approach (adapted from ISO, 2010). ........... 80
Figure 5.1. Conceptual gamified design, including challenges (a),
individualised progress feedback (b), and social persuasion feedback
(c) ............................................................................................................... 120
Figure 6.1 User-Centred Design Approach- Driving simulator experiment &
driver survey .............................................................................................. 134
Figure 6.2. Eco-safe driving companion ................................................................. 140
Figure 6.3. Different “faces” of the system – positive, neutral and negative.......... 141
Figure 6.4. Examples of advice and feedback messages......................................... 142
Figure 6.5. The CARRS-Q Advanced Driving Simulator. ..................................... 145
Figure 6.6. Examples of simulated driving scenarios. ............................................ 146
Figure 7.1. The CARRS-Q Advanced Driving Simulator ...................................... 179

Design and Development of an In-vehicle Human Machine Interface for Eco-Safe Driving xi
Figure 7.2. Examples of advice (upper panels) and feedback (lower panels)
messages..................................................................................................... 180
Figure 7.3. Experimental drives and procedures ..................................................... 181
Figure 7.4 The leader board screen presented to participants on iPad .................... 182

xii Design and Development of an In-vehicle Human Machine Interface for Eco-Safe Driving
List of Tables

Table 1.1. Research overview and relation to research questions and


publications .................................................................................................. 26
Table 3.1.Key Terminology in Driving Domain ....................................................... 60
Table 3.2. Driving Parameters Influencing Fuel Consumption and Safety ............... 65
Table 4.1. Demographics of Focus Group Participants ............................................. 83
Table 4.2. Summary of Thematic Analysis ............................................................... 85
Table 4.3. Selected Quotes Relevant to Theme 1 - "Perceived importance of
eco-safe driving behaviours" ....................................................................... 87
Table 4.4. Selected Quotes Relevant to Theme 2 - "Perceived usefulness of an
eco-safe in-vehicle HMI" ............................................................................. 89
Table 4.5. Selected Quotes Relevant to Theme 3 - "Intentions to use an eco-
safe in-vehicle HMI" .................................................................................... 93
Table 4.6. Selected Quotes Relevant to Theme 4 - "Eco-safe in-vehicle HMI
design characteristics".................................................................................. 99
Table 4.7. Selected Quotes Relevant to Theme 5 - "Potential problems
associated with the system" ....................................................................... 104
Table 4.8. Synthesis of Findings .............................................................................. 109
Table 6.1. Differences in driver acceptance by condition ....................................... 151
Table 6.2. Bivariate correlations between driver acceptance scales and
perceived trust ............................................................................................ 152
Table 6.3. Differences in driver workload by condition .......................................... 153
Table 6.4. Summary of means and standard deviations and significance tests
for eco-safe driving behaviour variables.................................................... 155
Table 6.5. Summary of significant post-hoc comparisons for acceleration
variables (Event 1: with traffic) ................................................................. 156
Table 6.6. Summary of significant post-hoc comparisons for acceleration
variables (Event 2: without traffic) ............................................................ 157
Table 6.7. Summary of significant post-hoc comparisons for deceleration
variables (long-range braking event) ......................................................... 158
Table 6.8. Means and standard deviations of perceptions of driving simulator
realism ........................................................................................................ 160
Table 7.1. Summary of subjective scales and internal reliability ............................ 184
Table 7.2. Intentions to use by system choice ......................................................... 185
Table 7.3. System choice by age and gender ........................................................... 186
Table 7.4. System choice by attitudes and perceptions ........................................... 187
Table 7.5. Incentives to motivate system use .......................................................... 188

Design and Development of an In-vehicle Human Machine Interface for Eco-Safe Driving xiii
Table 7.6. Means, standard deviations and results of statistical analyses for
impact of incentives on eco-driving behaviour .......................................... 190
Table 7.7. Amount willing to pay for system .......................................................... 191
Table 7.8. Factors effecting willingness to purchase ............................................... 192
Table 7.9. Impact of additional features on willingness to purchase ....................... 192

xiv Design and Development of an In-vehicle Human Machine Interface for Eco-Safe Driving
List of Abbreviations

ADAS Advanced Driver Assistance Systems


ARC Australian Research Council
CARRS-Q Centre for Accident Research and Road Safety-Queensland
C-ITS Cooperative Intelligent Transport System
HCI Human Computer Interaction
HMI Human Machine Interface
HUD Head-Up Display
IHBI Institute of Health and Biomedical Innovation
ISO International Organisation for Standardisation
ITS Intelligent Transport Systems
LED Light Emitting Diode
MSAQ Motion Sickness Assessment Questionnaire
QUT Queensland University of Technology
QUTHREC QUT Human Research Ethics Committee
SPSS Statistical Package for Social Science
TAM Technology Acceptance Model
V2I Vehicle-to-Infrastructure communication
V2V Vehicle-to-Vehicle communication
WHO World Health Organisation

Design and Development of an In-vehicle Human Machine Interface for Eco-Safe Driving xv
List of Definitions

Term Definition

Cooperative Intelligent Wireless communication of vehicle with other vehicles,


Transport System infrastructure or other parts of the road network.

Driver Assistance Level 1 and 2 of automation technologies such as automatic


Systems emergency braking.

Eco-driving Cost-effective driving style to reduce fuel consumption and


vehicle emissions
Eco-safe driving Driving style that reduces fuel consumption while improving
safe driving behavior.
Fuel consumption Total quantity of fuel used by a vehicle in a specific area and/or
time period
Fuel consumption rate Total quantity of fuel used by a vehicle per unit distance
(commonly expressed in litres/100 kilometres)
Fuel economy Inverse of fuel consumption rate, defined as distance travelled
per unit of fuel used (kilometers/litre)
Fuel efficiency Ratio of the work/energy output of an engine to the
work/energy input
Gamification Use of game design principles in a non-game context to boost
motivation and commitment to use the technology
Intelligent Transport Information and communication technologies in the field of
System road transport, including infrastructure, vehicles and users.
Acceptance Willingness and intention of individuals within the desired user
group to use technology for its intended purposes and in the
way it was designed
Acceptability Prospective judgement or potential acceptance of the system
when the driver has yet to experience using the system.

xvi Design and Development of an In-vehicle Human Machine Interface for Eco-Safe Driving
List of Publications

Dissertation Publications:

Vaezipour, A., Rakotonirainy, A., & Haworth, N. (2015). Reviewing in-vehicle


systems to improve fuel efficiency and road safety. Procedia Manufacturing, 3,
3192-3199. https://doi.org/10.1016/j.promfg.2015.07.869

Vaezipour, A. (2015) Design and development of an in-vehicle human machine


interface to improve fuel efficiency & safety. Presentation at Doctoral Consortium.
27th Conference of the Computer-Human Interaction Special Interest Group of
Australia on Computer-Human Interaction, 7-10 December 2015, Melbourne, Vic.
Vaezipour, A., Rakotonirainy, A., Haworth, N. (2016). Design of a gamified
interface to improve fuel efficiency and safe driving. In DUXU 2016: Design, User
Experience, and Usability: Novel User Experiences (Vol. 17, No. 9747, pp. 322-
332). Part of the Lecture Notes in Computer Science book series. Springer
International Publishing. https://doi.org/10.1007/978-3-319-40355-7_

Vaezipour, A., Rakotonirainy, A., Haworth, N., & Delhomme, P. (2017). Enhancing
eco-safe driving behaviour through the use of in-vehicle human-machine interface: A
qualitative study. Transportation Research Part A: Policy and Practice, 100, 247-
263. https://doi.org/10.1016/j.tra.2017.04.030

Vaezipour, A., Rakotonirainy, A., Haworth, N. & Delhomme, P. (2017). A


Simulator Evaluation of in-vehicle Human Machine Interface for Eco-Safe Driving.
Transportation Research Part A: Policy and Practice (under-review).

Vaezipour, A., Rakotonirainy, A., Haworth, N. & Delhomme, P. (2017). A


Simulator Evaluation of the Effect of Incentive on Adoption and Effectiveness of an
In-vehicle Eco-Safe Driving HMI. Journal of Safety Research (submitted).

Other Publication:

Orfila, O., Gruyer, D., Geoffroy, D., Glaser, S., Rakotonirainy, A., Vaezipour, A.,
Demmel, S. (2016) Immersive driving simulation architecture to support gamified
eco-driving instructions. In 23rd Intelligent Transportation Systems World Congress,
10-14 October 2016, Melbourne, Vic.

Design and Development of an In-vehicle Human Machine Interface for Eco-Safe Driving xvii
Statement of Original Authorship

The work contained in this thesis has not been previously submitted to meet
requirements for an award at this or any other higher education institution. To the best
of my knowledge and belief, the thesis contains no material previously published or
written by another person except where due reference is made. I have clearly stated
the contribution by others to jointly-authored works that I have included in my thesis.

Signature: QUT Verified Signature

Date: 25/03/2018

xviii Design and Development of an In-vehicle Human Machine Interface for Eco-Safe Driving
Acknowledgment

My research journey during the last three years was a roller-coaster of excitement and
emotions. As I am writing this, I am thinking of those who touched and inspired my
life in any way since I have started this research. All of this would have been
impossible without the support and care of many individuals along the way whom I
wish to acknowledge.
I wish to thank the Australian Research Council (No. DP140102895), Queensland
University of Technology (QUT), and Institute of Health and Biomedical Innovation
(IHBI) for funding support to carry out this research.
I wish to thank my supervisory team for the opportunity to be part of this exciting
project and make what was once my dream, a reality. To my principal supervisor,
Professor Andry Rakotonirainy – thank you for your support, guidance and for always
trusting in my abilities to complete this research. To my associate supervisor, Professor
Narelle Haworth – thank you for making time to review my research in your busy
schedule – your eyes for detail always amazed me. It was an honour to work with you
both and I am grateful for all I have learned from you.
I would also like to give a special thanks to Professor Patricia Delhomme. I can’t thank
you enough for the extraordinary support, encouragement and feedback you have
provided throughout my candidature. I hope one day I can visit you in France.
I wish to express my appreciation to those involved during my PhD milestones –
Professor James Freeman (Stage 2 proposal); Professor Daniel Johnson and Dr
Gregoire Larue (Confirmation seminar); and Professor Marcus Foth and Dr Ronald
Schroeter (Final seminar). Thank you for your constructive feedback and for sharing
your expertise.
I also wish to thank my colleagues at the Centre for Accident Research and Road
Safety – Queensland (CARRS-Q) and Queensland University of Technology. I could
not have asked for a more supportive and welcoming environment to undertake this
journey. Thank you for your contribution and sharing your expertise Ms Amy
Williamson, Dr Ioni Lewis, Dr Judy Fleiter, Dr Oscar Oviedo Trespalacios, Dr
Sebastien Demmel, Mr Andrew Haines, Ms Mahrokh Khazar, Mr Sepehr Ghasemi
Dehkordi, Dr Kirsten Vallmuur, Ms Wanda Griffin, Dr Angela Watson, Ms Kat
Bowman (Professional Editor), Ms Veronica Baldwin, Ms Ashlea Haddow, Ms
Mayuko Bock and Ms Andrea McCrindle.
A very special thank you to Ms Ian Ashworth and Ms Melissa Johnson at the Science
and Engineering Faculty at QUT for endless support, and to staff at the School of
Electrical Engineering and Computer Science – you’ve been incredibly amazing.
To my dear fellow postgraduate students who kept me going even on days I felt
hopeless, I cannot thank you enough. To Oscar for our fruitful research, career

Design and Development of an In-vehicle Human Machine Interface for Eco-Safe Driving xix
discussions in lunch time, sharing the struggle of those days of receiving comments on
thesis/paper submissions, and for always being there for a coffee break even when you
were busy. To Buthaina for your kindness, coffee and snack recharges. And thank you
to Teodora, Kalina, Fabius, Rusdi, Yusuf, Libi, Wahi, Peter, Kristi, Sepehr, Shamsi,
Heidy, Saif, Verity, Daniel, David, Klaire, Phuntsho, Raniya, Navid for the support
along the way.
To my lovely Mum and Dad, words simply cannot thank you for all that you have done
for me. Thank you for your encouragement, patience, positivity and love from miles
away, and your words of wisdom reminding me “don’t worry, everything will be ok”.
I would never have been able to achieve my goals without you, and for that I am
eternally grateful. I cannot wait to make up all those moments I could not spend with
you.
To my sister, Anousheh, for listening to all me phone calls all the way from Canada.
Thank you for making me laugh, I miss you every day. To my brother Ali and my
sister-in-law Nasim, and to Hossein, thank you for being supportive and making our
times as a family special. Thank you also to Bec, Steve, Nicky, Zac, Mickey, Phoebe
and B.J. cannot wait to spend more time with you.
Last but not least, I would like to thank David Soole for his unremitting
encouragement. The last years have been the most challenging days of my life, thank
you for your endless support during every stage of my research, all those weekends
and late night studying that we could have gone on adventures to find new bushwalks
and waterfalls, although they are all look the same!! I have never met anyone who
believes in me more, thank you for making me feel I could do this.

This too shall pass… 

xx Design and Development of an In-vehicle Human Machine Interface for Eco-Safe Driving
Chapter 1: Introduction

“We are running the most dangerous experiment in history right now,
which is to see how much carbon dioxide the atmosphere can handle
before there is an environmental catastrophe”

- Elon Musk

1.1 INTRODUCTORY COMMENTS

This chapter provides an overview of the program of research comprising this

thesis by publication. Specifically, this chapter outlines the research rationale,

including the problems associated with air pollution and traffic crashes (Section 1.2).

The chapter continues, highlighting the research gap, including the tendency for in-

vehicle systems to have focused on eco-driving or safe driving in isolation (Section

1.3), before providing an overview of the theoretical framework, including the

Technology Acceptance Model (TAM) and the user-centred design approach (Section

1.4). The objectives of the research, including research questions, are then outlined

(section 1.5), as well as the contributions of this research to the field (Section 1.6).

Finally, the scope of this thesis is outlined (Section 1.7), while Section 1.8 provides an

overview of the remaining chapters.

1.2 RESEARCH PROBLEM AND BACKGROUND

Road transport plays a significant role in various industries and mobility services

across the globe and has a vital impact on our daily lives. However, this reliance on

motorised vehicles has serious implications for human health and the environment,

including pollution and road traffic fatalities and injuries.

Chapter 1: Introduction 1
1.2.1 Environmental Concerns
While some politicians continue to express doubt regarding the existence of

climate change, and the topic continues to be fiercely debated among the general

population, there is a strong general consensus among the scientific community that

climate change and global warming is occurring and is unequivocal (Cook et al., 2016;

Oreskes, 2004). Indeed, the overwhelming scientific evidence suggests rapid upward

trends in human made emissions, atmospheric and ocean temperatures, and sea levels,

and there are growing concerns regarding shrinking ice sheets and sea ice, glacial

retreat, extreme weather events, and ocean acidification (NASA, 2017). Moreover,

reports suggest that 97% of all climate scientists agree that these trends are extremely

likely to have been greatly influenced by human behaviour, in particular the production

of greenhouse gases (Pachauri et al., 2014).

While emissions are produced from a range of activities across various

industries, road vehicles are responsible for 19% of all carbon dioxide (CO2) emissions

on a global scale, and have been intrinsically linked to rising global temperatures

(Matthews et al., 2009). Similarly, according to the Australian Authority Climate

Change (2014), 16% of Australia’s total emissions can be attributed to transport

vehicles, with light vehicles by far the largest contributor (Figure 1.1). Indeed, the

report shows that the Australian population travels more than 1.15 billion kilometres

a day, at an average of 49 kilometres per person per day.

2 Chapter 1: Introduction
Figure 1.1. Transport Emission by Mode in Australia, 2012
Source: Australian Authority Climate Change (2014)

Consequently, the ever-growing reliance on motor vehicles and resulting

gaseous and particulate pollution have severe impacts on human health, including risk

of lung and heart diseases. Indeed, a recent report by the World Health Organization

(2014) revealed that approximately 7 million deaths per year are caused by air

pollution, making air pollution the largest single environmental health threat on a

global scale. Thus, reducing the dependency on fossil fuels has been identified as

urgent public health concerns requiring increased research and industrial

developments (ETRAC, 2014).

1.2.2 Traffic Crashes


According to the World Health Organization (2015), approximately 1.25 million

road fatalities occur each year across the globe, with a further 50 million more people

injured in traffic crashes annually. In Australia in 2016, there were 1,295 road

fatalities, at a rate of 5.4 deaths per 100,000 of the population (Bureau of Infrastructure

Transport and Regional Economics, 2017b). While this represented a 7.5% increase

compared to 2015, a downward trend in traffic fatalities is evident over the last ten

years, with a reduction of 19.2% in average annual fatalities compared to 2007 (see

Chapter 1: Introduction 3
Figure 1.2). However, a similar downward trend is not observed in serious injury

crashes (see Figure 1.3).

Figure 1.2. Annual road fatalities in Australia, 2007-2016


Source: Bureau of Infrastructure Transport and Regional Economics (2017)

Figure 1.3. Annual serious injury crashes in Australia, 2002-2014


Source: Bureau of Infrastructure Transport and Regional Economics (2017)

There are also considerable social and economic costs associated with road

trauma. The World Health Organization (2015) has estimated that, on average, nations

spend approximately 3% of their GDP on costs associated with traffic crashes. In

Australia, most recent traffic crashes have been estimated to cost $17.85 billion per

4 Chapter 1: Introduction
year (Australian Automobile Association, 2017), with some having argued that such

figures represent an underestimation of the true social and economic costs of road

crashes (Tooth, 2010).

Research has shown a myriad of contributing factors associated with traffic

crashes. These have commonly been classified as either: (i) human errors, such as

inattention, lapses in judgement or traffic violations; (ii) environmental factors, such

as road design and weather conditions; or, (iii) vehicle factors, such as mechanical

issues (Treat et al., 1979). Of these factors, human errors have been reported as being

a critical factor in as much as 90% of crashes, a finding that has been well established

over a number of decades (Shinar, 1978; Singh, 2015; Treat et al., 1979).

Among the human factors associated with crashes are a range of behaviours such

as speeding, following too closely, driving while impaired, inattention, mobile phone

use while driving, as well as fatigue and monotony (Petridou & Moustaki, 2000).

Taken together, this research suggests a critical need for interventions aimed at

addressing risky driver behaviours, including both violations and errors, as well as

driving behaviours that increase the impact of vehicle use on the environment.

1.2.3 Eco-Driving and Eco-Safe Driving


Eco-driving has been demonstrated as a cost-effective approach for reducing fuel

consumption and vehicle emissions (see reviews Barkenbus, 2010; Jonkers, Wilmink,

Nellthorp, Guehnemann, & Olstam, 2016; McIlroy, Stanton, & Harvey, 2014;

Vaezipour, Rakotonirainy, & Haworth, 2015). Eco-driving is the outcome of decision-

making processes which aim to improve fuel economy and reduce emissions

including: (i) strategic decisions (e.g., vehicle selection, maintenance schedules); (ii)

tactical decisions (e.g., route selection, vehicle loading); and, (iii) operational

decisions (e.g., driving style) (Alam & McNabola, 2014).

Chapter 1: Introduction 5
While each of these decisions are important, the focus of this thesis is on

improving eco-driving style through operational decisions, which are associated with

relatively immediate impacts on fuel consumption and emissions (Martin, Chan, &

Shaheen, 2012; Young, Birrell, & Stanton, 2011). In contrast, strategic decisions, such

as purchasing new low-carbon emission or electric vehicles, are typically associated

with lower rates of implementation due to their higher costs to users, while tactical

decisions, such as planning the route to avoid congestion, may not be as feasible and

convenient for all drivers (Barkenbus, 2010).

Eco-driving style refers to a range of behaviours that reduce vehicle fuel

consumption and subsequent emissions. Such behaviours include smooth acceleration

and deceleration, anticipating the traffic flow, maintaining a constant speed,

appropriate gear choice, and cruising or using the momentum of the vehicle (Young et

al., 2011). Engaging in these behaviours can have benefits for both individuals,

through reductions in vehicle running and maintenance costs, as well as society,

through reduced CO2 emissions and other pollutants (Barkenbus, 2010).

While an eco-driving style can produce positive impacts on fuel efficiency, there

is also evidence to suggest it can potentially compromise driver safety (Haworth &

Symmons, 2001). For example, maintaining a constant cruising speed and choosing

the highest appropriate gear may increase the likelihood that a driver will decrease

headway to the vehicle in front (i.e., following distance) and reduce their ability to

brake appropriately, in turn increasing the risk of rear-end collisions (Young & Birrell,

2012; Young et al., 2011). In addition, such behaviour may also increase the likelihood

of a driver manoeuvring at inappropriately high speeds, such as cornering (CIECA,

2007). Taken together, these findings highlight that the development and evaluation of

eco-driving initiatives should be conducted with driver safety as a critical

6 Chapter 1: Introduction
consideration. As such, eco-safe driving can be defined as a range of driving

behaviours that reduce fuel consumption and subsequent emissions, while also

maintaining the safety of drivers.

A number of guidelines, public education campaigns, and driver training

initiatives have been developed to encourage eco-driving behaviour, with mixed

results (e.g. EcoDrivingUSA; ECOWILL, 2014; Graves, Jeffreys, & Roth, 2012). One

explanation for the inconsistent findings may be that, while many drivers are

ultimately aware of the range of eco-driving behaviours that reduce fuel consumption

and emissions, they lack the technical understanding of how and when to appropriately

perform these behaviours (Delhomme, Cristea, & Paran, 2013; Pampel, Jamson,

Hibberd, & Barnard, 2015). Similarly, the complexity of the driving task might mean

that drivers are not always aware of their actions, and in turn may not use their eco-

driving knowledge and skills to their full potential (Pampel et al., 2015). Alternatively,

drivers may make a conscious decision to drive in a manner that is not fuel efficient or

safe due to a variety of reasons they believe justify the behaviour in the given moment,

such as running late or enjoying the feeling of driving fast (Harvey, Thorpe, &

Fairchild, 2013).

Therefore, there is a need to provide drivers with reliable, real-time advice and

feedback to assist with not only their adoption of an eco-safe driving style, but their

ability to evaluate the effectiveness and positive impacts of such behaviour. While a

number of recent studies have examined independent strategies to address either eco-

driving or traffic safety (Saboohi & Farzaneh, 2009; Young et al., 2011), there is a

paucity of research exploring the benefits of technological initiatives targeted at both

eco-driving and road safety. As such, there is a need to explore the potential of such

Chapter 1: Introduction 7
approaches and the impact they may have on both environmental and road safety

outcomes.

1.2.4 In-Vehicle Technologies


Emerging technologies have provided opportunities for the transportation

industry and automotive manufacturers to, among other things, assist in reducing the

impact of motor vehicles on the environment (Fagnant & Kockelman, 2015; Wadud,

MacKenzie, & Leiby, 2016). However, improvements associated with some

technological initiatives, such as electric or automated vehicles, have been incremental

and have experienced numerous barriers to their widespread implementation, such as

higher costs to users and required changes to the road infrastructure.

The past decade has been witness to rapid advances in in-vehicle technologies

that provide drivers with real-time advice and feedback for improving fuel efficiency

and road safety. This has included Advanced Driver Assistance Systems (ADAS) and

in-vehicle human machine interfaces (HMIs). Some of these systems are built into

vehicles (i.e., dashboard components), while others are available as aftermarket

products (e.g., head-up-displays, smartphone applications) (see Hof et al., 2013 for a

comprehensive review).

Research has demonstrated the positive impacts of operational decisions (e.g.,

driving style) associated with using in-vehicle technologies and the subsequent

benefits on fuel efficiency. Indeed, previous studies have reported reductions in fuel

consumption of as much as 20% associated with the use of eco-driving in-vehicle

systems (Barth & Boriboonsomsin, 2009; Birrell, Fowkes, & Jennings, 2014;

Strömberg & Karlsson, 2013). Barth and Boriboonsomsin (2009) found that speed

feedback via in-vehicle dashboard displays can reduce fuel consumption by 10-20%,

depending on the driving scenario. Similarly, reductions of 6.8% in fuel consumption

8 Chapter 1: Introduction
have been observed when bus drivers received instantaneous feedback on their driving

through in-vehicle eco-driving systems (Strömberg & Karlsson, 2013).

A number of studies have also demonstrated the potential positive impacts of safe

driving in-vehicle systems on road safety outcomes (Demmel et al., 2011; GOA,

2013). Such systems include a variety of collision warning systems (e.g., adaptive

braking, blind-spot warning), as well as more targeted devices such as Intelligent

Speed Adaptation (ISA). These technological approaches provide real-time

information to drivers via wireless communication from vehicle-to-vehicle (V2V) or

vehicle-to-infrastructure (V2I) using telecommunication and localisation technologies

(e.g., GPS). Such systems are collectively referred to as cooperative intelligent

transport systems (C-ITS), and enable vehicles to wirelessly communicate with other

vehicles and infrastructure through the cooperative in-vehicle system (Bureau of

Infrastructure Transport and Regional Economics, 2017a).

To date, cooperative in-vehicle systems have typically been developed to address

safety concerns, and have largely ignored issues associated with eco-driving. One of

the main benefits of cooperative in-vehicle systems for eco-safe driving is supporting

anticipatory driving by providing continuous, real-time traffic information to drivers

(e.g., impending traffic signal changes, upcoming speed limit changes, route guidance

to avoid traffic congestion) (Rittger & Götze, 2017). However, such real-time traffic

information must be presented to drivers in an intuitive and acceptable manner and

ensure limited distraction from the driving task.

1.2.5 Driver Acceptance of Technology


Driver acceptance of technology is a critical factor associated with intentions to

use and actual use of in-vehicle systems (Horberry, Stevens, & Regan, 2014b). It has

been argued that simply demonstrating the effectiveness of an in-vehicle HMI to

Chapter 1: Introduction 9
positively impact behaviour does not guarantee it will be widely accepted and used by

drivers. Indeed, the use of such systems are typically voluntary, and prior research has

highlighted barriers to implementation and intentions to use. Specifically, Trübswetter

and Bengler (2013) identified a number of barriers associated with the use of in-vehicle

technologies, including perceived usefulness, functional limitations, system costs and

accessibility, issues related to trust in the system and privacy, undesirable system

feedback, and inadvertent negative consequences of system use (e.g., distraction).

Furthermore, barriers to the use of eco-driving initiatives have been reported even

among individuals who are concerned about the environment. For example, Barkenbus

(2010) reported that individuals were less likely to engage in emission reduction

activities when they perceived major contributors to emissions (e.g., large

corporations) were not taking steps to reduce their carbon footprint. However,

consumer resistance to innovation is not a new concept, leading many market

researchers to conclude that successful innovation is achieved by developing an

understanding of the users’ needs, as well as the barriers to use, and developing a

strategy to surpass these (Ram & Sheth, 1989).

Taken together, these findings highlight the fundamental need to adopt a user-

centred design approach (see Section 1.3.2 for a detailed explanation of this approach)

when developing such initiatives. Indeed, to date there is limited understanding of

which features and functions of in-vehicle HMIs increase the likelihood of driver

acceptance and intentions to use (Albert, Musicant, Oppenheim, & Lotan, 2016). That

is, it is important to carefully consider the impact of driver acceptance on system

implementation and effectiveness. Previous research has revealed that the most

effective system may not necessarily be the one associated with the highest levels of

driver acceptance, for example in the case of intelligent speed adaptation (Adell,

10 Chapter 1: Introduction
Varhelyi, & Nilsson, 2014). By adopting a user-centred design approach, practitioners

can increase the likelihood that drivers will perceive the system to be useful, usable

and easy to use, and subsequently increase driver intentions to use the system

(Vaezipour, Rakotonirainy, Haworth, & Delhomme, 2017).

1.2.6 Incentives, Gamification and Competition


One major challenge associated with in-vehicle technologies is to enhance driver

motivation to use these systems as part of their everyday driving. According to Ryan

and Deci (2000), there are two primary ways in which individuals are motivated,

namely intrinsic and extrinsic motivation. Intrinsic motivation refers to “doing

something because it is inherently interesting or enjoyable” (e.g., feels good or helps

a cause), while extrinsic motivation refers to “doing something because it leads to a

separable outcome” (e.g., reward or prize) (Ryan & Deci, 2000, p. 55). While both are

important and can motivate users, there is evidence to suggest that intrinsic

motivations alone are often insufficient to encourage eco-driving behaviour. Indeed,

qualitative research conducted in Newcastle, Australia revealed that while most people

expressed concern for the environment, convenience and saving time had a greater

influence on eco-driving behaviour (Harvey et al., 2013).

Similar findings have been observed in relation to motivations to use eco-driving

in-vehicle technology. Harvey et al. (2013) found that saving money through better

fuel economy was typically insufficient to motivate eco-driving, and that the voluntary

use of in-vehicle feedback devices is unlikely to be successful without the use of

extrinsic rewards. Thus, it appears that extrinsic rewards may be a critical component

in widespread implementation and use of eco-driving in-vehicle technology, in

particular when considering aftermarket products.

Chapter 1: Introduction 11
The use of both monetary and non-monetary (e.g., discounts, vouchers, prizes)

extrinsic incentives has been shown to be effective in improving a range of human

behaviours, from workplace and educational performance through to health

behaviours (Condly, Clark, & Stolovitch, 2003; Garbers & Konradt, 2014; Giles,

Robalino, McColl, Sniehotta, & Adams, 2014; Gneezy, Meier, & Rey-Biel, 2011;

Mortimer, Ghijben, Harris, & Hollingsworth, 2013). Moreover, incentives have also

been shown to be effective in encouraging safe driving behaviours (Tselentis, Yannis,

& Vlahogianni, 2016; Volkswagen, 2009), as well as eco-driving behaviour (Lai,

2015; Schall & Mohnen, 2017). However, concerns have been raised regarding the use

of extrinsic incentives, with some arguing that reward systems may diminish intrinsic

motivations and trap individuals in a ‘reward loop’ (Deci, Koestner, & Ryan, 2001;

Zichermann & Cunningham, 2011).

Qualitative research has suggested that incentives associated with money,

convenience and fun are typically most popular among drivers (Loumidi, Mittag,

Lathrop, & Althoff, 2011). Moreover, vehicle-related incentives, such as reductions in

insurance premiums and vehicle registration costs or fuel vouchers, are effective

approaches to structuring incentives (Michael Regan, Mitsopoulos-Rubens, Haworth,

& Young, 2002; Vaezipour et al., 2017). Interestingly, incentives associated with

competition, such as leader boards and challenges, have been found to be less popular,

however when explored further, competition with oneself was perceived as more of a

motivation than competing with other drivers (Loumidi et al., 2011).

Following on from this point, researchers and practitioners are increasingly

exploring the use of gamification principles to engage users and motivate system use

and behavioural change. Gamification refers to the “use of game design elements in a

non-game context” (Deterding, Sicart, Nacke, O'Hara, & Dixon, 2011, p. 11). With

12 Chapter 1: Introduction
respect to in-vehicle technology, gamification has most commonly involved the use of

tasks and challenges, performance feedback, scoring systems and leader boards.

Examples include the TomTom Curfer, Flo and Samsung S-Drive smartphone

applications. To date, the limited available research suggests that while gamification

may improve initial system engagement, there is a paucity of evidence that measure

long-term effects on system use or on eco-safe driving behaviour (Magana & Munoz-

Organero, 2015; Rapp, 2015). In any case, it has been acknowledged that the use of

gamification requires careful consideration of the potential for driver distraction from

the driving task (Diewald, Möller, Roalter, Stockinger, & Kranz, 2013).

1.3 RESEARCH GAP

As stated previously, the development of in-vehicle HMIs represents a

promising approach for providing relevant and continuous, real-time advice and

feedback to drivers. However, there is a paucity of research to date that has examined

the use of a single in-vehicle HMI to concurrently improve both eco-driving and safe

driving behaviour. Moreover, there is limited understanding of the features and

functions of such technology that are desired by drivers and that increase the likelihood

of acceptance and voluntary system use.

This current program of research attempts to address this gap in the research by

designing, developing and evaluating an in-vehicle HMI that provides real-time advice

and feedback to drivers regarding their eco-safe driving behavior (see Figure 1.4). In

addition, a user-centered design approach, whereby intended users are involved in all

phases of system development and evaluation (see following section for more

information), is adopted to increase the likelihood of driver acceptance of the system.

Thus, the primary challenge of this task is to develop a system that has a positive

impact on eco-safe driving, is appealing to drivers, and has minimal distractive impact.

Chapter 1: Introduction 13
Figure 1.4. Diagram of research gap addressed by this research

1.4 THEORETICAL FRAMEWORK

After a thorough examination of existing theories related to driver acceptance,

the Technology Acceptance Model (TAM) (Davis, 1989) was selected as the

theoretical framework for the current program of research. This theory was chosen due

to its proven ability to capture the concepts of perceived usefulness, usability, ease of

use and intention to use, as measures of driver acceptance. In addition, the user-centred

design approach to technology development was also used to guide both the design,

development and evaluation of the system. Finally, a number of attitudinal and

perception constructs, such as those relating to eco-safe driving, the environment and

technology, were also investigated. A more detailed explanation of the TAM and user-

centred design approach is presented in the following sections.

1.4.1 Technology Acceptance Model (TAM)


Recent developments in the field of in-vehicle technology have led to an

increased interest in designing various advanced driver assistance systems (ADAS) to

support driving tasks. While some systems have been integrated into new vehicles,

others have been incorporated into aftermarket products such as smartphone

applications and head-up displays (e.g. Navdy, Samsung S-Drive). However, this

14 Chapter 1: Introduction
growth in initiatives has given rise to a range of issues regarding poorly designed and

misused systems, which have consequently been plagued with problems such as

limited driver acceptance and poor implementation rates (Lee & Seppelt, 2009).

The effectiveness of an in-vehicle HMI is therefore highly dependent on driver

acceptance of the system, and in particular on the perceived usefulness, ease of use,

usability and intention to use the system. Broadly, driver acceptance has been defined

as “the degree to which an individual incorporates the system in his/her driving”

(Adell, 2009, p.31). Over the past few decades a number of theories and models of

technology acceptance have been developed. Among the most popular of the early

approaches was the Technology Acceptance Model, or TAM, which focuses on how

perceptions regarding the usefulness and ease of use of computer information systems

impact upon attitudes towards use, intentions to use and subsequent actual use (Davis,

1989).

Figure 1.5 provides a schematic representation of the relationship between the

various TAM constructs. Perceived usefulness refers to the degree to which an

individual perceives there to be meaningful and tangible benefits associated with using

the system, while perceived ease of use refers to the perceived degree of effort

associated with using the system (Davis, 1989). Both constructs are influenced by a

myriad of external factors, including but not limited to, device characteristics,

individual motivations and demographic characteristics, all of which influence the

likelihood of technology adoption (Davis, 1989; Ghazizadeh & Lee, 2014).

Chapter 1: Introduction 15
Figure 1.5. Technology Acceptance Model Framework
Source: Davis (1989)

Thus, the TAM not only specifies external factors that may predict technology

acceptance and subsequent usage, but can also provide insight into factors preventing

or restricting the adoption of a particular system. As such, the model can effectively

be used by researchers and practitioners to identify ways to improve a technological

system (Davis, 1989). More recently, a number of researchers have extended the TAM

to include a variety of additional external factors believed to underpin acceptance or

intentions to use, such as device characteristics, driver characteristics and driving

behaviour (Ghazizadeh & Lee, 2014).

The TAM has been previously validated in a range of studies of user acceptance,

including with in-vehicle systems, computer systems (e.g., websites, software), and

information systems (e.g., e-learning, mobile applications) (Taylor & Todd, 1995;

Venkatesh & Davis, 2000). Moreover, replication studies have suggested that the TAM

is “a valid and robust model” (King & He, 2006, p.740). However, the narrow focus

of the theory has resulted in many researchers adapting the theory in order to apply it

to various other contexts (Osswald, Wurhofer, Trösterer, Beck, & Tscheligi, 2012),

such as the TAM2 (Venkatesh & Davis, 2000), TAM3 (Venkatesh & Bala, 2008) and

more recently the Unified Theory of Acceptance and Use of Technology (UTAUT)

16 Chapter 1: Introduction
(Venkatesh, Morris, Davis, & Davis, 2003). The common factor among these models

is the behaviour analysed, namely that of technology usage and the knowledge of what

stimulates acceptance.

The importance of driver acceptance cannot be downplayed. According to Sauer

(1993, p.4), driver acceptance and subsequent adoption of a technological system is

“the single most important factor in determining success or failure of information

systems and technologies”. This point provides a meaningful segue into a discussion

of the importance of adopting a user-centred approach to system design to increase the

likelihood of driver acceptance of in-vehicle HMIs and motivate greater use (Maguire,

2001).

1.4.2 User-Centred Approach to Design


User-centred design refers to an iterative approach in which the needs,

motivations and limitations of the intended end users of a system are the driving force

behind the development of new technologies, and are considered at all stages of the

design process (Maguire, 2001). The fundamental importance of involving the user in

the design of new technology is not a new concept. Indeed, Norman and Draper (1986,

p.61) argued that “from the point of view of the user, the interface is the system …

[and] the needs of the users should dominate the design of the interface, and the needs

of the interface should dominate the design of the rest of the system”.

The International Organisation for Standardisation (ISO) (2010) has highlighted

a number of key principles and recommendations for user-centred system design.

Overall, these principles characterise the process of user-centred design into a number

of stages (see Figure 1.6). Specifically, it is argued that system developers must:

(i) analyse and comprehensively understand the context of the whole user

experience (e.g., tasks involved; organisational, social and

Chapter 1: Introduction 17
environmental context of use; needs, motivations and limitations of

intended users);

(ii) actively involve users in all stages throughout the development process;

(iii) conduct user-centred evaluations;

(iv) understand that the process is iterative and refine systems using the same

user-centred approach; and,

(v) adopt a multidisciplinary approach that includes a design team with a

variety of skills and perspectives.

Figure 1.6. The User-Centred Design Approach


Source: adapted from ISO (2010)

Therefore, the design and development of any in-vehicle HMI must consider the

perspectives of intended users in order to achieve a system that is perceived to be useful

and easy to use, and that will be associated with high intentions to use (Francois,

Osiurak, Fort, Crave, & Navarro, 2017). This research adopts a user-centred design by

including users at each stage of the design, development and evaluation, and by

making driver acceptance of the technology a focal point of the proposed evaluation.

18 Chapter 1: Introduction
1.5 RESEARCH OBJECTIVES

The overall objective of the current research is to design, develop and evaluate

an eco-safe in-vehicle HMI to reduce fuel consumption and promote safe driving

behaviour. Moreover, this research aims to identify feedback and device characteristics

associated with greater levels of driver acceptance in order to provide a foundation to

inform the design and development of such systems and encourage greater intentions

to use. This is achieved by adopting a user-centred design approach (ISO, 2010), for

the development of the system, as well as using the TAM as the theoretical framework

for the research (Davis, 1989).

In addition, an objective of this research is to develop an in-vehicle HMI that not

only has a positive impact on eco-safe driving, but that minimises adverse impacts on

driver distraction. That is, the research seeks to address a gap in the previous

development of such systems, by considering inadvertent negative impacts associated

with system use. This is also achieved through the adoption of a user-centred approach

to the design, development and evaluation of the system. Finally, this research aims to

examine the impact of incentives and social competition on eco-safe driving behaviour

and intentions to use the system. By examining theoretical construct such as driver

acceptance and incentives, this research acknowledges that system effectiveness, in

terms of impacts of behaviour, may not be sufficient in ensuring system use among

drivers given the documented barriers to implementation.

Taken together, this research seeks to answer the fundamental question: how can

an in-vehicle HMI make driving greener and safer? Specifically, the following

research questions are addressed in this program of research:

RQ1.What is the relevant information that should be conveyed to drivers in an in-

vehicle HMI to improve eco-safe driving?

Chapter 1: Introduction 19
RQ2. What impacts do advice and/or feedback characteristics of an in-vehicle eco-

safe HMI have on:

a. Driver acceptance of the technology?

b. Driver workload?

c. Eco-safe driving performance?

RQ3. What impacts do incentives associated with using an in-vehicle eco-safe HMI

have on eco-safe driving performance?

1.5.1 Research Approach


These research questions are addressed by three studies in this thesis. Due to the

multidisciplinary nature of the research, a mixed-method design consisting of both

qualitative and quantitative approaches is adopted (see Figure 1.7). The design and

development of the prototype system was informed by a critical review of the literature

(Study 1), as well as a two-phased qualitative research component, including focus

groups with drivers and an expert workshop, whereby various feedback and device

characteristics were explored (Study 2). Finally, a driving simulator experiment and

driver survey was conducted to assess the impact of advice and feedback messages

from the prototype system on real-time eco-safe driving behaviour, driver acceptance

and driver workload (Study 3). This study also sought to examine the additional effect

of incentives and social competition on eco-safe driving behaviour. More information

regarding the research design and methodology of each study can be in Figure 2.1 in

Chapter 2.

20 Chapter 1: Introduction
Figure 1.7. Mixed-Method Design Approach

1.6 RESEARCH CONTRIBUTION AND SIGNIFICANCE

The research questions and the research approach outlined in Section 1.4 resulted

in a number of contributions relevant to the field of automotive human-machine

interfaces and road safety. Firstly, this research led to the design, development and

evaluation of a new in-vehicle HMI to improve both eco-driving and safe driving. As

stated previously, existing in-vehicle HMIs have typically addressed these behaviours

in isolation, with concerns having been raised regarding the potential inadvertent

negative impacts of eco-driving in-vehicle HMIs on driver safety. By developing an

HMI that addresses both eco-driving and safe driving, while limiting driver distraction,

this research seeks to address this gap in existing systems that are available to drivers.

Secondly, this research contributed to a more comprehensive understanding of

driver needs and requirements associated with eco-safe in-vehicle HMI and barriers to

driver acceptance. This is particularly important given the paucity of information

regarding the features and functions of eco-safe in-vehicle systems that are desired by

drivers, as well as information on how to enhance driver acceptance and motivations

Chapter 1: Introduction 21
to use eco-safe in-vehicle systems. This research also explored the use of incentives

and competition to facilitate a higher degree of intention to use such systems. By

examining driver acceptance and motivation, this research acknowledges that system

effectiveness, in terms of impacts on driver behaviour, may not necessarily ensure high

levels of driver acceptance.

Thirdly, this multidisciplinary program of research contributes new and valuable

knowledge and enhances our understanding of the nature and content of design

features regarding the advice and feedback systems. Such information can be utilised

to improve design characteristics and subsequently increase driver acceptance,

minimise impacts of driver workload and improve behavioural outcomes. This

understanding also provides a foundation to inform the future design and development

of novel in-vehicle HMIs and identify the design characteristics of the most acceptable

an effective in-vehicle HMI for improving fuel efficiency and road safety.

This thesis added weight to the utility of the user-centred design approach, and

highlighted the validity and reliability of using the TAM as a theoretical framework

when developing and evaluating in-vehicle HMIs. Researchers and practitioners

should continue to apply these methods, including to the design process, associated

tools and qualitative and quantitative evaluation methodologies, to ensure that

achieving high levels of driver acceptance and minimal impacts on driver distraction

remain the focal points of in-vehicle HMI development.

This research contributes to the top transport-related research objectives, as

outlined by the Australian Government’s Science and Research Priorities (2015, p. 2).

Specifically, the report highlighted the importance of conducting research that

examines the development of “low cost, reliable, resilient and efficient transport

systems that meet the needs of businesses and enable sustainable mobility, while

22 Chapter 1: Introduction
lowering carbon emissions and other pollution … for domestic and global markets”.

In addition, this research addresses three of the Australian Research Council’s research

priorities (2015). That is, the development of this simple but innovative eco-safe

initiative will contribute to an ‘environmentally sustainable Australia’ by reducing

transport related emissions, ‘promote and maintain good health’ by reducing and

preventing the health impacts of transport pollution and road crashes, and assist in

creating ‘frontier technologies for building and transforming Australian industries’ by

developing an eco-safe in-vehicle system.

Finally, this multidisciplinary research is significant because it seeks to reduce

road trauma, fuel consumption and emissions using a single intervention, and thus

makes a significant contribution to the improvement of both road safety and the impact

of vehicle emissions on the environment. Moreover, the research seeks to increase our

understanding of the specific characteristics an eco-safe in-vehicle HMI should include

in order to motivate drivers to improve their eco-safe driving performance, and thus

will inform best-practice for the design and development of such systems.

1.7 DEMARCATION OF SCOPE

This current research is part of a larger Australian Research Council (ARC)

Discovery Grant (No. DP140102895) involving a multidisciplinary research team

titled “CoopEcosafe: A new cooperative, green and safe driving system”. The overall

project consists of three distinct phases:

 Phase I: Development of the CoopEcoSafe model. The aim of this phase is to

develop a mathematical model of driving behaviour that optimises both eco-

driving and safe driving behaviour. This phase is being conducted by another

PhD candidate at the Centre of Accident Research and Road Safety –

Queensland (CARRS-Q).

Chapter 1: Introduction 23
 Phase II: Design and development of the CoopEcoSafe HMI. The aim of this

phase is to design, develop and evaluate a cooperative in-vehicle HMI to

improve eco-safe driving style. This phase is the focus of the current thesis.

 Phase III: Estimation of the micro and macro benefits of the CoopEcoSafe

HMI. The aim of this phase is to evaluate the network-wide impact and long-

term benefits of the system, including willingness to use the system, as well as

impacts on fuel consumption, emissions and driving behaviour.

As stated, the current thesis focuses on the second phase of the larger program

of research. Specifically, this thesis adopts a user-centred design approach to design,

develop and evaluate a cooperative eco-safe in-vehicle HMI using the eco-safe driving

algorithm from Phase I. The current thesis does not attempt to model optimal fuel

consumption and safe driving behaviour or assess the more long-term benefits of using

the developed CoopEcoSafe system. As outlined above, these aspects are outside the

scope of the current research and are addressed in Phase I and III of the larger program

of research.

1.8 OUTLINE OF DISSERTATION

In Chapter 2, the research methodology for the qualitative and quantitative

components of the research are outlined. An overview of the specific procedures, data

analysis approaches and sample characteristics is also provided. Finally, the chapter

outlines the research questions and hypotheses addressed in studies one, two and three.

This thesis by publication contains a number of papers which are either

published or currently under review. In Chapter 3, a critical review of the literature is

conducted (Study 1), identifying the research gap, summarising existing approaches to

eco-driving, discussing potential conflicts with safety and reviewing evaluations of the

effectiveness of in-vehicle HMIs on fuel consumption and safety. This review seeks

24 Chapter 1: Introduction
to develop a comprehensive understanding of the relationship between eco-driving and

safety, as well as identifying the best practice principles associated with eco-safe in-

vehicle HMI systems to inform the subsequent studies.

Paper 1: Vaezipour, A., Rakotonirainy, A., & Haworth, N. (2015). Reviewing in-
vehicle systems to improve fuel efficiency and road safety. Procedia
Manufacturing, 3, 3192-3199. https://doi.org/10.1016/j.promfg.2015.07.869

In Chapters 4 and 5, a user-centred design approach is used to inform the

development of a prototype eco-safe in-vehicle HMI. This study utilised a qualitative

research approach, including user focus groups (Study 2).

Paper 2: Vaezipour, A., Rakotonirainy, A., Haworth, N., & Delhomme, P. (2017).
Enhancing eco-safe driving behaviour through the use of in-vehicle human-machine
interface: A qualitative study. Transportation Research Part A: Policy and
Practice, 100, 247-263. https://doi.org/10.1016/j.tra.2017.04.030

Paper 3: Vaezipour, A., Rakotonirainy, A., Haworth, N. (2016). Design of a gamified


interface to improve fuel efficiency and safe driving. In DUXU 2016: Design, User
Experience, and Usability: Novel User Experiences (Vol. 17, No. 9747, pp. 322-332).
Part of the Lecture Notes in Computer Science series. Springer International
Publishing. https://doi.org/10.1007/978-3-319-40355-7_

In Chapters 6 and 7, a user-centred evaluation included an experiment in a state-

of-the-art driving simulator and a driver survey (Study 3). This study adopts a mixed-

methods design approach, including objective measures of eco-safe driving

performance from the driving simulator and a driver survey examining a range of

subjective measures.

Paper 4: Vaezipour, A., Rakotonirainy, A., Haworth, N., & Delhomme, P. (2017). A
Simulator Evaluation of in-vehicle Human Machine Interface for eco-safe driving.
Transportation Research Part A: Policy and Practice (Under-review).

Paper 5: Vaezipour, A., Rakotonirainy, A., Haworth, N. (2017). Evaluating the effect
of incentive on eco-safe driving behaviour. Journal of Safety Research (Submitted).

Chapter 1: Introduction 25
Finally, Chapter 8 provides a discussion and concluding statements associated

with findings from all of the studies, including real-world implications, limitations and

directions for future research.

Table 1.1, provides an overview of the three studies performed as part of the

overall program of research, as well as their relation to the thesis chapters, research

questions and publications.

Table 1.1. Research overview and relation to research questions and publications

Studies Paper Title Research Questions


Study 1: Literature Paper 1: Reviewing in-vehicle RQ1. What is the relevant
review systems to improve fuel information that should be conveyed
(Chapter 3) efficiency and road safety to drivers in an in-vehicle HMI to
(published) improve eco-safe driving?

Study 2: Focus groups Paper 2: Enhancing eco-safe RQ1. What is the relevant
and expert workshop driving behaviour through the information that should be conveyed
(Chapters 4-5) use of in-vehicle human- to drivers in an in-vehicle HMI to
machine interface: A qualitative improve eco-safe driving? (from
study user perspective)
(published)

Paper 3: Design of a gamified


interface to improve fuel
efficiency and safe driving
(published)

Study 3: Driving Paper 4: Evaluating Eco-Safe RQ2. What impacts do advice


simulator experiment and in-vehicle Human Machine and/or feedback characteristics of an
driver survey Interface: A Simulator Study in-vehicle eco-safe HMI have on:
(Chapters 6-7) (under-review) a. Driver acceptance of the
technology?

b. Driver workload?

c. Eco-safe driving
performance?

Paper 5: A Simulator RQ3. What impacts do incentives


Evaluation of the Effect of associated with using an in-vehicle
Incentive on Adoption and eco-safe HMI have on eco-safe
Effectiveness of an In-vehicle driving performance?
Eco-Safe Driving HMI
(submitted)

26 Chapter 1: Introduction
Chapter 2: Research Design

2.1 INTRODUCTORY COMMENTS

This thesis uses a driver-centred approach to the design, development and

evaluation of an in-vehicle HMI, and this chapter describes the research design

methodology adopted to achieve the aims outlined in Chapter 1. This chapter aims to

provide a more comprehensive overview of the research methodologies undertaken as

part of this program of research, which is typically beyond the scope of a journal article

or conference paper methods section. The following sections provide an overview of

the three studies performed as part of the overall program of research, as well as

describing the methodology of each of the studies, including the design, participants

and procedures.

2.2 RESEARCH DESIGN AND METHODOLOGY

This research used a mixed-methods design consisting of both qualitative and

quantitative approaches. Figure 2.1 illustrates the user-centred design approach

adopted. User acceptance is critical for the success of any information system, and

thus, employing a user-centred approach and involving drivers in the design and

development of the system helps to ensure user requirements are considered as early

as possible in this design process (Francois et al., 2017).

Study 1 is a critical literature survey aimed at gaining a comprehensive

understanding of the relationship between eco-driving and safety, as well as

identifying eco-safe driving parameters for potential inclusion in the design and

development of an in-vehicle system. The review also sought to contribute to an

enhanced understanding of the influence of factors such as driver acceptance and

Chapter 2: Research Design 27


workload on the use of in-vehicle systems, as well as the potential motivational impact

of gamification and competition on rates of system use.

Figure 2.1. Research design as it relates to the user-centred design approach

28 Chapter 2: Research Design


Study 2 sought to expand upon this evidence in a two-phased qualitative

approach involving focus groups with drivers and a workshop with experts from a

range of relevant disciplines. This study also sought to identify user-preferred design

principles and feedback characteristics associated with eco-safe driving in-vehicle

HMIs. The findings from Study 1 and 2 were used to inform the development of an

eco-safe driving in-vehicle HMI to be evaluated in the final phase of the research.

Finally, in Study 3, a quantitative evaluation of the system was conducted,

including both a driving simulator experiment and driver survey. The driving simulator

experiment evaluated three variations of an eco-safe in-vehicle HMI, differing with

regard to the type of advice and feedback messages provided to drivers: (i) visual

advice only; (ii) visual feedback only; and (iii) visual advice and feedback.

Specifically, this study sought to examine the impact of advice and feedback on driver

acceptance of the technology, driver workload, and a range of eco-safe driving

behaviours. Moreover, the study investigated the impact of incentives associated with

using the system on eco-safe driving performance.

The following sections discuss the methodology of each of the three studies in

greater detail, including the design, participants, procedures and ethical considerations.

2.3 STUDY 1: LITERATURE REVIEW

Paper 1: Vaezipour, A., Rakotonirainy, A., & Haworth, N. (2015). Reviewing


in-vehicle systems to improve fuel efficiency and road safety. Procedia
Manufacturing, 3, 3192-3199. https://doi.org/10.1016/j.promfg.2015.07.869

Literature search criteria. This study reviewed existing literature on in-vehicle

HMI systems and their potential to improve eco-driving and safety. A particular focus

of the review was identifying eco-safe parameters for inclusion in such systems. In

addition, the review also sought to investigate the impact of a range of factors, such as

Chapter 2: Research Design 29


driver acceptance, workload, gamification principles and competition, on system use.

A number of keywords were used in combination with one another in order to identify

relevant studies, and are outlined below in relation to a range of topics:

 Eco-driving, green driving, fuel efficiency, fuel consumption, environment

 Traffic safety, road safety, traffic crash

 In-vehicle system, human machine interface

 Driver acceptance, usability, usefulness, ease of use, gamification

Papers published from 1970 onwards were considered for inclusion in the

review, as this was the year when driver feedback in-vehicle systems associated with

fuel consumption were first introduced. Published and grey literature from multiple

disciplines were included, however only papers written in English were considered.

The literature review paper was written and published in 2015, and so includes studies

up to that time.

Search Databases. Relevant papers were identified through searches of a

number of scientific databases (e.g., Google Scholar, Science Direct, Web of Science,

PsycINFO, Transportation Research Information Services). In addition, the references

cited in each identified paper were reviewed to discover additional relevant studies.

2.4 STUDY 2A: USER FOCUS GROUPS

Paper 2. Vaezipour, A., Rakotonirainy, A., Haworth, N., & Delhomme, P. (2017).
Enhancing eco-safe driving behaviour through the use of in-vehicle human-machine
interface: A qualitative study. Transportation Research Part A: Policy and
Practice, 100, 247-263. https://doi.org/10.1016/j.tra.2017.04.030

Paper 3. Vaezipour, A., Rakotonirainy, A., Haworth, N. (2016). Design of a gamified


interface to improve fuel efficiency and safe driving. In DUXU 2016: Design, User
Experience, and Usability: Novel User Experiences (Vol. 17, No. 9747, pp. 322-332).
Part of the Lecture Notes in Computer Science book series. Springer International
Publishing. https://doi.org/10.1007/978-3-319-40355-7_

30 Chapter 2: Research Design


Design. Consistent with a user-centred design approach, a series of focus groups

were conducted with drivers regarding their perceptions and attitudes towards various

design and feedback characteristics associated with eco-safe in-vehicle systems. Focus

groups have been demonstrated as an efficient approach to eliciting comprehensive,

in-depth discussions from participants (Braun & Clarke, 2006). The purpose of the

focus group sessions was to identify user-preferences in relation to these design and

feedback characteristics, in order to inform the design and development of an eco-safe

in-vehicle HMI for testing in Study 3.

Participants. Six focus groups were conducted with 34 licensed drivers from

Brisbane, Australia. Focus groups contained between five and six participants in each

session. The sample size was determined upon the occurrence of data saturation,

whereby limited new additional information was being obtained from each additional

focus group session (Guest, Bunce, & Johnson, 2006; O'Reilly & Parker, 2012;

Walker, 2012).

Participants were required to hold a current drivers licence to be eligible to

participate. There were no restrictions regarding age of driver, licence type or country

of origin of licence. The sample consisted of 53% males and had a mean age of 32.1

years (SD = 10.4 years; range = 19-61 years). The majority of participants (82%) held

an open drivers licence (i.e., unrestricted), while 12% were provisional licence holders

(i.e., unsupervised but restricted) and 6% held a learner’s licence. Roughly equivalent

numbers of drivers reported driving less than five hours per week (35%), six to ten

hours per week (29%) and more than eleven hours per week (35%). Moreover, almost

a third of participants (29%) reported being very experienced with technology, while

47% reported having average experience and 24% categorised themselves as

inexperienced.

Chapter 2: Research Design 31


Procedure. Participants were recruited from the city of Brisbane, in Queensland,

Australia, primarily through university email lists, snowball sampling methods, social

media posts and personal approaches at university campuses. Participants initially

completed a pen-and-paper survey that collected information regarding demographics,

driving experience and experience with technology. Focus groups ran for

approximately 60 minutes and participants received a movie voucher as

reimbursement for their time and contribution. Each focus group session was

conducted by the same facilitator, along with another researcher who took notes. The

focus group sessions were audio-recorded using a digital audio-recorder and later

transcribed by a professional stenographer. Upon the completion of transcription, all

audio files and identifying data were deleted or destroyed.

The facilitator began each focus group by describing a range of fuel-efficient and

safe driving behaviours. Participants were encouraged to discuss which of these

behaviours they would be interested in monitoring while driving. Following this, the

initial eco-safe HMI design concepts were presented to participants and briefly

described with the aid of images as stimuli (see Vaezipour, Rakotonirainy, & Haworth,

2016 ,Chapter 5 for more details). A series of open-ended questions elicited feedback

from participants regarding each in-vehicle interface. Questions were semi-structured

and based on the TAM focus group discussion guide cited in Mitsopoulos-Rubens and

Regan (2014).

A full list of the questions and prompts used in the focus group study can be

found in Appendix A. Questions were asked in relation to each of the concept HMIs

and covered the following topics: (i) perceived system acceptability (e.g., usefulness,

usability, ease of use); (ii) perceived impact on eco-safe driving behaviours; (iii)

32 Chapter 2: Research Design


perceived difficulties associated with the interface and ideas for improvements; and,

(iv) self-reported intentions to purchase and use.

Data Analysis. The focus groups transcriptions were imported into the Nvivo 11

software program. A thematic analysis was conducted to identify patterns across

responses and their relation to research questions (Braun & Clarke, 2006; Mark &

Yardley, 2004). All ideas, expressions, and words were classified into content themes

on the coding grid, with each theme consisting of units comprising one or more words

with similar meaning (Weber, 1990). Emerging themes that did not fit within this

framework were also considered during the coding process. This review and

refinement process led to the development of the final themes and sub themes.

A second researcher also coded the transcriptions in order to validate the initial

interpretations. Discrepancies were thoroughly discussed and guided the thematic

development, resulting in a set of coherent themes that reflected a comprehensive and

precise set of meanings for participant comments. Improvements to the themes were

made until the inter-coder agreement was 100%.

Commentary on the difference between Paper 2 and 3. Both Paper 2 and 3

report on findings from the focus group study. Paper 2, which is presented in Chapter

4, explores driver perceptions and attitudes toward eco-safe driving behaviour, as well

as driver acceptance of, and intention to use, various eco-safe HMI concept systems,

including attitudes towards various design and feedback characteristics. On the other

hand, Paper 3, which is presented in Chapter 5, examines the potential impact of

incentives and competition on driver acceptance of eco-safe in-vehicle HMI.

2.5 STUDY 2B: EXPERT WORKSHOP

Design. The second phase of the qualitative component of this research involved

a workshop with experts from a variety of disciplines (intelligent transportation

Chapter 2: Research Design 33


systems, traffic psychology, human computer interaction, user experience (UX) design

and industrial design). Expert workshops have been extensively used as part of the

design process for technological development and have been demonstrated as an

efficient approach to performing heuristic evaluations (Mankoff et al., 2003; Nielsen,

1994), to assess system usability prior to user testing. The primary aim of the

workshop was to elicit feedback associated with the first iteration of an eco-safe in-

vehicle system (see Figure 2.2 and video). This initial prototype system was designed

and developed based on the findings from the focus group study discussed in the

previous section.

Figure 2.2. First iteration of eco-safe in-vehicle HMI prototype

Link to video (https://vimeo.com/239547324 ) & QRCode

Specifically, the expert workshop aimed to identify design issues and generate

ideas for improving the system, through a process of brainstorming and constructive

criticism. A key focus was exploring alternative ways of displaying real-time eco-safe

34 Chapter 2: Research Design


driving advice and feedback to drivers. In addition, feedback and ideas for

improvement also focussed on the overall aesthetics of the design of the interface.

Participants & Procedure. Nine experts participated in the workshop. A

convenience sampling approach was adopted, with experts recruited from relevant

faculties and research centres from the Queensland University of Technology. A

multidisciplinary approach was adopted, with experts from a variety of disciplines

involved in the workshop, including intelligent transportation systems, traffic

psychology, human computer interaction, user experience (UX) design and industrial

design. The workshop ran for approximately two hours and participants received a $50

gift voucher as a reimbursement for their time and contribution.

The facilitator began the workshop by outlining the intended purpose of the eco-

safe in-vehicle HMI. Following this, a brief summary of findings from the focus group

study were provided with particular reference to how the findings had informed the

development of the prototype system. A short video detailing the prototype eco-safe

in-vehicle HMI, as well as a number of other examples of in-vehicle HMIs and

interactive systems, were also shown to the experts.

Participants then split into smaller groups to comprehensively discuss the

prototype eco-safe in-vehicle HMI. These discussions were open and fluid; the experts

were provided with a range of possible topics to discuss (see Appendix B), but also

encouraged to nominate their own discussion points. In addition, they were encouraged

to discuss potential problems with the prototype and suggest potential solutions and

improvements. A range of materials was provided to enable participants to visually

represent or create their ideas, including pens and markers, butchers paper, clay putty,

craft sticks, and a range of other assorted textiles. In addition, participants were

Chapter 2: Research Design 35


requested to detail each idea on an individual post-it note. The brainstorming session

lasted approximately 45 minutes.

Following the brainstorming session, the smaller groups came back together and

shared their ideas with the rest of the participants. A feedback grid was used to

systematically collate all feedback elicited during the workshop into four categories:

(i) positive aspects of the prototype; (ii) constructive criticisms; (iii) unique additional

ideas; and (iv) questions. Specifically, the feedback grid was drawn onto a large

whiteboard, with the idea-specific post-it notes scribed during the brainstorming

session placed into their respective categories. As each idea or concept was placed on

the feedback grid, similar notes from other groups were identified and also placed on

the grid, before the group discussed the idea or concept in more detail. This group

discussion lasted approximately one hour.

Data Analysis. A thematic analysis of the ideas and concepts from the feedback

grid was conducted to identify common themes and suggestions for improvements.

These themes were then assessed against findings from the focus groups as well as the

International Organisation of Standardisation’s guidelines for ergonomic aspects of in-

vehicle presentation for transport information and control systems. The system was

then refined further through an iterative process involving a series of rapid prototypes

(see Figures 2.3- 2.4 below), a pilot survey with CARRS-Q researchers, as well as

discussions with the PhD candidate supervisory team (Appendix C). Taken together,

these steps led to the design and development of the in-vehicle HMI prototype

evaluated in Study 3. The following section provides an overview of the prototype eco-

safe in-vehicle HMI, followed by an explanation of the system evaluation conducted

in the driving simulator (Study 3).

36 Chapter 2: Research Design


Figure 2.3. Examples of iterations of the eco-safe in-vehicle HMI prototype screens.

Figure 2.4. Examples of the eco-safe in-vehicle HMI prototype housing.

2.6 ECO-SAFE IN-VEHICLE HMI DESCRIPTION

The main goal of the in-vehicle HMI was to encourage the adoption of an eco-

safe driving style. The system was developed on the assumption of access to

technology that allows the use of real-time traffic information, such as vehicle-to-

vehicle (V2V) and vehicle-to-infrastructure (V2I) technology. This technology allows

the vehicle to communicate with traffic signals and stop signs, as well as other vehicles

on the road. The provision of real-time advice and feedback to drivers subsequently

promotes greater anticipation by drivers of changes in traffic conditions (Delhomme

et al., 2013; Popiv, 2012)

Chapter 2: Research Design 37


Overall, the system was designed to act as an eco-safe driving companion, with

features aimed at giving the system a sense of personality, an approach commonly

used in social robotics and persuasive technologies to create affective connections

between humans and technology (Fogg, 2002; Pieroni et al., 2015). The system

consisted of an Android smartphone (i.e., Samsung Galaxy S) which was placed into

a 3D printed housing and designed to represent the “head” of the system. The

smartphone screen was designed to represent the “face” of the system, with the screen

displaying “eyes” and a “mouth”, which change to assist in the provision of feedback.

As can be seen in Figure 2.3, the iterations of the eco-safe in-vehicle HMI

prototype screens were associated with both the ‘face’ and behavioural icons.

Iterations considered the number of elements included in the display as well as the

intuitiveness of the behavioural icons. As can be seen in Figure 2.4, the iterative

process to develop the housing involved initial pen-and-paper sketches (see Figure 2.4

a), followed by digitised illustrations (see Figure 2.4b). A 3D printer was then used to

develop a prototype of the housing (see Figure 2.4c), in which the smartphone and

LED lighting was inserted to create the final prototype (see Figure 2.4d-f).

The LED strips were controlled with Arduino UNO as a microcontroller of the

system connected to the network through an Ethernet cable. A unique IP address was

programmed into the module through another Ethernet cable in the driving simulator.

The component used includes; the Arduino UNO and its Ethernet module, Bread

board, MOSFET (IRF540), Resistor, RGB LED strip, and 12 Volt power supply. The

Pulse Wide Modulate (PWM) is used to control the brightness and intensity of the

LED colours. The LED strip is connected to the MOSFET (IRF540) Amplifier in order

to control the 12 Volt input with the PWM pin of the Arduino (Figure 2.5)

38 Chapter 2: Research Design


Figure 2.5. Diagram of the circuit

Examples of advice and feedback screens used by the system are shown in Figure

2.6 and 2.7. Advisory messages involved screens that presented an icon and simple

text (e.g., smooth braking) on a yellow background, and were projected to drivers

leading up to specific traffic events, such as when approaching traffic lights, to

encourage an anticipatory driving style.

Figure 2.6. Examples advice and feedback messages

Chapter 2: Research Design 39


Figure 2.7. Examples advice and feedback messages in context
Link to video (https://vimeo.com/240435824) & QRCode

Conversely, human-like feedback messages included the “face”, as well as a

smaller icon and illumination of the LED lights. This feedback differed depending on

whether the behavioural performance of the driver was appropriate (i.e., happy face

and a green light) or inappropriate (i.e., sad face and a red light). Feedback messages

were presented in real-time and were designed to enhance driver knowledge,

awareness and learning, and subsequently increase the likelihood that drivers would

be motivated to improve their driving performance (Gardner, Whittington, McAteer,

Eccles, & Michie, 2010; Shute, 2008). It is worth mentioning that the design of the

current system complies with Queensland traffic rules and recommendations of fuel

economy driver interfaces (Manser, Rakauskas, Graving, & Jenness, 2010).

2.7 STUDY 3: DRIVING SIMULATOR EXPERIMENT & DRIVER


SURVEY

Paper 4: Vaezipour, A., Rakotonirainy, A., Haworth, N. & Delhomme, P. (2017). A


Simulator Evaluation of in-vehicle Human Machine Interface for Eco-Safe Driving.
Transportation Research Part A: Policy and Practice (Under-review)

40 Chapter 2: Research Design


Paper 5: Vaezipour, A., Rakotonirainy, A., Haworth, N. & Delhomme, P. (2017). A
Simulator Evaluation of the Effect of Incentive on Adoption and Effectiveness of an
In-vehicle Eco-Safe Driving HMI. Journal of Safety Research (Submitted)

Method. The quantitative component of this research involved a driving

simulator experiment and driver survey. The driving simulator was used to objectively

evaluate the impact of the prototype eco-safe in-vehicle HMI on a range of eco-safe

behaviours, as well as the influence of incentives on eco-safe driving behaviour.

Driving simulators represent a safe and efficient way to study driving behaviour in a

more controlled environment, particularly when on-road studies are otherwise not

feasible due to safety considerations. In addition, the accompanying driver survey

involved a range of subjective measures designed to examine driver acceptance of the

technology, driver workload, and various perceptual and attitudinal constructs.

Participants were fitted with an accelerometer to measure foot movements. This

sensor was attached to the right foot of participants to measure the acceleration in x, y

and z axes of the foot while performing tasks in the driving simulator. In addition to

data relating to aspects of driving, Facelab eye tracking equipment recorded the

coordinates of participant eye and head movements as numerical data. Due to technical

issues associated with the equipment and subsequent impacts on the accuracy of the

data, an analysis of foot and eye movement is not included in this thesis.

Participants. Forty licensed drivers (20 females, 20 males), aged between 18 and

65 years (M = 31.35, SD = 12.18) were recruited from Brisbane, Australia. While 53

participants were initially recruited, 13 participants were excluded due to motion

sickness, dizziness and nausea, which is commonplace in simulator studies (Fisher,

Rizzo, Caird, & Lee, 2011). Data on the gender, age and licence type of the Queensland

driving population was used to inform sampling numbers. The final sample comprised

14 provisional licence holders (35%) and 26 open (i.e., unrestricted) licence holders

Chapter 2: Research Design 41


(65%). There was an even number of males and females across all age groups and

licence types. All provisional licence holders were aged 18-24 years, while 16 of the

open licence holders were aged 25-39 years and ten were aged 40-65 years. More

participants (62.5%) reported driving between 10,001 and 30,000 kilometres in the

previous year, with 35% reporting driving less than 10,000 kilometres.

Materials and procedure. Participants were primarily recruited through a media

release, the university webpage, and snowball sampling. Eligibility criteria included

holding a valid Australian driver licence (i.e. P1, P2 or open). The driving simulator

experiment and driver survey took approximately two hours to complete and

participants received a $70 gift voucher as reimbursement for their participation.

The quantitative study involved three phases of data collection (Figure 2.8): (i)

a pre-drive questionnaire, (ii) simulated driving scenarios with immediate post-drive

questionnaires, and (iii) a final post-drive questionnaire. Each of the questionnaires

was designed to elicit specific types of information.

Figure 2.8. Experimental Design and Procedure

42 Chapter 2: Research Design


Phase I: Pre-drive questionnaire

The pre-drive questionnaire was the most extensive of the three questionnaire

phases. A link to the questionnaire was emailed to participants, using the Qualtrics

survey software. In addition to socio-demographic information and transport and travel

characteristics, the questionnaire included a number of attitudinal and perception

measures associated with technology, the environment, eco-driving and safe driving.

A range of self-report measures were also collected, including safe driving and eco-

driving behaviour, experience with technology, and motivations for eco-driving and

safe driving.

Phase II: Simulated driving scenarios with immediate post-drive questionnaire

In the second phase, participants completed a practice drive to familiarise

themselves with the driving simulator followed by a number of experimental drives.

In the within-subject design, each participant was exposed to five conditions which

differed with regards to the level of advice and feedback received from the cooperative

eco-safe in-vehicle HMI. They were: (1) baseline; (2) advice only; (3) feedback only;

(4) advice and feedback; and, (5) self-selected incentive and competition. The order of

the first four conditions was counter-balanced among participants, while the incentive

and competition condition was always the last. By including the baseline condition in

the counterbalancing procedure, the potential influence of learning effects on

subsequent data was controlled for.

Each drive took approximately ten minutes to complete. Before each driving

condition, the researcher explained the in-vehicle eco-safe driving system using the

InVision user interface prototyping platform on an iPad (See Figure 2.9). In all

instances, participants were instructed to drive as they normally would and follow

signs to the airport.

Chapter 2: Research Design 43


Immediately following each condition, a post-drive questionnaire was

administered. This questionnaire sought to elicit feedback regarding driver acceptance

of each of the systems, including perceived usefulness, usability, ease of use and

intentions to use, as well as self-reported workload associated with using the system.

Phase III: Final post-drive questionnaire

The final post-drive questionnaire was designed to elicit more general attitudes

and perceptions regarding the eco-safe in-vehicle HMI. These included attitudes

toward feedback and device characteristics, normative influences, motivations and

incentives to use the system, attitudes toward trust and privacy of the system,

willingness to purchase the system, and perceptions of the realism of the simulated

driving experience.

Figure 2.9. Examples of digital prototypes screens presented on iPad


Link to screens (https://invis.io/YUA22Q36A)

44 Chapter 2: Research Design


Driving simulator and scenario. The experiment was conducted in the Centre

for Accident Research and Road Safety- Queensland (CARRS-Q) advanced driving

simulator. The driving simulator is a safe and convenient method for examining driver

behaviour, which was particularly pertinent given the potentially distracting effect of

the in-vehicle interface, which would make such a study dangerous in a less-controlled,

real-world environment.

The advanced driving simulator consists of a Holden Commodore passenger

vehicle with an automatic gearbox. The simulator uses the SCANeR™ studio from

OKTAL, linked to website software with eight computers and projectors and a six

degree of freedom (6DOF) motion platform that can move and twist in three

dimensions. When seated in the simulator vehicle, drivers are immersed in a virtual

environment that includes a 180° front field-of-view, simulated rear-view mirror

images, surround-sound for engine and environment noise, a real vehicle cabin and

simulated vehicle motion using a hydraulic system. The simulator collects driver

behaviour data such as speed, acceleration, braking, steering wheel and vehicle

positioning at a rate of 20 Hz.

SCANeR™studio vehicle mode with following engine specification used in the

experiment at driving simulator (see Figure 2. 10).

• Engine type (Otto cycle),

• Fuel density (amounts to 0.74 for EURO SUPER)

• Engine inertia properties, defined by a mass and a gyration radius.

• Curve of gas command opening functions of accelerator position.

• Mechanical characteristics of starter (Torque, Revolution speed ==> Starter

power)

Chapter 2: Research Design 45


Figure 2.10. SCANeR™ studio engine specification

Fuel consumption mapped, using torque relative to speed and using the look up

table (see Figure 2.11). Engine torque was defined using two different graphs.

Specifically, engine torque with full load was defined as the torque produced by the

engine at full throttle, while negative torque was defined as resistant torque produced

by the engine, as well as resistant torque produced by other devices (such as a water

pump or alternator), with both dependent on engine speed. Engine consumption at full

throttle referred to consumption, dependant on engine speed.

Figure 2.11. SCANeR™studio fuel consumption specification

46 Chapter 2: Research Design


In order to maximise the realism and familiarity of the driving experience, the

simulated scenario consisted of urban and suburban areas similar to the Brisbane CBD

and surrounds. Standard Australian speed limit signs were used, with speed limits of

40 km/h in urban CBD sections and 60 km/h in suburban areas. Participants were

required to drive in the left lane as they typically would on the Australian roads. The

simulated scenario was approximately seven kilometres long and programmed to

incorporate various traffic events including traffic lights, stop signs, pedestrian

crossings, a change of speed limit and following another vehicle. Each drive took

around ten minutes to complete and starting locations within the scenario were

counter-balanced across conditions to further reduce the impact of learning effects on

subsequent data.

Driving performance measures. Previous research has identified a number of

primary indicators for eco-safe driving behaviour, including deceleration, acceleration,

speed, headway, and appropriate gear changes (Ericsson, 2001). In the current study,

various traffic events were simulated to elicit these behaviours from drivers, with the

exception of appropriate gear changes due to the simulator vehicle having an automatic

gearbox. In addition, the use of an automatic vehicle has the benefit of reducing the

complexity in examining influences on fuel consumption, such that it can be more

reliably attributed to acceleration and deceleration behaviour. Moreover, the results

are more generalisable to other vehicles without gears, such as electric vehicles.

With regards to eco-driving variables, a range of data were collected, including

fuel consumption and smooth acceleration and deceleration. Fuel consumption (mL/s)

was estimated for the entire drive using the driving simulator software. Total travel

time (in seconds) was also measured. Acceleration and deceleration were both

measured in metres per second squared (m/s2). In addition, acceleration was measured

Chapter 2: Research Design 47


using pedal position, as a float value between 0 (released) and 1 (completely pressed).

Deceleration was also measured as applied force on the brake pedal (in Newtons, N),

as well as a measure of overall brake use calculated by taking the total area under the

curve when plotting brake pedal input by time. Algorithms that informed the

measurement of optimal eco-safe driving were calculated previously, the details of

which are out of scope for the current paper, but the methodology can be found in

(Nouveliere, Mammar, & Luu, 2012).

Acceleration data is reported for two separate events which required drivers to

accelerate and bring the vehicle up to complete stop at STOP sign. Both events

required drivers to turn right at a standard three-way intersection, with one event

requiring drivers to wait for traffic to pass before proceeding (in a 40 km/h zone),

whereas no traffic was present at the other intersection (60 km/h zone). These two

events were chosen to examine whether there are any impacts on acceleration

behaviour depending on whether a driver needs to wait for traffic.

Similarly, deceleration data is also reported for two separate events which

required drivers to decelerate and bring the vehicle to a complete stop. Both events

took place in a 40 km/h zone, with one event involving an approach to traffic signals

in the green phase at a standard four-way intersection, which changed to amber when

drivers were five seconds from the light. The other event involved a STOP sign at a

standard three-way intersection, with drivers required to wait for traffic to pass before

proceeding. These two events were chosen to investigate behavioural differences

associated with a long-range braking scenario in which drivers had an extended

amount of time to acknowledge a forced stop (e.g., STOP sign and traffic), compared

to a short-notice braking scenario (e.g., traffic light changing to red on approach).

48 Chapter 2: Research Design


With regards to safe driving variables, headway and speed were measured. A

single headway event required drivers to follow behind another vehicle travelling at

the speed limit (40 km/h). The presence of other traffic prevented drivers from being

able to overtake the vehicles in front of them. Headway was measured as time interval

to vehicle (TIV, see Equation 1), where Di is the relative distance to the vehicle in front

in meters and V is the actual speed of the simulator vehicle. This parameter is often

used in traffic safety regulations and Australian law mandates a minimum TIV of 2.5

seconds (Queensland Transport Main Road, 2017):

TIV =Di/V [Equation 1]

Finally, speed was monitored during the entire drive in both the 40 km/h and 60

km/h zones. Speeding behaviour was measured separately for each of the zones using

mean speeds when travelling in free-flowing traffic and the percentage of time spent

exceeding the speed limit. As a side note, previous research in simulated driving

experiments investigating driver distraction has indicated that lane keeping is an

automatic task for drivers and conscious attention to lane keeping does not necessarily

determine driver safety (Garrison & Williams, 2013). Therefore, lane keeping is not

included as a driver safety performance measure in this research.

Data Analysis. Data from the driving simulator was processed and extracted

using Matlab version R2017a. This extracted data and all data from the driver surveys

were checked, coded and entered into SPSS version 23. Data coding for each question,

and subsequent scales, were recorded in a code book. Due to the forced response nature

of the questionnaires and the choice response format of the vast majority of questions,

there was limited missing or invalid data. The data were checked for outliers, and

further checks were performed by examining descriptive statistics for all variables.

Chapter 2: Research Design 49


The data were checked to test for assumptions of parametric tests, including

normality, linearity, homoscedasticity, homogeneity of variance and multicollinearity.

To adhere with these assumptions, the Likert scale data were treated as interval in

nature. Inspection of standardised residual scores and scatterplots revealed that the

assumption of linearity was met. Violations of normality, homoscedasticity and

homogeneity of variance were relatively common. Normality was assessed through

inspection of skewness and kurtosis values, 95% trimmed means, visual inspection of

histograms and the Shapiro-Wilk statistic. Outliers did not appear to have an overly

influential impact on any of the analyses. Where outliers were identified, analyses were

performed both with and without the inclusion of outlier data. In all instances, there

was no impact of the outlier values and thus they were included in the analyses.

If data did not violate assumptions, parametric tests were performed (e.g.,

ANOVA, t-test), whereas non-parametric tests were performed in instances where

assumptions were violated (e.g., Friedman’s Test, Chi-square tests, Wilcoxon Signed

Ranks Test, Fisher’s Exact Test). Where appropriate, multiple analyses and post hoc

comparisons were adjusted using the Bonferroni correction. Effect sizes were

calculated whenever possible and interpreted according to the relevant criteria as

suggested in Cohen (1992). More comprehensive details regarding the specific

analyses performed can be found in the relevant sections of Papers 4 and 5 (see

Chapters 6 and 7, respectively).

Commentary on the differences between Paper 4 and 5. The main aim of Paper

4 was to evaluate the impact of advice and feedback messages from an eco-safe in-

vehicle HMI on eco-safe driving behaviour, driver acceptance and workload.

Specifically, the study sought to identify differences between systems that provided

advice only messages, feedback only messages, and advice and feedback messages.

50 Chapter 2: Research Design


In Paper 5, the impact of incentives and competition in relation to an in-vehicle

eco-safe driving HMI on subsequent eco-safe driving behaviours were examined. In

addition, a range of demographic, attitudinal and self-reported behaviour variables

were examined as possible factors associated with driver choice of a particular system.

2.8 ETHICS APPROVAL

Ethical clearance to conduct the study was obtained from the Queensland

University of Technology Human Research and Ethics Committee, in relation to both

the qualitative focus groups and expert workshop (Approval number 1500000683) and

driving simulator study and driver survey (Approval number 1600000651).

For all phases of the research, participants gave written consent to participate

and were informed that their participation was entirely voluntary and that data would

remain anonymous and reported on in an aggregate manner only. Participants were

also provided with information regarding the nature of the research, the contact details

of the research team, and were informed that they were free to withdraw from

participation at any time without penalty.

2.9 CONCLUDING REMARKS

This chapter has provided an overview of the overall research design and mixed-

method approach of this program of research, as well as an overview of the eco-safe

in-vehicle HMI developed as a part of this thesis. The next five chapters (Chapter 3-

7) present the five papers produced from three studies which were guided by the user-

centred design approach. Please note that due to the exploratory nature of Study 2b

(Expert Workshop), the results from this component of the research project are not

published as a separate paper, but rather discussed in Section 2.5. The list of references

from each paper is collated within one comprehensive bibliography as per Queensland

University of Technology guidelines for thesis-by-publications.

Chapter 2: Research Design 51


The following chapter represents the first publication from this thesis, a review

of the literature regarding the relationship between eco-driving and safety. In addition,

the chapter also reviews evaluations of in-vehicle HMI systems aimed at improving

eco-driving and road safety, highlights key driving parameters associated with eco-

safe driving behaviour, and examines the use of competition and persuasion to

motivate eco-safe driving.

52 Chapter 2: Research Design


Chapter 3: Reviewing In-vehicle Systems to
Improve Fuel Efficiency and
Road Safety (Paper 1)

“If I have seen further it is by standing on the shoulders of Giants.”

-Isaac Newton

3.1 INTRODUCTORY COMMENTS

According to the user-centred design approach, the first step involves

comprehensively understanding the context of system use (see Figure 3.1). As such,

this chapter presents a review of the literature (Study 1) regarding the relationship

between eco-driving and safety. Research evaluating the effectiveness of in-vehicle

HMI systems aimed at improving eco-driving and safety is then reviewed, focussing

on the key driving parameters highlighted by existing studies as being influential

toward eco-safe driving behaviour. Moreover, evidence relating to driver acceptance

of in-vehicle systems and the use of competition and persuasion to motivate eco-safe

driving are also reviewed.

This chapter seeks to address Research Question 1: “What is the relevant

information that should be conveyed to drivers in an in-vehicle HMI to improve eco-

safe driving?”. As such, this chapter seeks to identify the eco-safe driving parameters

that would be most efficacious for inclusion in the in-vehicle HMI to be developed as

part of the subsequent phases of this program of research. While prior research

reviewed as part of this chapter is undoubtedly important, this information will be

complementary to the data collected during the qualitative phase of this research

(Study 2) and will then be used to inform system development.

Chapter 3: Reviewing In-vehicle Systems to Improve Fuel Efficiency and Road Safety (Paper 1) 53
Since this is a thesis by publication, the chapters that follow represent papers that

have either been published or are currently under review. As such, each individual

paper also has a separate shorter literature review which complements the literature

review presented in this chapter.

Figure 3.1 User-Centred Design Approach - Critical Review

Taken from: Vaezipour, A., Rakotonirainy, A., & Haworth, N. (2015). Reviewing in-
vehicle systems to improve fuel efficiency and road safety. Procedia
Manufacturing, 3, 3192-3199. https://doi.org/10.1016/j.promfg.2015.07.869

Publication Status: Available online 23 October 2014

Journal Quality: Paper accepted for presentation at 6th International Conference on


Applied Human Factors and Ergonomics (AHFE 2015) and the Affiliated
Conferences, AHFE 2015. Peer-reviewed papers published in Open Access Journal of
Procedia Manufacturing rank Q4 (SCImago) in Artificial Intelligence and Industrial
and manufacturing Engineering.

Copyright: The publisher of this article (ELSEVIER B.V.) stated that authors can use
their articles in full or in part to include in their thesis of dissertation (provided that
this is not to be published commercially).

54 Chapter 3: Reviewing In-vehicle Systems to Improve Fuel Efficiency and Road Safety (Paper 1)
Statement of Contribution of Co-Authors for Thesis by Published Papers

The authors listed below have certified that:

 They meet the criteria for authorship in that they have participated in the
conception, execution, or interpretation, of at least that part of the publication
in their field of expertise;

 They take public responsibility for their part of the publication, except for the
responsible author who accepts overall responsibility for the publication;

 There are no other authors of the publication per these criteria;

 Potential conflicts of interest have been disclosed to (a) granting bodies (b) the
editor or publisher of journals or other publications, and (c) the head of the
responsible academic unit, and

 They agree to the use of the publication in the student’s thesis and its
publication on the QUT’s ePrints site consistent with any limitations set by
publisher requirements.

Contributor Statement of contribution

Atiyeh Vaezipour Read the academic papers for critical review,


write the manuscript and responded to
reviewer comments
Professor Andry Rakotonirainy Supervised the research, reviewed the
manuscript and supported the editorial process
Professor Narelle Haworth Supervised the research, reviewed the
manuscript and supported the editorial process

Principal Supervisor Confirmation: I have sighted email or other correspondence


from all co-authors confirming their certifying authorship.

Name: Professor Andry Rakotonirainy

QUT Verified Signature


Signature:

Date: 25/03/2018

Chapter 3: Reviewing In-vehicle Systems to Improve Fuel Efficiency and Road Safety (Paper 1) 55
3.2 ABSTRACT

Road transport plays a significant role in various industries and mobility services

around the globe and has a vital impact on our daily lives. However, it also has serious

impacts on both public health and the environment. In-vehicle feedback systems are a

relatively new approach to encouraging driver behaviour change for improving fuel

efficiency and safety in automotive environments. While many studies claim that the

adoption of eco-driving practices, such as eco-driving training programs and in-vehicle

feedback to drivers, has the potential to improve fuel efficiency, limited research has

integrated safety and eco-driving. Therefore, it is crucial to understand the human

factors related theories and practices which will inform the design of an in-vehicle

Human Machine Interface (HMI) that could provide real-time driver feedback and

consequently improve both fuel efficiency and safety. This paper provides a

comprehensive review of the current state of published literature on in-vehicle systems

to identify and evaluate the impact of eco-driving and safety feedback systems. This

paper also discusses how these factors may conflict with one another and have a

negative effect on road safety, while also exploring possible eco-driving practices that

could encourage more sustainable, environmentally-conscious and safe driving

behaviour. The review revealed a lack of comprehensive theoretical research

integrating eco-driving and safe driving, and no current available HMI covering both

aspects simultaneously. Furthermore, the review identified that some eco-driving in-

vehicle systems may enhance fuel efficiency without compromising safety. The review

has identified a range of concepts which can be developed to influence driver

acceptance of safety and eco-driving systems within the area of HMI. This can promote

new research aimed at enhancing our understanding of the relationship between eco-

driving and safety from the human factors viewpoint. This provides a foundation for

56 Chapter 3: Reviewing In-vehicle Systems to Improve Fuel Efficiency and Road Safety (Paper 1)
developing innovative, persuasive and acceptable in-vehicle HMI systems to improve

fuel efficiency and road safety.

3.3 INTRODUCTION

Road transport is increasing worldwide, with subsequent increases in greenhouse

gases and road trauma which have serious effects on both human health and the planet.

Transport safety and reduction in fossil fuels dependency have been identified as

urgent societal needs that should be met by scientific and technological researchers

and industrial developments (ETRAC, 2014). The number of studies endeavouring to

address the need to improve fuel efficiency and road safety is growing. To date,

reducing carbon emissions and making driving more ecological and safe have

generated different forms of intervention. Such interventions include: promotion of

electric cars, improving energy intensity of driving (Sivak, 2014), educational and in-

vehicle interventions. In-vehicle systems are fast becoming a key instrument in

encouraging the adoption of eco-driving practices and have the potential to lead to

reduction in fuel consumption and improved road safety by providing eco-driving

advice to drivers (ECOWILL, 2014). Eco-driving refers to cost-effective driving styles

that help to reduce fuel consumption and vehicle emissions.

Eco-driving has been trialled in a number of countries and there is growing

evidence suggesting it is a cost-effective approach for reducing fuel consumption.

However, little research has examined its effects on safety. Driving in a manner that is

eco-friendly and safe requires drivers to make driving task decisions, as well as

interactions with other road users. Theoretically, driving in 5th gear at 80 km/h

smoothly, without stopping, will result in very low fuel consumption and emissions.

However, increasing speed (e.g., exceeding the speed limit) in traffic streams increases

Chapter 3: Reviewing In-vehicle Systems to Improve Fuel Efficiency and Road Safety (Paper 1) 57
the risk of crashes and there may be conflicts between crash rates and fuel consumption

(Haworth & Symmons, 2001).

The past thirty years has seen increasingly rapid advances in the field of

Advanced Driving Assistance Systems (ADAS). Some systems have been integrated

into new vehicles, while others have been incorporated into aftermarket products and

other technologies such as smart phones and heads-up displays, which have been

brought into the vehicles to provide feedback to drivers. However, along with this

growth, there are increasing concerns regarding the use of these in-vehicle systems on

safety.

To date, there are a number of studies proposing independent strategies to

address safety or ecological driving (Saboohi & Farzaneh, 2009; Young et al., 2011).

However, few research studies have comprehensively integrated safety and ecological

in-vehicle systems (Young et al., 2011). Whilst some in-vehicle systems have been

shown to be effective in reducing fuel consumption, they do not address impacts on

safety and they have not rigorously considered the user acceptance and human factors

aspects during the design of new systems (Regan, 2001). A major challenge into the

future is to increase driver acceptance and safety, while also improving fuel efficiency.

This paper reviews the literature on in-vehicle systems which provide

information and feedback to drivers on eco-driving, both in terms of their effectiveness

for improving eco-driving while also investigating the road safety implications of these

systems. Moreover, the paper seeks to identify factors which can be developed to

influence eco-driving, safety and driver acceptance. The paper begins with a brief

description of the terminology and search methodology used to conduct the literature

review, before introducing the relationship between in-vehicle eco-driving and safety

58 Chapter 3: Reviewing In-vehicle Systems to Improve Fuel Efficiency and Road Safety (Paper 1)
systems, providing a critical review of the literature, and discussing future research

directions.

3.4 SEARCH METHODOLOGY

This study critically examines existing in-vehicle systems and their potential to

improve fuel efficiency and safety. In order to identify relevant studies, the initial

search was limited to the combination of keywords including, eco-driving, green

driving, fuel efficiency, road safety, crash, injury, and in-vehicle systems in the context

of driving. Papers published from 1970, when the first driver feedback related to fuel

consumption was introduced, were considered for the review. Both published and grey

literatures from multiple disciplines were included. Only papers written in English

were included. Relevant papers were identified by searching various scientific

databases (e.g., Google Scholar, Science Direct, Web of Science, PsycINFO,

Transportation Research Information Services). In addition, the references cited in

each identified paper were reviewed to discover additional sources of information.

Although we found that there are many factors affecting fuel consumption

studied in (Alwakiel, 2011; Bigazzi & Bertini, 2009; Demir, Bektaş, & Laporte, 2011),

in this review we have focused on those related to driver behaviour and specifically

assessing in-vehicle systems. Therefore, papers studied the other factors affecting fuel

consumption has been excluded in this review. Next, we define key terms that are used

in driving domain briefly in Table 3.1.

Chapter 3: Reviewing In-vehicle Systems to Improve Fuel Efficiency and Road Safety (Paper 1) 59
Table 3.1.Key Terminology in Driving Domain

Term Description
Fuel Total quantity of fuel consumed by a vehicle in specific area and time
consumption period (cited in (Haworth & Symmons, 2001))
Fuel
Total quantity of fuel consumed by a vehicle per unit distance (commonly
consumption
expressed in liters/100kilometers)
rate
Inverse of fuel consumption rate, define as distance travelled per unit of
Fuel economy
fuel consumed (kilometers/liter)(Haworth & Symmons, 2001)
Ratio of the work or energy output of an engine to the work or energy
Fuel efficiency
input (Haworth & Symmons, 2001)
Cost-effective driving style to reduce fuel consumption and vehicle
Eco-driving
emissions
Eco-safe driving Driving style that improves fuel efficiency without comprising safety
Using game design in a non-game context in an attempt to boost drivers'
Gamification motivation and commitment to use a new in-vehicle system (Diewald et
al., 2013)

3.5 RESULTS

3.5.1 Relationship Between Eco-Driving and Safety


Prior research has identified three broad eco-driving interventions designed to

improve fuel efficiency. The decision to perform eco-driving can be studied at three

levels (i) strategic decisions (e.g., selection of vehicle); (ii) tactical decisions (e.g.,

route selection and vehicle loading); and, (iii) operational decisions (driver style)

(Alam & McNabola, 2014). However, improving driving style has been shown to be

an important step to reducing fuel consumption and vehicle emissions compared to

other eco-driving decisions, such as vehicle maintenance (Martin et al., 2012).

According to Alam and McNabola (2014), many countries have adopted eco-driving

policies within the transport sector in a bid to reduce energy consumption and CO2

emissions. Therefore, the benefits of eco-driving are widespread over the past few

decades. Young and Birrell (2012) have suggested a number of practices for eco-

60 Chapter 3: Reviewing In-vehicle Systems to Improve Fuel Efficiency and Road Safety (Paper 1)
driving, including: planning ahead to avoid unnecessary stopping; choosing moderate

engine speeds and uniform control to achieve continuous speed; shifting to higher

gears as soon as possible, using positive (but not heavy) acceleration; avoiding sharp

braking and using engine braking for smooth deceleration.

However, whilst many existing eco-driving practices may improve fuel

efficiency, they may also compromise safety. Young and Birrell (2012) identified a

number of potential conflicts between safety and eco-driving practices. For example,

they argue that driving in fifth gear with a speed between 60 and 80km/h, without

stopping, will reduce fuel consumption and CO2 emissions, but may result in shorter

following distances and thus an increased risk of rear-end collisions. It has also been

argued that maintaining speed through intersections, instead of slowing down,

increases the likelihood that a driver will fail to detect other road users and that

distraction may result from using in-vehicle eco-driving systems (Haworth &

Symmons, 2001). In addition, one study undertaken in 2004 by the Turku University

in Finland (cited in CIECA, 2007), identified various situations where eco-driving

practices may compromise safety, including: Drifting around junctions and pedestrian

crossings in an attempt to avoid stopping; Reducing headway to the vehicle in front in

an effort to maximise speed homogeneity; coasting prematurely, disrupting the pattern

of traffic to the rear and increasing the risk of a rear-end collision; Rapid acceleration

to cruising speeds which may result in shorter safety margins to the vehicle in front;

Trying to stay in a high (fuel-efficient) gear, resulting in manoeuvring at

inappropriately high speeds (e.g., when cornering); and, Switching off the engine at

short stops, which may lead to the steering wheel locking in some vehicles.

Many of the safety issues cited above may occur while practicing eco-driving.

Young et al. (2011) reviewed a number of studies and concluded that whilst there are

Chapter 3: Reviewing In-vehicle Systems to Improve Fuel Efficiency and Road Safety (Paper 1) 61
large overlaps between eco-driving and safe driving behaviour, in some traffic

situations they are in conflict with one another (e.g., maintaining a constant speed may

cause late breaking and risk of rear-end collision). Similarly, Mark Symmons and

Haworth (2003) reported a negative relationship between fuel consumption and

crashworthiness. However, Wu, Zhao, and Ou (2011) argue that the solution may be

to develop feedback mechanisms from in-vehicle systems based on immediate traffic

conditions, such as headway spacing and traffic light. In a study conducted by

Strömberg, Karlsson, and Rexfelt (2015), 18 Swedish drivers were interviewed

regarding their views on eco-driving. The results revealed that drivers are not always

aware of their actions and may be less motivated to perform some eco-driving practices

due to disbelief in the benefits that could result from their actions. Therefore,

maintaining continuous and real-time feedback over time can increase motivation for

eco-safe behaviour (both green and safe). Indeed, Strömberg et al. (2015) suggest that

in-vehicle systems should be included in more vehicles to provide real-time and

continuous feedback to drivers on their actions.

However, most in-vehicle systems currently provide feedback on only one of

two important factors: (i) eco-driving or (ii) safety. The separation of these types of

feedback may encourage drivers to base their decision-making on only one of them.

Therefore, it is crucial to review the current state of knowledge on in-vehicle systems

and understand their impact on fuel consumption and safety. This will further inform

the design of in-vehicle systems that could provide real-time driver feedback and

consequently improve both fuel efficiency and safety. This paper further seeks to

identify eco-safe driving parameters, as a basis to promote positive driving behaviour

which can improve both fuel efficiency and road safety.

62 Chapter 3: Reviewing In-vehicle Systems to Improve Fuel Efficiency and Road Safety (Paper 1)
3.5.2 Effective Use of In-Vehicle Systems on Fuel Consumption and Safety
In-vehicle eco-driving systems are intended to support drivers to improve fuel

efficiency and reduce vehicle emissions (e.g., CO2) by giving real-time feedback on

fuel consumption. Horberry et al. (2014b) classify in-vehicle feedback systems in three

groups: informing, warning and intervening. Informing systems provide relevant

information to the driver that they might otherwise miss themselves. Alternatively,

warning systems alert the driver to take necessary action if they do not respond

appropriately to a particular situation, while intervening systems go one step further

and take control of the vehicle in a critical situation. This feedback can be presented

in the form of an auditory, visual or tactile alert via the in-vehicle system.

Driver feedback related to fuel consumption was first introduced in the early

1970s, after the first oil crisis, to improve fuel economy (e.g., manifold vacuum gauges

fitted in vehicles) (Van der Voort, Dougherty, & van Maarseveen, 2001). This

technology has continued to develop, for example in hybrid-electric vehicles such as

the Toyota Prius (1997) and the Honda Insight (1999), which included innovative

dashboard displays (Barkenbus, 2010). Various in-vehicle devices and technologies

continue to be developed to give eco-driving feedback to drivers and improve fuel

efficiency, such as dashboard displays (e.g., SmartGauge1, EcoAssist2), heads-up

displays and smart phone applications (e.g., Fiat Eco:Drive system3 and more; see Hof

et al. (2013)) for a comprehensive review of existing in-vehicle eco-driving systems

and functionalities).

1
http://smartdesignworldwide.com/work/ford-smart-gauge/
2
http://automobiles.honda.com/insight-hybrid/fuel-efficiency.aspx
3
http://www.fiat.com/ecodrive/

Chapter 3: Reviewing In-vehicle Systems to Improve Fuel Efficiency and Road Safety (Paper 1) 63
Evidence from current literature shows that eco-driving in-vehicle systems have

positive effects on fuel consumption and can improve fuel efficiency. Barth and

Boriboonsomsin (2009) found that speed feedback via in-vehicle dashboard displays

can reduce fuel consumption by 10-20% depending on the context of the driving

scenario. Similarly, reductions of 6.8% in fuel consumption have been observed when

bus drivers received instantaneous feedback on their driving through in-vehicle eco-

driving systems (Strömberg & Karlsson, 2013). However, Haworth and Symmons

(2001) advised that in-vehicle systems are not always safe to use if they cause

distraction to drivers. Overall, the majority of eco-driving systems have sought to

maximise environmental benefits without considering potential safety outcomes and

in most of the studies, the effect of safety has not been studied comprehensively, or in

some cases at all. Therefore, it is important to look at driver behaviour in different

driving contexts and highlight the parameters affecting fuel consumption and safety.

3.5.3 Driving Parameters Influencing Fuel Consumption and Safety


As mentioned above, different factors influence fuel consumption and safety. In

this section, we review those factors related specifically to driver behaviour in

response to advisory feedback from in-vehicle systems. Based on the review of the

literature, we have summarised a set of driving behaviour parameters that influence

fuel consumption and safety (see Table 3.2). These parameters are highly dependent

on the context of the surrounding traffic situation and affect both fuel consumption

and safety (Larue, Malik, Rakotonirainy, & Demmel, 2014).

64 Chapter 3: Reviewing In-vehicle Systems to Improve Fuel Efficiency and Road Safety (Paper 1)
Table 3.2. Driving Parameters Influencing Fuel Consumption and Safety

Driving Influence on Example of In-Vehicle


Parameters Fuel consumption Safety System Technologies

• Speed
Decrease fuel May increase risk of crashes Recommendation
Speed
consumption due to excessive speed • Intelligent Speed
Adoption
• Speed
May increase fuel
Recommendation
Following speed consumption in low Decrease risk of crashes
• Intelligent Speed
speeds
Adoption
Decrease fuel
Cruising speed Decrease risky manoeuvres • Cruise Control
consumption
Smooth Decrease fuel • Haptic Pedal
Decrease aggressive driving
acceleration consumption Feedback
Smooth Decrease fuel May increase risk of crashes • Haptic Pedal
deceleration consumption due to shorter headway Feedback
May increase risk of crashes
Increase fuel • Haptic Pedal
Sharp braking due to risk of rear-end
consumption Feedback
collision
Highest gear Decrease fuel May leads to less control of • Gear change advice
possible consumption the vehicle • Gear Shift Indicator
Decrease fuel
Idle time consumption (no
more than ~ 30 Sec)
• Collision
Prevent rear-end collision
Safe headway avoidance/warning
(TTC~ 2-4 Sec)
system
Decrease risk of crashes due
• Lane departure
Lane position to maintain the car in the
warning
lane
Increase fuel
Aggressive consumption due to
Increase risk of crashes
driving hard acceleration/
deceleration

A number of parameters emerged from studies examining the relationship

between advisory in-vehicle systems and fuel consumption, including: speed choice,

acceleration, deceleration, gear shifting and idling. These parameters are consistent

with earlier research conducted by Hooker (1988), suggesting that driver behaviour

has a significant effect on reducing fuel consumption. Not surprisingly, speed was

reported to be a main parameter in most studies. The field trial experiments conducted

Chapter 3: Reviewing In-vehicle Systems to Improve Fuel Efficiency and Road Safety (Paper 1) 65
by Barth and Boriboonsomsin (2009) found that providing speed recommendations

(based on real-time, dynamic traffic sensing and telematics data) to drivers via a

dashboard display reduced fuel consumption by 13% and reduced vehicle emissions,

compared to a control group. However, different studies have reported different

approaches to the provision of in-vehicle speed feedback, including: acceleration,

cruising and deceleration. For example, a study investigating the effect of an

accelerator advisory tool, which exercised resistance in the acceleration pedal when

drivers try to accelerate too rapidly, found a reduction in emissions during two of three

routes examined, as well as a significant reduction in strong/heavy acceleration

(Larsson & Ericsson, 2009). However, the study only found a small reduction in fuel

consumption when driving with only the acceleration advisory feedback activated.

The reviewed literature suggests that there are overlaps between eco-driving and

safety behaviour. An inspiring review conducted by Young et al. (2011), discussed

some of the in-vehicle systems that influence safe and/or eco-driving (e.g., satellite

navigation systems, congestion assistants, intelligent speed adoption). In fact, eco-safe

behaviours heavily depend on the driving context and there are several situations

where eco-driving and safety are in contrast with one another. For instance,

maintaining a constant speed in an attempt to avoid braking may compromise

headway, while travelling in the highest possible gear may adversely affect vehicle

control. Overall, safety parameters associated with in-vehicle systems that have been

studied by the literature include: safe headway, lane position, vehicle speed, and

aggressive driving. While all proponents of in-vehicle systems note the importance of

eco-driving and safety parameters, few examples were found in the literature assessing

the reality of actually applying them.

66 Chapter 3: Reviewing In-vehicle Systems to Improve Fuel Efficiency and Road Safety (Paper 1)
Although the reviewed studies acknowledged the positive effect of in-vehicle

systems, it is argued that more attention must be given when designing in-vehicle eco-

driving systems to the potential distraction and heightened mental workload that may

result from using these systems (McIlroy et al., 2014). Haworth and Symmons (2001)

warned that eco-driving feedback systems may have negative effects on safety if they

distract drivers. Indeed, it is important to bear in mind the drivers’ task demands before

thinking about designing such a device. Hence, a well-designed in-vehicle system can

support drivers to encourage both eco-driving and safe behaviour, if implemented with

careful consideration.

3.5.4 Driver Acceptance of In-Vehicle Systems


There have been a number of studies of in-vehicle systems which provide

feedback on vehicle fuel consumption and safety concerns (Barkenbus, 2010; Huang,

Roetting, McDevitt, Melton, & Courtney, 2005; Meschtscherjakov, Wilfinger,

Scherndl, & Tscheligi, 2009; Van der Voort et al., 2001). However, in-vehicle

feedback systems are only beneficial if accepted by drivers. Previous research suggests

that feedback can only advise, and will not necessarily stimulate sustainable ecological

behaviour unless the driver already possesses a strong inclination to drive in an eco-

friendly manner (Shipworth, 2000) and perceives in-vehicle feedback systems to be

effective. According to Adell (2009, p31), driver acceptance is “the degree to which

an individual incorporates the system in his/her driving”. This definition suggests that

acceptance is related to the actual use of the system by an individual. Similarly, Jamson

(2010) defines acceptance as how much the system would be used and drivers are

willing to buy it. It is important that the effective use of the in-vehicle system

understood and valued by drivers. And, this is highly depends on driver’s attitudes,

expectations and their experience using the in-vehicle systems (cited in Adell et al.,

Chapter 3: Reviewing In-vehicle Systems to Improve Fuel Efficiency and Road Safety (Paper 1) 67
2014). Consequently, in-vehicle systems will only be accepted if valued and satisfying

to use by the driver.

To date, many theories and models have been used and developed to describe

the acceptance of technology such as, Theory of Planned Behaviour (TPB) based on

the Theory of Reasoned Action; the Technology Acceptance Model (TAM); the

Unified Theory of Acceptance and Use of Technology (UTAUT) (Adell & Varhelyi,

2014; Horberry et al., 2014b). The main element in all these is the behaviour; which is

related to the use of the new system. However, Horberry et al. (2014b) presented

various existing theories and models behind driver acceptance of technology from the

perspectives of researchers, product designers and policymakers and argued that many

of these models were initially developed for non-driving context and there is no

specific and universally accepted model that widely agreed by all domains for

developing new driver acceptance technologies such as increasing safety. However,

(Horberry, Stevens, & Regan, 2014a) identified a number of general consensus in

regard to the acceptance of new technology, such as; (i) usefulness and ease of use the

system is two key factors in acceptance of the new technology; (ii) acceptance depends

on the individual factors such as gender, age, culture and personality; (iii) individuals

have different judgements, accordingly, drivers-centric view is the requirement to

predict and measure driver acceptance in individual level; (iv) for the system to be

acceptable, drivers do not need to like the system, but rather find it effective, however,

liking the system may increase the usability; (v) acceptance depends on the context of

the use and whether the use of the system is voluntary and influenced by social and

cultural norms. Last but not least, (vi) acceptance of the system may change over time,

depending on different context or as the drivers’ experience of the system develops. It

is also worth mentioning that, there is a lack of application of these models and

68 Chapter 3: Reviewing In-vehicle Systems to Improve Fuel Efficiency and Road Safety (Paper 1)
concepts in the context of eco-driving in-vehicle systems in the literature. Therefore,

better understanding of parameters influencing driver acceptance and applying them

can support the potential design of eco-safe in-vehicle system which is also accepted

by drivers (Ghazizadeh & Lee, 2014).

3.5.5 Social Concepts to Motivate Eco-Safe Driving


A large number of eco-safe driving behaviours are dependent on motivations,

beliefs and cultural background of the driver, all of which are psychologically and

socially complex issues. Rakotonirainy, Obst, and Loke (2012, p. 381) have indicated

that social norms are “rules and standards that are understood by members of a group,

and that guide and/or constrain their social behaviour”. Therefore, it is important to

investigate psychological and social concepts that motivate ecological behaviour, as

well as taking into account the real needs of drivers, levels of acceptance and the

usability of the system, from the preliminary stages of design and development.

Consequently, any theory or model of Human Machine Interface (HMI) in the context

of sustainable energy behaviour that fails to include social concepts could be argued

to lack a critical element. Whilst there exists different in-vehicle HMI for eco-driving

and safety (e.g. Pace, Ramalingam, & Roedl, 2007), there is little work on the use of

psychological and social concepts and there is a gap in the literature on the use of in-

vehicle systems to motivate sustainable, green and safe driving.

Very recently the use of gamification has emerged in the automotive industry,

primarily for its motivating and inspiring potential (Iacovides, 2009; Van Eck, 2006).

The term gamification refers to the use of game design, game playing techniques and

game mechanisms to engage users and motivate positive behaviour (Deterding, Björk,

Nacke, Dixon, & Lawley, 2013; Deterding et al., 2011). Recent studies outlined by

Jung, Schneider, and Valacich (2010) suggest that motivational affordance such as

Chapter 3: Reviewing In-vehicle Systems to Improve Fuel Efficiency and Road Safety (Paper 1) 69
gameful design can motivate positive involvement. Some recent gamification

applications have been reviewed (Bellotti, Berta, & De Gloria, 2014; Diewald et al.,

2013). However, the use of gamification in the automotive domain indicates some

challenges and limitations that may result, including distraction from the driving task

resulting in violations of traffic laws (Diewald et al., 2013). Therefore, it is critical to

evaluate any gamification application in the automotive domain before integrating it

into in-vehicle interfaces to ensure it does not interfere with safe driving behaviours or

increase the likelihood of driver distraction.

3.6 CONCLUSIONS

This paper presents a review of current state of published literature on eco-

driving and potential conflicts with safety, with a focus on feedback from in-vehicle

systems. There is evidence found in the literature that fuel efficiency is achieved as a

result of positive interactions between in-vehicle systems and drivers. However, the

review has also shown that much of the existing literature on in-vehicle system to date

applies only to eco-driving or safety and there is lack of research integrating both

comprehensively. While we argued that there are number of driving behaviour

parameters such as speed choice, deceleration, acceleration, idling, headway and lane

deviation which all contribute to eco-safe behaviour, few examples were found in the

literature assessing the reality of actually applying them. Nonetheless, we also found

there are still open questions with respect to driver acceptance of in-vehicle systems

to support driver performance for eco-safe driving and consensus is lacking. Taken

together, these findings enhance our understanding of factors such as acceptance which

play a significant role in sustainable use of the in-vehicle system and caution that we

cannot be sure the system will be beneficial if the acceptance of the system has not

been considered. In addition, this review supports the concept of using gamification to

70 Chapter 3: Reviewing In-vehicle Systems to Improve Fuel Efficiency and Road Safety (Paper 1)
motivate eco-safe cues among drivers. However, it is critical to evaluate any

gamification application in the automotive domain before integrating it into the in-

vehicle system to ensure it does not interfere with safe driving behaviours or distract

the driver.

The present study is the first step in an on-going program of research to develop

and design an innovative eco-safe in-vehicle feedback system to improve fuel

efficiency and safety which can maximise driver acceptance and reduce driver

distraction. Considerably more work will need to be done to understand driver

behaviour under technological instructions and ways to enhance drivers’ motivation to

use the in-vehicle system using gamification principles. More information on user

acceptance would help us to establish a greater degree of accuracy in design and

development of the in-vehicle system. Ensuring appropriate systems can enhance the

potential of using these technologies to promote environmentally friendly driving,

reducing emissions and save human life.

Chapter 3: Reviewing In-vehicle Systems to Improve Fuel Efficiency and Road Safety (Paper 1) 71
Chapter 4: Enhancing Eco-Safe Driving
Behaviour Using In-Vehicle
Human Machine Interface: A
Qualitative Study (Paper 2)

4.1 INTRODUCTORY COMMENTS

The second step of the user-centred design approach highlights the need to

specify user needs and requirements associated with system use (see Figure 4.1). As

such, this chapter describes the findings from a focus group study with drivers (Study

2a). The main aim of the paper presented in this chapter (Paper 2) is to report on driver

perceptions and attitudes toward eco-safe driving behaviour, including attitudes

towards in-vehicle HMI characteristics, as well as driver acceptance of, and intentions

to use, various eco-safe HMI concept systems.

Similar to Chapter 3, this chapter seeks to address Research Question 1: “What

is the relevant information that should be conveyed to drivers in an in-vehicle HMI to

improve eco-safe driving?”, but does so from a user perspective. That is, this chapter

seeks to identify user needs and requirements, in terms of opinions regarding overall

attitudes towards eco-safe driving behaviours, in-vehicle system usefulness in general,

as well as various characteristics associated with system design and how these may

influence technology acceptance and intentions to use. The findings of this paper are

intended to be taken together with the findings of the literature review presented in the

preceding chapter, as well as the findings from the expert workshop (Study 2b,

discussed in Section 2.5) and the following chapter, which also discusses findings from

the focus group study, but from a design perspective. Together, these findings used to

inform system development and evaluation, which is outlined in Chapters 6 and 7.

Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2) 73
Figure 4.1 User-Centred Design Approach- Driver Focus Groups

Taken from: Vaezipour, A., Rakotonirainy, A., Haworth, N., & Delhomme, P. (2017).
Enhancing eco-safe driving behaviour through the use of in-vehicle human-machine
interface: A qualitative study. Transportation Research Part A: Policy and
Practice, 100, 247-263. https://doi.org/10.1016/j.tra.2017.04.030

Publication Status: Accepted 28 April 2017

Journal Quality: Transportation Research Part A: Policy and Practice is a peer-


reviewed publication with international readership. The journal impact factor is 2.609,
and rank Q1 (SCImago) in Transportation, Management Science and Operations
Research, Civil and Structural Engineering.

Copyright: The publisher of this article (ELSEVIER B.V.) stated that authors can use
their articles in full or in part to include in their thesis of dissertation (provided that
this is not to be published commercially).

Statement of Contribution of Co-Authors for Thesis by Published Papers

The authors listed below have certified that:

 They meet the criteria for authorship in that they have participated in the
conception, execution, or interpretation, of at least that part of the publication
in their field of expertise;

74 Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2)
 They take public responsibility for their part of the publication, except for the
responsible author who accepts overall responsibility for the publication;

 There are no other authors of the publication per these criteria;

 Potential conflicts of interest have been disclosed to (a) granting bodies (b) the
editor or publisher of journals or other publications, and (c) the head of the
responsible academic unit, and

 They agree to the use of the publication in the student’s thesis and its
publication on the QUT’s ePrints site consistent with any limitations set by
publisher requirements.

Contributor Statement of contribution

Atiyeh Vaezipour Study design, recruitment, data collection, data


analysis, write the manuscript and responded to
reviewer comments
Professor Andry Rakotonirainy Supervised the research, reviewed the
manuscript and supported the editorial process
Professor Narelle Haworth Supervised the research, reviewed the
manuscript and supported the editorial process
Dr Patricia Delhomme Discussed the research design, reviewed the
manuscript and supported the editorial process

Principal Supervisor Confirmation: I have sighted email or other correspondence


from all Co-author confirming their certifying authorship.

Name: Professor Andry Rakotonirainy

QUT Verified Signature


Signature:

Date: 25/03/2018

Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2) 75
4.2 ABSTRACT

Background: The widespread reliance on motor vehicles has negative effects on both

the environment and human health. The development of an innovative in-vehicle

human-machine interface (HMI) has the potential to contribute to reducing traffic

pollution and road trauma.

Aim: A qualitative study, using a driver-centred design approach, was carried out to

test how best to provide ecological and safe (eco-safe) driving advice and feedback to

drivers on their driving style via an in-vehicle HMI.

Method: A total of 34 drivers (52.9% males), aged 19-61 years, participated in focus

groups which explored concepts from the Technology Acceptance Model (Davis,

1989).

Findings: Main themes emerging from the focus groups were: (i) perceived

importance of eco-safe driving behaviour; (ii) perceived usefulness of eco-safe in-

vehicle HMIs; (iii) intentions to use an eco-safe in-vehicle HMI; (iv) perceptions

towards eco-safe in-vehicle HMI design characteristics; and (v) potential problems

associated with using eco-safe in-vehicle HMIs.

Implications: This study provides the foundation to inform the design and

development of an evidence-based in-vehicle eco-safe HMI with high levels of driver

acceptance. Recommendations for future research are also discussed.

4.3 INTRODUCTION

Climate change and the impact of humankind on the environment is a prominent

public health issue requiring immediate action (Gore, 2006). Ambient air pollution

contributes to 3.7 million deaths each year worldwide, making it the single largest

threat to environmental health on a global scale (World Health Organization, 2014) .

76 Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2)
Our ever-increasing reliance on motor vehicles contributes to this problem by

increasing gaseous and particulate pollution, which in turn has severe impacts on

human health, such as increasing lung and heart disease. Moreover, 19% of all carbon

dioxide (CO2) emissions are caused by road vehicles, with such emissions intrinsically

linked to rising global temperatures (Matthews, Gillett, Stott, & Zickfeld, 2009).

There have been a number of recent developments within the transportation

industry aimed at, among other things, reducing the impact of motor vehicles on the

environment, such as electric or automated vehicles (Fagnant & Kockelman, 2015;

Wadud et al., 2016). However, improvements associated with these initiatives have

been relatively incremental due to the typically higher costs to users associated with

such options. On the other hand, eco-driving, which can broadly be defined as driving

behaviours aimed at reducing fuel consumption and subsequent emissions, has been

demonstrated as a promising and cost-effective approach (Barkenbus, 2010; Pampel

et al., 2015). The general objective of the present study is to identify user requirements

and perceived acceptability towards an in-vehicle HMI system for both eco-driving

and safe driving behaviours.

This paper is divided into four main sections, the first of which has three sub-

sections. The first section focuses on eco-driving as a climate change initiative, user-

centred design and driver acceptance, and the aims of the present study. The second

section outlines the study methodology, while the third section presents the results and

discusses the findings. The final section outlines the conclusions of the research and

its real-world implications.

4.3.1 Eco-Driving as a Climate Change Initiative


Prior research has identified three levels of decisions associated with eco-driving

(Alam & McNabola, 2014): (i) strategic decisions (e.g., vehicle selection, maintenance

Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2) 77
schedules); (ii) tactical decisions (e.g., route selection, vehicle loading); and, (iii)

operational decisions (e.g., driving style). While all of these decisions are important,

improving driving style is associated with relatively immediate impacts on fuel

consumption and emissions once the appropriate driving style is adopted (Martin et

al., 2012).

Despite these advantages, many drivers do not adopt or practice eco-driving.

Thus, Barkenbus (2010) described eco-driving as an underutilised initiative for

battling climate change, reporting that improvements in eco-driving have the potential

to reduce fuel consumption by up to 10%. As a result, governments in a number of

highly motorised countries have begun to develop and implement eco-driving policies

within the transport sector in a bid to reduce fuel consumption and subsequent

emissions (Alam & McNabola, 2014).

However, a number of studies have argued that eco-driving behaviour may at

times compromise safe driving. For example, maintaining a constant cruising speed

and choosing the highest appropriate gear, may increase the likelihood that a driver

will decrease headway to the vehicle in front and reduce their ability to brake

appropriately, in turn increasing the risk of rear-end collisions (Young & Birrell, 2012;

Young et al., 2011). In addition, such behaviour may also increase the likelihood of a

driver manoeuvring at inappropriately high speeds, such as when cornering (cited in

CIECA, 2007). These findings highlight that the development of any eco-driving

initiative should be conducted with driver safety as a critical consideration. That is, the

development of in-vehicle HMI systems should ultimately seek to address eco-safe

driving behaviours, defined as those behaviours that reduce fuel consumption and

subsequent emissions, while also considering the impact of system use on safe driving

behaviours.

78 Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2)
In recent years a number of guidelines (e.g. EcoDrivingUSA; ECOWILL, 2014),

public education campaigns and driver licence training (ECODRIVEN; Graves et al.,

2012 for examples) have been developed to encourage eco-driving behaviour among

the driving population, with mixed results. An explanation for this may be that while

many drivers are ultimately aware of the range of eco-driving behaviours that impact

upon fuel consumption and subsequent emissions, they lack the technical

understanding of how to appropriately perform these behaviours (Delhomme et al.,

2013; Pampel et al., 2015), or may make conscious decisions to drive in a manner that

is not fuel efficient or safe due to a variety of reasons they believe justify the behaviour

in the given moment, such as (Harvey et al., 2013) running late or enjoying the feeling

of driving fast. Alternatively, the complexity of the driving task might mean that

drivers are not always aware of their actions, and in turn may not use their eco-driving

knowledge and skills to their full potential (Pampel et al., 2015).

For this reason, the developments of in-vehicle human-machine interfaces

(HMI) represent a promising approach for providing relevant real-time advice and

feedback to drivers to assist them to better adopt eco-driving behaviours. Previous

research has highlighted the potential for eco-driving in-vehicle HMIs to have positive

impacts on fuel consumption and vehicle emissions (e.g. Ecodriver, 2011; Jonkers et

al., 2016; Larsson & Ericsson, 2009; Van der Voort et al., 2001). However, it is

important that the development of such initiatives adopt a user-centred design

approach and carefully consider the impact of issues associated with driver acceptance

on subsequent system effectiveness.

4.3.2 User-Centred Design and Driver Acceptance


The effectiveness of an in-vehicle HMI is highly dependent on driver acceptance

of the technology, and in particular on the perceived usefulness and intention to use

Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2) 79
the system. Driver acceptance has been defined as “the degree to which an individual

incorporates the system in his/her driving” (Adell, 2009, p. 31). This concept has also

been used to describe how much drivers would use the system and their willingness to

pay to purchase a system (Jamson, 2010). Adopting a user-centred design approach

can increase the likelihood of driver acceptance of an in-vehicle HMI and motivate

greater usage, in turn enhancing the effectiveness of a system (Maguire, 2001).

The International Organisation for Standardisation ISO (2010) has highlighted

four key principles and recommendations for user-centred design. Overall, these

principles characterise the process of user-centred design into a number of stages (see

Figure 4.2). Specifically, it is argued that system developers must (i) analyse and

comprehensively understand the context of the whole user experience, such as the

tasks involved, the organisational, social and environmental context of use, and the

needs, motivations and limitations of the intended users; (ii) actively involve users in

all stages throughout the development process; (iii) conduct user-centred evaluations;

(iv) understand that the process is iterative and refine systems using the same user-

centred approach; and (v) adopt a multidisciplinary approach that includes a design

team with a variety of skills and perspectives.

Figure 4.2. The User-Centred Design Approach (adapted from ISO, 2010).

80 Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2)
Therefore, the design and development of in-vehicle HMIs must carefully

consider user requirements (Vilimek & Keinath, 2014). Moreover, Norman and Draper

(1986, p. 61) argued that it is important for user requirements to be considered as early

in the design process as possible, stating that “from the point of view of the user, the

interface is the system, the needs of the users should dominate the design of the

interface, and the needs of the interface should dominate the design of the rest of the

system”.

Evaluating driver acceptance of in-vehicle HMIs is therefore critical. Indeed,

Sauer (1993, p. 4) argued that the acceptance and subsequent adoption of a

technological system is “the single, most important factor in determining success or

failure of information systems and technologies”. One of the most popular models used

to understand the determinants of user acceptance of technology is the Technology

Acceptance Model (TAM). This approach focuses on how perceptions regarding the

usefulness and ease of use of computer information systems impact upon attitudes

towards use, intentions to use and subsequent actual system use (Davis, 1989). The

TAM has been identified as being a highly reliable, valid and robust approach to

measuring instrument design (Taylor & Todd, 1995; Venkatesh et al., 2003).

4.3.3 Study Aims


This study addresses the scarcity of current research examining driver

acceptability of in-vehicle HMIs designed to enhance eco-driving. An exploratory

focus group study, based on the TAM, was conducted given that focus groups

represent an efficient approach to eliciting comprehensive, in-depth discussions from

participants (Braun & Clarke, 2006), while also maintaining a user-centred design

approach. This user-centred, qualitative method has also been argued to be an effective

approach for studying scarcely researched topics and phenomena (Given, 2015).

Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2) 81
This study aims to identify best-practice principles for the design and

development of in-vehicle HMI systems to improve eco-driving, while also

considering the potential implications of system use on safe driving behaviours and

driver acceptance of such systems. Secondly, the study aims to explore driver

perceptions and attitudes toward eco-safe behaviour, driver acceptance of, and

intention to use eco-safe HMI systems, as well as attitudes towards various design and

feedback characteristics.

4.4 METHOD

4.4.1 Participants
Participants were recruited from the city of Brisbane (Queensland, Australia)

primarily through university email lists, snowball sampling methods, social media

posts and personal approaches at university campuses, where flyers were distributed

in hardcopy. This study has a research ethics committee approval in accordance with

the Australian Code for the Responsible Conduct of Research (Approval number

1500000683).

A total of 34 drivers (52.9% males) agreed to participate across six focus group

sessions involving between five and six participants each. The sample size was

determined after data saturation, whereby limited new additional information was

being obtained from each additional focus group session (Guest et al., 2006; O'Reilly

& Parker, 2012; Walker, 2012). The advantages of qualitative methods for eliciting

rich, in-depth data have been clearly demonstrated in studies and the golden rule for

sample sizes consider this sample size is acceptable and that additional participants

would have been unlikely to have added new information (Delhomme & Meyer, 2002).

As can be seen in Table 4.1, the sample was aged between 19 and 61 years (M

= 32.11, SD = 10.44 years). The majority of participants (82.4%) held an Open driver’s

82 Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2)
licence (i.e., unrestricted), while 11.8% were Provisional licence holders (i.e.,

unsupervised but restricted) and 5.9% held a Learner’s licence. Roughly a third of

participants (29.4%) reported being very experienced with technology, while 47.1%

reported having average experience and 23.5% categorised themselves as

inexperienced.

Table 4.1. Demographics of Focus Group Participants

Variable No. of participants Variable No. of participants


Gender Age
Female 16 (47.1%) 18-29 16 (47.1%)
Male 18 (52.9%) 30-39 13 (38.2%)
40+ 5 (14.7%)
License type Motivation to improve fuel efficiency
Learner 2 (5.9%) Environment & money 24 (70.6%)
Provisional 4 (11.8%) Save money 8 (23.5%)
Open 28 (82.4%) Environment 1 (2.9%)
Nothing 1 (2.9%)
Average hours driven Experience with technology
<5 12 (35.3%) Experienced 10 (29.4%)
hours/week
6-10 10 (29.4%) Average 16 (47.1%)
hours/week
11+ 12 (35.3%) Inexperienced 8 (23.5%)

4.4.2 Procedure & Materials


Participants gave written consent to participate in audio-recorded focus group

discussions and were informed that their participation was entirely voluntary and

anonymous. Participants initially completed a pen-and-paper survey which collected

information regarding demographics, driving experience and experience with

technology. Focus groups ran for approximately 60 minutes and participants received

a movie voucher as reimbursement for their time and contribution. Each focus group

session was conducted by the same facilitator, together with another researcher who

took notes. The focus group sessions were audio-recorded and transcribed.

The facilitator began each focus group by outlining a range of fuel-efficient and

safe driving behaviours, including smooth acceleration and deceleration, fuel efficient

Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2) 83
and consistent speed choice, anticipating traffic flow, keeping a safe headway and

lane-keeping. Participants were encouraged to discuss which of these behaviours they

would be interested in monitoring while driving. Following this, a series of eco-safe

HMI concepts were presented to participants and briefly described with the aid of

images as stimuli (see Vaezipour et al., 2016 for more details).

A series of open-ended questions were then asked and participants provided

feedback regarding each in-vehicle interface, including what they believed to be the

key effective features and areas for improvement. Questions were semi-structured and

based on the TAM focus group discussion guide cited in Mitsopoulos-Rubens and

Regan (2014). Questions predominantly focused on acceptability, including: (i) user

perceptions of the conceptual designs (e.g., engagement, ease of understanding and

perceived ease of use); (ii) perceived impacts on eco-safe driving behaviours; (iii)

perceived difficulties associated with the interface and ideas for improvements; and,

(iv) intentions to use and purchase the interface.

4.4.3 Data Analysis


The audio-recorded focus groups were transcribed and imported into the Nvivo

11 software program based on principles of the TAM, as well as other key themes

evident in the existing literature (Miles, Huberman, & Saldana, 2013). A thematic

analysis was conducted to identify patterns across responses and their relation to

research questions (Braun & Clarke, 2006; Mark & Yardley, 2004).

The transcriptions of focus group sessions were coded using a thematic content

analysis (Weber, 1990), in order to reduce the information to more relevant,

manageable pieces of data. All ideas, expressions, and words were classified into

content themes on the coding grid. Each defined theme consisted of units comprising

one or more words with similar meaning (Weber, 1990). Emerging themes that did not

84 Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2)
fit within this framework were also considered during the coding process. This review

and refinement process led to the development of the final themes and sub themes.

Another researcher also coded the transcriptions in order to validate the initial

interpretations. Discrepancies were thoroughly discussed and guided the thematic

development, so as to generate a set of coherent themes that reflected a comprehensive

and precise set of meanings for the participants’ comments. Improvements to the

themes were made until the inter-coder agreement was 100%.

4.5 FINDINGS AND DISCUSSION

4.5.1 Number of Statements, Themes and Sub-Themes


The sample of 34 participants produced 273 statements, at a mean rate of 8.03

statements per participant. The statements were defined as participant attitudes and

perceptions in relation to each theme or sub theme. A total of five themes emerged

from the focus group discussions: (1) the perceived importance of monitoring eco-safe

driving behaviours; (2) perceived usefulness of eco-safe in-vehicle HMIs; (3)

intentions to use an eco-safe in-vehicle HMI; (4) perceptions toward eco-safe in-

vehicle HMI design and feedback characteristics, and (5) perceptions toward potential

problems associated with using eco-safe in-vehicle HMIs. Thirteen sub themes were

also identified (see Table 4.2).

Table 4.2. Summary of Thematic Analysis

Key themes and sub-themes Number of Frequency Age Gender


participants of
statements 18-29 30-39 40+ M F
1. Perceived importance of 10 (6M, 4F) 12 (4.3%) 6 3 3 8 4
eco-safe driving
behaviours
2. Perceived usefulness of 20 (11M, 9F) 30 (10.9%) 9 15 6 18 12
system
a. Save fuel and money 12 (6M, 6F) 16 (5.8%) 5 7 4 10 6
b. Improve safe driving 10 (4M, 6F) 14 (5.1%) 9 2 3 7 7
c. Help environment 3 (2M, 1F) 3 (16.1%) 2 1 0 2 1

Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2) 85
3. Intentions to use system
a. Driver type 22 (13M, 9F) 44 (16.1%) 21 13 10 26 18
b. Perceived ease of 6 (4M, 2F) 7 (2.5%) 5 1 1 4 3
use/usability
c. Accessibility and cost 15 (9M, 6F) 19 (6.9%) 14 5 0 12 7
d. Normative influences 4 (2M, 2F) 5 (1.8%) 4 1 0 3 2
4. Design characteristics
a. Device characteristics 18 (10M, 8F) 36 (13.1%) 20 15 1 24 12
b. Feedback 20 (12M, 8F) 34 (12.4%) 12 12 10 23 11
characteristics
5. Potential problems

a. Distraction 11 (6M, 5F) 15 (5.4%) 11 2 2 9 6


b. Privacy concern 12 (9M, 3F) 15 (5.4%) 9 5 1 12 3
c. Perceived accuracy 7 (4M, 3F) 7 (2.5%) 4 2 1 4 3
and trust
d. Novelty 7 (2M, 5F) 9 (3.2%) 7 0 2 4 5
e. Promote unsafe/illegal 5 (3M, 2F) 7(2.5%) 0 4 3 5 2
behaviour

A chi-square test of goodness of fit was performed to determine whether there

were significant differences in rates of responding by gender or age. Males cited

significantly more statements than females (χ2= 10.4, p < 0.01), however no significant

differences in response rates were observed by age (χ2 = 4.42, p = 0.11).

Tables 4.3 to 4.8 outline key statements from participants in relation to each of

the themes and sub-themes. Direct quotes are labelled with the corresponding age,

gender (i.e. F or M), and the number of years each participant had held their driver’s

licence (e.g., 34, F, 14) refers to a 34 years old female who obtained her driver’s licence

14 years ago.

4.5.2 Perceived Importance of Monitoring Eco-Safe Driving Behaviours


Participants were asked to discuss which behaviours they perceived as important

for improving fuel consumption and safe driving, specifically in relation to those

behaviours that could be monitored by an in-vehicle eco-safe HMI system. Ten

participants made a total of 12 statements regarding the perceived importance of

86 Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2)
monitoring eco-safe driving behaviours. Table 4.3 outlines the key relevant quotes

related to the perceived importance of eco-safe driving behaviours.

Table 4.3. Selected Quotes Relevant to Theme 1 - "Perceived importance of eco-safe


driving behaviours"

Subtheme Relevant Quotes

a. Fuel i. “I think monitoring the acceleration and deceleration would be effective because I don’t know myself
consumptio what's the right stopping distance and acceleration distance. I know it from a point of view of a
n passenger who drives with me, just from feedback, they believe that I stop and start quite quickly. So I
think monitoring that will be quite effective for my own personal benefit” (22, F, 1).

b. Safety i. “I think keeping a safe headway, that’s important, especially for safety. Not that I have an issue with
that myself, but I do have a couple of friends who do keep quite an unsafe distance between the cars
and they don’t realise themselves, so I think that would be effective” (22, F, 1). “I like having an
awareness of where I am in a lane, maybe some information about vehicles behind me and blind spots
will be really useful. I like the idea of where I am in a lane, behind me, in front of me, things I can’t
always see” (32, M, 15).

c. illusory i. “I think most people would think that they are a safe driver. If you ask people to rate their own driving,
superiority I’m pretty sure they’ll think they’re safe. Everyone thinks everyone else is a bad driver” (39, M, 21). “I
would say safety for everyone else because that’s the first thing I learned as a driver: don’t trust
anyone on the road – they can do the wrong thing. But for myself, I trust myself, it’s just whether or not
I can drive more fuel efficiently” (29, M, 10).

d. Monitoring i. “I think if it showed like they’re [other drivers] moving out of lanes, or they’re speeding or they’re
other braking suddenly, that kind of thing, as far as safety could be…more helpful to say that car ahead of
drivers you is doing something that you might need to be a little bit more aware of” (33, F, 13).

Overall, participants generally supported the notion that being able to monitor

eco-driving and safe driving behaviour is important and that an eco-safe in-vehicle

HMI system has the potential to improve their own driving by increasing their

awareness of their behaviour. Monitoring was argued to be particularly beneficial in

relation to behaviours that are typically difficult for one to subjectively monitor

themselves, such as acceleration and deceleration (Table 4.3, a.i), as well as

behaviours perceived to have a considerable impact on safety, such as maintaining a

safe headway, lane keeping and speeding (Table 4.3, b.i). Taken together, these

findings add support to the hypothesis that a well-designed, user-centric system would

be likely to be associated with relatively high levels of acceptance among drivers.

Interestingly, overall responses suggested some evidence of illusory superiority

effect (McKenna & Myers, 1997; Waylen, Horswill, Alexander, & McKenna, 2004).

Specifically, participants typically argued that, in relation to other drivers, it was most

Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2) 87
important to monitor safety behaviours, while in relation to their own behaviour, it was

most important to monitor fuel efficiency (Table 4.3, c.i). Drivers may perceive

themselves to be safer, better drivers than the average driver. While there is some

evidence to suggest that constructive feedback from an in-vehicle HMI that challenges

such biases may motivate behaviour change (Horrey, Lesch, Dainoff, Robertson, &

Noy, 2012), there is also the potential that such feedback may lead to cognitive

dissonance (Festinger, 1962) and result in users disengaging with the system,

particularly amongst those drivers who require behaviour change the most.

Following on from this bias, two participants also stated that they would appreciate

a system that could monitor the behaviours of other drivers, perhaps in an intuitive

manner, and warn them of potentially unsafe behaviour of other drivers (Table 4.3, d.i).

Thus, existing technologies, such as hazard detection and warning systems may be a

useful inclusion in the design of eco-safe in-vehicle HMI systems.

4.5.3 Perceived Usefulness of an Eco-Safe In-Vehicle System

Perceived usefulness can be defined as the degree to which an individual

believes that using a particular system (e.g., an eco-safe in-vehicle HMI) has the

potential to enhance their performance on a particular task (e.g., eco-safe driving). That

is, it relates to whether the system satisfies the needs and requirements of the user.

Overall, participants identified three broad ways they believed an eco-safe in-vehicle

HMI could be useful and help to improve their driving. This included: a) helping them

to save fuel and money; b) improving awareness and increasing safety; and, c)

supporting eco-driving and helping the environment.

88 Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2)
Table 4.4. Selected Quotes Relevant to Theme 2 - "Perceived usefulness of an eco-
safe in-vehicle HMI"

Relevant
Subtheme
Quotes

a. Save fuel i. “I’m interested in the fuel efficiency stuff because I don’t have anything like that in my car but if I did,
and money that would be useful” (32, M, 15). “Save money. Basically, it's the efficiency and eco-driving. Petrol is
expensive so if you do enough driving it [the in-vehicle system] will save you quite a bit of money (61,
M, 40). “With sudden stops and starts it's often more impact on brakes and more, sort of, wear and tear
on the mechanics of the car which there's a cost associated with, maintenance costs” (35, F, 17).“My
car doesn’t have any technology at all. So I’ll be interested in knowing how much petrol am I using or
when I went on a long drive, how much did I waste, so I’d be interested in it just for the feedback” (39,
M, 21).
ii. “I had this argument with one of my friends at work. He was saying “I’m a very good driver, I’ve never
had an accident in my life. But in terms of being eco-friendly, he was terrible. He was absolutely
terrible. I think these criteria would be good to get a sense of whether you’re driving in an eco-friendly
way” (34, M, 14).
iii. “If safe driving is more linked to maintenance costs and things like that, I think I would care more
about it” (32, M, 15). “Some people wouldn’t care about the environment but they might care about
how much money they saved” (39, M, 21).

b. Improving i. “I think it would be hard not to improve your driving if it’s giving you visual cues to say, you’re
safe driving stopping very aggressively and you go, am I really? I’ll try to stop less aggressive next time” (27, M,
9). “Could also be things that for years you thought was the right thing to do, and you’ve been doing it
but maybe it’s wrong. Maybe a system like this will show you - well actually this is how you safely
decelerate” (39, M, 21).
ii. “I think I’d personally improve my driving because it’s quite rare that I drive with somebody else in my
car, so I don’t have the opportunity to gain feedback on my own driving. And I guess that would be the
same for so many other people as well who are solo drivers and don’t also have passengers. This is a
good system to provide that feedback that they lack” (22, F, 1).
iii. “I’m also very interested in the safety. I want to know if I’m keeping to the right speed, because
obviously with children in the backseat, you want to make sure … My motivation would be more about
safe driving because I’m carrying two children” (35, F, 17).
iv. “I think it [monitoring safe driving] would be really good but I don’t think many people would like to
be told that you're not driving safely or feel that they’re being nagged about it. I think it needs to be like
a subconscious thing” (61, M, 40).
v. “If you’ve driven safely and you haven’t had a crash, why are you going to purchase a product? Or if
you’ve never had a speeding fine or a ticket or anything, why are you going to purchase a system that’s
going to tell you to drive safe when technically you’re driving safe already? (27, M, 9). “I don’t think
bad drivers would give a damn” (23, F, 7).

c. Help the i. “I like the idea of finding out the economic and environmental benefit of your behaviour” (53, F, 35).
environment “Carbon footprint. Yeah, tracing how you are doing in terms of your carbon footprint” (33, M, 15).

Save Fuel and Money


Twelve participants made a total of 16 statements regarding the perceived

usefulness of eco-safe in-vehicle HMI to save fuel and money, with the vast majority

arguing that such systems could increase their ability to monitor their fuel efficiency

and improve their eco-driving style. Ultimately, this was most often driven by the

desire to save money through reduced fuel and vehicle maintenance costs (Table 4.4,

a.i). A number of participants also suggested that such a system may be beneficial for

drivers who perceived themselves to be “good” drivers based on their crash or

infringement histories, but who had little insight into the eco-friendly nature of their

driving behaviour (Table 7, a.ii).

Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2) 89
The desire to use an eco-safe in-vehicle HMI to save fuel and money was noted

even among participants who expressed less concern with the potential for the system

to positively impact upon driving safety and/or the environment (Table 4.4, a.iii).

Personal monetary gains from using the system may be a particularly useful aspect of

the system to focus on when attempting to motivate more problematic and unsafe

drivers to use the system.

Improving Safe Driving


Ten participants made a total of 14 statements regarding the perceived usefulness

of eco-safe in-vehicle HMI to improve safe driving. A number of participants believed

such a system could enhance safety by increasing their awareness of their driving

behaviour, through an increased ability to monitor their behaviour, thus alerting them

to opportunities to improve their driving ability and overall safety (Table 4.4, b.i). This

was particularly pertinent amongst drivers who typically travelled alone and thus had

fewer opportunities to receive other forms of feedback regarding their driving

behaviour and for drivers without technology in their vehicle to help them ascertain

their fuel efficiency (Table 4.4, b.ii).

A number of participants also suggested that the presence of passengers, and in

particular children, would increase their motivation to use such an in-vehicle system

to improve their driving safety (Table 4.4, b.iii). However, some participants suggested

that direct feedback might not be well received by all drivers, given that it may conflict

with their own perceptions of their driving ability and levels of safety (Table 4.4, b.iv).

A number of participants noted that such a system might help drivers understand that

there is more to being a “good” driver than a clean crash or offence record, by

highlighting the efficiency, or lack thereof, of their driving (Table 4.4, b.v). However,

they also questioned whether such drivers would be interested in using such a system.

90 Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2)
Help the Environment
Overall, only three participants reported that they believed an eco-safe in-vehicle

HMI would be useful for helping the environment. Indeed, only three statements from

as many participants were made in relation to this subtheme (Table 4.4, c.i). This

finding is somewhat surprising given the media attention given to issues such as

climate change, fossil fuels and environmental sustainability. While further research

is required to more comprehensively investigate this issue, it may be that many drivers

are either uninformed or overwhelmed by the myriad of information in the media, or

have low perceptions regarding their ability to make a significant difference.

Alternatively, drivers may hold paradoxical attitudes whereby they believe that it is

important to protect the environment while also exhibiting behaviours that are not

congruent with this belief. Such paradoxical attitudes have been demonstrated

previously in the area of road safety, such as when drivers acknowledge that speeding

is dangerous while also reporting frequent speeding behaviour (Fleiter & Watson,

2006)

Following initial questions which discussed eco-safe in-vehicle HMIs in a

general sense, from which the preceding themes emerged, participants were presented

with a number of specific concept designs. The remaining themes outline participant

perceptions and attitudes toward these concepts in particular.

4.5.4 Intentions to Use an Eco-Safe In-Vehicle HMI


A number of factors were found to influence intentions to use an eco-safe in-

vehicle HMI system. These included: a) driver type; b) perceived ease of use and

usability; c) accessibility and cost; d) normative influences; and e) attitudes toward

safe and eco-friendly driving. These factors are discussed in more detail in the sections

below.

Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2) 91
Driver Type
Twenty-two participants made a total of 44 statements regarding the influence

of driver type on intentions to use an eco-safe in-vehicle HMI. Overall, participants

agreed that particular groups of drivers would be more likely to benefit from the use

of such a system compared to others. Indeed, a number of participants argued that

novice drivers could use the HMI to educate themselves from the early stages of

driving, improving their fuel efficiency and safety and increasing the likelihood that

an eco-safe driving style would be retained (Table 4.5, a.i).

In addition, a number of participants reported that parents of teenagers could use

the system to monitor the driving behaviour of their children (Table 4.5, a.ii). Such an

approach would serve a number of benefits, most notably motivating a key population

of drivers who typically are more resistant to voluntary use of such systems and

contributing to wide-scale system adoption among a new generation of drivers. This

also represents an opportunity for the use of gamification principles in the design of

the in-vehicle interface (Orfila et al., 2016; Steinberger, Schroeter, Foth, & Johnson,

2017), which could facilitate social persuasion among family and friends, such as

motivating behaviour change through competition and additional incentives (e.g. if

children produce better eco-safe driving scores parents agree to pay for fuel).

Moreover, parents who install such a system for their children may also be more likely

to use the system themselves when driving, further increasing the benefit of the system.

92 Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2)
Table 4.5. Selected Quotes Relevant to Theme 3 - "Intentions to use an eco-safe in-
vehicle HMI"

Subtheme Relevant Quotes

a. Driver type i. “I do agree that it could be used to educate people, like learners … if you had something like this,
it would be like, you’re being watched by big brother, and you cannot speed. But also that’s sort
of the highest risk age group as well, 17 years to 25” (35, F, 17).“It is good because if you learn
to drive efficiently when you first start driving, then you would probably do it for the rest of the
time you drive, which is very good.” (28, M, 10).
ii. “It will really work with families if they’ve got to share a car. Like if you are a parent and you
share your car with just say, two teenagers, you’d be like, ‘okay, you can drive the car but you
need to use the system. So, when you get home, I can see how you’re driving’. It would be a perfect
system to go, ‘how did my son drive’. Like did he just drive it super aggressively and he’s not going
to get away with that” (27, M, 9). “Parents putting it [the eco-safe in-vehicle system] in to watch
what the kids doing driving their car…I think parents may find it important for their children
learning to drive” (22, F, 5).
iii. “I think the people who would benefit most from this are people who think they’re good drivers but
don’t have any feedback … but then they are still going to go, the system is wrong, I’m good … I
don’t think it would improve my driving because I think I’d be interested to know that I wasn’t
leaving enough space, but I think because I’ve been driving for 14-15 years and not rear-ended
anyone. So, I’m pretty happy that I can drive at that level, whereas the system might not think I
can, which is fair enough but, I don’t think I’d pay it much attention” (30, M, 13).“I’d probably
say no. It’s beneficial but I don’t care. I will make right choices … I know I’m doing what I’m
doing. I know its fuel efficient, I don’t need to have a thing [in-vehicle system] tell me” (29, M, 10).
iv. “I’m not into technology … so I probably wouldn’t [use an eco-safe in-vehicle system]. Plus, I’ve
been driving for a long time … I think it’s brilliant for people who are comfortable with all this
new technology but for me, I’m just from that era … I know a lot of people that are of my age that
are very into new technology but I’m one of those people that can’t” (56, F, 36).
v. “I’ll probably suggest it to my boss to give out to all the different drivers and stuff just because
sometimes you have trades people getting back real quick from a delivery or something, but he
doesn’t know what they’ve been doing – flying and just throwing the vehicle around the corners
and stuff. So in that sense, I think it’s definitely got a lot of applications in businesses” (22, M, 4).
vi. “Maybe use it as punishment for people, if someone’s continuously a bad driver or been drunk or
reckless driving, throw it in their car and then the report is there for what they’re doing and then
police can see it. So a mandatory system for some people” (29, M, 10). “I think the most valuable
application for it [an eco-safe in-vehicle system] is mandatory use for bad drivers. The people that
need to monitor that sort of behaviour … safety-wise, people that would most benefit from it, I think
they’re doing the wrong thing now because they don’t care” (56, F, 36). “People with a lot of
traffic infringements, lots of speeding fines, who have low points, that kind of stuff” (22, M, 4).

b. Perceived ease i. “Depends how user friendly it is and how quickly I can do it [figure out how to use it]. If it’s going
of use and to be something that’s really time-consuming, maybe not [I wouldn’t use it]” (23, M, 1). “I would
usability use it, only if it’s easy to use. Like I use cruise control all the time because it’s really simple” (23,
F, 7).

c. Accessibility i. “I would use the system if I had access to it” (19, M, 1).
and cost ii. “It depends on how much it would cost. If it was free, yeah. If it cost money, probably not” (27, M,
9). “If it was supplied for free [I would use it]. I wouldn’t pay for it as an extra in my car” (29, M,
10). “I would consider it value adding for the car, like if I was looking at purchasing a brand new
car, and it had it already installed and ready to go, I’d pay a little bit [for the system]” (23, F, 7).
iii. “If it’s going to cost me $50 or $100, that’s a tank of fuel already. Like, unless it’s going to save
me $5 a tank or $10 a tank, I’m not going to do it” (27, M, 9). “If it had studies that said it was
going to save me like a $100, then I would probably pay like $10 or $20 to get it, but I’m not going
to pay more than it’s worth to have” (22, F, 5).
“I’d pay for it, if there were rewards, I’d pay for it and then maybe drive better to gain more
rewards” (30, M, 13).

d. Normative i. “If I did have friends using it that would make me want to use it” (23, M, 1). “I’d probably have a
influences try if a friend had tried it and said it's good or if I saw it on a TV show or a television ad” (31, F,
6).
ii. “If it came out in your insurance letters or when you renewed your licence, okay, check out this”
(23, F, 7).

Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2) 93
On the contrary, a number of participants expressed concerns that more

experienced drivers may be reluctant to accept feedback from an eco-safe in-vehicle

HMI system, particularly if it contradicts their own perceptions of their eco-safe

driving performance. Indeed, some participants suggested that they would ignore

feedback from the system if such a contradiction was observed (Table 4.5, a.iii). Such

statements may reflect a distrust in technology to accurately evaluate driving

behaviour with consideration of contextual and situational factors, or even a

subconscious unwillingness to receive objective feedback associated with their

optimism biases. Indeed, one participant mentioned her extensive previous driving

experience and lack of skill with technology as the primary reason she would be

uninterested in using such a system (Table 4.5, a.iv).

Overall, this finding and others related to apparent optimism biases among some

drivers, suggests a need to better educate drivers regarding relative versus absolute

risks. That is, it appears that many drivers believe that an absence of traffic crashes or

infringements from their driving history is evidence of their ability and safety as a

driver. While in some cases this is likely to be a true reflection, in others it may simply

be a function of the low absolute risk of involvement in a traffic crash in Australia

(Bureau of Infrastructure Transport and Regional Economics, 2015), with consistent

evidence highlighting the relative risks associated with a variety of illegal and unsafe

driving behaviours (Peden et al., 2004).

Another participant suggested that eco-safe in-vehicle HMIs may also be

beneficial for use among business vehicle fleets, in particular those industries where

driving between locations in a timely manner (e.g., delivery drivers) is encouraged

(Table 4.5, a.v). Given previous research demonstrating the increased risk of vehicle

crashes and traffic infringements among occupational drivers (Robb, Sultana,

94 Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2)
Ameratunga, & Jackson, 2008), the acceptability and effectiveness of such a system

in vehicle fleets is worthy of further research. Indeed, such an application may have

subsequent benefits if occupational drivers realise the positive impacts of the system

and become motivated to use the system in their personal vehicles. It is at this point

worth noting the snowball effect such a system would have in shared vehicles if all

drivers were motivated to use the system, which might be likely if significant others

(e.g., family members) are using the system and share positive experiences associated

with its use.

Finally, a number of participants argued that an eco-safe in-vehicle HMI system

would be a beneficial sanction for offending drivers. That is, participants suggested

making it mandatory for offending drivers to use the system in order to continue

driving or following a licence suspension (Table 4.5, a.vi). Such initiatives are already

evident in the area of road safety, such as the use of alcohol ignition interlocks (Elder

et al., 2011) and suggestions for the use of mandatory intelligent speed adaptation

(ISA) as a sanction for speeding offenders (Van der Pas, Kessels, Vlassenroot, & van

Wee, 2014).

Perceived Ease of Use and Usability


Six participants made a total of seven statements regarding the perceived ease of

use and usability of the concept eco-safe in-vehicle HMIs on intentions to use. The

degree to which participants perceived the systems as being easy to use, while also

being effective, efficient and engaging, was reported as having a positive relationship

with self-reported intentions to use such a system (Table 4.5, b.i). This finding is

consistent with previous research using the Technology Acceptance Model (TAM) to

predict intentions to use in-vehicle technology (Ghazizadeh & Lee, 2014). However,

given the demanding nature of the task of driving, the challenge remains of how to

Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2) 95
design a system such that it meets these user requirements without distracting the user

from the primary task (i.e., driving) and producing cognitive overload. Thus, the goal

should be to design a system that maintains functionality (e.g., effectiveness,

efficiency, user-satisfaction) while requiring minimal effort from the driver.

Cost and Accessibility


Fifteen participants made a total of 19 statements regarding the influence of cost

and accessibility of the concept eco-safe in-vehicle HMIs on intentions to use.

Participants reported a greater intention to use such a system if it was readily accessible

to them (Table 4.5, c.i). Moreover, the cost of the system appeared to heavily impact

upon perceptions of accessibility, with the majority of participants reporting that

reduced system purchase costs would be associated with greater intentions to use the

system (Table 4.5, c.ii). This finding is consistent with previous research (Stevens &

Burnett, 2014) arguing that cost, among others factors, is a key criteria of practical

acceptability.

Overall, participants were quite polarised regarding their willingness to pay for

such a system. While some argued that they would only be willing to use the system

if it was free, others reported a willingness to purchase the system, and perceived it as

a device that would add value to their vehicle (Table 4.5, c.ii). This finding has

important implications for the design of such a system. That is, the design of the system

as a smartphone application would facilitate low consumer purchase costs, but would

be associated with a number of additional issues related to the device characteristics,

such as automaticity of use. On the other hand, a stand-alone after-market or built-in

device would allow for the design of a system that addresses best practice device

characteristics, but would undoubtedly be more expensive and thus unattractive to

some potential users.

96 Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2)
Encouragingly, a number of participants argued that they would be willing to

invest in an eco-safe in-vehicle HMI if they were confident that the financial benefits

associated with using the system, such as saving money on fuel and vehicle

maintenance, would outweigh the costs of purchasing the system (Table 4.5, c.iii).

Financial incentives, such as rewards and prizes associated with eco-safe driving

behaviour, were also reported as being positively associated with intentions to use the

system (Table 4.5, c.iv). This finding is consistent with previous research that found a

positive impact of a reward strategy on fuel efficiency among bus fleets (Lai, 2015).

Such rewards could include discounts on vehicle insurance and/or registration, or

prizes associated with eco-safe driving behaviour, such as movie tickets, fuel vouchers

or other incentives.

Effectiveness of an incentives approach has been evidenced in a trial conducted

with Provisional drivers in Newcastle, Australia, as part of the Samsung S-Drive

smartphone application initiative, whereby young drivers earned rewards for safe

driving (Bennett, 2014). So, for incentives to be effective, the rewards need to be

meaningful to the user, both in terms of the degree to which the user wants the reward

and perceives the reward as being worth the time and effort required to earn it. Further

research should also investigate the long-term feasibility of such reward schemes.

Normative Influence
Four participants made a total of five statements regarding normative influences

on intentions to use the concept eco-safe in-vehicle HMIs. Participants highlighted the

influence of their peers and social network on the likelihood that they would use the

system, suggesting a greater intention to use the system themselves, if their peers used

and recommended it (Table 4.5, d.i). Insurance companies and transport authorities

were also reported as being able to influence intentions to use an HMI, with a number

Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2) 97
of participants suggesting that authoritative persuasion or recommendation could be

an effective approach to encouraging system use (Table 4.5, d.ii).

Taken together, these findings indicate the importance of a trustworthy source

from which drivers can obtain evaluative information regarding the utility of the

system. These findings are also consistent with previous research on the positive effect

of normative influences on technology acceptance. Furthermore, following on from

the previous section, the findings suggest a potential partnership with insurance

companies and transport authorities to offer discounts on insurance and/or vehicle

registration associated with system use and evidence of appropriate eco-safe driving

styles.

4.5.5 Eco-Safe In-Vehicle HMI Design Characteristics


Participants were also asked to discuss the specific design characteristics that

might influence their likelihood of using an eco-safe in-vehicle HMI. This included a)

various device characteristics, such as device type and automaticity, as well as b)

feedback characteristics, such as feedback type (e.g., auditory, visual), accuracy and

clarity of feedback and feedback customisation.

Device Characteristics
Eighteen participants made a total of 36 statements regarding their perceptions

toward the device characteristics of the concept eco-safe in-vehicle HMIs. There was

considerable variety in participant views regarding the most effective approach for

presenting feedback to drivers with regards to device type. Overall, the majority of

drivers reported a preference for one of three approaches: (i) an in-built system that

comes installed in a new vehicle or can be retrofitted to older vehicles; (ii) a stand-

alone, mobile system that can be fitted in vehicles (e.g., similar to many dash cams

and GPS systems); or (iii) a smartphone application that can be linked to the vehicle.

98 Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2)
Table 4.6. Selected Quotes Relevant to Theme 4 - "Eco-safe in-vehicle HMI design
characteristics"

Subtheme Relevant Quotes

a. Device i. “I think having it in-built [into the vehicle] will be much more effective because who’s going to opt
characteristics to go and actually purchase this system?” (22, F, 1). “I think it should already be in the car when
I’m getting the car. That might make me use it” (20, F, 3).
ii. “See my trouble is I drive a lot of different vehicles every day. Like it’s [the eco-safe-in-vehicle
system] going to have to be transferable from vehicle to vehicle” (30, M, 13). “I probably do think
that having a little additional system would be the easiest way to roll it out” (22, F, 1).
iii. “If it would look good in your car. Because a lot of people, like myself, I wouldn’t want something
that looks disgusting on my dash even if it is going to provide safe driving … I don’t need to have
something hanging out of my dash” (61, M, 40).
iv. “Well the most efficient way would be the whole thing would be phone linked. Like that would be
the easiest way to bring it out” (30, M, 13). “I think a phone is obviously the easiest way despite
all its possible downsides. The majority of people who drive have a phone … a two minute
download and you’ve got it” (22, M, 4).
v. “I’ve a little bit of an issue with the whole thing being at all phone linked. Like if you’re trying to
disengage phones from driving… in your mind you are linking a phone and a car as one thing”
(30, M, 13). “I definitely think scrapping the phone idea, I don’t think it should have any relation
to mobile phones at all” (22, F, 1).
vi. “For the system to function, it’s going to need to have a lot of sensors to know what it’s doing. Like
an iPhone’s not going to be able to tell you when you’re accelerating or decelerating” (30, M, 13).
vii. “I've got kids fighting in the back seat and I've just got to get to ballet. I'm just not going to stop
and do this and tap on my phone. But if it were automatically starting up in the car or somehow
connected to the car, it would be okay…. it shouldn’t be a process of having a setup stage as you
get into the vehicle. It should just be, switch it on when the car turns on” (35, F, 17). “It would
need to be quick to setup because if I was in any kind of hurry, and I jump in the car, and it was
going to take a minute to log on and do stuff to set it up, I’d be less likely to use it. But if it was
pretty quick… yeah that would be quite good” (39, M, 21).
viii. “You can combine it with GPS and all those kinds of things” (23, F, 7).
ix. “A lot of cars now have a little screen that will show you everything and if it just popped up on
that, for me, it’s just looking, it’s not on the windscreen, I’m just glancing to the side and that
would be fine” (35, F, 17).

b. Feedback i. “I think even having the options for each individual driver, to just like whether they want to
characteristics customise, semi-customise because everybody’s going to have different opinions and different
preferences … That just makes it a little bit more enjoyable, a little bit more relatable” (22, F, 1).
“If you can have the audio or the visual [feedback], or both of them, some people will like the
visual without the audio, or some will like the audio without visual. It needs to be personal” (61,
M, 40).
ii. “Both auditory and visual. Because just in case if you miss the visual, the audio will help you” (19,
M, 1).
iii. “I reckon visual is very quick, easier when you're driving … I think that’s very quick and easy to
communicate visually” (34, M, 14). “I’d prefer more audio stuff … I prefer to be told with a beep
or something like that that I’m too close to something or going outside my lane or something like
that”(32, M, 15).
iv. “I would like to have colours … so you don’t even need to look at it to know what's going on, you
just sort of see the colour and you know” (28, M, 10). “If you just keep a particular colour of light
say red and green. Like if you do bad red pops up and you're not actually looking at the colour but
you can still see” (26, M, 3).
v. “I keep on thinking about the night, about the lights [being distractive inside of the vehicle] in the
night time” (25, M, 6). “You don’t want to punish someone if you want them to keep using your
thing, so red is an aggressive colour” (23, M, 1).

Those participants who argued that the most effective approach would be for the

system to be in-built in new vehicles, with an option for retrofitting in older vehicles,

appeared to be primarily motivated by ease of accessibility. Indeed, these participants

generally suggested they would be unwilling to purchase such a system if it was not

built into their vehicle (Table 4.6, a.i). However, it was noted that this approach would

Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2) 99
not be advantageous for individuals who drive more than one vehicle (e.g., both

personal and work vehicles).

Instead, three participants suggested that the system should ideally be mobile

and transferable from one vehicle to another, such as a stand-alone mobile device,

similar to many existing dash-cams and GPS systems (Table 4.6, a.ii). However, a

number of participants suggested that such an approach would need to consider

aesthetics and not detract from the personality and overall look of their vehicle (Table

4.6, a.iii).

The majority of participants acknowledged that a smartphone application

approach would be the most readily accessible approach for a wider audience given

the proliferation of smartphones, as well as being transferable from vehicle to vehicle

(Table 4.6, a.iv). However, a major criticism of this approach was the perceived irony

of having an intervention designed to promote safety that simultaneously and actively

encouraged the use of a mobile phone while driving. Indeed, a number of participants

suggested that this represented a contradictory safety message and that mobile phones

should therefore have nothing to do with the system (Table 4.6, a.v). In addition,

participants raised concerns regarding the technological capacity for modern

smartphones to measure the specific behaviours of interest to an eco-safe in-vehicle

HMI, suggesting that additional measurement equipment would also be required.

While additional sensors, or connection to the on-board diagnostics (OBD) of a

vehicle, would be required to measure some eco-safe driving behaviours, statements

from some participants suggested they were not fully aware of the potential capacities

of their smart phones (Table 4.6, a.vi).

Regardless of the preferred approach, a common preference among participants

was for the system to operate with a high level of automaticity. That is, participants

100 Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A
Qualitative Study (Paper 2)
almost unanimously expressed a desire for the system to start automatically upon the

commencement of a trip and be straightforward to set up. Indeed, it was argued that a

less automatic system is more likely to conflict with daily life responsibilities and be

associated with lower rates of prolonged usage (Table 4.6, a.vii). Moreover, a number

of participants argued that such a system would be more appealing and useful if

combined with additional features, such as GPS, dash-cam and other in-vehicle

systems (Table 4.6, a.viii). That is, increasing the perceived utility of the system was

reported as being likely to enhance driver acceptance of the system and subsequent

intentions to use.

Interestingly, very few participants reported a preference for the use of head-up

displays (HUD), whereby visual information and feedback is projected onto the

windshield of the vehicle. A number of participants suggested that drivers are already

accustomed to looking at screens within the vehicle, whereas HUD may be more

distracting (Table 4.6, a.ix).While it remains unclear, this finding may represent a

reluctance among participants to try a new technology or a lack of familiarity regarding

the specific features and capabilities of this technology, however further research is

required to gain a more comprehensive understanding of the issue.

Overall, a high level of system automaticity appears fundamental to system

acceptance and use. In addition, such an approach is more likely to increase the

likelihood of long-term use and eventual adoption of use of the system as normative

behaviour. Enhancing the functionality and aesthetic design of the system would also

positively influence driver acceptance and subsequent rates of use.

Feedback Characteristics
Twenty participants made a total of 34 statements regarding their perceptions

toward feedback characteristics of the concept eco-safe in-vehicle HMIs. The majority

Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2) 101
of participants acknowledged that there are likely to be individual differences in

preferences regarding how feedback is provided by an eco-safe in-vehicle HMI and

that feedback would be most effective when an individual can reliably relate to the

information (Table 4.6, b.i). Such a finding highlights the importance for a system to

be readily customisable in regards to the manner in which feedback is received.

The primary debate regarding feedback characteristics centred on the most

preferred presentation format of feedback – namely auditory, visual or a mixture of

both. Participants varied considerably in their preferences, further highlighting the

importance of providing drivers with the ability to customise their feedback

preferences. That is, while some participants reported a preference for auditory

feedback, others suggested a preference for visual feedback (including the use of

ambient light). A number of participants even suggested that the most effective

approach may be a mixture of both auditory and visual feedback, such that they

complement one another to reduce driver distraction, as well as reducing the chance a

driver may miss the feedback information (Table 4.6, b.ii). Overall, the preference for

a particular approach appeared to be contingent on which feedback was perceived to

be most straightforward and easy to understand, and least distractive (Table 4.6, b.iii).

A number of participants were positive towards the use of ambient light

feedback, such as a device that glows particular colours in relation to various feedback

messages. Indeed, it was noted that such an approach may reduce driver distraction

associated with the system by allowing them to attend to the feedback information in

a largely peripheral manner (Table 4.6, b.iv). However, a number of participants

expressed concerns regarding the distractive qualities of ambient light at night, and the

emotional perception toward particular colours and the subsequent effect on

102 Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A
Qualitative Study (Paper 2)
acceptability (Table 4.6, b.v). These findings yet again highlight the importance of

providing drivers with the ability to customise their feedback preferences.

Taken together, participants desire an ability to customise the feedback

characteristics of the eco-safe in-vehicle system to suit their specific preferences. That

is, they believe the system should be flexible enough to adapt to the specific needs of

each individual user in various driving contexts.

4.5.6 Potential Problems Associated with Using an Eco-Safe In-Vehicle HMI


As has been mentioned in the previous sections, participants reported a number

of potential problems associated with using an eco-safe in-vehicle HMI, including a)

driver distraction, b) privacy concerns, c) perceived accuracy and trust in the system,

d) novelty, in particular the impact of novelty on prolonged use, and e) the promotion

of unsafe and illegal driving behaviours, such as mobile phone use while driving.

Distraction
Eleven participants made a total of 15 statements regarding the potential for use

of the concept eco-safe in-vehicle HMIs to result in driver distraction. Indeed, this was

the most common concern associated with the use of the system. As mentioned

previously, various device types and feedback approaches were argued to have the

potential to distract a driver’s attention from the driving task and thus reduce safety.

Participants highlighted the irony of this risk given that the system is designed to,

among other things, improve safety (Table 4.7, a.i). It appeared that feedback

preferences among participants were heavily influenced by concerns regarding the

potential distraction associated with that feedback. Specifically, a number of

participants suggested that both auditory and visual feedback had the potential to be

distracting or irritating to a driver and that distracting feedback may reduce the

likelihood of continued use of the system (Table 4.7, a.ii).

Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2) 103
Table 4.7. Selected Quotes Relevant to Theme 5 - "Potential problems associated
with the system"

Subtheme Relevant Quotes

a. Distraction i. “Information that’s coming up on the screen, that’s a distraction. That’s no different to texting while
you drive or any other distraction like that. To me it defeats the purpose of safety for something that’s
trivial” (39, M, 21). “I think that [an in-vehicle system] could be distracting … like I know for some
people it’s not distracting, but for other people, they can’t even use a GPS and look away from the
road for two seconds” (22, F, 5).
ii. “It can be irritating … I might turn it off” (51, F, 35).

b. Privacy i. “I’d never use it if it any of the information was sent to other people. Never ... If it ever goes to the
concern police or anything like that, if you’re really being unsafe, I just wouldn’t use it” (23, M, 1). “I think
also it would need to have no sort of tangible memory in regards to what you’ve been doing ... If it
[the eco-safe in-vehicle system] possibly became incriminating for me, then I won’t be using it” (30,
M, 13).

c. Perceived i. “It [the feedback] has to be realistic, that’s what I always say … Is it going to be, ‘you have to slow
accuracy down in a kilometre’ when you can easily slowdown in 200 metres and doesn’t make it like you’re
and trust in driving aggressively. I’m not going to drive like an 80 year old man, because you are going to lose
the system time in your day ... If the system’s going to tell you that that’s a safe distance, you’re going to be sitting
in traffic and people are going to be around you going, ‘what the hell is this guy doing? There's like
6 car lengths in front of him that anyone could be sitting in’ … If it keeps coming up when you feel
like, no, I’m not travelling too close, no I’m not decelerating too fast, if it’s not set to a standard that’s
not insanely slow driving [then I’d be more likely to use the system]” (27, M, 9).“I’d get really pissed
if I got penalised because I went down beside a car and it said that it was too close. I'm like, well, if
it’s not scraping, it’s not close, it fits” (29, M, 10).

d. Novelty i. “I think in the short term I’d use it. But long-term – it depends upon the finer elements such as alerts
and other little bits … as to whether I use it long term or not” (22, F, 1). “I don’t know for how long
[I would use it]. Like it will be interesting for a little bit, but I think you lose interest, especially if
you’ve fixed your bad habits” (27, M, 9). “I think for me, the problem will be the consistency of usage,
like how long are you going to use it for? There's always this period when people start using
[technology] and they are super excited and after a point they are like, ‘oh, it’s too much of a hassle’”
(22, M, 5).
ii. “Probably once I get a sense of how I was doing, kind of like using a pedometer, it’s like ‘okay, I walk
this many average steps, I don’t need to wear it anymore’. So then you go, ‘okay, I actually drive okay
… I really don’t need that feedback anymore’” (53, F, 35).

e. Promotion i. “I might try and see how I could push the system, like make it flash a lot that I’m doing things wrong
of unsafe and people would do that, like when you used to have the alcohol breath testers in the pub, you’d see
and illegal who can get the highest reading, it would be similar thing with this” (30, M, 13).“Good idea, if it
driving doesn’t get abused or misused. I can see my teenage son … use it for bad behaviour, instead of
behaviours improving, trying to compete in bad behaviour” (51, F, 35).

These findings are consistent with research which has questioned the impacts of

eco-driving and other types of in-vehicle systems on driver safety (Vaezipour et al.,

2015). Any additional interaction while driving may cause distraction due to increasing

the complexity of the main task and therefore it may be demanding regardless of

whether it is designed for eco-driving or safe driving style (Oviedo-Trespalacios,

Haque, King, & Washington, 2016). Therefore, these concerns are considered during

the design and development of in-vehicle systems, such that an eco-safe in-vehicle

system has minimal impact on driver safety. One potential approach to reducing the

104 Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A
Qualitative Study (Paper 2)
distracting qualities of such a system, which has been highlighted earlier, is ensuring

a high level of flexibility in the available feedback preferences, such that users can

readily customise such feedback according to their individual preferences and/or the

driving context.

Privacy Concerns
Twelve participants made a total of 15 statements regarding privacy concerns

associated with the concept eco-safe in-vehicle HMIs. The primary concern regarding

privacy was whether data collected by the system was accessible by authorities (e.g.,

law enforcement, insurance companies, or government agencies) and could

subsequently be used to incriminate them. Indeed, participants almost universally

reported that such a feature would result in them not wanting to use the system (Table

4.7, b.i).

According to Ghazizadeh and Lee (2014), perceptions of privacy can be

negatively influenced depending on the type of data being collected and who is able to

view the data. Privacy may also be a particular concern for specific groups of drivers

who may be more reluctant to be monitored due to their attitudes towards technology

(e.g., older drivers) or the impact that data sharing and accessibility may have on their

occupational security (e.g., commercial drivers) (Hickman & Hanowski, 2011).

Perceived Accuracy and Trust in the System


Seven participants made as many statements regarding their perceived trust in

the concept eco-safe in-vehicle HMIs, which appeared to be heavily dependent on

perceptions of the accuracy of the system assessments of their eco-safe driving

behaviour. Specifically, participants appeared to be most concerned that the system

would be too conservative in its measurements, resulting in drivers receiving feedback

Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2) 105
that their driving is unsafe or not eco-friendly, and that this feedback would be

contradictory to their own beliefs about the status of their behaviour (Table 4.7, c.i).

This finding has important implications for drivers who exhibit optimism biases

in relation to their eco-safe driving ability. That is, such a driver may be more likely

to attribute an observed disconnect between their perception of their eco-safe driving

and the behavioural measurements of an HMI to conservative feedback settings, when

in reality it is an accurate description of their driving performance. This highlights the

need for clear communication to the driver regarding how behaviour is measured such

that it is perceived to be transparent and legitimate. Overall, trust seems to be an

important concept in regards to self-reported intention to use an eco-safe in-vehicle

system.

Novelty
Seven participants made a total of nine statements regarding the potential impact

of novelty on continued, long-term use of the concept of eco-safe in-vehicle HMIs.

Indeed, a number of participants suggested that the novelty of the system may be short-

lived and that particular design characteristics, overall perceived usefulness and ease

of use (e.g., automaticity) would be crucial factors influencing the likelihood of longer-

term use (Table 4.7, d.i). In addition, three participants suggested that there may be a

reduction in perceived utility of the system associated with positive behaviour,

questioning what motives a driver would have to continue using the system if their

behaviour no longer resulted in feedback promoting behaviour change (Table 4.7, d.ii).

Promotion of Unsafe and Illegal Driving Behaviours


Two participants made a total of seven statements regarding the promotion of

unsafe and illegal driving behaviours associated with the concept of eco-safe in-

vehicle HMIs. As stated earlier, if a Smartphone application approach was adopted,

106 Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A
Qualitative Study (Paper 2)
that a mixed message would be sent regarding the dangers of mobile phone use while

driving. In addition, the HMI may inadvertently promote unsafe driving behaviours

among high-risk users, who may attempt to “get the worse score” possible. As a result,

a number of participants reported that feedback must be positively-geared in order to

minimise the likelihood of misuse of the system (Table 4.7, e.i). It may be necessary

to only provide positive feedback in order to prevent driver misuse. This may be

particularly relevant for younger drivers who may be more likely to abuse negative

feedback elements of the system.

4.6 CONCLUSION AND IMPLICATIONS

The purpose of the current study was to gather information to inform the design

and development of an eco-safe in-vehicle HMI. This study was exploratory in nature

and adopted a user-centred research approach to gain a comprehensive understanding

of the specific needs of drivers, including their motivations and requirements, as well

as factors that would represent limitations to use an eco-safe in-vehicle HMI. This

research adds to the existing body of research given that it is, to the authors’ best

knowledge, the first user-perception study of an in-vehicle HMI system designed to

address both eco-driving while improving safe driving in Australia. The methodology

is supported by the robust principles of the Technology Acceptance Model (TAM). A

synthesis of findings is presented in Table 4.8.

4.6.1 Implications for Practice


Overall, this study has highlighted the potential for an eco-safe in-vehicle HMI,

designed to provide information and feedback to drivers, to experience relatively high

levels of driver acceptance if carefully designed with user needs in mind. We found

that the perceived usefulness of an eco-safe in-vehicle HMI appears to be primarily

influenced by monetary savings associated with fuel and vehicle maintenance costs.

Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2) 107
However encouragingly, improving awareness of eco-safe driving style, and

supporting eco-driving and benefits to the environment also appear to influence

perceptions of usefulness, provided the system also saves the user money. We also

found important individual differences regarding motivations and attitudes toward

eco-safe driving. These factors can influence driver acceptance, as well as perceived

usefulness of an eco-safe in-vehicle HMI, and therefore must be considered.

The importance of considering user requirements and needs was also highlighted

when investigating perceptions towards a range of design and feedback characteristics.

Taken together, these findings highlight a fundamental need to develop an eco-safe in-

vehicle HMI that affords users the ability to customise feedback to suit their specific

preferences. That is, the system must be flexible enough to adapt to specific needs of

each individual user and in various driving contexts.

Moreover, a number of barriers to acceptability of eco-safe in-vehicle HMI

including trust and privacy were reported by participants. Specifically, some groups of

drivers may be more reluctant to be monitored due to their demographic and

personality characteristics and it may be useful for the system designers to develop

systems to remove or minimise these barriers. In general, distraction and the

inadvertent promotion of unsafe and illegal driving behaviours are perceived to be the

most likely potential problems associated with the use of an eco-safe in-vehicle HMI.

Thus, impacts on driver safety should be considered during the early stages of design

and development, such that an eco-safe in-vehicle HMI has minimal impact on driver

safety.

This study also highlighted a number of potential barriers to motivating more

problematic drivers to engage with such a system. However, much like many other

road safety initiatives, there are considerable gains to road safety, as well as the

108 Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A
Qualitative Study (Paper 2)
environment, which are readily achievable through the use of an eco-safe in-vehicle

HMI by members of the driving population and therefore an inability to engage more

problematic drivers would not necessarily represent ineffectiveness of such a system.

So, this study has highlighted the potential usefulness of an eco-safe in-vehicle

HMI for providing eco-safe advice and feedback to drivers. Moreover, it provides a

number of contributions to our understanding of eco-driving in-vehicle HMIs, as well

as some practical implications for designers and developers for increasing the level of

driver acceptance of in-vehicle systems to increase driver performance.

Table 4.8. Synthesis of Findings

Themes Findings
Perceived importance  General support of the notion that monitoring eco-safe driving behaviour is
of eco-safe driving important and that an in-vehicle HMI has the potential to improve driving
behaviours by increasing awareness of behaviour.
 Monitoring perceived to be particularly beneficial in relation to behaviours
that are difficult to subjectively monitor or that have considerable impacts
on safety.
 Some evidence of an illusory superiority bias, whereby individuals
overestimated their skills and abilities relative to other drivers, and believe
in is more important to monitor safe driving behaviours of other drivers
compared to themselves.
Perceived usefulness  Strong perceptions that an in-vehicle HMI could increase the ability to
of system monitor fuel efficiency and improve eco-driving style, most often driven
a. Save fuel and by the desire to save money through reduced fuel and vehicle maintenance
money costs.
 Some evidence that drivers believe that such a system may enhance insight
and increased awareness among drivers who do not drive in an eco-friendly
manner.
b. Improve safe  Strong perceptions that an in-vehicle HMI could enhance safety by
driving increasing awareness of driving behaviour through an increased ability to
monitor behaviour.
 Greater levels of reported motivation to use an in-vehicle HMI among
drivers who often drive with passengers, and in particular family.
 Concerns expressed regarding how feedback will be received by drivers
when it directly conflicts with their own perceptions of their driving ability.
c. Help  Limited perceptions that an in-vehicle HMI would be useful for helping the
environment environment. Reasons for this require more research, but may be associated
with low perceptions regarding their ability to make a significant difference
or paradoxical attitudes.
Intentions to use  Strong belief that particular groups of drivers would benefit more from the
system use of an in-vehicle HMI, such as novice drivers, parents of novice drivers,
a. Driver type corporate vehicle fleets and as a sanctioning option for offending drivers.
 Concerns expressed that more experienced drivers may be reluctant to
accept feedback from an in-vehicle HMI, particularly if it contradicts their
own perceptions of their driving performance.

Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2) 109
b. Perceived ease  Perceptions of an eco-safe in-vehicle HMI as being easy to use, effective
of use/usability and engaging observed as having a positive relationship with self-reported
intentions to use such a system.
 Concerns expressed regarding the potential for such a system to distract
drivers from the primary task of driving and lead to cognitive overload.
c. Accessibility and  Participants reported a greater intention to use an eco-safe in-vehicle HMI
cost if it was readily accessible to them, both in terms of access and
affordability.
 Mixed reports regarding willingness to pay for such a system, with greater
willingness reported if the financial benefits associated with using the
system were perceived as likely to outweigh the costs of purchasing the
system.
 Financial incentives, such as rewards and prizes associated with eco-safe
driving behaviour, were also reported as being positively associated with
intentions to use the system.
d. Normative  A number of normative influences were reported that would increase
influences likelihood to use of an eco-safe in-vehicle HMI, including peers, insurance
companies and transport authorities.
 The key element of normative influences appeared to be a trustworthy
source from which evaluative information regarding the utility of the
system could be obtained.
 These findings highlight the potential partnership with insurance
companies and transport authorities to offer incentives associated with
system use and appropriate behaviour.
Design  A well-designed, user-centric system is more likely to be associated with
characteristics relatively high levels of acceptance among drivers.
a. Device  Mixed views regarding the most effective approach for presenting feedback
characteristics to drivers, with three preferred approaches: (i) an in-built system; (ii) a
mobile system; or (iii) a smartphone application.
 Strong beliefs that the device, regardless of the preferred approach, should
operate with a high level of automaticity and would be more appealing and
useful if combined with additional features (e.g., GPS, dash-cam).
b. Feedback  Strong perceptions that an eco-safe in-vehicle HMI must be readily
characteristics customizable in regards to the manner in which feedback is received.
 Mixed views regarding the most appropriate presentation format of
feedback, with support for auditory, visual, ambient light or a mixture of
approaches. Preferences appeared to be largely dependent on which
approach individuals perceived as being less distracting.
Potential problems  Distraction was the most common concern associated with the use of an
a. Distraction eco-safe in-vehicle HMI. This finding has important implications for the
design and development of any system.
b. Privacy  Concerns expressed regarding accessibility of data collected by the system
concern to authorities for purposes of incrimination, and the security of any data
collected by the system.
c. Perceived  Concerns expressed regarding the accuracy of the system, such as
accuracy and conservative measurements. This finding highlights the need for clear
trust communication regarding how behaviour is measured such that it is
perceived to be transparent and legitimate.
d. Novelty  Some concerns expressed that the novelty of the system may be short-lived.
Reported that design characteristics, perceived usefulness and ease of use
would be crucial factors influencing the likelihood of longer-term use.
e. Promote  Concerns expressed that any system that incorporates a smartphone would
unsafe/illegal send a mixed message regarding the dangers of mobile phone use while
behaviour driving.
 Concerns expressed that gamification principles and scoring may
inadvertently promote unsafe driving among high-risk users who may
attempt to “get the worse score” possible, highlighting the need for
feedback to be positively-geared in order to minimise misuse.

110 Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A
Qualitative Study (Paper 2)
4.6.2 Limitation & Future Research
The present study is consistent with previous findings and provides evidence of

the applicability of the TAM in the context of an eco-safe in-vehicle HMI. The findings

suggest additional determinants to the TAM, including driver characteristics, design

characteristics of in-vehicle systems and normative influences on intentions to use an

eco-safe in-vehicle HMI, and future research should seek to evaluate the incorporation

of these determinants into the TAM, specifically in the context of an eco-safe in-

vehicle HMI.

A limitation of this study is that, although data saturation was reached in the

content of responses, the relatively small sample size may impact upon the

generalisability of the results. In addition, the sample consists of predominately

younger, urban drivers. However, given these characteristics, the sample is likely to

be more experienced with technology and as such represent the target audience of a

system like the one explored in this research. Nonetheless, future research will address

this limitation by conducting a user study with a larger, more representative sample.

Another limitation is the fact that user preferences may not necessarily reflect the most

effective characteristics of an eco-safe system. Larger scale, user-centred research is

required to objectively validate the findings of the current study, and represents the

next phase of the broader research project from which this study is derived.

Further research is also required to understand driver behaviour under

technological instruction and ways to enhance driver motivation to adopt an eco-safe

driving style. To explore these issues, we suggest future research should assess the use

of eco-safe in-vehicle HMIs in different settings. Future research should also

investigate the association between novelty and long-term use of in-vehicle HMIs to

gain a clearer understanding of this relationship and identify solutions to address any

issues.

Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A Qualitative
Study (Paper 2) 111
This study is the first stage of a larger project focused on in-vehicle HMI design

and funded by an Australian Research Council Discovery Grant. The findings of this

study will inform the design and development of an eco-safe in-vehicle HMI that will

be evaluated in a state-of-the-art driving simulator.

112 Chapter 4: Enhancing Eco-Safe Driving Behaviour Using In-Vehicle Human Machine Interface: A
Qualitative Study (Paper 2)
Chapter 5: Design of a Gamified Interface
to Improve Fuel Efficiency and
Safe Driving (Paper 3)

5.1 INTRODUCTORY COMMENTS

Similar to the preceding chapter, this chapter focuses on the second step of the

user-centred design approach. As such, the findings presented in this chapter aim to

complement the development of a comprehensive understanding of user needs and

requirements associated with system use. Specifically, this chapter describes

additional findings from the focus group study undertaken with drivers (Study 2a), but

focuses instead on issues associated with system design, development and

implementation. Thus, the findings of this chapter and the preceding chapter share the

same participants and procedures.

The main aim of the paper presented in this chapter (Paper 3) was to investigate

a number of concepts associated with system development. This included user

perceptions of the conceptual designs (e.g., how engaging the interface is, how easily

it is understood, perceived usability), as well as the potential impact of incentives and

social persuasion feedback on driver acceptance of the eco-safe in-vehicle HMI. In

addition, this study sought to explore the how particular design characteristics impact

driver acceptance of such interfaces.

This chapter also seeks to address Research Question 1: “What is the relevant

information that should be conveyed to drivers in an in-vehicle HMI to improve eco-

safe driving?”. It presents examples of a series of conceptual designs of in-vehicle

gamified HMIs, with a particular focus on competence, autonomy and relatedness, as

well as intrinsic versus extrinsic incentives and social persuasion feedback. These

Chapter 5: Design of a Gamified Interface to Improve Fuel Efficiency and Safe Driving (Paper 3) 113
findings, together with the findings from Chapters 3 and 4, are used to inform the

design and development of the system which is evaluated in Chapters 6 and 7.

Taken from: Vaezipour, A., Rakotonirainy, A., Haworth, N. (2016). Design of a


gamified interface to improve fuel efficiency and safe driving. In DUXU 2016: Design,
User Experience, and Usability: Novel User Experiences (Vol. 17, No. 9747, pp. 322-
332). Part of the Lecture Notes in Computer Science series. Springer International
Publishing. https://doi.org/10.1007/978-3-319-40355-7_31

Publication Status: Available online 22 June 2016

Journal Quality: Paper accepted for presentation at International conference of


Human Computer Interaction International (HCII 2016). Peer-reviewed papers
published in Lecture Notes in Computer Science series, rank Q2 (SCImago) in
computer science.

Copyright: The publisher of this article (Springer) stated that authors can use their
articles in full or in part to include in their thesis of dissertation (provided that this is
not to be published commercially).

Statement of Contribution of Co-Authors for Thesis by Published Papers

The authors listed below have certified that:

 They meet the criteria for authorship in that they have participated in the
conception, execution, or interpretation, of at least that part of the publication
in their field of expertise;

 They take public responsibility for their part of the publication, except for the
responsible author who accepts overall responsibility for the publication;

 There are no other authors of the publication per these criteria;

 Potential conflicts of interest have been disclosed to (a) granting bodies (b) the
editor or publisher of journals or other publications, and (c) the head of the
responsible academic unit, and

 They agree to the use of the publication in the student’s thesis and its
publication on the QUT’s ePrints site consistent with any limitations set by
publisher requirements.

114 Chapter 5: Design of a Gamified Interface to Improve Fuel Efficiency and Safe Driving (Paper 3)
Contributor Statement of contribution

Atiyeh Vaezipour Design study, recruitment, data collection, data


analysis, write the manuscript and responded to
reviewer comments
Professor Andry Rakotonirainy Supervised the research, reviewed the
manuscript and supported the editorial process
Professor Narelle Haworth Supervised the research, reviewed the
manuscript and supported the editorial process

Principal Supervisor Confirmation: I have sighted email or other correspondence


from all Co-author confirming their certifying authorship.

Name: Professor Andry Rakotonirainy

QUT Verified Signature


Signature:

Date: 25/03/2018

Chapter 5: Design of a Gamified Interface to Improve Fuel Efficiency and Safe Driving (Paper 3) 115
5.2 ABSTRACT

The increasing reliance on motor vehicles has negative effects on both human

health and the environment. Improving driving style has been shown to be a

particularly crucial and relatively quick step to reducing fuel consumption and vehicle

emissions. In this paper, a series of conceptual designs of in-vehicle gamified

interfaces were evaluated, with a particular focus on the ability to use such systems to

increase driver acceptance of feedback from such interfaces in order to promote eco-

safe driving. Self-determination theory (SDT) was used to inform the design of the

gamified interface concepts, with a particular focus on competence, autonomy and

relatedness, as well as intrinsic versus extrinsic incentives and social persuasion

feedback. The study adopts a user-centred design approach, utilising focus groups to

establish user needs and motivations to aid the design of a prototype system.

5.3 INTRODUCTION

An increasing reliance on motor vehicles subsequently results in an increase in

gaseous and particulate pollution, which has negative effects on both human health

and the environment. According to the World Health Organization (2014), ambient air

pollution attributes to 3.7 million deaths each year worldwide. Improving driving style

has been shown to be an effective way to reduce fuel consumption and vehicle

emissions (Martin et al., 2012), with the adoption of fuel-efficient driving behaviours

being reported as having the potential to reduce fuel consumption amongst the existing

vehicle fleet by 5-10% (Barkenbus, 2010).

Whilst there are large overlaps between eco-driving and safe driving behaviour,

in some traffic situations they are in conflict with one another (Haworth & Symmons,

2001; Vaezipour et al., 2015). In-vehicle Human Machine Interfaces (HMI) provide

an opportunity to provide real-time feedback to drivers based on the immediate traffic

116 Chapter 5: Design of a Gamified Interface to Improve Fuel Efficiency and Safe Driving (Paper 3)
conditions, support of the driver and navigation. These systems are developing rapidly,

having recently been used to promote fuel-efficient and safe driving.

This paper employed a user-centred design approach to inform the design of a

prototype eco-safe, gamified in-vehicle system. The use of game elements in the

driving context is becoming increasingly popular. The use of such principles as a

means to influence driver behaviours has been previously explored, including a

number of smartphone, app-based approaches, such as Driving Miss Daisy (Shi, Lee,

Kurczal, & Lee, 2012) , Learner Logbook (Fitz-Walter, Wyeth, Tjondronegoro, &

Scott-Parker, 2013) and VW Smile Drive (smiledrive.vw.com), as well as more

innovative approaches such as the Speed Camera Lottery (thefuntheory.com).

Furthermore, Rodríguez et al. (Rodríguez, Roa, Ibarra, & Curlango, 2014) sought to

prevent distractions through the use of ambient devices and haptic feedback.

The objective of the gamified interface is to motivate drivers to improve their

driving performance and cooperate with other drivers in order to improve fuel

efficiency and traffic safety. Of particular interest, is the acceptance and effectiveness

of providing driving performance feedback to drivers through this medium, both in

terms of individual feedback to the driver, as well as the exchange of social persuasion

feedback among other drivers. The study investigates a number of concepts, including

user perceptions of the conceptual designs (e.g., how engaging the interface is, how

easily it is understood, perceived usability), as well as perceived difficulties associated

with the interface and ideas for improvements.

Self-determination theory (SDT) (Deci & Ryan, 1985) was used to inform the

design of the game elements and interface concepts. According to SDT, three key

needs are essential for the achievement of psychological growth: competence,

relatedness and autonomy. Specifically, competence refers to an individual’s need to

Chapter 5: Design of a Gamified Interface to Improve Fuel Efficiency and Safe Driving (Paper 3) 117
master new tasks and learn new skills, relatedness to an individual’s need to experience

a sense of belonging and attachment to others, and autonomy to the need to feel in

control of one’s own behaviour and the subsequent outcomes. Not surprisingly, these

factors are often targeted in order to institute behaviour change, with high levels of

these factors argued to be most likely to result in intrinsically motivated behaviour. In

addition, the theory argues that both intrinsic (e.g., beliefs, expectations, and self-

identity) and extrinsic (e.g., incentives, social norms and pressure, cultural norms)

motivations can influence behaviour to varying degrees. Finally, social support is

argued to be important to psychological growth, such that encouraging interpersonal

relationships and social interactions (e.g., through group co-operation) can harness

feelings of relatedness and attachment with others.

Thus, according to SDT, a number of factors need to be considered when

developing in-vehicle systems for the purpose of facilitating intrinsically motivated

behaviours that are less amenable to change. These include: (i) avoiding external

rewards (e.g., prizes, cash incentives); (ii) providing positive feedback to drivers

regarding their behaviour related to relevant tasks (e.g., eco-safe driving); and, (iii)

fostering social interactions (e.g., through cooperative systems).

Therefore, this research explores the proposition that the autonomy associated

with an intrinsically motivated behaviour may be undermined by the provision of

external rewards, such that an individual’s perceived control over their behaviour is

diminished. In addition, this study investigates whether intrinsic motivations to

perform behaviour can be promoted through the provision of positive feedback

regarding task performance, aimed at improving feelings of competence. Finally, this

paper draws conclusions relating to user preferences for the design and development

of an in-vehicle gamified interface. The main focus is on the comparative effect of

118 Chapter 5: Design of a Gamified Interface to Improve Fuel Efficiency and Safe Driving (Paper 3)
individual versus social persuasion feedback in the driving context, which may be an

important element for designing in-vehicle interfaces.

5.4 CONCEPTUAL GAMIFIED DESIGN

In the focus groups, participants were presented with a number of conceptual

designs in the form of a paper prototype. These designs were displayed as images of a

smartphone depicting various screens associated with a hypothetical gamified eco-safe

in-vehicle interface. Participants were informed that these screens are only accessible

to drivers before the commencement or after the completion of a trip, in order to

minimise distraction while driving.

In Figure 5.1a, screens related to challenges and game elements reveal how a

user is able to choose between a variety of challenges, as well as view descriptions of

each challenge. The design was intended to provide users with challenges based on

different levels of difficulty to facilitate feelings of enhanced competence as they

progressed through levels and mastered various challenges, keeping in mind that prior

research has highlighted the need to balance difficulty and user skill in order to avoid

boredom (when challenges are too easy) and/or frustration (when challenges are too

difficult) (Przybylski, Rigby, & Ryan, 2010). The design also supported autonomy by

providing multiple options for challenges, allowing users to feel a greater sense of

control.

Chapter 5: Design of a Gamified Interface to Improve Fuel Efficiency and Safe Driving (Paper 3) 119
Figure 5.1. Conceptual gamified design, including challenges (a), individualised
progress feedback (b), and social persuasion feedback (c)

In Figure 5.1b, the screens display individualised performance indicators and

feedback. Specifically, these screens highlight a user’s points, level progress,

achievements, as well as more specific challenge and trip information (e.g., kilometres

travelled, proportion of correct/incorrect performance of the challenge behaviour).

Focus group participants were informed that these screens are only accessible upon

120 Chapter 5: Design of a Gamified Interface to Improve Fuel Efficiency and Safe Driving (Paper 3)
completion of a trip. Furthermore, they were shown additional concept screens (i.e., a

‘My Profile’ screen) outlining a variety of other information, such as total scores,

overall eco-driving and safety scores, a list of friends who also use the system, a list

of achievements earned and tips for increasing one’s total score. Given that research

has highlighted the importance of positive feedback to user experiences of competence

and motivation (Ryan & Deci, 2000), these screens were designed to give primarily

positively-geared feedback messages.

Finally, Figure 5.1c shows the social interaction features of the system. During

the trip a user can choose to have scores of other drivers also using the system in

proximity of their vehicle automatically projected onto the windshield (via heads-up

display, using vehicle-to-vehicle technology), showing comparative scores on overall

eco-driving and safety challenges. In addition, a number of screens are accessible after

a trip that reveal scores in comparison to friends and/or other drivers who use the

system. These components of the system were designed to provide users with the

ability to socially engage with other drivers and cooperate toward a common goal of

eco-safe driving.

5.5 METHOD

5.5.1 Participants
A total of 34 licensed drivers from the Australian state of Queensland were

recruited to participate in focus group discussions. The sample was roughly evenly

split on gender (16 female, 18 males), with an age range of 19-61 years (M= 32.11,

SD= 10.44). A convenience sampling approach was used, with participants being

university students or employees, their friends or associates. Focus groups were

approximately 60 minutes, with participants offered a movie voucher as

reimbursement for their time and contribution.

Chapter 5: Design of a Gamified Interface to Improve Fuel Efficiency and Safe Driving (Paper 3) 121
5.5.2 Procedure and Materials
Participants gave written consent to participate in an audio-recorded focus group

discussion. In total, 6 focus groups of 5-6 participants were conducted. The facilitator

began each focus group with a discussion about fuel-efficient and safe driving

behaviour, with participants encouraged to discuss their general concerns about eco-

safe driving and the behaviours they perceived as being important to monitor while

driving. The conceptualised gamified eco-safe in-vehicle system was than presented

to participants (see Figure 5.1), with the majority of the focus group devoted to

facilitator-guided, semi-structured discussions regarding participant perceptions

regarding issues such as usability, usefulness, and key effective features and areas for

improvement, particularly in relation to device and feedback characteristics, gamified

elements, and social persuasion feedback.

5.5.3 Data Analysis


Focus group sessions were audio-recorded, transcribed and coded using Nvivo

10. A thematic analysis was conducted to identify themes and patterns across

responses and their association with research questions. Following that, familiarisation

and ongoing interpretation of the data allowed initial codes to be generated, which in

turn were collated into a number of main themes. These included: (i) perceptions of

gamified concepts; (ii) incentives and motivations to use the system; and (iii) social

persuasion feedback.

5.6 FINDINGS

The research findings are presented in regards to three main themes discussed in

the previous section, namely: (i) perceptions of gamified concepts; (ii) incentives and

motivations to use the system; and (iii) social persuasion feedback.

122 Chapter 5: Design of a Gamified Interface to Improve Fuel Efficiency and Safe Driving (Paper 3)
5.6.1 User Perceptions of Gamified Concepts
Overall, there were mixed perceptions regarding the use of game elements for

an eco-safe in-vehicle system, with rather polarised perceptions regarding whether the

inclusion of such elements was positive or negative. Those at the most extreme

spectrum of less favourable attitudes suggested that the use of game elements would

be potentially detrimental to the overall perceived legitimacy of the system as a safety

intervention.

“I disagree with the game concept of it, it almost takes it away from safety and
driving and makes it not as serious” (F, 53).

More commonly, participants suggested that gamified elements simply didn’t appeal

to them. More specifically, many suggested that their preference would instead be for

a system that provided informative and personalised feedback regarding their eco-safe

driving behaviour, and in particular feedback designed to help them improve their eco-

safe driving. However, it is worth noting that these individuals often acknowledged

that some drivers may like, and benefit from, the gamified interface.

“I think I’d use it for feedback. I wouldn't use it for gaming. But like, if it just gave
me feedback after I finished driving, then yeah, I would use it” (F, 22).

“I don’t think I’d use the game features. I’d be interested in it helping me drive safer
and be more aware” (M, 32).

However, some people had positive perceptions towards gamified elements.

Specifically, a number of participants suggested that the system could be engaging and

decrease the negative effects of driving monotony.

“Probably more appealing as a game feature rather than just like “here’s buttons to
push to tell you something”. People might get a bit more into it as a game feature”
(M, 29).

Chapter 5: Design of a Gamified Interface to Improve Fuel Efficiency and Safe Driving (Paper 3) 123
“I think it’s going to be fun, because sometimes a driver is going to get bored so if
you have something fun, it makes your time pass smoothly” (M, 19).

It was argued that young drivers in particular may benefit from the increased

engagement with the driving task associated with a gamified system.

“The game thing could possibly be good for new, younger drivers when they’re first
starting to go out by themselves and hone their driving skills, once they get their
license” (F, 33).

“It would be good to have something a little bit fun in the car so for them [younger
drivers] the game might be good for making it, sort of a little bit fun and engaging”
(M, 36).

In addition, other participants highlighted that the game elements could facilitate

social persuasion among family and friends, suggesting they would be motivated to

challenge their peers and family.

“I could see it being engaging with the family, to see how other members of the family
are behaving. From that perspective, it would be beneficial” (F, 51).

As a result, the majority of participants suggested that the gamified elements

should represent more of a ‘background feature’ of the system, such that those users

who wanted to gamify the feedback could do so, while the default was to have

feedback provided in a more direct and informative manner.

“I think being able to turn that off would be good because I think personally I might
be interested in the game aspect for like a week or something and then I’ll probably
get over it and just want the information.” (F, 33).

“It would be good to have an option. If you’re into the game thing, I can compete
with my friends, same as any iPhone game” (M, 30).

124 Chapter 5: Design of a Gamified Interface to Improve Fuel Efficiency and Safe Driving (Paper 3)
Some participants questioned the likelihood of continued, long-term use of the

system, suggesting that the novelty of the gamified elements may be short-lived. In

addition, others suggested that frustration and discontinued use may result if the

challenges became too difficult to master.

“I think for me, the problem will be the consistency of usage. How long are you going
to use it for? There's always this period when people start using and then they are
like super excited and after a point where they like, “oh, it’s too much of a hassle”
(M, 22).

“I think it might improve my driving for a while but then, if I can’t reach any other
like challenges or achievements in the game, if it was a game, then I might just like
forget it and go back to my old driving” (F, 20).

Finally, a number of participants highlighted the importance of positively-geared

feedback in order to minimise the likelihood of misuse of the system and the

inadvertent encouragement of poor eco-safe driving behaviours.

“I might try and see how I could push the system, like make it flash a lot that I’m
doing things wrong and people would do that, like when you used to have the alcohol
breath testers in the pub, you’d see who can get the highest reading, it would be
similar thing with this” (M, 30).

“Good idea, if it doesn’t get abused or misused. I can see my teenage son … use it
for bad behaviour, instead of improving, trying to compete in bad behaviour” (F,
51).

5.6.2 Incentives and Motivations to Use the System


Participants reported polarised views regarding the factors that would motivate

them to use the system. On the one hand, a number of participants suggested that

intrinsic incentives, such as the knowledge that using the system was improving their

eco-safe driving behaviour and helping the environment, would be sufficient to

motivate them to use the system.

Chapter 5: Design of a Gamified Interface to Improve Fuel Efficiency and Safe Driving (Paper 3) 125
“I think for me personally, an estimate of how much I saved, money and emissions,
that would actually be the only thing important to me. If you give me a cinema
voucher after one year, great, but I wouldn’t really go for it” (M, 28).

“The value has really got to come from the saving, the personal saving, not the
[external] reward” (F, 35).

“I care about the environment. That would be my motivation. And any extraneous
stuff like, yeah might be good, but my interest in this is from an environmental point
of view” (M, 39).

“I wouldn’t need exterior rewards, you can get rewarded for just being a decent
driver. I’d just use it for plain curiosity” (M, 23).

Overall, the intrinsic incentives associated with game elements, such as

achievements and leader boards, appeared to have little impact on reported motivations

to use the system among the participants, even though they were typically interested

in receiving feedback regarding their eco-safe driving performance.

“I don’t want a trophy, but I want to know when I'm driving efficiently” (M, 28).

“I don’t care if my phone’s saying congratulations, you got a point and a star. I don’t
care” (M, 29).

Interestingly, a number of participants noted that it is important to ensure that

the system was accommodating, noting that different users will be motivated by

different incentives, and that the system must suit an individual’s needs.

“If you had a series of what the rewards are, from the point of view of I want to save
money, you want to save the planet, you want to compete with your mates then you
set it up into different things. The technology underneath is the same but you've got
to suit you” (M, 61).

“I’d rather reward myself … maybe you could customise it in a way that mattered to
you” (M, 39).

126 Chapter 5: Design of a Gamified Interface to Improve Fuel Efficiency and Safe Driving (Paper 3)
On the other hand, a number of participants argued that intrinsic incentives

would not be sufficient to motivate them to use the system. Instead, these participants

suggested that extrinsic rewards would be necessary to motivate them to engage with

the system.

“If you’re going to use something that takes your time, and I don’t feel like there’s a
benefit to myself or anything for using it, I don’t care if I can see what kind of driver
I am … I need an incentive … I need money or something for free or a discount” (M,
29).

“There's got to be some sort of a reward in it for me” (M, 61).

In particular, amongst those participants who advocated extrinsic incentives,

many suggested the most effective rewards would be those related to the driving task,

such as fuel vouchers or insurance and vehicle registration discounts. It was also noted

that the rewards would have to outweigh the effort associated with using the system

and altering one’s driving.

“I think the incentive would have to be car-related, like insurance [discounts],


because that’s going to appeal to all audiences” (F, 22).

“If my insurance [company] said that if you used this you get a percentage off on
your thing [premiums], I’d be like “yeah, why not”” (M, 29).

“It would have to be ongoing beneficial money value for me in fuel or rego ... it
would have to be something that would keep you driving well for the points to be
able to get the reward … and the rewards need to be worth it” (M, 27).

Finally, one participant noted that the focus on extrinsic incentives may result in

only short-term behaviour change, particularly if there is no incentive to continue

engaging in particular eco-safe driving behaviours once a challenge has been mastered.

Interestingly, this individual also reported that they perceived extrinsic rewards as

Chapter 5: Design of a Gamified Interface to Improve Fuel Efficiency and Safe Driving (Paper 3) 127
important to increasing their motivation to use the system, suggesting paradoxical

attitudes.

“Isn’t it meant to change your habits driving, not just make you try to win a game. I
mean, if there's good prizes, I’m going to drive really well to get to the targets and
then, that’s it” (M, 27).

5.6.3 Social Persuasion Feedback


Participants were also divided regarding their attitudes towards the social

persuasion feedback aspects of the concept system. Some drivers suggested that they

would be motivated to compete with friends and family in an attempt to determine who

is a better driver, while others noted that they would be curious to compare their eco-

safe driving ability against other drivers. There was some evidence that having family

or peers who use the system would increase the likelihood that participants would use

the system themselves.

“If my friends all use it I will use it. Because I want to compete in the scores. So, I
am more into the social side” (F, 25).

“I would use it to see other people’s scores” (M, 29).

“People are not only trying to better themselves but they can see their friends, which
more and more people are going to go see if they can out do their friends” (F, 56).

However, among these participants, some voiced concern about being judged by

other drivers, highlighting that negative feedback may reduce motivation to use the

system, particularly among peers and family.

“I think the social interaction part is very good. So, you will see what the
other person’s score is, but you don't want people judging you, saying “this
car is really bad and he's not a good driver”” (M, 26).

128 Chapter 5: Design of a Gamified Interface to Improve Fuel Efficiency and Safe Driving (Paper 3)
“I think if you’ve got an awful score, you just turn it off. You’d never do it. If you’ve
got a great score, you’d leave it on. You don’t really want all your friends to get to
know you’re a shit driver” (M, 61).

Other participants were less interested in the social persuasion feedback,

suggesting that they were more interested in competing with themselves and focusing

on improving their behaviour in comparison to prior levels.

“It’s only if you can beat yourself. I don’t want to compare with other people” (F,
35).

5.7 DISCUSSION AND DESIGN IMPLICATIONS

The findings from this study provide important implications for the design of

future gamified eco-safe in-vehicle systems and highlight the importance of adopting

a user-centred design approach from an early stage of design, through to development

and implementation of any system.

Overall, participants held less than favourable attitudes regarding the

gamification of an eco-safe in-vehicle system, instead reporting a preference for more

informative and personalised feedback. However, the potential for such elements to

increase engagement with the driving task and facilitate social persuasion feedback led

many to argue that game elements should represent a ‘background feature’ of the

system, such that those users who wanted to gamify the feedback could do so. This

finding highlights that users desire a sense of autonomy and control over the system,

such that they can modify the system to best suit their specific individual needs.

Perhaps the most important goal is to provide feedback in an engaging manner

and allowing users to choose if, and how, they want to gamify the experience for

themselves. This is likely to differ dependent on user demographics and personality

characteristics, such as age, gender, sociability and previous experience with games

Chapter 5: Design of a Gamified Interface to Improve Fuel Efficiency and Safe Driving (Paper 3) 129
and game play. It is worth mentioning that participant responses did not show large

differences between age and gender, but rather were relatively divided based on life

stage (e.g., novice vs experienced drivers; parents of young children vs parents of

teenagers), experience with using the technology and purpose of driving.

These findings provide support for the ideas of Deterding (2011) who argued

that “a game element may be both intrinsically and extrinsically motivating for certain

people in certain situations at certain times” (Cited in Seaborn & Fels, 2015, p. 20).

This finding therefore suggests that it is important to design in-vehicle systems in a

way that allows personalisation and customisation, to accommodate individual users.

This finding is consistent with those of previous studies (Seaborn & Fels, 2015).

Furthermore, consistent with prior research, a number of participants

highlighted the importance of balancing task difficulty and user skill. Thus, careful

consideration must be given to the development of challenges in order to maintain

engagement and avoid situations where users become bored, such as when challenges

are too easy, or frustrated, such as when challenges are too difficult.

According to SDT, intrinsic motivation is crucial for long-term behavioural

change. The results of the focus study provided some evidence that intrinsic incentives

would be sufficient to motivate user engagement with the system, however many

participants still reported that extrinsic rewards would be necessary. Interestingly, a

paradoxical attitude was noted whereby a participant reported a need for extrinsic

rewards to engage with the system while also acknowledging that a focus on extrinsic

incentives may result in only short-term behaviour change, particularly if there is no

incentive to continue engaging in particular eco-safe driving behaviours once a

challenge has been mastered. Further research is required to more comprehensively

130 Chapter 5: Design of a Gamified Interface to Improve Fuel Efficiency and Safe Driving (Paper 3)
understand this issue, however participants did again note the importance of autonomy

and control in deciding the incentives that are most suitable to them.

The theory also argues that social support is important to psychological growth,

and that encouraging interpersonal relationships and social interactions can foster

feelings of relatedness. Overall, a majority of participants expressed favourable

attitudes towards the social persuasion feedback aspects of the concept system. This

finding suggests that the inclusion of elements such as group messages, blogs,

connectivity to social networks, and chat functions may all increase a user’s motivation

to engage with the system.

Finally, a number of feedback characteristics were noted as being important, and

should be considered in future research and system development. These included

providing positively-geared feedback in order to minimise the likelihood of misuse of

the system and the inadvertent encouragement of poor eco-safe driving behaviours.

5.8 CONCLUSION

This paper presents the findings from a user-centred design focus group study

which sought to investigate user perceptions of a conceptualised gamified eco-safe in-

vehicle system. Results suggest that many users hold less than favourable attitudes

towards gamified elements being more than a background feature to such a system,

however support the functionality for users to gamify the system if they wish. Thus,

consistent with SDT’s concept of autonomy, users desire the ability to customise

system features depend on their specific needs. Moreover, to meet a user’s competence

needs, challenges must be designed such that they balance task difficulty and user skill

and experience levels, while the inclusion of social persuasion feedback aspects

appears to be an effective method for meeting a user’s need for relatedness. More

research is required to comprehensively understand the role of intrinsic versus

Chapter 5: Design of a Gamified Interface to Improve Fuel Efficiency and Safe Driving (Paper 3) 131
extrinsic incentives on motivations to engage with in-vehicle system. Future research

should seek to develop a more realistic prototype for further evaluation in regards to

usability, paying particular attention to the impact of system use on safety outcomes,

such as driver distraction and the inadvertent promotion of unsafe or unlawful driving

behaviours.

132 Chapter 5: Design of a Gamified Interface to Improve Fuel Efficiency and Safe Driving (Paper 3)
Chapter 6: A Simulator Evaluation of In-
vehicle Human Machine
Interface for Eco-Safe Driving
(Paper 4)

6.1 INTRODUCTORY COMMENTS

The final step of the user-centred design approach relates to system evaluation

against user requirements (see Figure 6.1). As such, this chapter describes the findings

from a subjective and objective evaluation of the prototyped eco-safe in-vehicle HMI

(Study 3). This evaluation included an experiment conducted in a state-of-the-art

driving simulator and a driver survey. The use of a driving simulator allows for an

objective evaluation of the impact of system use on behaviour in a safe and controlled

environment, where on-road studies may not feasible due to safety considerations

The eco-safe in-vehicle HMI evaluated in this chapter was designed and

developed based on the findings from the preceding chapters (Study 2a), as well as a

series of design iterations and an expert workshop (Study 2b, discussed in Section 2.5).

The main aim of the paper presented in this chapter (Paper 4) is to evaluate three

prototypes of an eco-safe in-vehicle HMI. The systems varied in regard to the types of

advice and/or feedback messages provided to drivers: (i) visual advice only; (ii) visual

feedback only; (iii) visual advice and feedback.

This chapter seeks to address Research Question 2: “What impacts do advice

and/or feedback characteristics of an in-vehicle eco-safe HMI have on driver

acceptance of technology (RQ2a), driver workload (RQ2b), and eco-safe driving

performance (RQ2c)?”. That is, this chapter seeks to objectively evaluate the

effectiveness of the prototype eco-safe in-vehicle HMI on a range of eco-safe driving

Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4) 133
behaviours. In addition, the accompanying driver survey involved a range of subjective

measures examining driver acceptance of the technology and driver workload. The

findings discussed in this chapter are extended upon in Chapter 7, where an analysis is

reported detailing the additional influence of incentives and competition associated

with system use on subsequent eco-safe driving behaviour.

Figure 6.1 User-Centred Design Approach- Driving simulator experiment &


driver survey

Taken from: Vaezipour, A., Rakotonirainy, A., Haworth, N. & Delhomme, P. (2017).
A Simulator Evaluation of in-vehicle Human Machine Interface for Eco-Safe Driving.
Transportation Research Part A: Policy and Practice.

Publication Status: Under-review

Journal Quality: Transportation Research Part A: Policy and Practice is a peer-


reviewed publication with international readership. The journal impact factor is 2.609,
and rank Q1 (SCImago) in Transportation, Management Science and Operations
Research, Civil and Structural Engineering.

Copyright: The publisher of this article (ELSEVIER B.V.) stated that authors can use
their articles in full or in part to include in their thesis of dissertation (provided that
this is not to be published commercially).

134 Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4)
Statement of Contribution of Co-Authors for Thesis by Published Papers

The authors listed below have certified that:

 They meet the criteria for authorship in that they have participated in the
conception, execution, or interpretation, of at least that part of the publication
in their field of expertise;

 They take public responsibility for their part of the publication, except for the
responsible author who accepts overall responsibility for the publication;

 There are no other authors of the publication per these criteria;

 Potential conflicts of interest have been disclosed to (a) granting bodies (b) the
editor or publisher of journals or other publications, and (c) the head of the
responsible academic unit, and

 They agree to the use of the publication in the student’s thesis and its
publication on the QUT’s ePrints site consistent with any limitations set by
publisher requirements.

Contributor Statement of contribution

Atiyeh Vaezipour Study Design, recruitment, data collection, data


analysis, write the manuscript.
Professor Andry Rakotonirainy Supervised the research, reviewed the
manuscript and supported the editorial process
Professor Narelle Haworth Supervised the research, reviewed the
manuscript and supported the editorial process
Dr Patricia Delhomme Discussed the research design, reviewed the
manuscript and supported the editorial process

Principal Supervisor Confirmation: I have sighted email or other correspondence


from all Co-author confirming their certifying authorship.

Name: Professor Andry Rakotonirainy

QUT Verified Signature


Signature:

Date: 25/03/2018

Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4) 135
6.2 ABSTRACT

This paper investigates the use of In-vehicle human machine interfaces (HMIs)

for improving eco-safe driving style, which have been demonstrated as being effective

for reducing fuel consumption and emissions, as well as improvements in road safety.

A total of forty drivers participated in a simulated driving experiment to evaluate three

variations of an eco-safe in-vehicle HMI: visual advice only; visual feedback only; or

visual advice and feedback. Subjective and objective measures were analysed,

including fuel consumption, eco-safe driving behaviour, driver acceptance, and

workload. All system types were associated with the relatively high levels of driver

acceptance, with the advice only system accepted the most. While all system types

produced relatively low levels of workload for drivers, systems involving feedback

significantly increased the workload associated with using the interface. The findings

suggest that the combined advice and feedback system has the potential to

simultaneously reduce fuel consumption and emissions and improve eco-safe driving

behaviour. Specifically, both advice and feedback appeared to be critical in

encouraging positive changes in eco-safe driving behaviour. This study can inform the

design and development of future in-vehicle HMIs to improve eco-safe driving that

are accepted by drivers and have minimal adverse impacts on driver workload.

6.3 INTRODUCTION

Emerging technologies have provided opportunities for the transportation

industry and automotive manufacturers to assist in reducing the impact of motor

vehicles on the environment, such as electric or automated vehicles (Fagnant &

Kockelman, 2015; Wadud et al., 2016). However, improvements associated with such

benefits are typically incremental and have numerous barriers to their widespread

136 Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4)
implementation, such as higher costs to users, acceptance, laws about responsibility,

and required changes to the road infrastructure.

Eco-driving interventions have been demonstrated to be a cost-effective and

quicker approach to reducing fuel consumption and emissions (see review Barkenbus,

2010; Jonkers et al., 2016; McIlroy et al., 2014; Vaezipour et al., 2015). The term eco-

driving refers to three levels of decision-making, including: (i) strategic decisions (e.g.,

vehicle selection, maintenance schedules); (ii) tactical decisions (e.g., route selection,

vehicle loading); and, (iii) operational decisions (e.g., driving style) (Alam &

McNabola, 2014). While all of these are undoubtedly important, this paper focuses on

improving operational decisions, given that the adoption of a more appropriate eco-

driving style can have immediate impacts on reducing fuel consumption and emissions

(Martin et al., 2012).

A number of studies have demonstrated that the use of in-vehicle HMIs has the

potential to support and encourage eco-driving and safe driving (Ecodriver, 2011;

Huang et al., 2005; Kurani, Stillwater, Jones, & Caperello, 2013; McIlroy, Stanton,

Godwin, & Wood, 2017). However, existing systems currently provide advice or

feedback on either eco-driving or safety in isolation, with a paucity of research

examining both of these factors being integrated in a single system. Thus, in this paper

we focus on eco-safe driving, which can be defined as driving behaviours aimed at

reducing fuel consumption and emissions such as smooth acceleration and deceleration

and anticipating the traffic flow to minimise the use of brake and accelerator

(Barkenbus, 2010; Delhomme et al., 2013; Sivak & Schoettle, 2012), as well as safe

driving behaviour including adherence with legal speed limits and keeping a safe

following distances.

Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4) 137
Moreover, current research has typically focussed on evaluating the impact of

these systems on eco-driving or safe driving behaviours, with considerably less

research investigating the driver acceptance of such systems. To date, the impact of

advice and feedback from in-vehicle HMIs on eco-safe driving is not well understood.

Driver acceptance has been defined as the degree to which individuals voluntarily

incorporate the system into their driving, and is associated with factors such as

perceived usefulness, usability and ease of use, as well as intentions to use (Adell,

2009). Indeed, as stated by Van Der Laan, Heino, and De Waard (1997, p. 1), “it is

unproductive to invest effort in designing and building an intelligent co-driver if the

system is never switched on, or even disabled”. This suggests that it critical to examine

the factors that motivate acceptance when conducting evaluations of technological

systems.

On the other hand, a number of studies have also examined the potential risk of

in-vehicle HMI usage on distracting a driver’s attention from the driving task and thus

reducing driver safety (Vaezipour et al., 2015). It is argued that distraction may occur

due to the increasing complexity and workload demand of the driving task when

interacting with a system (Oviedo-Trespalacios et al., 2016). One potential approach

to reducing the distractive qualities of such a system is to ensure that drivers are given

a high degree of flexibility in customising the manner in which advice and feedback is

presented to them, as well as design characteristics of the system (Vaezipour et al.,

2017).

Taken together, there is a need for an effective in-vehicle HMI to motivate eco-

safe driving. However, it is crucial that the development of such a system adopts a

user-centred design approach, whereby drivers’ needs, motivations and limitations

represent the driving force behind the development of the technology and are

138 Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4)
considered at all stages of the design process (Maguire, 2001). Such an approach will

increase the likelihood of driver acceptance of the technology and avoid costs

associated with system rejection (Francois et al., 2017; Vaezipour et al., 2015). In

addition, the design and development of such a system must seek to minimise driver

workload, in order to maintain driver safety associated with use of the system.

This study forms part of a larger project aimed at the design and development of

an in-vehicle HMI to reduce fuel consumption and emissions and improve safe driving.

The aim of current study was to evaluate the impact of a prototype in-vehicle HMI

designed to encourage improvements in eco-safe driving style While the focus of the

current paper is to evaluate an in-vehicle HMI in a simulated driving experiment, a

brief overview of the design and development process for the system is provided in the

following section in order to give the reader some contextual background.

6.4 ECO-SAFE IN-VEHICLE HMI DESIGN PROCESS AND


DESCRIPTION

The overarching goal of the in-vehicle HMI was to encourage the adoption of an

eco-safe driving style. The system was developed on the assumption of access to

technology that allows the use of real-time traffic information, such as vehicle-to-

vehicle (V2V) and vehicle-to-infrastructure (V2I) technology, summarised as V2X

technology. This technology allows the vehicle to communicate with traffic signals

and stop signs, as well as other vehicles on the road. The provision of real-time advice

and feedback to drivers subsequently promotes greater anticipation by drivers of

changes in traffic conditions (Delhomme et al., 2013; Popiv, 2012).

The in-vehicle HMI evaluated in the current study was developed through an

iterative design process, following the key principles of a user-centred design approach

adopted from the International Organisation for Standardisation (ISO, 2010) and also

Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4) 139
complies with Australia road rules (Queensland Transport Main Road, 2017). Given

the importance of involving users in all stages of the design and development process

(Francois et al., 2017), a series of focus groups was then conducted with drivers in

Australia to determine driver requirements and characteristics of an in-vehicle HMI

(Vaezipour et al., 2016; Vaezipour et al., 2017). The system was then refined further

through an iterative process involving a series of rapid prototypes. Finally, a

multidisciplinary workshop was conducted with experts from a variety of disciplines,

including intelligent transportation systems, traffic psychology, human computer

interaction, user experience (UX) design and industrial design. Iterations considered

the number of elements included in the display as well as the intuitiveness of the

behavioural icons. Expert workshops have been extensively used as part of the design

process for technological development and have been demonstrated as an efficient

approach to performing heuristic evaluations (Mankoff et al., 2003; Nielsen, 1994), to

assess system usability prior to user testing.

The system was designed to act as an eco-safe driving companion, with features

aimed at giving the system a sense of personality, an approach commonly used in

social robotics and persuasive technologies to create affective connections between

humans and technology (Fogg, 2002; Pieroni et al., 2015) . The system prototyped

using a smartphone screen which was placed into a 3D printed housing and designed

to represent the “head” of the system. This housing was then placed into a mounting

bracket on the windscreen (see Figure 6.2).

Figure 6.2. Eco-safe driving companion

140 Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4)
The smartphone was the main computational part of the system, with an

application used to communicate with the driving simulator computers, and

subsequently present advice and feedback to drivers that mimicked real-time

information. The smartphone screen was designed to represent the “face” of the

system, with the screen displaying “eyes” and a “mouth”, which change to assist in the

provision of feedback (see Figure 6.3). In addition, an Arduino UNO is used as the

microcontroller of an RGB LED strip of lights that are placed inside the housing and

change colour to represent appropriate (green) and inappropriate (red) behavioural

performance.

Figure 6.3. Different “faces” of the system – positive, neutral and negative
Advice messages were designed to guide drivers prior to key traffic situations

Examples of advice and feedback screens used by the system are shown in Figure

6.4. Advisory messages involved screens that presented an icon and simple text (e.g.,

smooth braking) on a yellow background, and were projected to drivers leading up to

specific traffic events, such as when approaching traffic lights, to encourage an

anticipatory driving style. Conversely, feedback messages included the “face”, as well

as a smaller icon and illumination of the LED lights. This feedback differed depending

on whether the behavioural performance of the driver was good (i.e., happy face and

a green light) or poor (i.e., sad face and a red light).

Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4) 141
Figure 6.4. Examples of advice and feedback messages.
Ctrl + click to see the video or QRCode

Advisory messages were designed to guide drivers prior to key traffic situations

using the driving simulator map for real-time traffic scenarios (e.g., traffic lights, speed

limit changes, vehicle following scenarios) and encourage anticipatory driving, which

is a fundamental aspect of eco-safe driving. Conversely, human-like feedback

messages were provided in real-time and were designed to enhance driver knowledge,

awareness and learning, and subsequently increase the likelihood that drivers would

be motivated to improve their driving performance (Gardner et al., 2010; Shute, 2008).

6.5 RESEARCH HYPOTHESES

This paper examines the impact of advice and feedback from an in-vehicle eco-

safe HMI on driver acceptance of the technology, driver workload, and eco-safe

driving behaviours. The four conditions evaluated in the driving simulator were based

on the level of advice and feedback received from an eco-safe in-vehicle HMI: (1)

baseline where no messages were presented to drivers, however the inactive system

142 Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4)
was present in the vehicle during the drive (2) advice only which encourage

anticipatory driving (3) feedback only which indicates behavioural performance (4)

advice and feedback. Specifically, we test the following hypotheses:

H1: Driver acceptance of the in-vehicle eco-safe HMI increases with the amount

of information conveyed to drivers.

H2: Driver workload associated with the in-vehicle eco-safe HMI increases with

the amount of information conveyed to drivers.

H3: Eco-safe driving behaviour (i.e., fuel consumption, smooth acceleration and

deceleration, speeding and headway) will improve in association with all three

variations of the in-vehicle eco-safe HMI, with improvements increasing with the

amount of information conveyed to drivers.

In relation to each hypothesis, the advice/feedback system represents the most

informative system, followed by feedback only system and finally the advice only

system. The first hypothesis is based on the assumption that greater provision of

information will increase the perceived legitimacy and trust associated with the

system, however it is also acknowledged that a potential caveat of such an assumption

results when the feedback provided by the system does not correspond with a drivers’

perceptions of their own behavioural performance, and that in such instances these

drivers may report lower levels of acceptance associated with conditions involving

feedback (Festinger, 1962).

6.6 METHOD

6.6.1 Participants
Forty participants (20 females, 20 males), aged between 18 and 65 years (M =

31.35, SD = 12.18) were recruited from Brisbane, Australia. While 53 participants

Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4) 143
were initially recruited, 13 participants were excluded due to motion sickness,

dizziness and nausea that is commonplace in simulator studies (Fisher et al., 2011).

Data on the gender, age and licence type of the Queensland driving population was

used to inform sampling numbers. The final sample comprised of 14 provisional

licence holders (35%) and 26 open (i.e., unrestricted) licence holders (65%). There

was an even number of males and females across all age groups and licence types. All

provisional licence holders were aged 18-24 years, while 16 of the open licence holders

were aged 25-39 years and ten were aged 40-65 years. The majority of participants

(62.5%) reported driving between 10,001 and 30,000 kilometres in the previous year,

while 35% reporting driving less than 10,000 kilometres.

Participants were primarily recruited through a media release, the university

webpage, social media posts, and snowball sampling. Eligibility criteria included

holding a valid driver licence (i.e. P1, P2 or open). The study was conducted in

accordance with the Australian Code for the Responsible Conduct of Research (QUT

Ethic approval number 1600000651). Participants gave written consent to participate

and were informed that their participation was entirely voluntary and that data would

remain anonymous and reported on in an aggregate manner only.

6.6.2 Driving simulator and scenario


The experiment was conducted in the Centre for Accident Research and Road

Safety- Queensland (CARRS-Q) advanced driving simulator. The driving simulator

represented a safe and convenient method for examining driver behaviour, particularly

pertinent given the potentially distracting effect of the in-vehicle interface, which

might be risky in a less-controlled, real-world environment.

The advanced driving simulator consists of a Holden Commodore passenger

vehicle with an automatic gearbox (see Figure 6.5). The simulator uses the SCANeR™

144 Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4)
studio from OKTAL, linked to website software with eight computers and projectors

and a six degree of freedom (6DOF) motion platform that can move and twist in three

dimensions. When seated in the simulator vehicle, drivers are immersed in a virtual

environment that includes a 180° front field-of-view, simulated rear-view mirror

images, surround-sound for engine and environment noise, a real vehicle cabin and

simulated vehicle motion using a hydraulic system. The simulator collects driver

behaviour data such as speed, acceleration, braking, steering wheel and vehicle

positioning at a rate of 20 Hz.

Photos by Sonja de Sterke

Figure 6.5. The CARRS-Q Advanced Driving Simulator.

In order to maximise the realism and familiarity of the driving experience, the

simulated driving scenario consisted of urban and suburban areas similar to the

Brisbane city area and surrounds. Standard Australian speed limit signs were used,

with speed limits of 40 km/h in urban city areas and 60 km/h in suburban areas (see

Figure 6.6). Participants were required to drive in the left lane as they typically would

on Australian roads. The simulated driving scenario was approximately seven

kilometres long and programmed to incorporate traffic events including traffic signals,

stop signs, pedestrian crossings, a change of speed limit and following another vehicle.

In all instances, participants started their drive within the city area and were instructed

to drive as they normally would and follow signs to the airport.

Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4) 145
Figure 6.6. Examples of simulated driving scenarios.
6.6.3 Experimental Design and Procedure
In the within-subjects design each participant was exposed to all four conditions.

The order of the conditions was counter-balanced among the participants. As stated

previously, the four conditions evaluated in the driving simulator were based on the

level of advice and feedback received from an eco-safe in-vehicle HMI (i.e. baseline,

advice only, feedback only, advice and feedback). By including the baseline condition

in the counterbalancing procedure, the potential influence of learning effects on

subsequent data was controlled for.

The current study involved three phases of data collection: (i) a pre-drive

questionnaire, (ii) simulated driving scenarios with immediate post-drive

questionnaires, and (iii) a final post-drive questionnaire. In the first phase, a link to a

pre-drive questionnaire was emailed to participants, using the Qualtrics survey

software, to collect demographic data and transport and travel characteristics of the

sample, self-reported experience with technology and a range of attitudinal and

perception scales. In the second phase, upon arrival at the driving simulator,

participants were advised of the experimental procedure and completed a practice

drive to familiarise themselves with the driving simulator. Each participant then

completed a total of four experimental drives, with each drive taking approximately

ten minutes to complete. Before each driving condition, the researcher explained the

in-vehicle eco-safe driving system using the InVision user interface prototyping

platform on an iPad (Appendix G). Immediately following each condition, participants

146 Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4)
completed a short questionnaire designed to assess perceived workload and in-vehicle

HMI acceptance.

In the final phase of the study, all participants completed a further questionnaire

designed to collect data related to overall attitudes and perceptions of the device and

simulated driving experience. The driving simulator experiment and driver survey took

approximately two hours to complete and participants received a $70 gift voucher as

reimbursement for their participation.

6.6.4 Questionnaire measures


A range of scales were utilised in order to measure participant acceptance of the

cooperative eco-safe in-vehicle HMI. The Van Der Laan Acceptance Scale (Van Der

Laan et al., 1997) was used to measure perceived usefulness (across five items) and

general satisfaction with the system (across four items), with items measured on a +2

to -2 scale and composite scores for each scale summed and averaged. Scales were

also used to measure perceived ease of use (ten items) and perceived usefulness (13

items), which were adapted from Staubach, Schebitz, Köster, and Kuck (2014), and

developed based on the Technology Acceptance Model (Davis, 1989). Items on each

of the scales were measured on a 5-point Likert scale and summed to form composite

scores for each scale.

System usability was measured using the System Usability Scale (SUS), which

elicits subjective assessments of usability across ten items (Bangor, Kortum, & Miller,

2008). Each item is scored on a 5-point Likert scale, with composite scores ranging

from 0-100. Research has indicated that SUS scores of 70 represent average usability

(Lewis & Sauro, 2009). Finally, workload was measured using the Raw NASA-TLX

(NASA-RTLX), which elicits subjective ratings of workload for mental demand,

physical demand, temporal demand, performance, effort and frustration (Hart, 2006;

Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4) 147
Hart & Staveland, 1988). Each item is scored on a 0-100 scale, with scores across all

six items summed and then averaged. A review of studies using the NASA-TLX,

conducted by Grier (2015), found an average mean score of 41.52 on the scale is

relation to driving a car. Reliability analyses revealed relatively strong internal

reliability across the scales, with Cronbach’s alpha values ranging from .67-.94 (see

Appendix F).

6.6.5 Driving performance measures


Previous research has identified a number of primary indicators for eco-safe

driving behaviour, including deceleration, acceleration, speed, headway, and

appropriate gear changes (Ericsson, 2001). In the current study, various traffic events

where simulated to elicit these behaviours from drivers, with the exception of

appropriate gear changes due to the simulator vehicle having an automatic gearbox. In

addition, the use of an automatic vehicle has the benefit of reducing the complexity in

examining influences on fuel consumption, such that it can be more reliably attributed

to acceleration and deceleration behaviour. Moreover, the results are more

generalisable to other vehicles without gears, such as electric vehicles.

A range of eco-driving variables were measured including fuel consumption and

smooth acceleration and deceleration. Fuel consumption (mL/s) was estimated for the

entire drive using the driving simulator software. Total travel time (in seconds) was

also measured. Acceleration and deceleration were both measured in metres per second

squared (m/s2). In addition, acceleration was measured using pedal position, as a float

value between 0 (released) and 1 (completely pressed). Deceleration was also

measured as applied force on the brake pedal (in Newtons, N), as well as a measure of

overall brake use calculated by taking the total area under the curve when plotting

brake pedal input by time. Algorithms that informed the measurement of optimal eco-

148 Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4)
safe driving were calculated previously, the details of which are out of scope for the

current paper, but the methodology can be found in Nouveliere et al. (2012).

Acceleration data is reported for two separate events which required drivers to

accelerate and bring the vehicle up to complete stop at a STOP sign. Both events

required drivers to turn right at a standard three-way intersection, with one event

requiring drivers to wait for traffic to pass before proceeding (in a 40 km/h zone),

whereas no traffic was present at the other intersection (60 km/h zone). These two

events were chosen to examine whether there are any impacts on acceleration

behaviour depending on whether a driver needs to wait for traffic.

Similarly, deceleration data is also reported for two separate events which

required drivers to decelerate and bring the vehicle to a complete stop. Both events

took place in a 40 km/h zone, with one event involving an approach to traffic signals

in the green phase at a standard four-way intersection, which changed to amber when

drivers were five seconds from the signal. The other event involved a STOP sign at a

standard three-way intersection, with drivers required to wait for traffic to pass before

proceeding. These two events were chosen to investigate behavioural differences

associated with a long-range braking scenario in which drivers had an extended

amount of time to acknowledge a forced stop (e.g., STOP sign and traffic), compared

to a short-notice braking scenario (e.g., traffic light changing to red on approach).

With regards to safe driving variables, headway and speed were measured. A

single headway event required drivers to follow behind another vehicle travelling at

the speed limit (40 km/h). The presence of other traffic prevented drivers from being

able to overtake the vehicles in front of them. Headway was measured as time interval

to vehicle (TIV, see Equation 1), where Di is the relative distance to the vehicle in

front in meters and V is the actual speed of the simulator vehicle (this parameter is

Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4) 149
often used in traffic safety regulations and recent Australian road rules mandates a

minimum TIV of 2.5 secs):

𝑇𝐼𝑉 =Di/V [Equation 1]

Finally, speed was monitored during the entire drive in both the 40 km/h and 60

km/h zones. Speeding behaviour was measured separately for each of the zones using

mean speeds when travelling in free-flowing traffic and the percentage of time spent

exceeding the speed limit.

6.6.6 Data Analysis


Data collected from the driving simulator was processed and extracted using

Matlab version R2017a. To assess the impact of advice and feedback from the

cooperative eco-safe in-vehicle HMI on the driver acceptance, workload and eco-safe

driving behaviour variables, a series of one-way repeated measures ANOVAs were

conducted, reporting on the multivariate statistic Wilks’ Lambda to account for

violations of the assumption of homogeneity of variance. Post hoc comparisons were

adjusted using the Bonferroni correction and the strength of effect sizes were

interpreted according to Cohen (1992) criteria (.01 = small, .06 = moderate, .14 =

large). When assumptions of parametric testing were violated, data were analysed

using Friedman’s Test, with Wilcoxon Signed Ranks Tests to assess pairwise

comparisons and adjusted using the Bonferroni correction. In instances where the

results of non-parametric and parametric tests were identical, the results of the

parametric analyses are reported.

6.7 RESULTS

6.7.1 H1: Driver acceptance of the technology


Driver acceptance was measured using four scales to assess perceived usability,

ease of use, usefulness, and system satisfaction. The mean scores and standard

150 Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4)
deviations for the scales across the various conditions are presented in Table 6.1. When

considering mean scores on the SUS, it can be seen that all conditions were perceived

as having relatively average usability. Statistical analyses revealed a significant effect

of condition on perceived usability of the system. Assessing the pairwise comparisons,

it was found that drivers reported significantly higher levels of usability associated

with the advice only condition compared to the feedback only condition (p = .010).

Table 6.1. Differences in driver acceptance by condition

M(SD)

Scales Advice Feedback Advice/ Results


Only Only Feedback
Usability of the system (SUS) c 72.5 66.63 68.19 λ = .80, F (2, 38) = 4.80, p =
(14.43) (15.43) (13.67) .014, ηp2 = .20
Perceived ease of use of the 36.70 32.90 33.43 λ = .76, F (2, 38) = 6.08, p =
systemb (5.07) (6.56) (5.92) .005, ηp2 = .24
Perceived usefulness of the 43.78 41.55 43.83 χ2 (2, n = 40) = 3.65, p = .161
systemb (9.18) (8.40) (7.03)
Usefulness (Van der Laan .51 .52 .66 χ2 (3, n = 40) = 35.1, p =.825
Scale)a (.72) (.79) (.74)
Satisfaction (Van der Laan .48 .03 .12 λ = .79, F (2, 38) = 5.12, p =
Scale)a (.72) (.89) (.86) .011, ηp2 = .21
N = 40.
a
Adapted from van der Laan, et al. (1997). b Adapted from Staubach, et al. (2104).
c
Adapted from Bangor et al. (2008), Ranges for the scales were as follows: SUS (0-100); ease of use (10-50); usefulness 1 (13-
65); usefulness 2 (-2 to +2); and, satisfaction (-2 to +2), with higher scores representing greater driver acceptance.

While mean scores for ease of use suggested that drivers appeared to find each

system relatively easy to use, there was a significant effect of condition. Specifically,

drivers perceived the advice only condition to be significantly easier to use than both

the feedback only (p = .007) and advice/feedback conditions (p = .004). Results

revealed no significant effect of condition on perceived usefulness of the system,

regardless of the scale used. Nonetheless, mean scores on both scales suggested above

average perceived usefulness in each condition, such that drivers found each of the

systems as useful as one another.

There was however, a significant effect of condition on self-reported satisfaction

with the system. Specifically, drivers reported significantly higher levels of system

satisfaction associated with the advice only condition compared to both the feedback

Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4) 151
only (p = .008) and advice/feedback condition (p = .035). Looking at the mean scores,

it can be seen that drivers had relatively neutral satisfaction levels associated with the

feedback only and advice/feedback conditions.

To investigate whether discrepancies between perceived driving performance

and feedback received from the system may have had an impact on driver acceptance,

bivariate correlations between perceived trust in the system and driver acceptance were

carried out. The perceived trust scale involved 10 items assessing the perceived

reliability, integrity and trustworthiness of the information provided by the system. As

can be seen in Table 6.2, medium to strong significant positive correlations were

observed, suggesting that those drivers who perceived the system to be more reliable

and precise were more likely to find the system acceptable to use. Such a finding lends

evidence to the assumption that drivers who perceive feedback provided by the system

to be incongruent with perceptions of their own behavioural performance are more

likely to report lower levels of acceptance associated with the system.

Table 6.2. Bivariate correlations between driver acceptance scales and perceived
trust

Bivariate correlation with perceived trust


Scale Advice only Feedback Only Advice/Feedback
Usefulness (Van der Laan Scale)a .48** .53** .64**
Satisfaction (Van der Laan Scale)a .39* .49** .46**
Perceived ease of use of the systemb .29 .48** .57**
Perceived usefulness of the systemb .46** .54** .66**
Perceived usability of the system (SUS) c .37** .34* .57**
N = 40. * < .05 ** < .01
a
Adapted from van der Laan, et al. (1997). b Adapted from Staubach, et al. (2104). c Adapted from Bangor et al. (2008).

Taken together, these findings suggest relatively strong driver acceptance of all

systems, with the exception of system satisfaction, regardless of the types of messages

presented to drivers. Nonetheless, there were a number of significant preferences

towards advice messages. Thus, the results of the current study do not support H1.

Indeed, feedback messages were more likely to be perceived as distracting, difficult to

152 Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4)
understand, imprecise and untrustworthy, which had negative impacts on overall driver

acceptance levels.

6.7.2 H2: Workload


The mean scores and standard deviations for the NASA-RTLX across the

various conditions are presented in Table 6.3. When considering mean scores on the

scale, it can be seen that relatively low levels of workload were associated with each

of the conditions, with mean scores increasing contingent on the amount of

information presented to drivers. Indeed, there was a significant main effect of

condition on workload, Wilks’ Lambda = .56, F (3, 37) = 9.87, p = < .001, ηp2 = .44.

Largely consistent with H2, as the amount of information presented to drivers

increased, so too did workload associated with using the system. Indeed, pairwise

comparisons revealed significantly greater levels of workload associated with the

advice/feedback condition compared to both the advice only (p < .001) and baseline

condition (p < .001). Similarly, significantly greater levels of workload were

associated with the feedback only condition compared to both the advice only (p =

.002) and baseline condition (p < .001). No significant differences were observed

between the feedback only and advice/feedback conditions, or between the advice only

condition and baseline.

Table 6.3. Differences in driver workload by condition

Condition Mean (SD) Significant post-hoc p


comparison
Baseline 27.32 (12.87) Baseline < Advice/feedback p < .001
Advice only 28.36 (14.55) Baseline < Feedback only p < .001
Feedback only 35.69 (14.35) Advice only < Advice/feedback p < .001
Advice/feedback 36.18 (14.60) Advice only < Feedback only p < .002

Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4) 153
6.7.3 H3: Eco-safe driving behaviour
The mean and standard deviations for driving behaviour measures across the

various conditions, and results of each driving behaviour are presented in the right-

hand side of are presented in Table 6.4.

Fuel consumption and trip time


As can be seen in Table 6.4, there was a significant effect of condition on both

fuel consumption and trip time. Specifically, drivers used significantly less fuel in the

advice/feedback condition compared to either the advice only (p = .001) or feedback

only conditions (p = .005), or the baseline (p < .001). In addition, drivers took

significantly longer to complete the drive in the feedback only and advice/feedback

conditions compared to the baseline (p < .001 in both instances), while trip times were

also significantly longer in the feedback only condition compared to the advice only

condition (p = .004).

154 Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4)
Table 6.4. Summary of means and standard deviations and significance tests for eco-safe driving behaviour variables
M(SD)
Variable Baseline Advice Only Feedback Only Advice/Feedback Results
Fuel consumption (mL/sec) 359.9 (49.44) 354.2 (43.27) 350.7 (49.21) 335.1 (40.56) λ = .47, F (3, 37) = 13.98, p < .001, ηp 2 = .53
Total trip time (secs) 508.6 (77.09) 528.8 (65.83) 564.4 (70.73) 549.5 (82.09) λ = .55, F (3, 37) = 10.17, p < .001, ηp 2 = .45
Event1: with traffic
Accelerator pedal position (0-1) .14 (.04) .13 (.04) .11 (.06) .12 (.04) χ2 (3, n = 40) = 35.1, p < .001
SD Accelerator pedal position .07 (.04) .07 (.04) .08 (.05) .05 (.03) χ2 (3, n = 40) = 23.64, p < .001
Max. Accelerator pedal position .25 (.08) .25 (.09) .26 (.13) .20 (.07) χ2 (3, n = 40) = 23.22, p < .001
Event 2: without traffic
Accelerator pedal position (0-1) .22 (.06) .18 (06) .17 (.06) .16 (.06) χ2 (3, n = 40) = 37.47, p < .001
SD Accelerator pedal position .06 (.03) .05 (.04) .05 (.03) .04 (.03) χ2 (3, n = 40) = 16.47, p = .001
Max. Accelerator pedal position .30 (.08) .24 (.08) .24 (.07) .22 (.07) χ2 (3, n = 40) = 39.6, p < .001
Short-notice braking event
Brake use (N) 17.61 (6.37) 18.43 (6.97) 19.54 (7.46) 16.43 (7.07) χ2 (3, n = 34) = 2.97, p = .397
SD Brake use (N) 22.42 (8.99) 22.43 (8.20) 22.74 (9.57) 18.10 (6.81) χ2 (3, n = 35) = 1.59, p = .661
Max. Brake force (N) 74.20 (31.22) 72.13 (23.78) 78.58 (33.70) 62.37 (22.67) χ2 (3, n = 32) = 7.74, p = .052
Brake use area under curve 442.96 (220.75) 458.86 (181.34) 473.79 (225.04) 418.04 (179.03) χ2 (3, n = 33) = 1.07, p = .784
Long-range braking event
Brake use (N) 19.38 (9.43) 19.24 (9.58) 18.53 (8.72) 15.70 (7.67) χ2 (3, n = 37) = 7.05, p = .070
SD Brake use (N) 22.38 (6.67) 21.14 (7.69) 20.48 (6.15) 16.77 (5.23) χ2 (3, n = 30) = 12.60, p = .006
Max. Brake force (N) 67.62 (27.50) 71.07 (27.50) 72.43 (25.37) 54.63 (18.94) χ2 (3, n = 33) = 11.99, p = .007
Brake use area under curve 470.69 (215.18) 487.83 (248.81) 487.79 (210.75) 383.32 (153.78) χ2 (3, n = 35) = 8.25, p = .041
40 km/h Speed zone
Mean speed (km/h) 37.21 (1.91) 38.01 (1.80) 37.65 (1.49) 36.51 (1.47) λ = .49, F (3, 37) = 12.96, p < .001, ηp 2 = .51
% Time speeding 7.94 (5.71) 8.08 (4.68) 6.26 (4.30) 5.51 (4.33) λ = .64, F (3, 32) = 5.94, p = .002, ηp 2 = .36
Max. speed (km/h) 53.14 (7.91) 55.91 (6.69) 56.28 (7.85) 48.61 (6.61) λ = .48, F (3, 37) = 13.16, p < .001, ηp 2 = .52
60 km/h Speed zone
Mean speed (km/h) 58.33 (2.41) 56.86 (2.31) 56.65 (2.16) 57.18 (2.13) λ = .49, F (3, 34) = 11.62, p < .001, ηp 2 = .51
% Time speeding 26.06 (13.31) 16.66 (12.21) 16.48 (10.71) 18.09 (12.92) λ = .49, F (3, 37) = 12.96, p < .001, ηp 2 = .51
Max. speed (km/h) 64.29 (5.77) 62.25 (3.89) 61.77 (6.14) 62.76 (4.73) χ2 (3, n = 40) = 11.64, p = .009
Vehicle headway
Time interval to vehicle (secs) 6.47 (1.52) 7.01 (1.46) 6.95 (1.56) 7.55 (1.70) λ = .56, F (3, 37) = 9.71, p < .001, ηp 2 = .44

Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4) 155
Acceleration
The means and standard deviations for the acceleration variables, across the

various conditions, are presented in Table 6.4, as well as results of the significance

tests. As stated previously, acceleration data was collected for two separate events.

The first event required drivers to accelerate from a STOP sign after waiting for traffic

to pass by. In this event, there were significant main effects of condition for mean

pedal position, standard deviation of the pedal position and maximum pedal position.

As can be seen in Table 6.5, post-hoc comparisons revealed that drivers

demonstrated smoother acceleration in both the advice/feedback and feedback only

conditions, compared to the baseline, as well as smoother acceleration in the feedback

only condition compared to the advice only condition. In addition, there was

significantly less variability in pedal position in the advice/feedback condition

compared to both the feedback only and advice only conditions, and a marginally

significant difference compared to the baseline (z = -2.57, p = .010). Finally, maximum

acceleration was significantly lower in the advice/feedback condition compared to all

other conditions.

Table 6.5. Summary of significant post-hoc comparisons for acceleration variables


(Event 1: with traffic)
Variable Significant post-hoc comparisons Z p
Accelerator pedal position (0-1a) Baseline > Feedback only -4.11 < .001
Baseline > Advice/feedback -3.40 .001
Advice only > Feedback only -3.55 < .001
SD Accelerator pedal position Advice only < Advice/feedback -2.98 .003
Feedback only < Advice/feedback -4.44 < .001
Max. Accelerator pedal position Baseline > Advice/feedback -3.90 < .001
Advice only > Advice/feedback -3.53 < .001
Feedback only > Advice/feedback -3.55 < .001
a
float value between 0 (pedal released) and 1 (pedal completely pressed)

In the second acceleration event, drivers were not required to wait for traffic to

pass before accelerating from the STOP sign. In this event, there were once again

156 Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4)
significant main effects of condition on mean pedal position, standard deviation of the

pedal position, and maximum pedal position. As can be seen in Table 6, post-hoc

comparisons revealed that, compared to the baseline, drivers in all conditions

demonstrated smoother acceleration, less variability in pedal position, and lower

maximum acceleration.

Table 6.6. Summary of significant post-hoc comparisons for acceleration variables


(Event 2: without traffic)
Variable Significant post-hoc comparisons Z p
Accelerator pedal position (0-1a) Baseline > Advice only -4.50 < .001
Baseline > Feedback only -4.33 < .001
Baseline > Advice/feedback -4.96 < .001
SD Accelerator pedal position Baseline > Advice only -3.27 .001
Baseline > Feedback only -2.68 .007
Baseline > Advice/feedback -4.07 < .001
Max. Accelerator pedal position Baseline > Advice only -4.12 < .001
Baseline > Feedback only -4.02 < .001
Baseline > Advice/feedback -5.26 < .001
a
float value between 0 (pedal released) and 1 (pedal completely pressed)

Deceleration
The means and standard deviations for the deceleration variables, across the

various conditions, are also presented in Table 6.4, along with results of the

significance tests. In the short-notice braking event (i.e., on approach to a green light

changing to red) there were no significant main effects of condition on any of the

deceleration variables, including mean brake use, standard deviation of brake use,

maximum brake force, and area under the brake use curve. Conversely, in the long-

range braking event (i.e., approaching a STOP sign with traffic passing through the

intersection), significant effects of condition were found for standard deviation of

brake use, maximum brake force and area under the brake use curve. There was

however only a marginally significant effect on mean brake use.

Table 6.7 presents a summary of the significant post-hoc comparisons for

deceleration variables in the long-range braking event. As can be seen, there was

significantly less variability in braking in the advice/feedback condition compared all

Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4) 157
other conditions, as well as significantly lower maximums in the advice/feedback

condition compared to both the feedback only and advice only conditions. Finally,

braking was significantly smoother in the advice/feedback condition compared to the

feedback only condition.

Table 6.7. Summary of significant post-hoc comparisons for deceleration variables


(long-range braking event)
Variable Significant post-hoc comparisons Z p
SD Brake use (Na) Baseline > Advice/feedback -3.95 < .001
Advice only > Advice/feedback -3.14 .002
Feedback only > Advice/feedback -3.50 < .001
Max. Brake force (Na) Advice only > Advice/feedback -2.69 .007
Feedback only > Advice/feedback -3.78 < .001
Brake use area under curve Feedback only > Advice/feedback -2.84 .004
a
Applied force on the brake pedal in Newtons

Speeding and headway


The means and standard deviations for the speed and time interval to vehicle

variables, across the various conditions, are presented in Table 6.4. Mean speeds in all

conditions, and in both speed zones, were under the speed limit. Nonetheless, a number

of significant main effects of condition were observed in relation to the speed

variables. Specifically, a significant effect on free-flowing mean speeds was found for

both the 40 km/h zone, and 60 km/h zone. In the 40 km/h zone, participants drove at

significantly lower speeds in the advice/feedback condition compared to both the

feedback only and advice only conditions (p < .001, in both instances), as well as the

baseline (p = .021). In addition, participants drove at significantly higher speeds in the

advice only condition compared to the baseline (p = .037). In the 60 km/h zone, mean

speeds were significantly lower in all conditions when compared to the baseline (p <

.005, in all instances).

The descriptive data for percentage of time spent speeding shows that speeding

behaviour was relatively infrequent in 40 km/h zones, ranging from 5.5-8.1% across

the conditions. Conversely, speeding behaviour was more prevalent in the 60 km/h

158 Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4)
zone, ranging from 16.5-18.1% in the system conditions, and 26.1% in the baseline

condition. However, maximum speeds were considerably higher, relative to the speed

limit in question, in the 40 km/h zone compared to the 60 km/h zone, ranging between

8.6-16.3 km/h above the speed limit in the 40 km/h zone, compared to 1.8-4.3 km/h

over in the 60 km/h zone.

Significant effects were also observed in relation to the percentage of time spent

over the speed limit in both the 40 km/h zone, and 60 km/h zone. Specifically, drivers

spent significantly less time exceeding the speed limit in the 40 km/h zone in the

advice/feedback condition compared to either the advice only condition (p = .007) or

baseline (p = .006), while in the 60 km/h zone, participants spent significantly less

time exceeding the speed limit in all conditions compared to the baseline

(AdviceFeedback/Baseline: p = .006; Feedback only/Baseline: p < .001; Advice

only/Baseline: p = .012).

There was also a significant main effect of condition on maximum speed in both

the 40 km/h zone, and in the 60 km/h zone. In the 40 km/h zone, maximum speeds

were significantly lower in the advice/feedback condition compared to all other

conditions (AdviceFeedback/Baseline: p < .001; Feedback only/Baseline: p < .001;

Advice only/Baseline: p = .007). Conversely, in the 60 km/h zone, mean speeds were

significantly lower in both the feedback only and advice only conditions, compared to

the baseline (Feedback only/Baseline: z = -3.12, p = .002; Advice only/Baseline: z = -

2.66, p = .008).

Finally, a significant effect of condition was also observed in relation to mean

Time Interval Vehicle. The pairwise comparisons, revealed that headway was

significantly larger in the advice/feedback condition compared to the baseline,

however no other significant differences were observed. When analysing the means, it

Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4) 159
can be seen that drivers typically kept a considerable distance from the vehicle in front,

with the lowest average distance being 6.5 seconds in the baseline condition, which is

still much larger than the recommended 2.5 seconds under Australian road rules.

Overall, the findings were largely consistent with H3, which hypothesised that

as the amount of information presented to drivers through advice and feedback

messages from the system increased, so too would the positive impact of eco-safe

driving behaviour. The advice/feedback system had the most positive impacts on all

eco-safe driving behaviours measured, with the exception of acceleration, where both

the advice/feedback and feedback only systems had the most pronounced impacts.

6.7.4 Realism of the simulated driving experience


Taken together, the results provide evidence that the cooperative in-vehicle, eco-

safe driving HMI has the potential to positively impact eco-safe driving behaviour,

particularly when the system involves both advice and feedback messages. However,

the impact appeared least pronounced for braking and deceleration behaviour. When

exploring data regarding participant perceptions of the realism of the driving simulator

experience (see Table 6.8), it was observed that participants had relatively neutral

views regarding how realistic they found the simulated driving experience. More

specifically, while they tended to have relatively good experiences with steering and

speed maintenance, a greater number of participants reported a considerably poorer

experience with the brakes in the simulator.

Table 6.8. Means and standard deviations of perceptions of driving simulator realism
Variablea Mean SD
How realistic was the simulator experience? 3.05 1.06
Experience with the brakes 2.33 0.73
Experience with the steering 3.65 0.77
Experience with maintaining desired speed 3.75 0.81
a
All items scored on a 5-point Likert scale. Realism item: 1 = very unrealistic to 5 very realistic; experience items: 1 = very poor
to 5 = very good.

160 Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4)
6.8 DISCUSSION

The primary aim of the current research was to evaluate a new eco-safe in-

vehicle HMI designed to reduce fuel consumption and emissions while improving safe

driving. Specifically, the research sought to examine the influence of advice and/or

feedback messages from an in-vehicle HMI on driver acceptance, workload and eco-

safe driving behaviour.

This research has shown that driver acceptance of prototyped eco-safe in-vehicle

HMIs, were relatively strong, regardless of the types of messages presented to drivers.

That is, drivers reported each system was usable, easy to use and useful. These findings

are encouraging and further strengthens the idea of adopting user-centred design

approach for developing new technologies (Francois et al., 2017; Maguire, 2001;

Vilimek & Keinath, 2014).

Participants generally reported greater levels of usability, ease of use and

satisfaction in association with the advice only condition. Drivers also reported that

feedback messages were more distracting, and frustrating than advice messages.

Specifically, it appears that while constructive feedback from an in-vehicle HMI may

motivate behaviour change (Horberry et al., 2014b), if feedback provided by eco-safe

driving in-vehicle HMI does not match the drivers’ own perceptions of their driving

performance, it may also lead to cognitive dissonance (Festinger, 1962). Thus, the

results of the current study do not support hypothesis 1, which hypothesised that

greater driver acceptance would be observed when drivers where presented with

greater levels of information.

Additional analyses supported this notion, observing medium to strong

significant positive correlations between perceived trust and driver acceptance

measures, suggesting that drivers who perceived the system to be less reliable and less

Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4) 161
precise were more likely to find the system less acceptable to use. These findings are

consistent with previous research which has highlighted the relationship between

driver acceptance of negative feedback from an in-vehicle HMI and perceptions of

their own driving behaviour, as well as trust as a predictor of technology acceptance

(Ghazizadeh & Lee, 2014; Kim & Malhotra, 2005).These findings highlight the need

to carefully consider the manner in which feedback, and in particular negative

feedback, is presented to drivers for it to be well received and perceived as acceptable.

A key component of achieving this goal is likely to be establishing high levels of

perceived accuracy of system feedback. The inclusion of both advisory messages that

notify drivers regarding what behaviours need to be performed, as well as behavioural

feedback messages, appeared to represent the most comprehensive and effective

provision of information to drivers. These findings are consistent with previous studies

highlighting the importance of performance feedback as a key motivator of behaviour

change (e.g. Gardner et al., 2010; Hattie & Timperley, 2007; Webb, Sniehotta, &

Michie, 2010).As such, future design and development of feedback messages should

ensure comprehension and reliability of the feedback messages, while also minimising

the degree to which the message is overwhelming and distracting (Shute, 2008).

Achieving this balance is likely to build trust in the system, subsequently reducing the

likelihood of frustration and feelings of being judged when using the technology.

Moreover, it may be beneficial to tailor and personalised the level of advice

and/or feedback provided to drivers dependent on their level of eco-safe driving skills,

as objectively measured by the system. Such a recommendation is consistent with

previous research in the area of computer-based instruction for education (Magaña &

Muñoz-Organero, 2016; Shute, 2008; Taubman-Ben-Ari & Yehiel, 2012; Van

Huysduynen, Terken, Meschtscherjakov, Eggen, & Tscheligi, 2017). This may include

162 Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4)
options for greater customisation of the feedback characteristics and more contextual

information. Specifically, there is unlikely to be a one-size-fits-all approach to the

provision of feedback, and as such users should have considerable control over the

characteristics in which feedback is provided to them. Such a recommendation is

highlighted in the International Organisation for Standardisation (ISO/TR), and

supported by previous research (Fors, Kircher, & Ahlstrom, 2015; Steinberger,

Schroeter, Foth, et al., 2017; Vaezipour et al., 2017).

In addition, the level of feedback provided may need to be expanded to include

more comprehensive information. This include information on precisely how eco-safe

driving behaviours are measured by the system, specific data on behavioural

performance (such as fuel consumption, speed profiles, deceleration and acceleration

data), and a more in-depth description of what is required from the driver to improve

their behaviour. Given the extensive nature of such information, this feedback should

be included in the back-end of the system, or available for viewing in a compatible

smartphone application in order to minimise the distractive impact of system use on

driver performance (Steinberger, Proppe, Schroeter, & Alt, 2016; Stevens, Quimby,

Board, Kersloot, & Burns, 2002).

Findings regarding the impact of advice and/or feedback messages on driver

workload were largely consistent with hypothesis 2, such that increases in workload

were associated with increases in the amount of information drivers were required to

attend to and process. Interestingly, while the conditions involving feedback produced

significantly greater workload than the advice only condition and baseline, no

significant mean differences were observed between the two conditions involving

feedback, or between the advice only condition and baseline, suggesting that advice

messages produced minimal additional workload for drivers.

Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4) 163
While consistent with hypotheses, these findings have implications for overall

safety associated with using the system when feedback messages are included, and in

particular the potential distractibility of the system. That said, according to previous

research, the mean scores for the two conditions involving feedback still represent

relatively low levels of workload (Grier, 2015). Moreover, previous research has

demonstrated that merely requesting individuals to drive in an eco-friendly manner,

without even introducing a driver support system, can lead to increased driver

workload due to the increased focus on one’s own actions (Birrell, Young, & Weldon,

2010; Pampel et al., 2015). Nonetheless, in moving to emerging technologies, future

research should investigate potential avenues for reducing the driver workload

associated with feedback messages in emergency situations (e.g., the prioritisation of

information when a potential hazard is detected). This is particularly important since

distraction may occur due to the increasing complexity and workload demand of the

driving task when interacting with a system (Oviedo-Trespalacios et al., 2016).

The overarching aim of the cooperative eco-safe driving in-vehicle HMI was to

improve eco-driving behaviour, while maintaining driver safety. It was hypothesised

that conditions involving feedback messages would have a greater impact on driving

behaviour, compared to advice messages, given that the feedback sought to notify

drivers of inappropriate behavioural performance and the need for improvement.

Overall, the findings were largely consistent with hypothesis 3. The advice/feedback

condition compare to baseline was associated with reduction in fuel consumption

(6.7%), smoother acceleration (33%) and smoother deceleration (25.1%), as well as

reduction in speeding behaviour (2% with greater reductions observed for time spent

speeding up to 49.8% reduction) and improvements in vehicle headway (16.7%).

These results are consistent with previous research, which has typically reported

164 Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4)
reductions in fuel consumption ranging from 5% to 15%, with some reporting

reductions up to 30% depending on traffic conditions (see review see review of 19

studies by Alam & McNabola, 2014; Barkenbus, 2010; McIlroy et al., 2017).

The feedback only condition was also associated with significant improvements

in smooth acceleration and some measures of speeding behaviour. Conversely, the

advice only condition was only associated with significant improvements in

acceleration in non-complex scenarios and on some measures of speeding behaviour.

In addition, the conditions involving feedback messages were both associated with

significant increases in total trip time. This finding is consistent with previous research

that suggests when improving eco-driving, drivers often engage in anticipation of

traffic environment, drive slower and pay greater attention to the driving task (Pampel

et al., 2015).

There were a number of interesting findings regarding the influence of more

complex driving scenarios, such as traffic and changing conditions, on the

effectiveness of the various systems. Specifically, while significantly smoother

acceleration was observed for all conditions, compared to the baseline, in the non-

traffic event, only the conditions involving feedback messages were associated with

improved acceleration in the traffic event. Similarly, lower than baseline means were

observed for all conditions across all deceleration variables in association with both

the short-notice and long-range braking events. While these findings were only

statistically significant for the long-range braking event, the differences observed in

the short-notice braking event appear to be meaningful, and it is likely that the large

variances in the data restricted the number of statistically significant findings observed.

Taken together, these findings suggest that more complex driving scenarios may

benefit from behavioural feedback. Given the prevalence of traffic on urban and

Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4) 165
suburban streets, and the unpredictable nature of driving in general, improvements in

acceleration and deceleration associated with such feedback could produce

considerable reduction in fuel consumption, emission and environmental benefits

(Fors et al., 2015; Graving, Manser, & Becic, 2010; McIlroy et al., 2017).

Overall, speed data suggested compliant behaviour among the sample of drivers,

with mean speeds for all conditions, and in both speed zones, under the speed limit.

Data for the percentage of time spent speeding also showed considerably less frequent

speeding in the 40 km/h zone compared to the 60 km/h zone. This finding is consistent

with other self-reported and objective measures of speeding behaviour in these zones

among Queensland drivers (Soole, 2012; Transport and Main Roads, 2014). However

maximum speeds were considerably higher in the 40 km/h zone compared to the 60

km/h zone, relative to the speed limit in question. Taken together, the speed data

suggests that while drivers speed more in the 60 km/h zone, a number of short but

serious infringements were committed in the 40 km/h zone.

The findings of this research complement those of earlier studies demonstrating

the potential positive benefits of cooperative driving support systems on fuel

consumption and reduce emissions (McIlroy et al., 2017), smoother acceleration and

deceleration (Graving et al., 2010; Van der Voort et al., 2001), accelerator pedal

pressure (Azzi, Reymond, Merienne, & Kemeny, 2011), and speed behaviour

(Steinberger, Schroeter, & Watling, 2017; Trommer & Höltl, 2012). Moreover, this

research has practical implications for designers in the automotive industry attempting

to develop systems to improve eco-safe driving behaviour using in-vehicle technology

with a high level of driver acceptance.

166 Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4)
6.9 CONCLUSION

The findings from the current experiment suggests that a cooperative eco-safe

driving in-vehicle HMI involving both advice and feedback messages has the potential

to improve eco-safe driving, while also being accepted by drivers and producing

minimal adverse impacts on workload. While it appears that feedback messages are

critical to the effectiveness of the system, the influence of advice messages cannot be

downplayed. Indeed, the advice/feedback condition outperformed the feedback only

condition on the vast majority of eco-safe driving behaviour measures and thus it

appears that the effectiveness of the system relies on both message types being

presented to the driver, rather than one or the other. Taken together, the results provide

evidence that driver acceptance is a crucial antecedent for encouraging voluntary

system use and for widespread system implementation, however does not equate to

system effectiveness. Indeed, system effectiveness, in terms of positive impacts on

behaviour and behavioural outcomes, is ultimately more important from a practical

perspective. That is, driver acceptance and system effectiveness are two separates, but

complementary, requirements of a successful system.

The current paper highlights the potential benefits associated with such systems,

however the challenge remains regarding how best to provide real-time feedback to

drivers that is perceived to be precise and which encourages voluntary behavioural

change with minimal increases in driver workload. This is particularly important given

that drivers in the current study suggested that the relatively minimalistic feedback that

was provided was somewhat judgemental. Moreover, perceived trust in the reliability

of the feedback being presented appears to be fundamental to achieving optimal levels

of driver acceptance. Future research should more comprehensively examine optimal

Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4) 167
ways for providing feedback to drivers and seek to further refine the design and

development of such a system for improving eco-safe driving behaviour.

6.10 LIMITATIONS

While the findings of this study are undoubtedly encouraging, a number of

limitations must be acknowledged. The resulting sample was relatively small in size

which may impact on generalisability of the results. That said, this is an acceptable

sample size in driving simulator studies (Čegovnik, Stojmenova, Jakus, & Sodnik,

2018; H. Jamson, Hibberd, & Merat, 2015; Saifuzzaman, Zheng, Haque, &

Washington, 2015; Staubach et al., 2014). Nonetheless, future research should address

this limitation by conducting a user study with a larger, more representative sample.

As with all research conducted in a simulated driving environment, there are

concerns regarding the real-world generalisability of the findings. Participants

reported neutral perceptions regarding the realistic nature of the simulated driving

experience, good experiences with the steering and speed maintenance aspects of the

simulator, suggesting some evidence of the generalisability of the findings. However,

a number of participants also reported poor experiences with the brakes in the

simulator, which may help to explain the variability and the lack of significant

differences observed in the deceleration data. These issues, while not explored in

detail, may have included a lack of responsiveness or hypersensitivity and could have

certainly affected subsequent data and analyses. Notwithstanding these limitations,

prior research has demonstrated generalisability of findings from a simulated driving

environment to the real-world content (Charlton, 2009; Krause, Yilmaz, & Bengler,

2014; Santos, Merat, Mouta, Brookhuis, & De Waard, 2005). Lastly, the short duration

of the driving experiment did not allow for an assessment of the long-term uptake of

the system, nor the effects associated with driver habituation.

168 Chapter 6: A Simulator Evaluation of In-vehicle Human Machine Interface for Eco-Safe Driving (Paper 4)
Chapter 7: A Simulator Study of the Effect
of Incentives on Adoption and
Effectiveness of an In-Vehicle
Driving HMI (Paper 5)

7.1 INTRODUCTORY COMMENTS

This chapter builds on the findings presented in the previous chapter and

additionally focuses on the final step of the user-centred design approach related to

system evaluation. In this chapter, additional findings from the driving simulator

experiment (Study 3) are discussed. Specially, the main aim of the paper presented in

this chapter (Paper 5) was to explore the impact of incentives and competition

associated with using the in-vehicle eco-safe driving HMI on subsequent eco-safe

driving behavioural performance. In addition, a range of demographic, attitudinal and

self-reported behaviour variables were examined as possible factors associated with

preferred system choice among drivers.

This chapter seeks to address Research Question 3: “What impact do incentives

associated with using an in-vehicle eco-safe HMI have on eco-safe driving

performance?”. Despite increasing levels of interest in in-vehicle technology, low

levels of willingness to purchase among drivers suggests that a major challenge relates

to enhancing motivation to use these systems as part of their everyday driving. Thus,

this chapter seeks to objectively evaluate the influence of incentives and competition

associated with system use on eco-safe driving behaviour, as well as assess intentions

to use and willingness to purchase the system, both of which play crucial roles in

sustained voluntary use of in-vehicle HMIs.

Chapter 7: A Simulator Study of the Effect of Incentives on Adoption and Effectiveness of an In-Vehicle Driving
HMI (Paper 5) 169
Taken from: Vaezipour, A., Rakotonirainy, A., Haworth, N. & Delhomme, P. (2017).
A Simulator Evaluation of the Effect of Incentive on Adoption and Effectiveness of
an In-vehicle Eco-Safe Driving HMI. Journal of Safety Research

Publication Status: Submitted

Journal Quality: Journal of Safety research is a peer-reviewed publication with


international readership. The journal impact factor is 1.841, and rank Q1 (SCImago)
in safety, risk, reliability and quality.

Copyright: The publisher of this article stated that authors can use their articles in full
or in part to include in thesis of dissertation (provided that this is not to be published
commercially).

Statement of Contribution of Co-Authors for Thesis by Published Papers

The authors listed below have certified that:

 They meet the criteria for authorship in that they have participated in the
conception, execution, or interpretation, of at least that part of the publication
in their field of expertise;

 They take public responsibility for their part of the publication, except for the
responsible author who accepts overall responsibility for the publication;

 There are no other authors of the publication per these criteria;

 Potential conflicts of interest have been disclosed to (a) granting bodies (b) the
editor or publisher of journals or other publications, and (c) the head of the
responsible academic unit, and

 They agree to the use of the publication in the student’s thesis and its
publication on the QUT’s ePrints site consistent with any limitations set by
publisher requirements.

170 Chapter 7: A Simulator Study of the Effect of Incentives on Adoption and Effectiveness of an In-Vehicle
Driving HMI (Paper 5)
Contributor Statement of contribution

Atiyeh Vaezipour Study Design, recruitment, data collection, data


analysis, write the manuscript.
Professor Andry Rakotonirainy Supervised the research, reviewed the
manuscript and supported the editorial process
Professor Narelle Haworth Supervised the research, reviewed the
manuscript and supported the editorial process
Dr Patricia Delhomme Discussed the research design, reviewed the
manuscript and supported the editorial process

Principal Supervisor Confirmation: I have sighted email or other correspondence


from all Co-author confirming their certifying authorship.

Name: Professor Andry Rakotonirainy

QUT Verified Signature


Signature:

Date: 25/03/2018

Chapter 7: A Simulator Study of the Effect of Incentives on Adoption and Effectiveness of an In-Vehicle Driving
HMI (Paper 5) 171
7.2 ABSTRACT

Rapid developments in transportation technologies, such as in-vehicle human-

machine interfaces (HMI), have the potential to improve driving behaviour. However,

the use of such approaches is typically voluntary and there are numerous barriers to

their widespread implementation. The aim of the current paper is to investigate the

impact of an incentive and competition to use an in-vehicle HMI on subsequent eco-

safe driving behaviour, as well as assess intentions to use and willingness to purchase

the system, both of which play crucial roles in sustained voluntary use of in-vehicle

HMIs. Forty drivers participated in a driving simulator experiment and driver survey.

Three variations of an eco-safe driving in-vehicle HMI were evaluated (advice only,

feedback only, combined advice and feedback), followed by an incentive and

competition condition. The findings revealed the 4.7% reduction in fuel consumption

associated with an addition of incentive and competition on eco-safe driving

behaviour. Moreover, there was some evidence to suggest that a range of extrinsic and

intrinsic incentives may be beneficial for increasing intentions to use such a system.

The implications of these findings for the design and development of such systems, as

well as their implementation in real-world settings, is discussed.

7.3 INTRODUCTION

Traffic crashes contributes to more than a million fatalities and 50 million

injuries annually worldwide (World Health Organization, 2015). In addition, the use

of motorised road transport is a leading contributor to the production of pollutant gases,

including CO2. Indeed, in 2013, transportation alone represented 16% of all emissions

in Australia, with light vehicles the largest contributor amongst vehicle fleets

(Australian Authority Climate Change, 2014). Thus, motorised vehicles pose a

significant threat to the environmental and human health.

172 Chapter 7: A Simulator Study of the Effect of Incentives on Adoption and Effectiveness of an In-Vehicle
Driving HMI (Paper 5)
Emerging technologies are evolving at an increasing pace and, in particular,

there have been rapid advances in the field of in-vehicle technologies, including

aftermarket products such as smartphone applications and smarter human machine

interfaces (HMI). Technology has been designed to provide real-time advice and

feedback to drivers with the aim of encouraging improvements in either safe driving

or eco-driving. For example, a number of studies have demonstrated the effectiveness

of such systems in improving safe driving behaviour (Adell, Várhelyi, & Fontana,

2011; Eby et al., 2015; Steinberger, Schroeter, & Watling, 2017), as well as in reducing

fuel consumption and emissions (Dahlinger & Wortmann, 2016; McIlroy et al., 2017;

Staubach et al., 2014; Van der Voort et al., 2001). However, there is a paucity of

research exploring in-vehicle HMIs that integrate eco-driving and safe driving. As

such, this paper focuses on eco-safe driving, defined as a range of driving behaviours

that reduce fuel consumption and subsequent emissions while also maintaining safety.

These includes, smooth acceleration and deceleration and anticipating the traffic flow

to minimise the use of brake and accelerator (Barkenbus, 2010; Delhomme et al., 2013;

Sivak & Schoettle, 2012), as well as safe driving behaviour including adherence with

legal speed limits and keeping a safe following distance.

Moreover, the use of such systems is typically voluntary, with evidence

suggesting fundamental barriers to their widespread implementation and low

intentions to use such technology, even among individuals who are concerned about

the environment. For example, individuals are less likely to be motivated to engage in

emission reduction activities when they perceive major contributors to greenhouse

emissions (e.g., major corporations) as not doing their part (Barkenbus, 2010). There

is also limited understanding of the features and functions of such technology desired

by drivers which increase the likelihood of acceptance and usage (Albert et al., 2016).

Chapter 7: A Simulator Study of the Effect of Incentives on Adoption and Effectiveness of an In-Vehicle Driving
HMI (Paper 5) 173
Therefore, despite increasing levels of interest in in-vehicle technology, low levels of

willingness to purchase among drivers suggests that a major challenge is enhancing

their motivation to use these systems as part of their everyday driving.

According to Ryan and Deci (2000, p. 55), motivation is either intrinsic or

extrinsic. Specifically, intrinsic motivation refers to “doing something because it is

inherently interesting or enjoyable” (e.g., feels good or helps a cause), while extrinsic

motivation refers to “doing something because it leads to a separable outcome” (e.g.,

reward or gain). While both are important and can motivate users, some research

suggests that intrinsic motivations alone are often insufficient to encourage eco-safe

driving behaviour. Indeed, qualitative research conducted in Newcastle, Australia

revealed that while most people expressed concern for the environment, convenience

and saving time had a greater influence on eco-driving behaviour (Harvey et al., 2013).

Similar findings are observed when examining motivations to use eco-driving

in-vehicle technology. Harvey et al. (2013) found that the incentive to save money

through better fuel economy was typically insufficient to motivate eco-driving, and

that the voluntary use of in-vehicle feedback devices is unlikely to be successful

without the use of extrinsic rewards. Moreover, prior research by Vaezipour,

Rakotonirainy, Hawoth and Delhomme (2017), revealed paradoxical attitudes

whereby intrinsic motivations, such as improving the environment and road safety,

were reported as motivations for system use, however many participants also reported

that extrinsic rewards would be necessary to motivate actual long-term system use

(Vaezipour et al., 2017). Thus, it appears that extrinsic rewards (e.g., money, insurance

discounts) may be a critical component of widespread implementation and use of eco-

safe driving in-vehicle technology, in particular aftermarket products.

174 Chapter 7: A Simulator Study of the Effect of Incentives on Adoption and Effectiveness of an In-Vehicle
Driving HMI (Paper 5)
Both monetary and non-monetary (e.g., discounts, vouchers, prizes) extrinsic

incentives have been shown to be effective in improving a range of human behaviours

including energy efficiency in the home (Guerassimoff & Thomas, 2015), performance

in the workplace (Condly et al., 2003; Garbers & Konradt, 2014), educational

performance (Gneezy et al., 2011), and a range of health behaviours, such as physical

activity, smoking cessation, vaccinations and blood donation (Giles et al., 2014;

Mortimer et al., 2013). Incentives have also been shown to be effective in improving

safe driving behaviour, such as the Speed Camera Lottery in Stockholm, Sweden (see

Volkswagen, 2009), and a variety of Pay As You Drive vehicle insurance schemes

(Tselentis et al., 2016). There is also evidence to suggest that reward systems can be

used to encourage eco-driving behaviour, even when an in-vehicle device is not used

(Lai, 2015; Schall & Mohnen, 2017).

Although limited, previous research has demonstrated the potential impact of

incentives on both the use of eco-driving and safe driving in-vehicle technology and

subsequent behaviour associated with the use of such systems. Indeed, a number of

studies have shown that incentives can increase willingness to purchase and intentions

to use these systems (Regan et al., Chorlton, Hess, Jamson, & Wardman, 2012; 2002;

Vaezipour et al., 2017), as well as improve behavioural outcomes associated with such

systems (Reagan, Bliss, Van Houten, & Hilton, 2013). Moreover, incentives can also

increase the sustained, long-term use of eco-driving in-vehicle systems (Dahlinger &

Wortmann, 2016).

Qualitative research has suggested that incentives associated with money,

convenience and fun are typically most popular among drivers (Loumidi et al., 2011).

Moreover, vehicle-related incentives, such as reductions in insurance premiums and

vehicle registration costs or fuel vouchers, are effective approaches to structuring

Chapter 7: A Simulator Study of the Effect of Incentives on Adoption and Effectiveness of an In-Vehicle Driving
HMI (Paper 5) 175
incentives (Regan et al., 2002; Vaezipour et al., 2017). Interestingly, incentives

associated with competition, such as leader boards and challenges, have been found to

be less popular, however when explored further, competition with oneself was

perceived as more of a motivation compared to competing with other drivers (Loumidi

et al., 2011).

Following on from this point, researchers and practitioners are increasingly

exploring the use of gamification principles to engage users and motivate system use

and behavioural change. Gamification refers to the “use of game design elements in a

non-game context” (Deterding et al., 2011, p. 11). With respect to in-vehicle

technology, gamification has most commonly involved the use of tasks and challenges,

performance feedback, scoring systems and leader boards. Examples include the

TomTom Curfer, Flo and Samsung S-Drive smartphone app. To date, the limited

available research suggests that while gamification may improve initial system

engagement, there is a paucity of evidence to suggest long-term effects on system use

or on eco-safe driving behaviour (Magana & Munoz-Organero, 2015; Rapp, 2015). In

any case, it has been acknowledged that the use of gamification requires careful

consideration of the potential for driver distraction from the driving task (Diewald et

al., 2013).

However, concerns have been raised regarding the use of extrinsic incentives,

with some arguing that reward systems may diminish intrinsic motivations (Deci et al.,

2001), and trap individuals in a ‘reward loop’ (Zichermann & Cunningham, 2011).

Thus, this paper seeks to explore the impact of incentives to motivate the use of an

eco-safe driving in-vehicle HMI, as well as on subsequent eco-safe driving behaviours.

176 Chapter 7: A Simulator Study of the Effect of Incentives on Adoption and Effectiveness of an In-Vehicle
Driving HMI (Paper 5)
7.4 STUDY AIMS

The current study is part of a larger program of research aimed at improving eco-

safe driving through the development of an in-vehicle HMI. The design and

development of the prototype eco-safe in-vehicle HMI evaluated in this paper adopted

a user-centred design approach according to the International Organisation for

Standardisation’s (2010) guidelines for human-centred design of interactive systems.

User-centred design refers to an iterative approach in which the needs, motivations and

limitations of the intended end-users of a system represent the driving force behind the

development of new technologies, and are considered at all stages of the design process

(Maguire, 2001). Previous research has outlined the design and development process

for creating the prototype eco-safe in-vehicle HMI, as well as reported on the impact

of the system on driver acceptance of the technology, workload and eco-safe driving

behaviours (see Vaezipour et al., 2016; Vaezipour et al., 2017).

The aim of the current paper is to investigate the impact of incentive and

competition on adoption and effectiveness of an eco-safe in-vehicle HMI. In addition,

the influence of variations of an in-vehicle HMI (i.e. advice only, feedback only,

combined advice and feedback) messages on system preference is investigated, as are

factors associated with willingness to purchase such a system. The impact of

demographic, attitudinal and self-reported behaviour variables on system choice and

intentions to use the system, are also explored.

7.5 METHOD

7.5.1 Participants
Participants were primarily recruited through a media release, the university

webpage, and snowball sampling. Eligibility criteria included holding a valid

Australian driver licence (i.e. provisional or open). While 53 participants were initially

Chapter 7: A Simulator Study of the Effect of Incentives on Adoption and Effectiveness of an In-Vehicle Driving
HMI (Paper 5) 177
recruited, 13 participants were excluded due to motion sickness, dizziness and nausea

that is commonplace in driving simulator studies (Fisher et al., 2011). Therefore, 40

participants (20 female, 20 male), aged between 18 and 65 years (M = 31.35, SD =

12.18) took part in the experiment.

The sample comprised of 14 provisional (i.e., unsupervised but restricted)

licence holders (35%) and 26 open (i.e., unrestricted) licence holders (65%). There

was an equivalent number of males and females across all age groups and licence

types. All provisional licence holders were aged 18-24 years, while 16 of the open

licence holders were aged 25-39 years and ten were aged 40-65 years. Participants

(62.5%) reported driving between 10,001 and 30,000 kilometres in the previous year,

with 35% reporting driving less than 10,000 kilometres.

7.5.2 Driving Simulator and Scenario


The experiment was conducted in the Centre for Accident Research and Road

Safety- Queensland (CARRS-Q) advanced driving simulator. The driving simulator

represented a safe and convenient method for examining driver behaviour, which was

particularly pertinent given the potentially distracting effect of the in-vehicle interface,

which may be risky in a less-controlled, real-world environment.

The advanced driving simulator consists of a Holden Commodore passenger car

with an automatic gearbox (see Figure 7.1). The simulator uses the SCANeR™ studio

from OKTAL, linked to website software with eight computers and projectors and a

six degree of freedom (6DOF) motion platform that can move and twist in three

dimensions. When seated in the simulator vehicle, drivers are immersed in a virtual

environment that includes a 180° front field-of-view, simulated rear-view mirror

images, surround-sound for engine and environment noise, a real vehicle cabin and

simulated vehicle motion using a hydraulic system. The simulator collects driver

178 Chapter 7: A Simulator Study of the Effect of Incentives on Adoption and Effectiveness of an In-Vehicle
Driving HMI (Paper 5)
behaviour data such as speed, acceleration, braking and vehicle positioning at a rate of

20 Hz.

Photos by Sonja de Sterke

Figure 7.1. The CARRS-Q Advanced Driving Simulator

To maximise the realism and familiarity of the driving experience, the simulated

scenario consisted of urban and suburban areas similar to the Brisbane city area and

surrounds. Standard Australian speed limit signs were used, with speed limits of 40

km/h in urban sections and 60 km/h in suburban areas. The simulated scenario was

approximately seven kilometres long and programmed to incorporate traffic events

including signalised and non-signalised intersections, speed limit changes and a

vehicle following event. In all instances, participants were instructed to drive as they

normally would and to follow posted signs to the airport. Each drive took around ten

minutes to complete and starting locations within the scenario were counter-balanced

across conditions to further reduce the impact of learning effects on subsequent data.

7.5.3 Experimental Design


In the within-subjects design, each participant was exposed to five conditions.

Driving conditions differed with regards to the type of messages presented to drivers

via the in-vehicle HMI: (i) baseline (no messages); (ii) advice only; (iii) feedback only;

(iv) advice and feedback; and (v) self-selected user choice condition involving

incentive and competition. The order of the first four conditions were counter-

balanced, while the incentive and competition condition always occurred last.

Chapter 7: A Simulator Study of the Effect of Incentives on Adoption and Effectiveness of an In-Vehicle Driving
HMI (Paper 5) 179
Examples of the advice and/or feedback screens used by the system are shown

in Figure 7.2. Advisory messages involved screens that presented an icon and simple

text (e.g., smooth braking) on a yellow background, and were projected to drivers

leading up to specific traffic events, such as when approaching traffic lights, to

encourage an anticipatory driving style. Conversely, feedback messages included a

“face”, as well as a smaller icon and illumination of the LED lights. This feedback

differed depending on whether the behavioural performance of the driver was

appropriate (i.e., happy face and a green light) or inappropriate (i.e., sad face and a red

light). In the baseline condition, no messages were presented to drivers; however, the

inactive system was present in the vehicle during the drive.

Figure 7.2. Examples of advice (upper panels) and feedback (lower panels)
messages

 HMI Goal: Reduce fuel consumption and improve safe driving specifically,

promote regularity and anticipatory driving by reducing the level of

inappropriate acceleration and deceleration, reduce speed violation, reduce

inappropriate following distance.

 Required action by drivers: smooth acceleration, smooth braking, adhere to

speed limit and keeping the safe headway in different traffic context.

 Real-time advice: promoting anticipatory driving by providing real-time

advice to driver before they approach the traffic situation using V2X

technology.

180 Chapter 7: A Simulator Study of the Effect of Incentives on Adoption and Effectiveness of an In-Vehicle
Driving HMI (Paper 5)
 Real-time performance feedback: provide feedback to driver regarding their

success or failure in achieving the goal.

7.5.4 Procedure
The study was conducted in accordance with the Australian Code for the

Responsible Conduct of Research (QUT Ethics approval number 1600000651).

Participants gave written consent to participate and were informed that their

participation was entirely voluntary and that data would remain anonymous and

reported on in an aggregate manner only.

The current study involved three phases of data collection (Figure 7.3): (i) a pre-

drive questionnaire; (ii) a series of simulated driving scenarios and immediate post-

drive questionnaires; and, (iii) a final post-experiment questionnaire.

Figure 7.3. Experimental drives and procedures

In phase one the questionnaire was emailed to participants using the Qualtrics

survey software. In phase two, participants completed a practice drive to familiarise

themselves with driving simulator, followed by a total of five experimental driving

conditions with immediate post-drive questionnaires. Before each driving condition,

Chapter 7: A Simulator Study of the Effect of Incentives on Adoption and Effectiveness of an In-Vehicle Driving
HMI (Paper 5) 181
the researcher explained the in-vehicle eco-safe driving system using the InVision user

interface prototyping platform on an iPad (see Appendix G).

Prior to the final drive, participants were shown a mock leader board and advised

that they were currently ranked seventh out of all participants who had taken part in

the experiment (see Figure 7.4) and asked to self-select from the three systems they

had experienced (i.e., advice only, feedback only, advice/feedback) for the last drive.

Participants were then informed that if they successfully improved their eco-safe

driving performance and reached the top rank on the leader board they would receive

an additional $10 reward. Participants were advised to self-select their preferred

system type for the final drive. All drivers received the additional incentive, regardless

of their performance in the final condition for ethical reasons. This methodology was

hypothesised to encourage participants to respond naturally to driving under the

influence of a monetary incentive combined with competition with other drivers.

Figure 7.4 The leader board screen presented to participants on iPad

In phase three, following the five conditions, participants filled out a final post-

drive questionnaire. Despite the claim made in phase two, all participants received a

$70 gift voucher (including the $10 bonus) as reimbursement for their two hours

182 Chapter 7: A Simulator Study of the Effect of Incentives on Adoption and Effectiveness of an In-Vehicle
Driving HMI (Paper 5)
participation in the driving simulator experiment, which was a requirement of the

ethics approval for this study.

7.5.5 Driving Behaviors Measures


Previous research has identified a number of primary indicators for eco-safe

driving behaviour, including deceleration, acceleration, speed, headway (Ericsson,

2001). In the current study, various traffic events where simulated to elicit these

behaviours from drivers. Fuel consumption and trip time were measured as a further

indicator of eco-driving behaviour.

Fuel consumption (mL/s) and total travel time (in seconds) were estimated for

the entire drive by the driving simulator software. A total of four traffic events required

drivers to decelerate and bring the vehicle to a complete stop, including on approach

to traffic signals in various phases (e.g., amber, red), and STOP sign. Following these

five events, drivers were then required to accelerate and bring the vehicle up to speed.

Deceleration and acceleration were both measured in metres per second squared

(m/s2). In addition, braking was also measured as applied force on the brake pedal (in

Newtons, N), while acceleration was measured using pedal position (a float value

between 0 and 1).

With regards to safe driving variables, headway and speed were measured. A

single headway event required drivers to follow behind another vehicle travelling at

the speed limit (40 km/h). The presence of other traffic prevented drivers from being

able to overtake the vehicles in front of them. Headway was measured as time interval

to vehicle (TIV) and was calculated using the following equation, where Di is the

relative distance to the vehicle in front in meters and V is the actual speed of the

simulator vehicle (this parameter is often used in traffic safety regulations and recent

Australian law mandates a minimum TIV of 2.5 secs): 𝑇𝐼𝑉 =Di/V [1]

Chapter 7: A Simulator Study of the Effect of Incentives on Adoption and Effectiveness of an In-Vehicle Driving
HMI (Paper 5) 183
Finally, speed was monitored during the whole drive in both the 40 km/h and 60

km/h zones. Speeding behaviour was measured separately for each of the zones using

mean speeds when travelling in free-flowing traffic (in km/h) and the percentage of

time spent exceeding the speed limit.

7.5.6 Subjective Measures


In addition to the objective driving behaviour data collected from the driving

simulator, subjective data was also collected through the three questionnaires. This

included a range of attitude and perception scales, and self-reported behavioural

variables. A brief summary of the scales used and internal validity of scales using

Cronbach’s α are presented in Table 7.1. Cronbach’s α coefficients ranged from .74 to

.90, suggesting moderate to strong internal reliability on these scales.

Table 7.1. Summary of subjective scales and internal reliability

No.
Factors Scale α
items
Pre-trip questionnaire
Attitudes toward technologya 6 1 Strongly disagree - 5 Strongly agree .74
Experience with technology 7 1 No experience - 5 Expert .79
Attitudes toward environmentb 15 1 Strongly disagree - 5 Strongly agree .84
Attitudes toward eco-driving 10 1 Strongly disagree - 5 Strongly agree .84
Perceptions of safe driving 9 1 Strongly disagree - 5 Strongly agree .78
Self-reported safe driving (DBQ)c 13 1 Never - 6 All the time .70
Immediate post-trip questionnaire
Intention to use the system 3 1 Strongly disagree - 5 Strongly agree .83-.90
Post-trip questionnaire
Perceived trust in the system 10 1 Strongly disagree - 5 Strongly agree .88
a
Adopted from (Weyer, Fink, & Adelt, 2015).
b
Adopted from the Environmental Attitudes Inventory (Milfont & Duckitt, 2010).
c
Driving Behaviour Questionnaire Adopted from af Wåhlberg, Dorn, and Kline (2011).

7.5.7 Data Analysis


Data collected from the driving simulator was processed and extracted using

Matlab version R2017a. To assess whether there were differences between drivers who

chose each type of system based on key demographic, attitudinal or self-reported

behaviour variables, a series of one-way between-groups ANOVAs were conducted,

with a Bonferroni correction for multiple comparisons. In instances involving

184 Chapter 7: A Simulator Study of the Effect of Incentives on Adoption and Effectiveness of an In-Vehicle
Driving HMI (Paper 5)
categorical independent variables, Fisher’s Exact test was conducted. To assess the

impact of incentives, competition on eco-safe driving behaviour, a series of paired-

samples t-tests were performed, again with a Bonferroni adjustment. Where data

violated the assumption of parametric tests, Wilcoxon signed ranks tests were

performed.

Effect sizes were calculated to provide an indication of the magnitude of the

effects observed, and were interpreted using Cohen (1992) criteria. For the one-way

between-groups ANOVAs and paired samples t-tests, eta squared (η2) was calculated

(.01 = small effect, .06 = moderate effect; .14 = large effect), while for the Fisher’s

Exact test, effect size was calculated using Cramer’s V (.1 = small effect, .3 = moderate

effect; .5 = large effect). Finally, for the Wilcoxon signed rank test, effect size was

calculated using r (.1 = small effect, .3 = moderate effect; .5 = large effect).

7.6 RESULTS

7.6.1 System Choice and Intention to Use


As stated, in the self-select condition, participants chose which of the systems

they had previously experienced (i.e., advice only, feedback only or advice feedback)

that they would prefer to use again in the final drive. A total of 15 participants (37.5%)

chose the advice only system, while 11 (27.5%) chose the feedback only system, and

14 (35%) chose the advice/feedback system (see Table 7.2). There was no significant

difference in the number of participants who chose each condition, Q (2, n = 40) =

.650, p = .723.

Table 7.2. Intentions to use by system choice

System type n (%) M SD


Advice only 15 (37.5%) 9.80 3.23
Feedback only 11 (27.5%) 8.18 2.96
Advice/feedback 14 (35.0%) 10.79 2.42

Chapter 7: A Simulator Study of the Effect of Incentives on Adoption and Effectiveness of an In-Vehicle Driving
HMI (Paper 5) 185
Intention to use was measured using three items, with the resulting scale having

a range from 3 (representing low intentions to use) to 15 (representing high intentions

to use). Intentions to use was measured for each condition. The means and standard

deviations for intentions to use by system choice are shown in Table 7.2. There was

no significant main effect of condition on self-reported intention to use, F (2, 37) =

2.50, p = .096, ηp2 = .12. Specifically, an inspection of the means suggests that drivers

who chose each system reported only moderate intentions to actually use the system.

7.6.2 Factors Associated with System Choice


Table 7.3 shows data for system choice by age and gender. Driver age was

categorised into two groups, to represent younger drivers (18-24 years) and other

drivers (25+ years), with a Fisher’s Exact Test revealing a marginally significant age

difference associated with system choice (p = .049, Cramer’s V = .39). The advice only

system was the most popular choice among drivers aged 25 years and older, being

chosen by half of all the drivers in this age group, and the least popular choice among

younger drivers at just 14.3%. Conversely, the advice/feedback system was the most

popular choice among younger drivers, being chosen by 57.1% of this age group,

compared to 23.1% of drivers aged 25 years and above. This data reveals that younger

drivers are more susceptible to feedback messages, with 85.7% of younger drivers

choosing a system involving such messages, compared to just half of the drivers aged

25 years and over. No significant gender differences were observed, (p = .852,

Cramer’s V = .11).

Table 7.3. System choice by age and gender

Age Gender
System type n 18-24 25+ Male Female
Advice only 15 2 (14.3%) 13 (50%) 8 (40%) 7 (35%)
Feedback only 11 4 (28.6%) 7 (26.9%) 6 (30%) 5 (25%)
Advice/feedback 14 8 (57.1%) 6 (23.1%) 6 (30%) 8 (40%)
Total 40 14 26 20 20

186 Chapter 7: A Simulator Study of the Effect of Incentives on Adoption and Effectiveness of an In-Vehicle
Driving HMI (Paper 5)
Group differences on a range of attitudinal, perception and self-reported

behavioural variables were also explored. The means and standard deviations, as well

as the test results, are outlined in Table 7.4. For all scales, with the exception of the

self-reported safe driving behaviour scale (DBQ), a higher score is representative of

more positive attitudes or perceptions, or more previous experience with technology.

Only one significant between-group difference was observed, for self-reported safe

driving, F (2, 37) = 4.89, p = .013, η2 = .209. Pairwise comparisons revealed that

drivers who chose the advice only system reported significantly safer driving

behaviour compared to drivers who chose the feedback only system. Both this finding,

and the marginally significant difference for age, becomes non-significant once

adjusting for the Bonferroni correction (α = .006).

Table 7.4. System choice by attitudes and perceptions

Advice Feedback Advice/


Variable (range)
only only feedback
Self-reported safe driving (13-65) 22.20 (5.74) 28.27 (4.76) 24.79 (3.91)
Attitudes toward eco-driving (10-50) 37.07 (5.20) 35.36 (7.72) 38.21 (5.51)
Perceptions of safe driving (9-45) 34.47 (6.16) 30.55 (4.25) 34.07 (3.75)
Attitudes toward environment (15-75) 62.40 (6.83) 57.36 (8.05) 61.43 (7.64)
Attitudes toward technology (6-30) 20.27 (3.71) 21.45 (4.61) 20.36 (2.62)
Experience with technology (7-35) 21.60 (5.59) 23.36 (3.80) 20.00 (4.39)
Perceived trust in the system (10-50) 33.40 (7.00) 34.18 (6.16) 35.64 (5.24)
Advice only (n = 15), feedback only (n = 11), advice feedback (n = 14).

There were no significant differences between drivers who chose each type of

system based on attitudes toward eco-driving, F(2, 37) = 0.68, p = .514, η2 = .035,

perceptions of safe driving, F(2, 37) = 2.32, p = .113, η2 = .111, attitudes toward the

environment, F(2, 37) = 1.55, p = .225, η2 = .078, attitudes toward technology, F(2,

37) = 0.39, p = .678, η2 = .021, experience with technology, F(2, 37) = 1.55, p = .225,

η2 = .077, or perceived trust, F(2, 37) = 0.48, p = .621, η2 = .025.

Chapter 7: A Simulator Study of the Effect of Incentives on Adoption and Effectiveness of an In-Vehicle Driving
HMI (Paper 5) 187
7.6.3 Impact of Incentives and Motivations on Intentions to Use
Participants were asked to report the extent to which a range of incentives

(including both intrinsic and extrinsic rewards) and motivations would increase their

likelihood of using the system. Table 7.5 displays participant responses to the items,

which were measured on a 5-point Likert scale (1 = strongly disagree to 5 = strongly

agree). Extrinsic rewards, in the form of vehicle-related savings such as saving money

on fuel and vehicle maintenance as a result of adopting an eco-driving style (item 2),

or prizes for appropriate behaviour (item 1), were reported as being very likely to

increase intentions to use the system. With regards to prizes, these included both those

related to the vehicle, such as reduced insurance premiums or fuel vouchers, as well

as other prizes, such as cash or movie vouchers.

Table 7.5. Incentives to motivate system use

M SD
Extrinsic rewards
1. Prizes associated with “good” driving (e.g., reduced insurance, fuel vouchers) 4.02 1.00
2. Saving money on fuel and vehicle maintenance 4.20 0.91
Intrinsic rewards
3. Reducing CO2 emissions and helping the environment 3.97 0.92
4. Improving the safety of myself, my passengers and other drivers 4.13 0.88
5. Competing with friends/family through leader boards 3.30 1.20
6. Competing with other drivers through leader boards 3.03 1.14
7. Competing with myself using score boards to improve my driving 3.63 1.15
8. Driving tasks and challenges to progress through (e.g., like a game) 3.38 1.17
n = 40.

Intrinsic rewards were also reported as being highly likely to increase intentions

to use the system. These included reducing the impact of the vehicle on the

environment (item 3) and improving road safety (item 4). More neutral responses were

observed in relation to the impact of motivating factors on intentions to use, such as

competition with either friends and family (item 5) or other drivers in general (item 6),

however competing with oneself (item 7) received slightly greater support. There were

also neutral responses regarding the impact of gamification principles, such as

completing tasks and challenges (item 8), on intentions to use the system.

188 Chapter 7: A Simulator Study of the Effect of Incentives on Adoption and Effectiveness of an In-Vehicle
Driving HMI (Paper 5)
7.6.4 Impact of Incentives and Competition on Eco-Safe Driving Behaviours
As stated in Section 7.5.4, in the final incentive condition of the experiment,

participants were shown a mock leader board that showed their position as seventh

based on scores derived from their eco-safe driving performance. Participants were

subsequently offered an additional cash incentive if they were able to progress to the

top of the leader board by the end of the final drive. Thus, this condition was designed

to measure the impact of both incentives (in the form of a cash prize) and competition

with other drivers on eco-safe driving performance, utilising gamification principles.

Having determined that there were no significant differences between

participants who choose each of the three system types in the final drive, data for these

groups were combined to form the incentives condition. Data for the ‘no-incentive’

condition was collated using the paired data for each participant from the

corresponding condition associated to the system they chose in the incentive condition

(e.g., those who chose the advice only system had their data paired to their advice only

condition drive).

As can be seen in Table 7.6, there were minimal significant effects associated

with the addition of incentives and competition. Indeed, the significant effects

observed were for fuel consumption, Z = -2.69, p = .007, r = .30, standard deviation in

accelerator pedal position in the traffic event, Z = -3.16, p = .002, r = .35, mean speed

in the 60 km/h zone, Z = -2.21, p = .027, r = .25, maximum speed in the 60 km/h zone,

Z = -2.19, p = .028, r = .24, and mean TIV, t = -2.26, p = .029, η2 = .116. In each

instance, behaviour improved in the incentives condition compared to the no-incentive

condition, and effects were of a moderate magnitude. All of these findings, with the

exception of standard deviation in accelerator pedal position in the traffic event,

become non-significant once adjusted for the Bonferroni correction (α = .002).

Chapter 7: A Simulator Study of the Effect of Incentives on Adoption and Effectiveness of an In-Vehicle Driving
HMI (Paper 5) 189
Table 7.6. Means, standard deviations and results of statistical analyses for impact of incentives on eco-driving behaviour
Incentive No incentive Test Significance Effect
Variable n M (SD) n M (SD) statistic (p value) size
Trip time (secs) 40 536.72 (80.99) 40 536.96 (65.16) t = -0.03 .980 η2 = .000
Fuel consumption (mL/sec) 40 327.94 (49.82) 40 344.17 (44.42) Z = -2.69** .007 r = .30
Scenario with traffic
Accelerator pedal position (0-1) 40 .13 (.06) 40 .13 (.06) Z = 0.34 .737 r = .04
SD Accelerator pedal position 40 .05 (.03) 40 .07 (.05) Z = -3.16** .002 r = .35
Max. Accelerator pedal position 40 .21 (.08) 40 .23 (.12) Z = -1.52 .130 r = .17
Scenario without traffic
Accelerator pedal position (0-1) 40 .16 (.08) 40 .17 (.06) t = -0.77 .444 η2 = .005
SD Accelerator pedal position 40 .05 (.04) 40 .05 (.04) Z = 0.47 .638 r = .05
Max. Accelerator pedal position 40 .23 (.11) 40 .23 (.08) Z = -1.09 .276 r = .12
Short-notice braking event
Brake use (N) 40 22.11 (14.17) 40 20.32 (14.86) Z = 1.06 .288 r = .12
SD Brake use (N) 40 24.72 (19.83) 40 23.52 (18.64) Z = 0.32 .747 r = .04
Max. Brake force (N) 40 71.47 (52.93) 40 74.38 (43.18) Z = -1.32 .188 r = .15
Brake use area under curve 40 566.35 (420.86) 40 499.29 (400.56) Z = 1.26 .206 r = .14
Long-range braking event
Brake use (N) 40 21.24 (12.78) 40 19.13 (9.92) Z = 0.48 .628 r = .05
SD Brake use (N) 40 24.99 (17.66) 40 21.73 (10.27) Z = 0.51 .610 r = .06
Max. Brake force (N) 40 78.15 (60.74) 40 67.50 (32.68) Z = 0.29 .769 r = .03
Brake use area under curve 40 543.47 (325.15) 40 496.96 (258.00) Z = 0.73 .468 r = .08
40 km/h Speed zone
Mean speed (km/h) 40 37.33 (1.96) 40 37.49 (1.97) t = -0.43 .669 η2 = .011
% Time speeding 37 7.50 (7.38) 38 6.89 (5.08) Z = 0.50 .615 r = .06
Max. speed (km/h) 40 51.60 (8.56) 40 53.37 (9.40) t = -0.96 .345 η2 = .003
60 km/h Speed zone
Mean speed (km/h) 39 56.21 (2.69) 40 57.04 (2.33) Z = -2.21* .027 r = .25
% Time speeding 24 16.74 (15.78) 31 16.03 (12.07) Z = 0.67 .506 r = .09
Max. speed (km/h) 40 60.94 (5.02) 40 63.01 (5.17) Z = -2.19* .028 r = .24
Vehicle headway
Time interval to vehicle (secs) 40 6.65 (1.87) 40 7.19 (1.59) t = -2.26* .029 η2 = .116
Z: Wilcoxon signed rank test; t: paired-samples t-test.
* p < .05 **; p < .01

190Chapter 7: A Simulator Study of the Effect of Incentives on Adoption and Effectiveness of an In-Vehicle Driving HMI (Paper 5)
7.6.5 Willingness to Purchase the System
Given that the likely implementation of a cooperative eco-safe in-vehicle HMI

such as the one examined in the current study is likely to be voluntary, it is important

to determine how much drivers might be willing to pay for such a system. As can be

seen in Table 7.7, roughly 60% of participants reported some willingness to pay for

the system. Among those willing to pay for the system, a price tag between $50 and

$100 was reported as being most reasonable for each type of system, with slightly

more drivers willing to pay more for the advice only system. That said, approximately

40% of drivers reported that they would not be willing to pay for the system in its

current form. When willingness to pay was recoded as a dichotomous (i.e., yes/no)

variable, there was no significant between-group differences in willingness to pay

dependant on system type, Q (2, n = 40) = .296, p = .867.

Table 7.7. Amount willing to pay for system

Advice only Feedback only Advice feedback


N % N % N %
Not willing to pay 15 37.5 16 40.0 15 37.5
Less than $50 9 22.5 9 22.5 7 17.5
$51-100 9 22.5 12 30.0 14 35.0
$101-200 7 17.5 3 7.5 3 7.5
More than $200 0 0 0 0 1 2.5

A number of additional items were also used to explore willingness to purchase

(see Table 7.8). As can be seen, there were relatively neutral perceptions regarding the

likelihood that having the system in-built in to new or used vehicles would increase

the willingness to purchase the vehicle, regardless of the system. Moreover,

participants reported being relatively unlikely to purchase any of the systems if they

perceived them as being unlikely to result in greater savings in fuel and vehicle costs

than the purchase price.

Chapter 7: A Simulator Study of the Effect of Incentives on Adoption and Effectiveness of an In-Vehicle Driving
HMI (Paper 5) 191
Table 7.8. Factors effecting willingness to purchase

Advice only Feedback only Advice/feedback


M SD M SD M SD
I would be more likely to buy a new/use
3.15 1.12 3.05 1.13 3.20 1.14
vehicle if system was in-built into it
I would be willing to pay more for the
system than I would save in fuel and 2.45 1.06 2.33 1.02 2.45 1.09
vehicle costs

Finally, the influence of the inclusion of additional technologies into the design

and development of the system on willingness to purchase was also explored. As can

be seen in Table 7.9, the inclusion of GPS/navigation systems, dash cams and

communication capabilities were all reported as being likely to somewhat increase

willingness to purchase, with navigation features perceived as a particularly appealing

additional feature.

Table 7.9. Impact of additional features on willingness to purchase

Additional technology M SD
GPS/navigation 4.13 .94
Dashcam 3.57 .87
Bluetooth/communication 3.65 .96
Items measured on 5-point Likert scale (1 = strongly disagree to 5 = strongly agree).

7.7 DISCUSSION AND CONCLUSION

The aim of the present paper was to explore the impact of incentives and

competition associated with the use of an eco-safe driving in-vehicle HMI on

subsequent eco-safe driving behaviour. In addition, the influence of advice and

feedback messages from the HMI on system preference and factors associated with

willingness to purchase such a system were also investigated.

A range of demographic, attitudinal and self-reported behaviour variables were

examined as possible factors to explain why drivers chose a particular system. Overall,

there was limited evidence of significant differences between drivers who chose each

of the three systems. Indeed, after adjusting the alpha level to account for family-wise

error rate, there were no significant findings. Only age and self-reported safe driving

192 Chapter 7: A Simulator Study of the Effect of Incentives on Adoption and Effectiveness of an In-Vehicle
Driving HMI (Paper 5)
reached statistical significance, with older drivers who perceived their driving to be

safer significantly more likely to choose the advice only system. Given that the

Bonferroni adjustment may lead to conservative interpretations of the results due to

increased Type II error (Field, 2013), it is still worth discussing the potential

implications of these results. That is, there is some evidence to suggest that individuals

with greater driving experience and who perceive themselves to be better drivers may

prefer advice messages, or alternatively dislike receiving feedback messages. One

possible explanation for this preference may be that they do not perceive negative

feedback as accurately reflecting their perceived driving performance and ability

which may results in cognitive dissonance (Festinger, 1962) and led to disliking the

system.

Overall, drivers reported moderate intentions to actually use the in-vehicle eco-

safe system in the long term. This finding is generally consistent with previous

research which has highlighted the difficulty in achieving widespread voluntary use of

in-vehicle technology, such as the system examined in the current study (Harvey et al.,

2013). Nonetheless, drivers still reported a relatively strong willingness to pay for such

a system, with 60% of participants suggesting they would pay up to $100. That said,

approximately 40% of drivers reported that they would not be willing to pay for the

system in its current form, suggesting that the inclusion of additional technologies into

the design and development of the system, such as GPS/navigation systems, dash cams

and communication capabilities, all represent potential directions for future research.

Overall, incentives and competition, as explored in the current study, appear to

have had minimal impact on further improving driving behaviour associated with the

use of the cooperative eco-safe in-vehicle HMI. While this appears to raise concerns

regarding the cost-effectiveness of such an approach, there are a number of factors to

Chapter 7: A Simulator Study of the Effect of Incentives on Adoption and Effectiveness of an In-Vehicle Driving
HMI (Paper 5) 193
consider. For example, vehicle-related incentives, such as reduced insurance

premiums, vehicle registration costs or fuel vouchers, may be a more effective

incentive for encouraging improved behaviour and system adherence (Lai, 2015;

Schall & Mohnen, 2017; Tselentis et al., 2016). As such, the small amount of cash

offered may not have been a sufficient or appropriate incentive to encourage behaviour

change. Future research should investigate the impact of a variety of other incentive

types, and in particular vehicle-related incentives.

Secondly, the use of incentives may be more effective in encouraging greater

intentions to use or actual rates of use, compared to improving driving behaviour

associated with system use. Indeed, the impact of incentives was positive on intentions

to use and rates of use (Chorlton et al., 2012; Michael Regan et al., 2002; Vaezipour

et al., 2017). This study was only able to subjectively explore the potential impact of

incentives on intentions to use. The findings suggest strong support for both extrinsic

(e.g., savings in fuel and vehicle maintenance costs, prizes) and intrinsic incentives

(e.g., environmental benefits, improved road safety). Future research should seek to

extend on the current findings by examining the impact of incentives on actual rates

of use. Indeed, given the difficulties that have been demonstrated with regards to

widespread system implementation and voluntary use (Harvey et al., 2013), an

argument could be made the greatest benefits of incentives may in fact be associated

with increasing system usage rates, as opposed to additional behavioural outcomes.

Finally, drivers in this sample displayed generally exemplary eco-safe driving

behaviour in the non-incentive condition. As such, the degree to which behavioural

improvement was possible in the incentive condition may have been somewhat

diminished. In addition, drivers in the current study also generally had strong positive

attitudes toward the environment and eco-driving, positive perceptions of safe-driving,

194 Chapter 7: A Simulator Study of the Effect of Incentives on Adoption and Effectiveness of an In-Vehicle
Driving HMI (Paper 5)
as well as relatively high levels of previous experience with technology. Future

research should investigate whether the use of incentives can improve intentions to use

and subsequent behaviour of more problematic drivers, in particular those who display

worse eco-safe driving performance and less positive attitudes toward eco-driving and

the environment.

There was also limited evidence supporting the impact of competition on either

intention to use the system or eco-safe driving behaviour. Neutral responses were

expressed regarding the incentivising nature of competition with either friends or

family or other drivers in general, or the inclusion of gamification components such as

tasks and challenges. Such a finding is consistent with prior research which suggests

typically neutral attitudes toward competition as an incentive (Loumidi et al., 2011).

That said, drivers appeared to be slightly more incentivised by competition with

themselves. Taken together with previous findings from a qualitative study (Vaezipour

et al., 2016), it appears that gamification features such as leader boards, are optimally

included as an optional component of the system, rather than a mandatory feature.

A number of limitations must be addressed with regards to the current study.

Firstly, given that the incentives condition was always the last drive, there is the

possibility that fatigue and learning effects influenced the results. In addition, the cash

amount may not have been large enough to be seen as worth the effort, or may not

have been representative of the type of incentive required to encourage behaviour

change among the drivers. In addition, it is possible that what is perceived as the most

appropriate incentive differs between individuals. Moreover, as with all research

conducted in a simulated driving environment, there are concerns regarding the real-

world generalisability of the findings. Specifically, participants reported neutral

perceptions regarding the realistic nature of the simulated driving experience.

Chapter 7: A Simulator Study of the Effect of Incentives on Adoption and Effectiveness of an In-Vehicle Driving
HMI (Paper 5) 195
196 Chapter 7: A Simulator Study of the Effect of Incentives on Adoption and Effectiveness of an In-Vehicle
Driving HMI (Paper 5)
Chapter 8: Discussion and Conclusion

8.1 INTRODUCTORY COMMENTS

The overall objective of the current research was to design, develop and evaluate

a cooperative eco-safe in-vehicle Human Machine Interface (HMI) to reduce fuel

consumption and promote safe driving behaviour. Specifically, the in-vehicle HMI

proposed in this research was designed to minimise adverse impacts on driver

workload given that the high workload could distract drivers. Moreover, this research

aimed to identify in-vehicle HMI characteristics associated with greater levels of

driver acceptance in order to inform the design and development of such systems and

encourage drivers’ intentions to use them. Finally, this research aimed to examine the

impact of incentives and competition on eco-safe driving behaviour and intentions to

use the system.

Three studies were conducted using a mixed-methods design consisting of both

qualitative and quantitative approaches. In Study 1, a critical review of the literature

was conducted to gain a comprehensive understanding of the relationship between eco-

driving and safety and to highlight gaps in the research. In addition, the review sought

to identify best-practice principles associated with the design and development of in-

vehicle HMIs, and in particular those aimed at impacting eco-safe driving.

Study 2 used a dual-phase qualitative approach to explore novel ways to provide

real-time eco-safe driving advice and feedback to drivers via an in-vehicle HMI. This

included a series of focus groups with drivers, as well as an expert workshop. This

study was guided by the user-centred design approach and the TAM, focusing on

driver acceptance. Information gathered from this iterative methodology informed the

Chapter 8: Discussion and Conclusion 197


development of a prototype cooperative eco-safe in-vehicle HMI to be evaluated in the

next phase of the research.

Finally, Study 3 involved a quantitative evaluation of the in-vehicle HMI,

including a driving simulator experiment and a driver survey. The study evaluated

three prototypes of an eco-safe in-vehicle HMI, varied in the type of messages

provided to drivers through the system: (i) visual advice only; (ii) visual feedback only;

and, (iii) visual advice and feedback. In the simulator experiment, objective measures

were analysed to assess the impact of the system on fuel consumption and eco-safe

driving behaviours, including acceleration, deceleration, speed and headway.

Conversely, in the driver survey, several subjective measures assessed driver

acceptance, workload, self-reported behaviours, attitudes and perceptions, and

demographics. In addition, Study 3 examined the influence of incentives and

competition on eco-safe driving behaviour and intentions to use the system.

This chapter discusses the overall results from the current program of research

in relation to the research questions outlined in Section 1.5. The implications of the

research for the environment, road safety, and human machine interaction are

discussed, and the strength and limitations of the research are outlined. Finally, the

thesis concludes with a number of recommendations for future research.

8.2 DISCUSSION OF FINDINGS

Research Question 1: What is the relevant information that should be conveyed to


drivers in an in-vehicle HMI to improve eco-safe driving?

The critical review (Study 1) highlighted a range of driving behaviours and

behavioural outcomes that can be supported by in-vehicle systems. These include

driving speed, smooth deceleration and acceleration, and vehicle headway (Barth &

Boriboonsomsin, 2009; Birrell & Young, 2011; Boriboonsomsin, Vu, & Barth, 2010;

198 Chapter 8: Discussion and Conclusion


Staubach et al., 2014). Given this evidence, it appears that such information would be

useful for inclusion and monitoring in an eco-safe in-vehicle HMI to reduce fuel

consumption and improve safe driving.

The review of the research also revealed that, to date, most practitioners have

focused on either eco-driving or safe driving behaviours in isolation when developing

in-vehicle systems, with a paucity of systems addressing eco-safe driving. This

represents a significant gap in the research and suggests that an integrated system

addressing eco-safe driving behaviours, as developed and evaluated in the current

program of research, could have considerable positive impacts on reducing fuel

consumption and improving eco-safe driving behaviour and represent a more

parsimonious approach to improving this range of behaviours within the vehicle.

In addition, driver acceptance of technology has been identified as a vital factor

associated with in-vehicle systems that can significantly influence their ability to

achieve optimal benefits (Adell, 2010). Indeed, as stated by (Van Der Laan et al., 1997,

p. 1), “it is unproductive to invest effort in designing and building an intelligent co-

driver if the system is never switched on, or even disabled”. As such, this research

project sought not only to evaluate the effectiveness of the in-vehicle system on eco-

safe driving behaviours, but also explore ways to increase driver acceptance of the

technology. This was achieved by adopting a user-centred design approach, whereby

users were involved in each phase of the design process, including developing a

comprehensive understanding of their needs and requirements prior to initial system

development.

Specifically, the findings of the literature review (Study 1) helped to inform the

focus group study (Study 2a), which more directly examined the needs and

requirements of the users. Both of these studies attempted to address Research

Chapter 8: Discussion and Conclusion 199


Question 1, with the focus group study approaching the research question more

directly from the driver’s perspective. The focus group study revealed that drivers are

generally supportive of in-vehicle systems that increase their ability to monitor their

own eco-safe driving behaviour. This was particularly true in relation to behaviours

that are difficult for drivers to subjectively monitor themselves, and behaviours

perceived to have a considerable impact on eco-safe driving such as smooth

acceleration and deceleration, maintaining headway and driving speed as well as the

real-time monitoring of upcoming traffic information and the surrounding

environment. These parameters support earlier studies which suggest the provision of

real-time traffic information to drivers promotes greater anticipation of changes in

traffic by drivers and benefits on fuel efficiency and safety (Delhomme et al., 2013;

Popiv, 2012; Rommerskirchen, Helmbrecht, & Bengler, 2014).

This finding is not only encouraging, but is also consistent with previous studies

that have suggested drivers are generally open-minded and receptive to in-vehicle

systems aimed at monitoring driving behaviour (Horberry et al., 2014b; Michael

Regan et al., 2002). It also represents an important step in establishing driver

acceptance of such systems. In addition, the relatively consistent findings regarding

driver acceptance of various eco-driving and safe driving feedback systems also

provides evidence of the openness of drivers to behaviour monitoring using in-vehicle

systems (Adell, Várhelyi, & Hjälmdahl, 2008; Meschtscherjakov et al., 2009;

Roetting, Huang, McDevitt, & Melton, 2003; M Symmons, Rose, & Van Doorn,

2009).

Consistent with previous research, responses from the focus groups regarding

the monitoring of behaviour suggested evidence of an illusory superiority effect

(McKenna & Myers, 1997; Waylen et al., 2004). Specifically, drivers showed a

200 Chapter 8: Discussion and Conclusion


tendency to perceive themselves as safer, better drivers than others. This finding has

important implications for system development, such that although constructive

feedback from an in-vehicle HMI that challenges such biases may motivate behaviour

change (Horrey et al., 2012), it may also lead to cognitive dissonance and result in

users disengaging with the system (Festinger, 1962). Moreover, such biases may be

particularly likely among those drivers who require behaviour change the most. These

results provide further support for the notion that in-vehicle systems which personalise

driver feedback may be more effective in motivating behaviour change (Cialdini,

2001; Gilman et al., 2015; Magaña & Muñoz-Organero, 2016; Taubman-Ben-Ari &

Yehiel, 2012; Van Huysduynen et al., 2017).

The following research questions discuss the findings from a subjective and

objective evaluation of the three prototyped eco-safe in-vehicle HMI systems (i.e.,

advice only, feedback only, advice/feedback). This evaluation included an experiment

conducted in a state-of-the-art driving simulator combined with a driver survey, which

sought to evaluate the impact of the in-vehicle HMI on driver acceptance, driver

workload and eco-safe driving behavioural performance.

Research Question 2a: What impact do advice and/or feedback characteristics of an


in-vehicle eco-safe HMI have on driver acceptance of the technology?

Findings from the driving simulator experiment (Study 3) suggest that all three

prototyped visual eco-safe in-vehicle HMIs evaluated in this research (i.e., advice

only, feedback only, and advice/feedback) were associated with relatively high levels

of driver acceptance. That is, drivers reported that each system was usable, useful and

easy to use. Nonetheless, participants reported the greatest levels of driver acceptance

and satisfaction in association with the advice only condition, with drivers reporting

that feedback messages were more distracting and frustrating than advice messages.

Chapter 8: Discussion and Conclusion 201


Consistent with previous research, the research also suggested a number of

potential problems, such as privacy and trust concerns, can negatively influence driver

acceptance of an in-vehicle system (Ghazizadeh & Lee, 2014; Hickman & Hanowski,

2011; Kim & Malhotra, 2005). Specifically, drivers who perceive a system to be less

reliable and less precise are more likely to find the system less acceptable, and such

attitudes may be more likely among drivers who perceive feedback messages as not

accurately corresponding with their own perceptions of their driving performance. For

this reason, many practitioners recommend the adoption of the user-centred design

approach and an analysis of driver acceptance when developing a new system

(Horberry et al., 2014b; Maguire, 2001; Vilimek & Keinath, 2014). These findings are

discussed further in Section 8.3.

Taken together, these findings highlight the importance of carefully considering

the manner in which feedback is provided to ensure it is widely accepted by drivers.

A key component of achieving this goal is likely to be establishing high levels of

perceived accuracy of system feedback. The inclusion of both advisory messages that

notify drivers regarding what behaviours need to be performed, as well as behavioural

feedback messages, appeared to represent the most comprehensive and effective

provision of information to drivers. These findings are consistent with previous studies

highlighting the importance of performance feedback as a key motivator of behaviour

change (e.g. Gardner et al., 2010; Hattie & Timperley, 2007; Webb et al., 2010).

For this reason, it is recommended that the provision of more in-depth feedback

be explored to further increase the impact of feedback on eco-safe driving behaviour.

This may include providing information regarding why a particular behaviour is

important, as well as specific instructions about how to engage in behavioural

modifications to improve driving performance. However, it is recommended that such

202 Chapter 8: Discussion and Conclusion


in-depth and detailed information should only be considered in the back-end of a

system (e.g., to be accessed post-drive), in order to minimise the distractive impact of

system use on driver performance (Steinberger et al., 2016; Stevens et al., 2002).

From a user-centred design approach, the finding that all three system variations

were associated with relatively high levels of driver acceptance is encouraging. Driver

acceptance is a crucial antecedent for encouraging voluntary system use and for

widespread system implementation, however does not equate to system effectiveness.

Indeed, system effectiveness, in terms of positive impacts on behaviour and

behavioural outcomes, is ultimately more important from a practical perspective. That

is, driver acceptance and system effectiveness are two separate, but complementary,

requirements of a successful system. The findings regarding the effectiveness of the

system on behavioural outcomes are discussed in the following sections.

Research Question 2b: What impact do advice and/or feedback characteristics of an


in-vehicle eco-safe HMI have on driver workload?

Similar to the findings for driver acceptance, all three system variations were

observed as being associated with acceptable levels of driver workload, when

compared to previous research on workload during driving tasks (Grier, 2015). That

is, the advice and feedback configuration did not have a negative impact on reported

subjective driver workload to an extent that raised concerns about its effect on driver

safety. This is particularly important since distraction may occur due to the increasing

complexity and workload demand of the driving task when interacting with a system

(Oviedo-Trespalacios et al., 2016).

Nonetheless, the findings regarding the impact of advice and feedback messages

on driver workload were largely consistent with the hypotheses, as workload increased

contingent on the amount of information presented to drivers. Such a finding is also

Chapter 8: Discussion and Conclusion 203


consistent with previous research (e.g. Regan, Lee, & Young, 2008; Rouzikhah, King,

& Rakotonirainy, 2013; Wickens, 2002). Indeed, as noted previously, drivers reported

that they found feedback messages more distracting and difficult to understand,

compared to advice messages.

These findings highlight the potential to improve the feedback messages, with

the challenge appearing to be identifying how to provide enough information to ensure

comprehension of the message, while minimising the degree to which the message is

overwhelming and distracting (Shute, 2008). A possible approach may be to tailor the

level of advice and feedback provided to drivers dependent on their level of eco-safe

driving skills, as objectively measured by the system. Such a recommendation is

consistent with previous research in the area of computer-based instruction (Brouwer

et al., 2015; Shute, 2008; Stillwater & Kurani, 2013).

That said, future research should seek to evaluate the use of the system in more

complex traffic environments to ensure the system does not distract drivers and ensure

driver safety is maintained, particularly during hazardous situations on the road.

Research Question 2c: What impact do advice and/or feedback characteristics of an


in-vehicle eco-safe HMI have on eco-safe driving performance?

The overarching aim of the cooperative eco-safe in-vehicle HMI was to improve

eco-driving behaviour, while maintaining driver safety. It was hypothesised that

conditions involving feedback messages would have a greater impact on driving

behaviour, compared to advice messages, given that the feedback sought to notify

drivers of inappropriate behavioural performance and the need for improvement. This

is consistent with previous research highlighting the importance of constant reminders

and performance feedback for behaviour change (Gardner et al., 2010; Hattie &

Timperley, 2007; Pampel et al., 2015; Webb et al., 2010).

204 Chapter 8: Discussion and Conclusion


Overall, the findings largely supported this hypothesis, such that systems

involving feedback messages outperformed the advice only system. However, the

importance of advice messages cannot be downplayed, such that the advice/feedback

system consistently demonstrated the greatest positive impact on eco-safe driving

performance. This suggests that advice and feedback messages serve unique and

complimentary functions in improving eco-safe driving behaviour. Indeed, advice

messages may serve to enhance driver awareness of the need to adopt or maintain an

eco-safe driving style and promote anticipatory driving, which is the key factor in eco-

safe driving (Bengler, Drüke, Hoffmann, Manstetten, & Neukum, 2017; Orfila, Saint

Pierre, & Messias, 2015), while feedback messages may provide regular indications

of areas for behavioural improvement.

With regards to the advice/feedback system, improvements were observed

across the range of eco-safe driving behaviours. Specifically, when compared to

baseline, fuel consumption was reduced by 6.7%, acceleration by 14.3% to 33.3%, and

deceleration by 5.6% to 25.1%. Moreover, overall vehicle speeds were reduced by up

to 2%, with greater reductions observed for time spent speeding (up to 49.8%

reduction) and maximum speeds (up to 8.5% reduction). Finally, vehicle headway was

improved by 16.7%. These results are consistent with previous research, which has

typically reported reductions in fuel consumption ranging from 5% to 15%, with some

reporting reductions up to 30% depending on traffic conditions (see review see review

of 19 studies by Alam & McNabola, 2014; Barkenbus, 2010; McIlroy et al., 2017).

Moreover, there was evidence from the current study to suggest that drivers

particularly benefitted from receiving behavioural feedback messages in more

complex driving scenarios. Specifically, larger impacts on deceleration were observed

during short-notice braking scenarios, while greater impacts on acceleration behaviour

Chapter 8: Discussion and Conclusion 205


were observed during higher traffic scenarios. Such findings suggest that the positive

impact of in-vehicle HMIs may be greatest when the driving task demands a greater

workload from the driver such as limited visibility of surrounding traffic environment

(e.g. driver assistance at traffic light) (Preuk, Stemmler, & Jipp, 2016; Rittger, 2015;

Rittger & Götze, 2017).

Taken together with the findings regarding driver acceptance and workload, the

results of the current program of research are very encouraging. Specifically, this

research suggests that a cooperative eco-safe driving in-vehicle HMI involving both

advice and feedback messages has the potential to improve eco-safe driving, while also

being accepted by drivers and producing minimal adverse impacts on driving. While

the research has demonstrated that such an approach can be effective, previous

research has highlighted that system effectiveness does not guarantee that drivers will

voluntarily adopt the technology and express intentions to use the initiative (Horberry

et al., 2014b; Jamson, 2006; Jamson, Chorlton, & Carsten,2012). This presents an

ongoing challenge for such initiatives.

Research Question 3: What impact do incentives associated with using an in-vehicle


eco-safe HMI have on eco-safe driving performance?

Overall, incentives and competition, at least as explored in the current study,

appeared to have had minimal impact on further improving eco-safe driving behaviour

associated with system use. That said, the addition of incentives and competition did

result in a significant 4.7% further reduction in fuel consumption. With respect to the

addition of incentives associated with using in-vehicle HMIs, previous research has

found that a combination of real-time feedback and incentive have positive impact on

driving behaviour (Merrikhpour, Donmez, & Battista, 2014; Ian J. Reagan, Bliss,

Houten, & Hilton, 2013).

206 Chapter 8: Discussion and Conclusion


Several factors that may explain the lack of other significant findings. Firstly,

the small amount of money offered may not have been a sufficient or appropriate

incentive to encourage behaviour change. Indeed, prior research has suggested that

vehicle-related incentives, such as reduced insurance premiums, vehicle registration

costs or fuel vouchers, are typically the most effective types of incentives (Lai, 2015;

Schall & Mohnen, 2017; Tselentis et al., 2016). Secondly, the positive impacts on eco-

safe driving behaviour observed in the non-incentive condition, may have diminished

the possibility to observe significant further improvements in behaviour.

Finally, the use of incentives may be more effective in encouraging greater

intentions to use or actual rates of use, compared to improving driving behaviour

associated with system use (Chorlton et al., 2012; Michael Regan et al., 2002). The

current program of research revealed only moderate intentions among drivers to

actually use the eco-safe driving in-vehicle HMI in the long term. While disappointing,

this finding is generally consistent with previous research which has highlighted the

difficulty in achieving widespread voluntary use of in-vehicle technology (Harvey et

al., 2013; S. Jamson, 2006; S. Jamson et al., 2012). Nonetheless, there were some

encouraging findings, with two-thirds of surveyed drivers reporting a willingness to

pay for such a system, and increased intentions to use associated with the inclusion of

additional technologies to the system, such as navigation, dash-cams and

communication capabilities (Hipp, Bengler, Kressel, & Feit, 2018).

These findings suggest that incentives may be required to increase intentions to

use, with strong support for both extrinsic (e.g., savings in fuel and vehicle

maintenance costs, prizes) and intrinsic incentives (e.g., environmental benefits,

improved road safety). However, there was limited evidence supporting the impact of

competition and gamification, at least as studied in the current research, on either

Chapter 8: Discussion and Conclusion 207


intention to use or further behavioural improvements. Such a finding is consistent with

prior research which suggests typically neutral attitudes toward competition as a sole

incentive (Loumidi et al., 2011; Musicant, Lotan, & Grimberg, 2015; Toledo, Lotan,

Caird, Horrey, & Trick, 2016).

8.3 GENERAL RECOMMENDATIONS

As described in Section 1.4, the Technology Acceptance Model and user-centred

design approach were used as the theoretical framework for this program of research.

While not addressed in the discussion of findings related to the research questions, the

research methodologies of Study 2 in particular assisted with the identification of user

needs and requirements regarding specific design and device characteristics associated

with eco-safe in-vehicle HMIs. This information was subsequently used to inform the

design and development of the system evaluated in Study 3.

Device type and accessibility. The majority of drivers reported a preference for

one of three approaches: (i) an in-built system installed in new vehicles or retrofitted

to older vehicles; (ii) a stand-alone device (e.g., similar to dash cams and GPS

systems); or (iii) a smartphone application that can be linked to the vehicle. Overall,

preferences for each approach were inexplicably linked to perceptions of accessibility.

Specifically, those who preferred an in-built system or stand-alone device were

attracted to their ability to be fitted to the vehicle and to start automatically upon

ignition, with minimal hassle to the driver. Conversely, the comparatively higher costs

associated with purchase were noted as disadvantages, relative to a smartphone

application.

Stand-alone devices and smartphone applications were acknowledged as more

advantageous for individuals who drive more than one vehicle (e.g., both personal and

work vehicles) and require system portability. However, in relation to the smartphone

208 Chapter 8: Discussion and Conclusion


application, serious concerns were raised regarding the perceived irony of having an

intervention designed to promote safety that simultaneously and actively encouraged

the use of a mobile phone while driving. Thus, the stand-alone device, while not

without concerns regarding purchase costs, was deemed to be the most effective

approach when developing the prototype system evaluated in Study 3.

Automaticity and additional features. Regardless of the preferred approach, a

common preference among participants was for the system to operate with a high level

of automaticity. That is, participants almost unanimously expressed a desire for the

system to start automatically upon the commencement of a trip and be straightforward

to set up. A less automatic system was perceived as more likely to conflict with daily

life responsibilities and was associated with lower rates of prolonged usage. Moreover,

users reported higher intentions to use the system, due to increased perceptions of

system usefulness, when combined with additional features, such as navigation, dash-

cams, communication capabilities and other in-vehicle systems.

Customisability of feedback characteristics. The primary debate regarding

feedback characteristics centred on the most preferred presentation format of feedback

– namely auditory, visual or a mixture of both. While participants varied considerably

in their preferences, visual feedback, including the use of ambient light, was perceived

to have the least impact on driver distraction by allowing users to attend to the feedback

information in a largely peripheral manner. Nevertheless, the results clearly suggest

that a one-size-fits-all approach is inappropriate in relation to such a system, and thus

it is recommended that the system should provide drivers with the ability to customise

their feedback preferences. Such a recommendation is consistent with the International

Organisation for Standardisation (ISO/TR) and is supported by previous research (Fors

et al., 2015; Staubach et al., 2014; Steinberger, Schroeter, Foth, et al., 2017).

Chapter 8: Discussion and Conclusion 209


Gamification. Overall, participants held less than favourable attitudes regarding

the gamification of an eco-safe in-vehicle system, instead reporting a preference for

more informative and personalised feedback. However, the potential for such elements

to increase engagement with the driving task and facilitate social persuasion feedback

led many to argue that game elements should represent a ‘background feature’ of the

system, such that those users who wanted to gamify the feedback could do so. This

finding highlights that users desire a sense of autonomy and control over the system,

so that they can modify the system to best suit their specific individual needs.

8.4 RESEARCH IMPLICATIONS

The primary contribution of the current program of research is that it

demonstrates a cooperative eco-safe driving in-vehicle HMI involving both advice and

feedback messages has the potential to improve eco-safe driving, while also being

accepted by drivers and producing minimal adverse impacts on driver workload.

Moreover, the adoption of a user-centred approach has enhanced our

understanding of the relationship between eco-driving and safety from the human

factors viewpoint. A comprehensive insight has been gained into the user needs and

requirements associated with cooperative eco-safe in-vehicle HMIs. The present

research adds weights to the user-centred design approach and the importance of the

Technology Acceptance Model from the early stages of design and development

(Francois et al., 2017; Horberry et al., 2014b). This understanding provides a

foundation for the future design and development of innovative, persuasive and

acceptable in-vehicle HMI systems to improve fuel efficiency and road safety. For

example, the findings highlighted the barriers to driver acceptance and variation in

user preferences regarding a range of device characteristics, highlighting a

fundamental need to develop systems that afford the ability for users to customise and

210 Chapter 8: Discussion and Conclusion


personalise the feedback to suit their specific preferences (Brouwer et al., 2015;

Gilman et al., 2015; Magaña & Muñoz-Organero, 2016).

While somewhat mixed results were observed regarding intentions to voluntarily

use the system, there were positive findings regarding the potential for vehicle-related

incentives to increase intentions to engage with the system. Such incentives could

include reduced insurance premiums or vehicle registration costs. This highlights the

potential to partner with insurance companies and transport authorities to offer

discounts on insurance and/or vehicle registration associated with system use and

evidence of appropriate eco-safe driving styles. Such an approach is likely to have a

positive effect on use of the technology.

Moreover, this research suggests that the support provided to drivers by an eco-

safe in-vehicle HMI that provides real-time advice and feedback should be particularly

enticing for a range of specific populations. For example, widespread adoption of such

an initiative in commercial fleets has the potential to produce significant benefits for

companies, in terms of reducing their carbon footprint and reducing costs associated

with fuel consumption and traffic crashes. In addition, the approach would be

beneficial for use with novice drivers, to enhance and normalise an eco-safe driving

style from the early stage of their driving experience. The system also represents a

viable sanctioning option for traffic offenders. Such a finding has important

environmental and road safety implications, including the ability to produce reductions

in fuel consumption and emissions, as well as enhancing driver safety.

Taken together, the findings of this research suggest an important role of in-

vehicle HMIs for reducing fuel consumption (up to 6.7% reduction) which is

consistent with previous research that typically reported reductions in fuel

consumption ranging from 5% to 15% (Alam & McNabola, 2014; Ecodriver, 2011; M

Chapter 8: Discussion and Conclusion 211


Symmons et al., 2009). Improvement was also observed in regards to reduction in

overall time spent speeding (up to 49.8% reduction), which is a major factor in serious

and fatal traffic crashes (Australian Transport Council, 2014; World Health

Organization, 2015). Thus, the findings of this research further support an important

the value of eco-safe in-vehicle HMI as a targeted intervention for a range of public

health issues that can help save lives by combatting pollution and traffic crashes.

That said, comprehensive behavioural and attitudinal change will most likely be

achieved through a multi-faceted approach including both technological solutions, as

well as driver training, education and enforcement. Future research should therefore

seek to evaluate such approaches, including alternative HMI designs to those studied

in this program of research.

8.5 THEORETICAL IMPLICATIONS

Previous research into driver acceptance of in-vehicle technology reveals that

acceptance is a complex construct that is highly dependent on a driver’s attitudes,

expectations and their experiences using in-vehicle systems (Adell & Varhelyi, 2014).

In-vehicle systems are typically only accepted if they are valued by the driver and

satisfying to use. Research has suggested that feedback can only advise, and will not

necessarily stimulate, sustainable ecological behaviour, unless the driver already

possesses a strong inclination to drive in an eco-friendly manner and perceives the

system to be effective (Shipworth, 2000). Indeed, it is noted that acceptance levels

towards a system are fluid and may fluctuate over time depending on different contexts

or driver experiences with the system. Taken together, these findings suggest that a

driver-centric view is necessary to enhance understanding and measure driver

acceptance at an individual level (Horberry et al., 2014a).

212 Chapter 8: Discussion and Conclusion


Therefore, the contribution of this thesis is twofold. Firstly, the present program

of research adopted the TAM and user-centred design approach as the theoretical

framework. This approach further supports the applicability of the TAM in the context

of an eco-safe in-vehicle HMI and confirmed the complexity of factors that influence

driver acceptance. Moreover, this research identified additional determinants to the

TAM, including driver characteristics, design characteristics of in-vehicle systems,

motivations and normative influences on intentions to use an eco-safe in-vehicle HMI.

A number of barriers to driver acceptability of eco-safe in-vehicle systems emerged

from this research, including perceived trust in the system, privacy concerns, novelty

and distraction, all of which require distinctive consideration for inclusion in the TAM.

Secondly, this combination of findings provides a theory based methodological

tool to support designers of products for incorporating new technologies in vehicles,

as well as for decision makers to better understand contextual factors that may limit

the acceptance of strategies to promote eco-safe driving. Future research should seek

to evaluate the incorporation of these determinants into the TAM, specifically in the

context of an eco-safe in-vehicle HMI. In light of emerging technologies, these

findings require further validation in order to investigate how driver acceptance of

technology improves using longitudinal methodologies.

8.6 STRENGTHS AND LIMITATIONS

This research adds to the existing body of research as it is, to the authors’ best

knowledge, the first user-centred study of an in-vehicle HMI system designed to

address both eco-driving and safe driving internationally. Furthermore, the research

methodology is grounded in the robust principles of the Technology Acceptance

Model and user-centred design approach. The research employed a mixed-method

design approach, drawing on the strengths of in-depth, subjective approaches such as

Chapter 8: Discussion and Conclusion 213


focus groups and workshops, as well as a range of objective approaches including a

driving simulator study and driver survey.

The state-of-the art driving simulator used in this program of research is a high-

fidelity simulator with a motion platform and six degrees of freedom that can move

and twist in three dimensions. This simulator has been successful in developing traffic

flow models that performed better than previous models in explaining real world

driving data (Saifuzzaman et al., 2015).

While the findings of this study are undoubtedly encouraging, a number of

limitations must be acknowledged. With respect to the focus group study, a

convenience sampling approach was adopted. The resulting sample was relatively

small in size and consisted of predominantly younger, urban drivers. This may impact

the generalisability of the results for the focus group. That said, given these

characteristics, the sample is likely to be more experienced with technology and as

such represent the target audience of a system like the one explored in this research.

Nonetheless, future research should address this limitation by conducting a user study

with a larger, more representative sample.

As with all research conducted in a simulated driving environment, there are

concerns regarding the real-world generalisability of the findings. Overall, participants

had relatively neutral perceptions regarding the realistic nature of the simulated driving

experience, suggesting some evidence of the generalisability of the findings.

Moreover, prior research has demonstrated generalisability of findings from a

simulated driving environment to the real-world content (Charlton, 2009; Krause et

al., 2014; Santos et al., 2005).

The short duration of the driving experiment did not allow for an assessment of

the long-term uptake of the system, nor the effects associated with driver habituation.

214 Chapter 8: Discussion and Conclusion


In addition, driving scenarios did not examine the more complex traffic scenarios

where eco-driving and safe driving behaviour are in conflict with each other, such as

cornering or the presence of other risky drivers on the road.

While this study has provided evidence to suggest such a system can be used

without having negative impact on driver workload, these findings were based on self-

reported subjective data rather than actual objective workload measurements such as

glance behaviour of participants. Moreover, a number of other potential concerns

associated with use of the system were expressed by participants. This included

privacy concerns, perceived accuracy of the system and subsequent trust, and the

impact of novelty on prolonged use. Privacy concerns are particularly relevant if

partnerships are established with insurance companies or traffic authorities for the

purpose of incentives to use the system, and the potential for data from the system to

be passed onto law enforcement agencies and used to incriminate them.

Perceived accuracy of feedback provided by the system, and the subsequent

impact on trust, was a common concern among drivers. Indeed, trust was an important

factor associated with intentions to use the system. Specifically, participants appeared

to be most concerned that the system would be too conservative in its measurements,

resulting in drivers receiving feedback that their driving is unsafe or not eco-friendly,

and that this feedback would be contradictory to their own perceptions of their

behaviour. This finding has important implications for drivers who exhibit optimism

biases in relation to their eco-safe driving ability, such that they are more likely to

attribute an observed disconnect to perceived system inaccuracies. This highlights the

need for clear communication to the driver regarding how behaviour is measured so

that it is perceived to be transparent and legitimate.

Chapter 8: Discussion and Conclusion 215


Lastly, this research states that using an in-vehicle HMI leads to a reduction in

fuel consumption and micro-level emissions. However, it is worth mentioning overall

levels of emissions across the population may increase (e.g., at the macro-economic

level) due to an increasing demand for mobility and population. This increase in

demand is known as the rebound effect (Rubin, 2007; Saunders, 1992). This further

highlights an importance of coupling technological advancements with other policies

and interventions that limit the resource use.

8.7 DIRECTIONS FOR FUTURE RESEARCH

The present program of research supports the applicability of the TAM in the

context of a cooperative eco-safe in-vehicle HMI. The findings also suggest a number

of potential additions to the theory, including driver characteristics, design

characteristics, motivation and normative influences on intentions to use. Future

research should seek to evaluate the incorporation of these determinants into the TAM,

specifically in the context of an eco-safe in-vehicle HMI.

While associated with adequate driver acceptance and impacts on driver

workload, the findings suggest that future research could examine even more ways for

providing feedback messages to drivers, including information on eco-routing choice,

which may also have significant negative impact on fuel consumption and emissions.

This may include options for greater customisation of the feedback characteristics and

more contextual information. For example, the advice/feedback could be adjusted and

personalised based on the improvements in drivers’ skills. There is unlikely to be a

one-size-fits-all approach to the provision of feedback, and as such users should have

considerable control over the characteristics of the feedback provided to them. In

particular, future research should explore the provision of more comprehensive

information, such as how eco-safe driving behaviours are measured by the system, and

216 Chapter 8: Discussion and Conclusion


a more detailed description of what is required from the driver to improve their

behaviour. Given the extensive nature of such information, this feedback should be

included in the back-end of the system and presented after driving, perhaps in the form

of a compatible smartphone application.

In addition, future research should explore the integration of additional

technologies in the system design. This may include navigation, dash-cams and

communication capabilities, as well as hazard detection. Such an approach serves a

number of benefits, including improving the perceived usefulness of the device in the

long-term and minimising the number of separate systems in the vehicle. Moreover,

future research should explore the complimentary impact of other behavioural and

attitudinal changes interventions, such as driver training, education, enforcement, and

alternative HMI designs solutions.

Further research is also required to more comprehensively explore the impact of

competition on driver behaviour, and the potential for this approach to be used to

enhance voluntary use of in-vehicle HMIs in long term. Future research should

evaluate the use of eco-safe in-vehicle HMIs in different contexts, including in

organisational fleets, among novice drivers, and under well-established incentive

schemes. Indeed, future research should investigate the impact of a variety of other

incentive types, and in particular vehicle-related incentives, on not only eco-safe

driving behaviour, but also intentions to use the system. Such research should exhibit

a particular focus on more problematic drivers or as a tool to train inexperience drivers,

including those who display poorer eco-safe driving performance and/or have less

positive attitudes toward eco-safe driving and the environment.

Finally, given the rapid nature of automotive technology developments and the

rising level of automation, future research needs to examine more closely the impact

Chapter 8: Discussion and Conclusion 217


of the accuracy of the advice/feedback from in-vehicle HMI and the subsequent impact

on trust in this technology. This would be a fruitful area for further work.

8.8 CONCLUDING REMARKS

A number of emerging technologies, such as electric and automated vehicles and

in-vehicle systems (e.g., head-up-displays, V2V and V2I) are currently gaining more

widespread evaluation and implementation. Nonetheless, there remains a need for even

greater efforts to develop interventions to reduce the human impact on climate change

and protect the earth for future generations (Guitton, 2017). The motorised transport

sector represents an area of human activity that is particularly amenable to changes

that could have considerable environmental benefits. Moreover, this sector

additionally impacts public health through fatalities and injuries in traffic crashes.

The primary aim of the current research was to develop and evaluate a

cooperative eco-safe in-vehicle HMI designed to reduce fuel consumption and

emissions while improving safe driving. Taken together, the results of the research

provide evidence that a cooperative eco-safe driving in-vehicle HMI involving both

advice and feedback messages has the potential to improve eco-safe driving, while also

being accepted by drivers and producing minimal adverse impacts on driver workload.

However, the effectiveness of the system, and even the driver acceptance of the

system, as observed in the current research, does not ensure that users will voluntarily

adopt the system as part of their driving routine.

Thus, future research should comprehensively explore the use of incentives as a

means to motivate drivers to use the system. While evidence from the current research

suggests that these may have limited additional impact on eco-safe driving

performance, driver feedback suggested that incentives represent a plausible approach

to increasing system usage rates. Vehicle-related incentives appear to be most

218 Chapter 8: Discussion and Conclusion


effective, suggesting that widespread implementation may be achievable if incentive

schemes can be established with insurance companies or traffic authorities.

Chapter 8: Discussion and Conclusion 219


Bibliography

Adell, E. (2009). Driver experience and acceptance of driver assistance system- a case of speed
adaption. (PhD), Lund University, Sweden.
Adell, E. (2010). Acceptance of driver support systems. Paper presented at the Proceedings of
European Conference on Human Centred Design for Intelligent Transport Systems :
April 29 - 30 2010 Berlin, Germany, Berlin, Germany.
Adell, E., & Varhelyi, A. (2014). Modeling acceptance of driver assistance systems:
Application of the unified theory of acceptance and use of technology. In T. H. Michael
A. Regan & S. Alan (Eds.), Driver Acceptance of New Technology : Theory,
Measurement and Optimisation. Farnham, Surrey, England: Ashgate Publishing, Ltd.
Adell, E., Várhelyi, A., & Fontana, M. d. (2011). The effects of a driver assistance system for
safe speed and safe distance – A real-life field study. Transportation Research Part C:
Emerging Technologies, 19(1), 145-155. doi:10.1016/j.trc.2010.04.006
Adell, E., Várhelyi, A., & 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), 445-458. doi:10.1016/j.trf.2008.04.003
Adell, E., Varhelyi, A., & Nilsson, L. (2014). The definition of aceptance and acceptability. In
T. H. Michael A. Regan & S. Alan (Eds.), Driver Acceptance of New Technology :
Theory, Measurement and Optimisation (pp. 11-21). Farnham, Surrey, England:
Ashgate Publishing, Ltd.
af Wåhlberg, A., Dorn, L., & Kline, T. (2011). The Manchester Driver Behaviour
Questionnaire as a predictor of road traffic accidents. Theoretical Issues in Ergonomics
Science, 12(1), 66-86.
Alam, M. S., & McNabola, A. (2014). A critical review and assessment of Eco-Driving policy
& technology: Benefits & limitations. Transport Policy, 35, 42-49.
doi:10.1016/j.tranpol.2014.05.016
Albert, G., Musicant, O., Oppenheim, I., & Lotan, T. (2016). Which smartphone's apps may
contribute to road safety? An AHP model to evaluate experts' opinions. Transport
Policy, 50, 54-62. doi:10.1016/j.tranpol.2016.06.004
Alwakiel, H. N. (2011). Leveraging Weigh-In-Motion (WIM) data to estimate link-based
heavy-vehicle emissions. Portland State University, USA.
Australian Authority Climate Change. (2014). Light vehicle emissions standards for Australia.
Retrieved from climatechangeauthority.gov.au
Australian Automobile Association. (2017). Cost of Road Trauma in Australia. Retrieved from
www.aaa.asn.au
Australian Goverment. (2015). Science and Research Priorities.
Australian Research Council. (2015). Fact Sheet: Science and Research Priorities. Retrieved
from science.gov.au
Australian Transport Council. (2014). National Road Safety Action Plan 2015 - 2017.
Azzi, S., Reymond, G., Merienne, F., & Kemeny, A. (2011). Eco-driving performance
assessment with in-car visual and haptic feedback assistance. Journal of Computing
and Information Science in Engineering, 11(4), 041005. doi:10.1115/1.3622753
Bangor, A., Kortum, P. T., & Miller, J. T. (2008). An empirical evaluation of the system
usability scale. International Journal of Human–Computer Interaction, 24(6), 574-594.
doi:10.1080/10447310802205776
Barkenbus, J. N. (2010). Eco-driving: An overlooked climate change initiative. Energy Policy,
38(2), 762-769. doi:10.1016/j.enpol.2009.10.021
Barth, M., & Boriboonsomsin, K. (2009). Energy and emissions impacts of a freeway-based
dynamic eco-driving system. Transportation Research Part D: Transport and
Environment, 14(6), 400-410. doi:10.1016/j.trd.2009.01.004
Bibliography 221
Bellotti, F., Berta, R., & De Gloria, A. (2014). A Social Serious Game Concept for Green,
Fluid and Collaborative Driving. In A. De Gloria (Ed.), Applications in Electronics
Pervading Industry, Environment and Society (pp. 163-170): Springer, Cham.
Bengler, K., Drüke, J., Hoffmann, S., Manstetten, D., & Neukum, A. (2017). UR: BAN Human
Factors in Traffic: Springer Vieweg, Wiesbaden.
Bennett, C. (2014). Smartphone app trial rewards safe drivers. Retrieved from
theherald.com.au/story/2127656/smartphone-app-trial-rewards-safe-drivers/
Bigazzi, A. Y., & Bertini, R. L. (2009) Adding green performance metrics to a transportation
data archive. Transportation Research Record (pp. 30-40).
Birrell, S. A., Fowkes, M., & Jennings, P. A. (2014). Effect of using an in-vehicle smart driving
aid on real-world driver performance. IEEE Transactions on Intelligent Transportation
Systems, 15(4), 1801-1810.
Birrell, S. A., & Young, M. S. (2011). The impact of smart driving aids on driving performance
and driver distraction. Transportation Research Part F: Traffic Psychology and
Behaviour, 14(6), 484-493.
Birrell, S. A., Young, M. S., & Weldon, A. M. (2010). Delivering Smart Driving Feedback
through a Haptic Pedal. Contemporary Ergonomics and Human Factors, 431-439.
Boriboonsomsin, K., Vu, A., & Barth, M. (2010). Eco-driving: pilot evaluation of driving
behavior changes among US drivers. University of California Transportation Center.
Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative research
in psychology, 3(2), 77-101. doi:doi.org/10.1191/1478088706qp063oa
Brouwer, R. F. T., Stuiver, A., Hof, T., Kroon, L., Pauwelussen, J., & Holleman, B. (2015).
Personalised feedback and eco-driving: An explorative study. Transportation Research
Part C: Emerging Technologies, 58(Part D), 760-771. doi:10.1016/j.trc.2015.04.027
Bureau of Infrastructure Transport and Regional Economics. (2015). Road Trauma Australia.
Retrieved from Canberra ACT: bitre.gov.au
Bureau of Infrastructure Transport and Regional Economics. (2017a). Costs and benefits of
emerging road transport technologies (146). Retrieved from Canberra ACT:
bitre.gov.au
Bureau of Infrastructure Transport and Regional Economics. (2017b). Road Rrauma Australia
2016 statistical summary, BITRE, Canberra ACT.
Čegovnik, T., Stojmenova, K., Jakus, G., & Sodnik, J. (2018). An analysis of the suitability of
a low-cost eye tracker for assessing the cognitive load of drivers. Applied Ergonomics,
68, 1-11.
Charlton, S. G. (2009). Driving while conversing: Cell phones that distract and passengers who
react. Accident Analysis & Prevention, 41(1), 160-173.
Chorlton, K., Hess, S., Jamson, S., & Wardman, M. (2012). Deal or no deal: Can incentives
encourage widespread adoption of intelligent speed adaptation devices? Accident
Analysis & Prevention, 48, 73-82.
Cialdini, R. B. (2001). Influence: Science and Practice, Allyn & Bacon.
CIECA. (2007). CIECA Eco-driving project: Eco-driving in category B driver training & the
driving test. Retrieved from cieca.eu/project/
Cohen, J. (1992). A power primer. Psychol Bull, 112(1), 155-159.
Condly, S. J., Clark, R. E., & Stolovitch, H. D. (2003). The Effects of Incentives on Workplace
Performance: A Meta‐analytic Review of Research Studies 1. Performance
Improvement Quarterly, 16(3), 46-63.
Cook, J., Oreskes, N., Doran, P. T., Anderegg, W. R. L., Verheggen, B., Maibach, E. W., . . .
Rice, K. (2016). Consensus on consensus: a synthesis of consensus estimates on human-
caused global warming. Environmental Research Letters, 11(4), 048002.
doi:doi.org/10.1088/1748-9326/11/4/048002
Dahlinger, A., & Wortmann, F. (2016). Towards the Design of Eco-Driving Feedback
Information Systems–A Literature Review. Multikonferenz Wirtschaftsinformatik
(MKWI) 2016, 2, 801-812.

222 Bibliography
Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of
information technology. MIS Quarterly, 319-340. doi:doi.org/10.2307/249008
Deci, E. L., Koestner, R., & Ryan, R. M. (2001). Extrinsic rewards and intrinsic motivation in
education: Reconsidered once again. Review of Educational Research, 71(1), 1-27.
Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation and self-determination in human
behavior: Springer Science & Business Media.
Delhomme, P., Cristea, M., & Paran, F. (2013). Self-reported frequency and perceived
difficulty of adopting eco-friendly driving behavior according to gender, age, and
environmental concern. Transportation Research Part D: Transport and Environment,
20. doi:10.1016/j.trd.2013.02.002
Delhomme, P., & Meyer, T. (2002). La recherche en Psychologie sociale: projets, méthodes
et techniques. Paris: A. Colin.
Demir, E., Bektaş, T., & Laporte, G. (2011). A comparative analysis of several vehicle
emission models for road freight transportation. Transportation Research Part D:
Transport and Environment, 16(5), 347-357. doi:10.1016/j.trd.2011.01.011
Demmel, S., Gruyer, D., Besnier, J., Ben Jemaa, I., Pechberti, S., & Rakotonirainy, A. (2011,
5-9 June 2011 ). Collision warning dissemination in vehicles strings: an empirical
measurement. Paper presented at the Intelligent Vehicles Symposium (IV), 2011 IEEE,
Baden.
Deterding, S. (2011). Situated motivational affordances of game elements: A conceptual model.
Paper presented at the Workshop at CHI, Vancouver, BC, Canada.
Deterding, S., Björk, S. L., Nacke, L. E., Dixon, D., & Lawley, E. (2013). Designing
gamification: creating gameful and playful experiences. Paper presented at the CHI' 13
Extended Abstracts on Human Factors in Computing Systems, Paris, France.
Deterding, S., Sicart, M., Nacke, L., O'Hara, K., & Dixon, D. (2011). Gamification. using
game-design elements in non-gaming contexts. Paper presented at the CHI'11 Extended
Abstracts on Human Factors in Computing Systems, Vancouver, BC, Canada.
Diewald, S., Möller, A., Roalter, L., Stockinger, T., & Kranz, M. (2013). Gameful design in
the automotive domain: review, outlook and challenges. Paper presented at the
Proceedings of the 5th International Conference on Automotive User Interfaces and
Interactive Vehicular Applications, Eindhoven, The Netherlands.
Eby, D. W., Molnar, L. J., Zhang, L., Louis, R. M. S., Nicole Zanier, L. P., & Kostyniuk.
(2015). Keeping Older Adults Driving Safely: A Research Synthesis of Advanced In-
Vehicle Technologies. Retrieved from University of Michigan Transportation Research
Institute and the ATLAS Center.
ECODRIVEN. European Ecodriving & Traffic Safety Campaigns. Retrieved from
ecodrive.org
Ecodriver. (2011). Ecodriver EU-FP & projects. Retrieved from htttp://ertico.com/ecodriver
EcoDrivingUSA. Alliance of Automobile Manufacturers. Retrieved from
ecodrivingUSA.com
ECOWILL. (2014). Ecodriving. Retrieved from ecodrive.org
Elder, R. W., Voas, R., Beirness, D., Shults, R. A., Sleet, D. A., Nichols, J. L., & Compton, R.
(2011). Effectiveness of Ignition Interlocks for preventing alcohol-impaired driving and
alcohol-related crashes: A Community Guide Systematic Review. American Journal of
Preventive Medicine, 40(3), 362-376. doi:10.1016/j.amepre.2010.11.012
Ericsson, E. (2001). Independent driving pattern factors and their influence on fuel-use and
exhaust emission factors. Transportation Research Part D:Transport and
Environment, 6(5), 325-345. doi:10.1016/S1361-9209(01)00003-7
ETRAC. (2014). European Road Transport Research Advisory Council. Retrieved from
ertrac.org
Fagnant, D. J., & Kockelman, K. (2015). Preparing a nation for autonomous vehicles:
opportunities, barriers and policy recommendations. Transportation Research Part A:
Policy and Practice, 77, 167-181. doi:10.1016/j.tra.2015.04.003
Festinger, L. (1962). A theory of cognitive dissonance (Vol. 2): Stanford University Press.
Bibliography 223
Field, A. (2013). Discovering statistics using SPSS. London: SAGE publications.
Fisher, D. L., Rizzo, M., Caird, J., & Lee, J. D. (2011). Handbook of Driving Simulation for
Engineering, Medicine, and Psychology. USA: CRC Press.
Fitz-Walter, Z., Wyeth, P., Tjondronegoro, D., & Scott-Parker, B. (2013). Driven to drive:
Designing gamification for a learner logbook smartphone application. Paper presented
at the Proceedings of the First International Conference on Gameful Design, Research,
and Applications, Toronto, Ontario, Canada.
Fleiter, J. J., & Watson, B. C. (2006). The speed paradox: The misalignment between driver
attitudes and speeding behaviour. Journal of the Australasian College of Road Safety,
17(2), 23-30.
Fogg, B. J. (2002). Persuasive technology: using computers to change what we think and do.
Ubiquity, 2002(December), 2. doi:10.1145/764008.763957
Fors, C., Kircher, K., & Ahlstrom, C. (2015). Interface design of eco-driving support systems
- Truck drivers' preferences and behavioural compliance. Transportation Research Part
C: Emerging Technologies, 58, 706-720. doi:10.1016/j.trc.2015.03.035
Francois, M., Osiurak, F., Fort, A., Crave, P., & Navarro, J. (2017). Automotive HMI design
and participatory user involvement: review and perspectives. Ergonomics, 60(4), 541-
552. doi:10.1080/00140139.2016.1188218
Garbers, Y., & Konradt, U. (2014). The effect of financial incentives on performance: A
quantitative review of individual and team‐based financial incentives. Journal of
Occupational and Organizational Psychology, 87(1), 102-137.
Gardner, B., Whittington, C., McAteer, J., Eccles, M. P., & Michie, S. (2010). Using theory to
synthesise evidence from behaviour change interventions: The example of audit and
feedback. Social Science & Medicine, 70(10), 1618-1625.
doi:10.1016/j.socscimed.2010.01.039
Garrison, T. M., & Williams, C. C. (2013). Impact of relevance and distraction on driving
performance and visual attention in a simulated driving environment. Applied Cognitive
Psychology, 27(3), 396-405.
Ghazizadeh, M., & Lee, J. D. (2014). Modeling driver acceptance: from feedback to monitoring
and mentoring system. In A. Stevens, M. Regan, & T. Horberry (Eds.), Driver
Acceptance of New Technology: Theory, Measurement and Optimisation (pp. 51-70).
Farnham, Surrey, England: Ashgate Publishing, Ltd.
Giles, E. L., Robalino, S., McColl, E., Sniehotta, F. F., & Adams, J. (2014). The effectiveness
of financial incentives for health behaviour change: systematic review and meta-
analysis. PloS One, 9(3), e90347.
Gilman, E., Keskinarkaus, A., Tamminen, S., Pirttikangas, S., Röning, J., & Riekki, J. (2015).
Personalised assistance for fuel-efficient driving. Transportation Research Part C:
Emerging Technologies, 58, 681-705. doi:10.1016/j.trc.2015.02.007
Given, L. M. (2015). 100 questions (and answers) about qualitative research. London: SAGE
Publications.
Gneezy, U., Meier, S., & Rey-Biel, P. (2011). When and why incentives (don't) work to modify
behavior. The Journal of Economic Perspectives, 25(4), 191-209.
GOA. (2013). Intelligent Transport Systems: vehicle-to-Vehicle: Technologies Expected to
Offer Safety benefits but a Variety of Deployment Challenges Exist. Retrieved from
www.gao.gov
Gore, A. (2006). An inconvenient truth: The planetary emergency of global warming and what
we can do about it: Rodale.
Graves, G., Jeffreys, I., & Roth, M. (2012). RACQ EcoDrive Research Study: final report.
Retrieved from racq.com.au
Graving, J. S., Manser, M. P., & Becic, E. (2010). Reduction in fuel consumption depends on
the fuel economy display and driver sex: An observed interaction. Paper presented at
the Adjunct Proceedings of the Second International Conference on Automotive User
Interfaces and Interactive Vehicular Applications (AutomotiveUI 2010), Pittsburgh,
Pennsylvania.
224 Bibliography
Grier, R. A. (2015). How high is high? A meta-analysis of NASA-TLX global workload scores.
Paper presented at the Proceedings of the Human Factors and Ergonomics Society
Annual Meeting, Los Angeles.
Guerassimoff, G., & Thomas, J. (2015). Enhancing energy efficiency and technical and
marketing tools to change people's habits in the long-term. Energy and Buildings, 104,
14-24. doi:10.1016/j.enbuild.2015.06.080
Guest, G., Bunce, A., & Johnson, L. (2006). How many interviews are enough? An experiment
with data saturation and variability. Field methods, 18(1), 59-82.
Guitton, M. J. (2017). Rethinking the impact of new technologies in a climate change era:
Challenges or opportunities? Computers in Human Behavior, 76, 1-2.
doi:10.1016/j.chb.2017.06.005
Hart, S. G. (2006). NASA-task load index (NASA-TLX): 20 years later. Paper presented at the
Proceedings of the human factors and ergonomics society annual meeting, Los Angeles.
Hart, S. G., & Staveland, L. E. (1988). Development of NASA-TLX (Task Load Index):
Results of empirical and theoretical research. Advances in Psychology, 52, 139-183.
Harvey, J., Thorpe, N., & Fairchild, R. (2013). Attitudes towards and perceptions of eco-
driving and the role of feedback systems. Ergonomics, 56(3), 507-521.
doi:doi.org/10.1080/00140139.2012.751460
Hattie, J., & Timperley, H. (2007). The Power of Feedback. Review of Educational Research,
77(1), 81-112. doi:10.3102/003465430298487
Haworth, N., & Symmons, M. (2001). The relationship between fuel economy and safety
outcomes: Monash University Accident Research Centre.
Hickman, J. S., & Hanowski, R. J. (2011). Use of a video monitoring approach to reduce at-
risk driving behaviors in commercial vehicle operations. Transportation Research Part
F: Traffic Psychology and Behaviour, 14(3), 189-198. doi:10.1016/j.trf.2010.11.010
Hipp, E., Bengler, K., Kressel, U., & Feit, S. (2018). The Research Initiative UR:BAN. In K.
Bengler, J. Drüke, S. Hoffmann, D. Manstetten, & A. Neukum (Eds.), UR:BAN Human
Factors in Traffic: Approaches for Safe, Efficient and Stress-free Urban Traffic (pp. 3-
26). Wiesbaden: Springer Fachmedien Wiesbaden.
Hof, T., Conde, L., Garcia, E., Iviglia, A., Jamson, S., Jopson, A., . . . Zlocki, A. (2013). D11.1:
A state of the art review and user’s expectations. Retrieved from ecoriver-project.eu
Hooker, J. (1988). Optimal driving for single-vehicle fuel economy. Transportation Research
Part A: General, 22(3), 183-201.
Horberry, T., Stevens, A., & Regan, M. (2014a). Driver acceptance of new technology:
Synthesis and perspective. In T. H. Michael A. Regan & S. Alan (Eds.), Driver
Acceptance of New Technology : Theory, Measurement and Optimisation (pp. 335-
348). Farnham, Surrey, England: Ashgate Publishing, Ltd.
Horberry, T., Stevens, A., & Regan, M. (2014b). Driver acceptance of new technology: Theory,
measurement and optimisation. Farnham, Surrey, England: Ashgate Publishing, Ltd.
Horrey, W. J., Lesch, M. F., Dainoff, M. J., Robertson, M. M., & Noy, Y. I. (2012). On-board
safety monitoring systems for driving: review, knowledge gaps, and framework.
Journal of safety research, 43(1), 49-58. doi:10.1016/j.jsr.2011.11.004
Huang, Y., Roetting, M., McDevitt, J. R., Melton, D., & Courtney, T. K. (2005). In-vehicle
safety feedback. Professional Safety, 50(1), 20-27.
Iacovides, I. (2009). Exploring the link between player involvement and learning within digital
games. Paper presented at the Proceedings of the 23rd British HCI group annual
conference on people and computers: Celebrating people and technology, Cambridge,
United Kingdom.
ISO. (2010). International Organisation for Standardization. Ergonomics of Human-System
Interaction – Part 210: Human-Centered Design for Interactive Systems. ISO 9241-
210. Geneva: Beuth.
Jamson, H., Hibberd, D., & Merat, N. (2015). Interface design considerations for an in-vehicle
eco-driving assistance system. Transportation Research Part C: Emerging
Technologies.
Bibliography 225
Jamson, S. (2006). Would those who need ISA, use it? Investigating the relationship between
drivers’ speed choice and their use of a voluntary ISA system. Transportation Research
Part F: Traffic Psychology and Behaviour, 9(3), 195-206.
doi:10.1016/j.trf.2005.11.002
Jamson, S. (2010). Acceptability data - what should or could it predict?. Paper presented at the
International Seminar on Acceptance, Paris, France.
Jamson, S., Chorlton, K., & Carsten, O. (2012). Could Intelligent Speed Adaptation make
overtaking unsafe? Accident Analysis & Prevention, 48, 29-36.
doi:10.1016/j.aap.2010.11.011
Jonkers, E., Wilmink, I., Nellthorp, J., Guehnemann, A., & Olstam, J. (2016). D54.1: Costs
and benefits of green driving support systems. Retrieved from ecodriver-project.eu
Jung, J., Schneider, C., & Valacich, J. (2010). Enhancing the motivational affordance of
information systems: The effects of real-time performance feedback and goal setting in
group collaboration environments. Management Science, 56(4), 724-742.
Kim, S. S., & Malhotra, N. K. (2005). A longitudinal model of continued IS use: An integrative
view of four mechanisms underlying postadoption phenomena. Management Science,
51(5), 741-755. doi:10.1287/mnsc.1040.0326
King, W. R., & He, J. (2006). A meta-analysis of the technology acceptance model.
Information & Management, 43(6), 740-755.
Krause, M., Yilmaz, L., & Bengler, K. (2014). Comparison of real and simulated driving for a
static driving simulator. Adv. Hum. Aspects Transp.: Part II, 8, 29.
Kurani, K. S., Stillwater, T., Jones, M., & Caperello, N. (2013). Ecodrive I-80: A Large Sample
Fuel Economy Feedback Field Test Final Report. Institute of Transportation Studies
Report ITS-RR-13-15, University of California, Davis.
Lai, W.-T. (2015). The effects of eco-driving motivation, knowledge and reward intervention
on fuel efficiency. Transportation Research Part D: Transport and Environment, 34,
155-160. doi:10.1016/j.trd.2014.10.003
Larsson, H., & Ericsson, E. (2009). The effects of an acceleration advisory tool in vehicles for
reduced fuel consumption and emissions. Transportation Research Part D: Transport
and Environment, 14(2), 141-146. doi:doi.org/10.1016/j.trd.2008.11.004
Larue, G. S., Malik, H., Rakotonirainy, A., & Demmel, S. (2014). Fuel consumption and gas
emissions of an automatic transmission vehicle following simple eco-driving
instructions on urban roads. IET Intelligent Transport Systems, 8(7), 590-597.
Lee, J. D., & Seppelt, B. D. (2009). Human factors in automation design. Springer handbook
of automation, 417-436.
Lewis, J., & Sauro, J. (2009). The factor structure of the system usability scale. Human
centered design, 5619, 94-103.
Loumidi, A. K., Mittag, S., Lathrop, W. B., & Althoff, F. (2011). Eco-driving incentives in the
North American market. Paper presented at the Proceedings of the 3rd International
Conference on Automotive User Interfaces and Interactive Vehicular Applications.
Magana, V. C., & Munoz-Organero, M. (2015). GAFU: Using a Gamification Tool to Save
Fuel. IEEE Intelligent Transportation Systems Magazine, 7(2), 58-70.
doi:10.1109/mits.2015.2408152
Magaña, V. C., & Muñoz-Organero, M. (2016). Artemisa: A Personal Driving Assistant for
Fuel Saving. IEEE Transactions on Mobile Computing, 15(10), 2437-2451.
doi:10.1109/TMC.2015.2504976
Maguire, M. (2001). Methods to support human-centred design. International Journal of
Human-Computer Studies, 55(4), 587-634. doi:10.1006/ijhc.2001.0503
Mankoff, J., Dey, A. K., Hsieh, G., Kientz, J., Lederer, S., & Ames, M. (2003). Heuristic
evaluation of ambient displays. Paper presented at the Proceedings of the SIGCHI
conference on Human factors in computing systems, Lauderdale, FL, USA.
Manser, M. P., Rakauskas, M., Graving, J., & Jenness, J. (2010). Fuel economy driver
interfaces: develop interface recommendations (Report on Task 3). National Highway
Traffic Safety.
226 Bibliography
Mark, D. F., & Yardley, L. (2004). Content and thematic analysis. California: SAGE.
Martin, E., Chan, N., & Shaheen, S. (2012). Understanding how ecodriving public education
can result in reduced fuel use and greenhouse gas emissions. Paper presented at the
91st Annual Meeting of the Transportation Research Board, Washington, DC, January.
Matthews, H. D., Gillett, N. P., Stott, P. A., & Zickfeld, K. (2009). The proportionality of
global warming to cumulative carbon emissions. Nature, 459(7248), 829-832.
doi:10.1038/nature08047
McIlroy, R. C., Stanton, N. A., Godwin, L., & Wood, A. P. (2017). Encouraging eco-driving
with visual, auditory, and vibrotactile stimuli. IEEE Transactions on Human-Machine
Systems.
McIlroy, R. C., Stanton, N. A., & Harvey, C. (2014). Getting drivers to do the right thing: a
review of the potential for safely reducing energy consumption through design. IET
Intelligent Transport Systems, 8(4), 388-397. doi:10.1049/iet-its.2012.0190
McKenna, F. P., & Myers, L. B. (1997). Illusory self‐assessments—Can they be reduced?
British Journal of Psychology, 88(1), 39-51. doi:10.1111/j.2044-8295.1997.tb02619.x
Merrikhpour, M., Donmez, B., & Battista, V. (2014). A field operational trial evaluating a
feedback–reward system on speeding and tailgating behaviors. Transportation
Research Part F: Traffic Psychology and Behaviour, 27(Part A), 56-68.
doi:10.1016/j.trf.2014.09.002
Meschtscherjakov, A., Wilfinger, D., Scherndl, T., & Tscheligi, M. (2009). Acceptance of
future persuasive in-car interfaces towards a more economic driving behaviour. Paper
presented at the Proceedings of the 1st International Conference on Automotive User
Interfaces and Interactive Vehicular Applications, Essen, Germany.
Miles, M. B., Huberman, A. M., & Saldana, J. (2013). Qualitative data analysis: A methods
sourcebook. London: SAGE Publications.
Milfont, T. L., & Duckitt, J. (2010). The environmental attitudes inventory: A valid and reliable
measure to assess the structure of environmental attitudes. Journal of Environmental
Psychology, 30(1), 80-94. doi:10.1016/j.jenvp.2009.09.001
Mitsopoulos-Rubens, E., & Regan, M. (2014). Measuring acceptability through questionnaires
and focus groups. In A. Stevens, M. Regan, & T. Horberry (Eds.), Driver Acceptance
of New Technology : Theory, Measurement and Optimisation (pp. 89-104.). Farnham,
Surrey, England: Ashgate Publishing, Ltd.
Mortimer, D., Ghijben, P., Harris, A., & Hollingsworth, B. (2013). Incentive‐based and non‐
incentive‐based interventions for increasing blood donation. The Cochrane Library.
Musicant, O., Lotan, T., & Grimberg, E. (2015). Potential of group incentive schemes to
encourage use of driving safety apps. Transportation Research Record: Journal of the
Transportation Research Board(2516), 1-7.
NASA. (2017, June 15). Global Climate Change: Vital Signs of the Planet. Retrieved from
climate.nasa.gov/evidence/
Nielsen, J. (1994). Heuristic evaluation. Usability inspection methods, 17(1), 25-62.
Norman, D. A., & Draper, S. W. (1986). User centered system design: New perspectives on
Human-Computer Interaction. London: Lawrence Erlbaum Associates.
Nouveliere, L., Mammar, S., & Luu, H. T. (2012, 11-14 April 2012). Energy saving and safe
driving assistance system for light vehicles: Experimentation and analysis. Paper
presented at the Proceedings of 2012 9th IEEE International Conference on
Networking, Sensing and Control, Beijing, China.
O'Reilly, M., & Parker, N. (2012). ‘Unsatisfactory Saturation’: a critical exploration of the
notion of saturated sample sizes in qualitative research. Qualitative Research, 190-197.
doi:10.1177/1468794112446106
Oreskes, N. (2004). Beyond the ivory tower. The scientific consensus on climate change.
Science, 306(5702), 1686. doi:10.1126/science.1103618
Orfila, O., Gruyer, D., Geoffroy, D., Glaser, S., Rakotonirainy, A., Vaezipour, A., & Demmel,
S. (2016). Immersive driving simulation architecture to support gamified eco-driving

Bibliography 227
instructions. Paper presented at the In 23rd Intelligent Transportation Systems World
Congress, Melbourne, Australia.
Orfila, O., Saint Pierre, G., & Messias, M. (2015). An android based ecodriving assistance
system to improve safety and efficiency of internal combustion engine passenger cars.
Transportation Research Part C: Emerging Technologies, 58(Part D), 772-782.
doi:10.1016/j.trc.2015.04.026
Osswald, S., Wurhofer, D., Trösterer, S., Beck, E., & Tscheligi, M. (2012). Predicting
information technology usage in the car: towards a car technology acceptance model.
Paper presented at the Proceedings of the 4th International Conference on Automotive
User Interfaces and Interactive Vehicular Applications, Portsmouth New Hampshire.
Oviedo-Trespalacios, O., Haque, M. M., King, M., & Washington, S. (2016). Understanding
the impacts of mobile phone distraction on driving performance: A systematic review.
Transportation Research Part C: Emerging Technologies, 72, 360-380.
doi:10.1016/j.trc.2016.10.006
Pace, T., Ramalingam, S., & Roedl, D. (2007). Celerometer and idling reminder: persuasive
technology for school bus eco-driving. Paper presented at the CHI'07 Extended
Abstracts on Human Factors in Computing Systems, San Jose, CA, USA.
Pachauri, R. K., Allen, M. R., Barros, V. R., Broome, J., Cramer, W., Christ, R., . . . Dasgupta,
P. (2014). Climate change 2014: synthesis report. Contribution of Working Groups I,
II and III to the fifth assessment report of the Intergovernmental Panel on Climate
Change. Geneva, Switzerland: IPCC.
Pampel, S. M., Jamson, S., Hibberd, D. L., & Barnard, Y. (2015). How I reduce fuel
consumption: An experimental study on mental models of eco-driving. Transportation
Research Part C: Emerging Technologies, 58, 669-680. doi:10.1016/j.trc.2015.02.005
Peden, M., Scurfield, R., Sleet, D., Mohan, D., Hyder, A. A., Jarawan, E., & Mathers, C. D.
(2004). World report on road traffic injury prevention: World Health Organization
Geneva.
Petridou, E., & Moustaki, M. (2000). Human Factors in the Causation of Road Traffic Crashes.
European Journal of Epidemiology, 16(9), 819-826.
Pieroni, M., Rizzello, L., Rosini, N., Fantoni, G., De Rossi, D., & Mazzei, D. (2015, 14-16
Dec. 2015). Affective Internet of Things: Mimicking human-like personality in
designing smart-objects. Paper presented at the 2015 IEEE 2nd World Forum on
Internet of Things (WF-IoT), Milan, Italy.
Popiv, D. (2012). Enhancement of driver anticipation and its implications on efficiency and
safety. Doctoral dissertation, Faculty Mech. Eng., TU München, Munich, Germany.
Preuk, K., Stemmler, E., & Jipp, M. (2016). Does surrounding traffic benefit from an assisted
driver with traffic light assistance system? Transportation Research Part F: Traffic
Psychology and Behaviour, 43, 302-314. doi:10.1016/j.trf.2016.09.008
Przybylski, A. K., Rigby, C. S., & Ryan, R. M. (2010). A motivational model of video game
engagement. Review of General Psychology, 14(2), 154.
Queensland Transport Main Road. (2017). Road Rules for Everuday Driving. Retrieved from
https://www.qld.gov.au/transport/safety/rules/road/distances#trans1
Rakotonirainy, A., Obst, P., & Loke, S. W. (2012). Socially aware computing constructs.
International Journal of Social and Humanistic Computing, 1(4), 375-395.
Ram, S., & Sheth, J. N. (1989). Consumer resistance to innovations: the marketing problem
and its solutions. Journal of Consumer Marketing, 6(2), 5-14.
Rapp, C. (2015). Gamification as a tool to encourage eco-driving. Blekinge Institute of
Technology, Karlskrona, Sweden.
Reagan, I. J., Bliss, J. P., Houten, R. V., & Hilton, B. W. (2013). The Effects of External
Motivation and Real-Time Automated Feedback on Speeding Behavior in a Naturalistic
Setting. Human Factors, 55(1), 218-230. doi:10.1177/0018720812447812
Reagan, I. J., Bliss, J. P., Van Houten, R., & Hilton, B. W. (2013). The effects of external
motivation and real-time automated feedback on speeding behavior in a naturalistic
setting. Hum Factors, 55(1), 218-230. doi:10.1177/0018720812447812
228 Bibliography
Regan, M. (2001). Intelligent transport systems : Safety and human factor issues Noble Park,
Vic: Royal Automobile Club of Victoria.
Regan, M., Lee, J., & Young, K. (2008). Driver distraction: Theory, effects, and mitigation:
CRC Press.
Regan, M., Mitsopoulos-Rubens, E., Haworth, N., & Young, K. (2002). Acceptability of in-
vehicle intelligent transport systems to Victorian car drivers. Retrieved from
monash.edu
Rittger, L. (2015). Driving Behaviour and Driver Assistance at Traffic Light Intersections.
(Doctoral Thesis), Universität Würzburg, Fakultät für Humanwissenschaften.
Retrieved from nbn-resolving.org/urn:nbn:de:bvb:20-opus-117646
Rittger, L., & Götze, M. (2017). HMI Strategy–Recommended Action UR: BAN Human
Factors in Traffic (pp. 119-150): Springer Vieweg, Wiesbaden.
Robb, G., Sultana, S., Ameratunga, S., & Jackson, R. (2008). A systematic review of
epidemiological studies investigating risk factors for work-related road traffic crashes
and injuries. Injury Prevention, 14(1), 51-58. doi:10.1136/ip.2007.016766
Rodríguez, M. D., Roa, R. R., Ibarra, J. E., & Curlango, C. M. (2014). In-car ambient displays
for safety driving gamification. Paper presented at the Proceedings of the 5th Mexican
Conference on Human-Computer Interaction.
Roetting, M., Huang, Y.-H., McDevitt, J. R., & Melton, D. (2003). When technology tells you
how you drive: Truck drivers’ attitudes towards feedback by technology.
Transportation Research Part F: Traffic Psychology and Behaviour, 6(4), 275-287.
doi:10.1016/j.trf.2003.09.001
Rommerskirchen, C., Helmbrecht, M., & Bengler, K. (2014). The Impact of an Anticipatory
Eco-Driver Assistant System in Different Complex Driving Situations on the Driver
Behavior. IEEE Intelligent Transportation Systems Magazine, 6(2), pp 45-56.
Rouzikhah, H., King, M., & Rakotonirainy, A. (2013). Examining the effects of an eco-driving
message on driver distraction. Accident Analysis & Prevention, 50, 975-983.
doi:10.1016/j.aap.2012.07.024
Rubin, J. (2007). The Efficiency Paradox. CIBC World Markets, StrategEcon.
Ryan, R. M., & Deci, E. L. (2000). Intrinsic and extrinsic motivations: Classic definitions and
new directions. Contemporary Educational Psychology, 25(1), 54-67.
doi:10.1006/ceps.1999.1020
Saboohi, Y., & Farzaneh, H. (2009). Model for developing an eco-driving strategy of a
passenger vehicle based on the least fuel consumption. Applied Energy, 86(10), 1925-
1932. doi:10.1016/j.apenergy.2008.12.017
Saifuzzaman, M., Zheng, Z., Haque, M. M., & Washington, S. (2015). Revisiting the Task–
Capability Interface model for incorporating human factors into car-following models.
Transportation Research Part B: Methodological, 82, 1-19.
Santos, J., Merat, N., Mouta, S., Brookhuis, K., & De Waard, D. (2005). The interaction
between driving and in-vehicle information systems: Comparison of results from
laboratory, simulator and real-world studies. Transportation Research Part F: Traffic
Psychology and Behaviour, 8(2), 135-146. doi:10.1016/j.trf.2005.04.001
Sauer, C. (1993). Why information systems fail: a case study approach. UK: Alfred Waller
Ltd., Publishers.
Saunders, H. D. (1992). The Khazzoom-Brookes Postulate and Neoclassical Growth. The
Energy Journal, Volume 13(Number 4), 131-148.
Schall, D. L., & Mohnen, A. (2017). Incentivizing energy-efficient behavior at work: An
empirical investigation using a natural field experiment on eco-driving. Applied
Energy, 185, 1757-1768. doi:10.1016/j.apenergy.2015.10.163
Seaborn, K., & Fels, D. (2015). Gamification in theory and action: A survey. International
Journal of Human-Computer Studies, 74, 14-31. doi:10.1016/j.ijhcs.2014.09.006
Shi, C., Lee, H. J., Kurczal, J., & Lee, A. (2012). Routine driving infotainment app:
Gamification of performance driving. Paper presented at the Adjunct Proceedings of

Bibliography 229
the 4th International Conference on Automotive User Interfaces and Interactive
Vehicular Applications, Portsmouth, New Hampshire.
Shinar, D. (1978). Psychology on the road. The human factor in traffic safety: New York, USA:
John Wiley and Sons.
Shipworth, M. (2000). Motivating home energy action. Australian Greenhouse Office.
Shute, V. J. (2008). Focus on formative feedback. Review of Educational Research, 78(1), 153-
189. doi:10.3102/0034654307313795
Singh, S. (2015). Critical reasons for crashes investigated in the national motor vehicle crash
causation survey. Retrieved from Transportation Research Board (TRB) 97th Annual
Meeting, Washington, D.C: trid.trb.org
Sivak, M. (2014). Making Driving less Energy Intensive than Flying. Retrieved from
umich.edu
Sivak, M., & Schoettle, B. (2012). Eco-driving: Strategic, tactical, and operational decisions of
the driver that influence vehicle fuel economy. Transport Policy, 22, 96-99.
Soole, D. W. (2012). The relationship between drivers’ perceptions toward police speed
enforcement and self-reported speeding behaviour. (Master by Research), Queensland
University of Technology.
Staubach, M., Schebitz, N., Köster, F., & Kuck, D. (2014). Evaluation of an eco-driving
support system. Transportation Research Part F: Traffic Psychology and Behaviour,
27, 11-21. doi:10.1016/j.trf.2014.09.006
Steinberger, F., Proppe, P., Schroeter, R., & Alt, F. (2016). CoastMaster: An Ambient
Speedometer to Gamify Safe Driving. Paper presented at the Proceedings of the 8th
International Conference on Automotive User Interfaces and Interactive Vehicular
Applications, Ann Arbor, MI, USA.
Steinberger, F., Schroeter, R., Foth, M., & Johnson, D. M. (2017). Designing gamified
applications that make safe driving more engaging. Paper presented at the Proceedings
of CHI 2017: Conference on Human Factors in Computing Systems.
Steinberger, F., Schroeter, R., & Watling, C. N. (2017). From road distraction to safe driving:
Evaluating the effects of boredom and gamification on driving behaviour, physiological
arousal, and subjective experience. Computers in Human Behavior, 75, 714-726.
doi:doi.org/10.1016/j.chb.2017.06.019
Stevens, A., & Burnett, G. (2014). Designig In-Vehicle Technology for Usability. In A.
Stevens, M. Regan, & T. Horberry (Eds.), Driver acceptance of new technology:
Theory, measurement and optimisation (pp. 253-267). Farnham, Surrey, England:
Ashgate Publishing, Ltd.
Stevens, A., Quimby, A., Board, A., Kersloot, T., & Burns, P. (2002). Design guidelines for
safety of in-vehicle information systems. TRL Limited.
Stillwater, T., & Kurani, K. S. (2013). Drivers discuss ecodriving feedback: Goal setting,
framing, and anchoring motivate new behaviors. Transportation Research Part F:
Traffic Psychology and Behaviour, 19, 85-96.
Strömberg, H., & Karlsson, I. M. (2013). Comparative effects of eco-driving initiatives aimed
at urban bus drivers–Results from a field trial. Transportation Research Part D:
Transport and Environment, 22, 28-33.
Strömberg, H., Karlsson, I. M., & Rexfelt, O. (2015). Eco-driving: Drivers’ understanding of
the concept and implications for future interventions. Transport Policy, 39, 48-54.
Symmons, M., & Haworth, N. (2003). Are crashworthiness and fuel economy necessarily
conflicting outcomes? Road & Transport Research Journal, 12(1), 49-55.
Symmons, M., Rose, G., & Van Doorn, G. (2009). Ecodrive as a road safety tool for Australian
conditions. Retrieved from
infrastructure.gov.au/roads/safety/publications/2009/pdf/eco_drive.pdf
Taubman-Ben-Ari, O., & Yehiel, D. (2012). Driving styles and their associations with
personality and motivation. Accident Analysis & Prevention, 45, 416-422.
Taylor, S., & Todd, P. (1995). Assessing IT usage: The role of prior experience. MIS Quarterly,
561-570. doi:10.2307/249633
230 Bibliography
Toledo, T., Lotan, T., Caird, J., Horrey, W., & Trick, L. (2016). Feedback Technologies to
Young Drivers. Handbook of Teen and Novice Drivers: Research, Practice, Policy, and
Directions.
Tooth, R. (2010). The cost of road crashes: A review of key issues. Sydney: Macquarie
University, Department of Economics.
Transport and Main Roads. (2014). Queensland Speed Management Strategy 2010–2013:
Safer speeds, safer driving, for safer communities. Retrieved from tmr.qld.gov.au
Treat, J. R., Tumbas, N., McDonald, S., Shinar, D., Hume, R. D., Mayer, R., . . . Castellan, N.
(1979). Tri-level study of the causes of traffic accidents: final report. Executive
summary. Retrieved from Indiana University: hdl.handle.net/2027.42/64993
Trommer, S., & Höltl, A. (2012). Perceived usefulness of eco-driving assistance systems in
Europe. IET Intelligent Transport Systems, 6(2), 145-152.
Trübswetter, N., & Bengler, K. (2013). Why should I use ADAS? Advanced driver assistance
systems and the elderly: knowledge, experience, and usage barriers. Paper presented at
the Proc. 7th Int. Driving Symposium on Human Factors in Driver Assessment,
Training and Vehicle Design.
Tselentis, D. I., Yannis, G., & Vlahogianni, E. I. (2016). Innovative Insurance Schemes: Pay
as/how You Drive. Transportation Research Procedia, 14, 362-371.
doi:10.1016/j.trpro.2016.05.088
Vaezipour, A., Rakotonirainy, A., & Haworth, N. (2015). Reviewing in-vehicle systems to
improve fuel efficiency and road safety. Procedia Manufacturing, 3, 3192-3199.
doi:10.1016/j.promfg.2015.07.869
Vaezipour, A., Rakotonirainy, A., & Haworth, N. (2016). Design of a Gamified Interface to
Improve Fuel Efficiency and Safe Driving. In A. Marcus (Ed.), DUXU 2016: Design,
User Experience, and Usability: Novel User Experiences. Lecture Notes in Computer
Science (pp. 322-332). Switzerland: Springer International Publishing.
Vaezipour, A., Rakotonirainy, A., Haworth, N., & Delhomme, P. (2017). Enhancing eco-safe
driving behaviour through the use of in-vehicle human-machine interface: A qualitative
study. Transportation Research Part A: Policy and Practice, 100, 247-263.
doi:10.1016/j.tra.2017.04.030
Van Der Laan, J. D., Heino, A., & De Waard, D. (1997). A simple procedure for the assessment
of acceptance of advanced transport telematics. Transportation Research Part C:
Emerging Technologies, 5(1), 1-10.
Van der Pas, J. W. G. M., Kessels, J., Vlassenroot, S. H., & van Wee, B. (2014). The pros and
cons of Intelligent Speed Adaptation as a restrictive measure for serious speed
offenders. Transportation Research Part A: Policy and Practice, 67, 158-174.
doi:10.1016/j.tra.2014.07.001
Van der Voort, M., Dougherty, M. S., & van Maarseveen, M. (2001). A prototype fuel-
efficiency support tool. Transportation Research Part C: Emerging Technologies, 9(4),
279-296. doi:10.1016/S0968-090X(00)00038-3
Van Eck, R. (2006). Digital game-based learning: It's not just the digital natives who are
restless. EDUCAUSE review, 41(2), 16.
Van Huysduynen, H. H., Terken, J., Meschtscherjakov, A., Eggen, B., & Tscheligi, M. (2017).
Ambient Light and its Influence on Driving Experience. Paper presented at the
Proceedings of the 9th ACM International Conference on Automotive User Interfaces
and Interactive Vehicular Applications (AutomotiveUI ’17), September 24–27, 2016,
Oldenburg, Germany.
Venkatesh, V., & Bala, H. (2008). Technology acceptance model 3 and a research agenda on
interventions. Decision sciences, 39(2), 273-315.
Venkatesh, V., & Davis, F. D. (2000). A theoretical extension of the technology acceptance
model: Four longitudinal field studies. Management Science, 46(2), 186-204.
Venkatesh, V., Morris, M. G., Davis, G. B., & Davis, F. D. (2003). User acceptance of
information technology: Toward a unified view. MIS Quarterly, 27, 425-478.

Bibliography 231
Vilimek, R., & Keinath, A. (2014). User-centred design and evaluation as a prerequisite for the
success of disruptive innovations: An electric vehicle case study. In M. Regan, T.
Horberry, & A. Stevens (Eds.), Driver acceptance of new technology: Theory,
measurement and optimisation (pp. 169-186). Farnham, Surrey, England: Ashgate
Publishing Limited.
Volkswagen. (2009). TheFunTheory. Retrieved from thefuntheory.com
Wadud, Z., MacKenzie, D., & Leiby, P. (2016). Help or hindrance? The travel, energy and
carbon impacts of highly automated vehicles. Transportation Research Part A: Policy
and Practice, 86, 1-18. doi:10.1016/j.tra.2015.12.001
Walker, J. L. (2012). Research column. The Use of Saturation in Qualitative Research.
Canadian Journal of Cardiovascular Nursing, 22(2). doi:10.1007/s11274-015-1903-5
Waylen, A. E., Horswill, M. S., Alexander, J. L., & McKenna, F. P. (2004). Do expert drivers
have a reduced illusion of superiority? Transportation Research Part F: Traffic
Psychology and Behaviour, 7(4), 323-331. doi:10.1016/j.trf.2004.09.009
Webb, T. L., Sniehotta, F. F., & Michie, S. (2010). Using theories of behaviour change to
inform interventions for addictive behaviours. Addiction, 105(11), 1879-1892.
doi:10.1111/j.1360-0443.2010.03028.x
Weber, R. P. (1990). Basic content analysis. Thousand Oaks, United States: SAGE.
Weyer, J., Fink, R. D., & Adelt, F. (2015). Human–machine cooperation in smart cars. An
empirical investigation of the loss-of-control thesis. Safety Science, 72, 199-208.
doi:10.1016/j.ssci.2014.09.004
Wickens, C. D. (2002). Multiple resources and performance prediction. Theoretical Issues in
Ergonomics Science, 3(2), 159-177. doi:10.1080/14639220210123806
World Health Organization, W. (2014). Ambient (outdoor) air quality and health. Retrieved
from who.int
World Health Organization, W. (2015). Global status report on road safety 2013: Supporting
a decade of action WHO Library Cataloguing Retrieved from who.int.
Wu, C., Zhao, G., & Ou, B. (2011). A fuel economy optimization system with applications in
vehicles with human drivers and autonomous vehicles. Transportation Research Part
D: Transport and Environment, 16(7), 515-524.
Young, M., & Birrell, S. (2012). Ecological IVIS design: Using EID to develop a novel in-
vehicle information system. Theoretical Issues in Ergonomics Science, 13(2), 225-239.
doi:10.1080/1463922X.2010.505270
Young, M., Birrell, S., & Stanton, N. (2011). Safe driving in a green world: A review of driver
performance benchmarks and technologies to support ‘smart’ driving. Applied
Ergonomics, 42(4), 533-539. doi:10.1016/j.apergo.2010.08.012
Zichermann, G., & Cunningham, C. (2011). Gamification by design: Implementing game
mechanics in web and mobile apps. Canada: O'Reilly Media, Inc.

232 Bibliography
Appendices

A. FOCUS GROUPS MATERIAL


 


      

 






 



 


Appendices 233

 




 





 



 



 
   
      
      
      


 



234 Appendices

Appendices 235
236 Appendices
Appendices 237
238 Appendices
Appendices 239
Appendices 241
242 Appendices
Appendices 243
244 Appendices
B. EXPERT WORKSHOP MATERIALS


 





 





 





Appendices 245
246 Appendices






Appendices 247
248 Appendices
Appendices 249
250 Appendices
Appendices 251






252 Appendices
Appendices 253
254 Appendices
C. PILOT HMI ICONS

Appendices 255
D. PROGRESSION OF DEVICE DESIGN

256 Appendices
E. SIMULATOR STUDY SURVEY
















Appendices 257





























258 Appendices
Appendices 259
260 Appendices




















Appendices 261
262 Appendices
Appendices 263
264 Appendices
Appendices 265
266 Appendices
Appendices 267















268 Appendices




Appendices 269




270 Appendices
Appendices 271
272 Appendices
Appendices 273
F. ADDITIONAL DATA ANALYSIS (STUDY 3)

Table E.1. Extended demographic characteristics of the sample.

Demographic characteristics No. % Mean SD


Gender
Female 20 50.0
Male 20 50.0
Age 31.35 12.18
18-24 14 35.0
25-39 16 40.0
40-65 10 25.0
Licence type
Provisional (P) 14 35.0
Open (O) 26 65.0
Years held licence 12.03 12.19
Employment statusa
Employed 26 65.0
Student 21 52.5
Currently not employed 2 5.0
Educational attainment
High school (Year 10) 12 30.0
Trade qualification/TAFE 4 10.0
Tertiary degree 24 60.0
a
A total of nine (9) participants reported being employed and studying, while none reported being retired.

Table E.2. Transport and travel characteristics of the sample.

Driving characteristics No. %


Average hours driven per weeka
0-5 17 42.5
6-10 16 40.0
11-20 6 15.0
21-30 1 2.5
Kilometres driven in past yeara
< 10,000 14 35.0
10,001-20,000 15 37.5
20,001-30,000 10 25.0
Type of vehicle most commonly drivena
Passenger car 33 82.5
SUV/4WD 5 12.5
Motorcycle 2 5.0
Ownership of vehicle most commonly driven
Privately owned 36 90.0
Company owned 4 10.0
Transmission type of vehicle most commonly driven
Automatic 28 70.0
Manual 12 30.0
Type of fuel most commonly used in vehiclea
E10 Unleaded petrol 6 15.0
Regular unleaded petrol 21 52.5
Premium unleaded petrol 11 27.5
Diesel 2 5.0
Average proportion of time driven on road types
Urban (city, suburban) 58.8
Rural (country) 9.9
Motorway/highway/freeway 34.3
a
No participants reported driving more than 30 hours per week on average, driving more than 30,000kms in the previous year,
driving a heavy vehicle or riding a moped/scooter, or using a vehicle powered by LPG (gas) or electric vehicle.

274 Appendices
Table E.3. Experience with technology of the sample.
Number (%)
Technological device None Beginner Intermediate Advanced Expert
Personal computer/laptop 0 (0%) 1 (2.5%) 8 (20%) 21 (52.5%) 10 (25%)
Smartphone 0 (0%) 1 (2.5%) 12 (30%) 21 (52.5%) 6 (15%)
Tablet 4 (10%) 2 (5%) 14 (35%) 17 (42.5%) 3 (7.5%)
Video game console/device 10 (25%) 8 (20%) 12 (30%) 4 (10%) 6 (15%)
In-vehicle safety technology 5 (12.5%) 12 (30%) 18 (45%) 3 (7.5%) 2 (5%)
In-vehicle eco-driving technology 10 (25%) 12 (30%) 13 (32.5%) 4 (10%) 1 (2.5%)
Other in-vehicle technology 7 (17.5%) 10 (25%) 14 (35%) 5 (12.5%) 4 (10%)

Table E.4. Internal reliability of pre- and post-trip driver survey scales for simulator
study (Study 3).

Scale No. of items Cronbach’s α


Pre-trip questionnaire
Attitudes toward technology 6 .74
Attitudes toward the environment 15 .84
Attitudes towards eco-driving 10 .88
Perceptions of eco-driving behaviour 6 .75
Perceptions of safe driving behaviour 9 .78
Experience with technology 7 .79
Self-reported safe driving (DBQ) 13 .70
Post-trip questionnaire
Attitudes towards the device feedback characteristics 10 .83
Normative influences on system use 6 .90
Perceived trust in the system 10 .88
Perceived privacy of the system 5 .95
a
There were a total of four (4) driving conditions: baseline, advice only, feedback only, and advice and feedback. Only the NASA-
RTLX was administered following the baseline condition.

Table E.5. Internal reliability of immediate post-trip driver survey scales for
simulator study (Study 3).

Cronbach’s α
Scale No. of Advice Feedback Advice/
items only Only Feedback
Usefulness (Van der Laan Scale)a 5 .85 .88 .88
Satisfaction (Van der Laan Scale)a 4 .83 .91 .88
Perceived ease of use of the systemb 10 .84 .87 .83
Perceived usefulness of the systemb 13 .94 .92 .88
Intention to use the system 3 .83 .90 .86
Perceived usability of the system (SUS) c 10 .84 .86 .84
Workload (NASA-RTLX) d 6 .76 .69 .70
a
Adapted from van der Laan, et al. (1997), b Adapted from Staubach, et al. (2104), c Adapted from Bangor et al. (2008).
d
Workload was also measured following the baseline drive, with the scale having a Cronbach’s alpha of .67.

Table E.6. Attitudinal and perception constructs across the sample.

Construct Mean SD
Attitudes toward technology 3.44 .60
Attitudes toward the environment 4.05 .50
Attitudes toward eco-driving 3.70 .60
Perceptions of eco-driving 3.99 .56
Perceptions of safe driving 3.69 .56

Appendices 275
Table E.7. Attitudes toward device characteristics from simulator study (Study 3).

M SD
Auditory advice/feedback would make the system easier to understand 3.10 .93
Auditory advice/feedback would be distracting and/or annoying 3.38 .95
I would prefer advice/feedback to be speech rather than auditory tones (e.g., beep) 2.98 1.10
I would prefer to only have visual advice/feedback 3.45 1.01
I would prefer to have both visual and auditory advice/feedback 3.72 1.13
It would be important to allow users to customise the auditory advice/feedback of 3.87 .72
the system to suit their preferences
I would prefer is the system was an application on my smartphone 3.58 .90
It would be important to allow users to customise the visual advice/feedback of the 3.90 .74
system to suit their preferences
The system made me feel like I had a driving companion 3.12 1.04
The system made me feel like I was being judged 3.53 .99
The novelty of this system would wear off quickly 3.35 .83
The system would keep my attention 3.47 .88
I had difficulty seeing the visual display 2.23 .83
The visual display was clear 4.00 .72
The behavioural icons were difficult to understand 2.10 .74
The behavioural icons were intuitive and made sense 3.87 .82
The light was distracting and/or annoying 2.83 1.15
The light helped make me aware when advice/feedback was being presented 3.88 .85
I did not like the face of the system 2.28 .88
I would want to be able to customise the features of the system 3.70 .91
I think there will be no difference between my perceived driving behaviour and 2.88 .85
that of the vehicle monitoring device
The system seeks to improve both fuel efficiency and safe driving 3.82 .71
The system is more about improving fuel efficiency than safe driving 2.90 .96
The system is more about improving safe driving than fuel efficiency 2.93 .86

Table E.8. Means and standard deviations for driver acceptance measures.
Scale/Item Description Mean (SD)
Advice only Feedback only Advice/feedback
Perceived usefulness (van der Laan)
Useful-useless 0.53 (.96) 0.52 (1.11) 0.67 (.97)
Bad-gooda 0.75 (.81) 0.27 (1.01) 0.40 (.87)
Effective-unnecessary 0.10 (1.01) 0.27 (.99) 0.50 (1.06)
Assisting-worthless 0.48 (1.01) 0.58 (.84) 0.65 (0.80)
Raising alertness-sleep inducing 0.70 (.72) 0.93 (.83) 1.05 (.78)
Perceived satisfaction (van der Laan)
Pleasant-unpleasant 0.55 (.88) 0.15 (1.03) 0.15 (1.00)
Nice-annoying 0.53 (.99) -0.03 (1.07) 0.10 (1.11)
Irritating-likeablea 0.48 (.93) 0.02 (1.03) 0.00 (1.01)
Undesirable-desirablea 0.35 (.74) 0.00 (.85) 0.23 (.86)
a
Items were reverse coded.

276 Appendices
Table E.9. Means and standard deviations for driver acceptance measures.
Scale/Item Description Mean (SD)
Advice Feedback Advice/
only only feedback
Perceived ease of use
I found the system to be intuitive to use 3.60 (.71) 3.45 (.88) 3.63 (.77)
Using the system distracted me from driving 3.35 (.86) 2.85 (1.00) 2.78 (.89)
The feedback of the system was precise enough 3.23 (.80) 2.65 (1.15) 2.68 (.92)
My interaction with the system was clear and understandable 3.80 (.76) 3.48 (.91) 3.73 (.88)
It was easy for me to follow the information provided by the system 3.75 (.87) 3.23 (1.07) 3.35 (1.10)
Interacting with the system was frustrating 3.73 (.85) 2.93 (1.02) 3.08 (1.19)
Interacting with the system was comfortable 3.68 (.80) 3.55 (1.01) 3.33 (1.19)
The feedback from the system was encouraging 3.40 (.81) 2.90 (.96) 2.90 (1.13)
Learning to use the system was easy for me 4.20 (.69) 1.03 (.83) 4.00 (.75)
Overall, I find the system easy to use 3.98 (.83) 3.85 (.86) 3.98 (.73)
Perceived usefulness
I believe the system can help reduce fuel consumption and emissions 3.63 (.71) 3.38 (.87) 3.55 (.71)
Using the system increased my awareness of eco-driving 3.43 (.93) 3.45 (.93) 3.58 (.78)
Using the system helped me accelerate in a more fuel-efficient way 3.43 (.87) 3.20 (.94) 3.50 (.88)
Using the system helped decelerate in a more fuel-efficient way 3.35 (.95) 3.10 (.96) 3.08 (.92)
Using the system increased my awareness of fuel efficiency 3.33 (1.07) 3.28 (.93) 3.53 (.78)
Using the system helped me to avoid speeding 3.58 (.84) 3.45 (1.01) 3.68 (.89)
Using the system helped me keep a safe headway 3.33 (.83) 3.05 (.96) 3.40 (.84)
Using the system helped me to anticipate the flow of traffic 3.13 (1.04) 2.85 (.86) 3.13 (1.04)
Using the system helped me use the vehicle’s momentum more 3.15 (.98) 2.90 (.78) 2.90 (.90)
Using the system helped me to improve my awareness while driving 3.48 (.96) 3.53 (.91) 3.65 (.77)
Using the system helped me to enhance my safety 3.33 (1.02) 3.08 (.94) 3.40 (.90)
Using the system helped me to save fuel 3.20 (.88) 3.03 (.77) 3.28 (.55)
Perceived usability (SUS)
I think that I would like to use the system frequently 1.95 (1.15) 1.83 (.98) 1.80 (1.07)
The system looks unnecessarily complex 0.98 (.89) 1.08 (.86) 1.15 (.74)
The system looks very easy to use 3.20 (.72) 3.00 (.85) 3.07 (.69)
I would need the support of a technician to be able to use the system 0.67 (.92) 0.88 (.88) 0.93 (.92)
The various functions in the system seem well integrated 2.57 (.87) 2.40 (.96) 2.57 (.84)
I think there appears to be too much inconsistency in the system 1.40 (.93) 2.03 (1.00) 2.00 (.96)
I imagine most people would learn to use the system very quickly 3.10 (.90) 3.13 (.82) 3.02 (.77)
The system looks very awkward to use 1.10 (.87) 1.40 (1.06) 1.15 (1.05)
I believe that I could very confidently use the system 3.20 (.82) 2.90 (.87) 3.05 (.75)

Appendices 277
G. HMI EXPLANTION FROM SIMULATOR STUDY (STUDY 3)

278 Appendices
Driving events Eco-safe driving HMI behaviour

Default HMI screen: Eyes blinking

Advice: Smooth braking, shown 30


meters from the intersection in the
following driving scenarios:
-Deceleration when approaching a
traffic light (Green-Amber- Red)
-Deceleration when approaching a
traffic light (Red- Green)
-Deceleration when approaching a
zebra crossing
-Deceleration when approaching an
intersection with a STOP sign
Advice: Smooth acceleration, shown
3 seconds before the driver is required
to start driving after each complete
stop

Advice: Keep to speed limit, shown


whenever there is a change in speed
zone (40 km/h - 60 km/h)

Appendices 279
Advice: Keep safe headway, shown
whenever a driver approaches other
vehicles in flowing traffic

Feedback: Correct braking, shown


when the deceleration curve is within
the defined fuel-efficient parameters

Feedback: Incorrect braking,


shown when the deceleration curve is
outside the defined fuel-efficient
parameters

Feedback: Correct acceleration,


shown when the acceleration curve is
within the defined fuel-efficient
parameters

Feedback: Incorrect acceleration,


shown when the acceleration curve is
outside the defined parameters

280 Appendices
Feedback: Headway warning,
shown when the ‘time interval
vehicle’ with the car in the front is
less than 2.5 seconds

Feedback: Speed warning, shown


when the driver exceeds the speed
limit (plus 9% tolerance) in any speed
zones

Appendices 281