Fullpaper Format

12th ITS European Congress, Strasbourg, France, 19-22 June 2017

Paper ID # (TS27)

User Acceptance of Driverless Shuttles Running in an Open and Mixed Traffic Environment

Sina Nordhoff* a,b, Bart van Arem a, Natasha Merat c, Ruth Madigan c, Lisa Ruhrort b, Andreas Knie b, and Riender Happee a b

a TU Delft, 2600 GA Delft, The Netherlands Innovation Centre for Mobility and Societal Change – InnoZ, Euref-Campus 16, 10829 Berlin, Germany c University of Leeds, United Kingdom

Abstract User acceptance is a vital requirement for the success of automated vehicles that have been extensively addressed in current research in various acceptance studies. These studies have contributed to our understanding of potential acceptance factors of automated vehicles. However, the main focus of these studies has been on conventional vehicles with steer and pedals that were tested in artificial and simulated environments with a focus on in-vehicle technology. Furthermore, they list and identify a number of potential acceptance factors, but without making a systematic use of validated models to systematically present these factors. The current research addresses these gaps in research by investigating user acceptance of driverless shuttles in public transport in an open and mixed traffic environment on real semi-public roads in Berlin-Schöneberg. Results indicate that the acceptance and use of these driverless shuttles is influenced by their perceived usefulness, ease of use and social influence. Keywords: automated public transport, user acceptance


1. Introduction Automated driving has received a lot of attention from media and different stakeholders who are involved with the development of automated driving technology. Public exposure to various forms of automated vehicles has slowly started to increase in recent years, with several pilot projects and test runs of automated vehicles occurring Europe. This makes it possible to examine the user acceptance of automated vehicles in practice, which is defined as key prerequisite to ensure that the investments on automated vehicles are likely to pay off (Milakis, Snelder, Van Arem, Van Wee & De Almeida Correia, 2017). An example for a test field that has been opened very recently for the trialing of automated driving technology is the EUREF Campus in Berlin Schöneberg. The EUREF Campus is a 5.5 hectare large semi-public domain that is the host to approximately 100 companies and research facilities with 1.400 employees at the moment. With the agglomeration of sustainable and economic companies, it strives to develop innovations for smart sustainable cities. Since December 19, 2016, the driverless shuttle by the U.S. based startup Local Motors operates on the EUREF-Campus to provide a shuttle service for the employees of the EUREF campus as well as national and international guests and other interested users. The driverless shuttle, for which the term 4P vehicle was coined in a previous study (Nordhoff, Van Arem & Happee, 2016), is a SAE level 4 automation vehicle that does not have a steering wheel or pedals, and that requires some supervision either from an operator inside the vehicle or in an external control room. P stands for pod-like to clearly distinguish this type of automated vehicle from the classical and traditional automated automobile that is equipped with standard manual controls. The definition of the 4P vehicle corresponds with the recent update of the automation taxonomy by the Society of Automotive Engineers (SAE) proposed in the SAE taxonomy J3016 from September 2016. In this updated version, 4P’s technically correspond with SAE level 4 Automated Driving System-dedicated vehicles (ADS-DVs) that cannot be driven in automated mode beyond their operational design domain (ODD) (SAE, 2016). The driverless shuttle that operates on the EUREF campus runs with an operator on board to supervise its operations and intervene when necessary. The shuttle currently operates at a speed of 8 km/h on the campus, which allows a maximum speed of 10 km/h. 1.1. Research gaps The last four years have seen a high number of research studies investigating the acceptance of automated vehicles (e.g., AAA, 2016; Abraham, Lee, Brady, Fitzgerald, Mehler, Reimer & Coughlin, 2016; Bansal, Kockelman, & Singh, 2016; Bazilinskyy, Kyriakidis, & De Winter, 2015; Krueger, Rashidi & Rose, 2016; Fraedrich, Cynganski, Wolf & Lenz, 2016; Kyriakidis, Happee & De Winter, 2015; Hohenberger, Sporrle & Welpe, 2016; Madigan et al., 2016; König & Neumayr, 2017; Nees, 2016; Sanaullah, Hussain, Chaudhry, Case & Enoch, 2017; Schoettle & Sivak, 2015; Schoettle & Sivak, 2016; 2


Souders & Charness, 2016; Tennant, Howard, Franks, Bauer, Stares, Pansegrau, Stysko-Kunkowska & Cuevas-Badallo, 2016; Zmud, Sener & Wagner, 2016). These studies contribute to current research on acceptance of automated vehicles by identifying a number of potential acceptance factors that can be mainly summarized as socio-demographic characteristics, mobility characteristics, psychological characteristics, functional-utilitarian and symbolic-affective motives. The next step in research is to make a systematic use of validated models to organize the factors influencing user acceptance of automated vehicles. Furthermore, a large number of these acceptance studies have focused on automated vehicles with steer and pedals that have been tested in artificial environments. There has been research on drivers’ interaction with and acceptance of an assisted (e.g. Hoedemaker & Brookhuis, 1998; Rudin-Brown & Parker, 2004), partially (e.g. Louw & Merat, 2017; Van den Beukel & Van der Voort, 2017) or highly automated driving system in driving simulators (Merat & Jamson, 2009; Merat, Jamson, Lai & Carsten, 2012). These studies have in common that they focus on automated vehicles that can be controlled by standard manual controls by the driver to take back vehicle control and be in the loop again. As a result, there is limited knowledge on the factors that drive acceptance of automated vehicles without steer and pedals in real environmental conditions. One possible explanation is their low dispersion in consumer markets. The highest level that is currently being commercialized is SAE level 2 automation (Smith, 2016), while the next step involves adding more automated features (e.g. automated lane overtaking) to realize SAE level 3 (conditional automation). The use of driverless vehicles as feeder modes to public transport, which are the focus of this research here, has mainly been restricted to a few pilot projects worldwide (e.g. City Mobil2, WEpods) and thus only a few people have ever experienced them in daily conditions in real and complex environments, such as the EUREF Campus in Berlin Schöneberg. For these reasons, there is a lack of knowledge of the factors which will affect whether or not people will use these vehicles. 1.2. Research objectives To address these gaps in research, the 4P Acceptance Model was built as a synthesis of current acceptance studies on automated vehicles and other domains to explain, predict and optimize acceptance of driverless vehicles in public transport technology (Nordhoff, Van Arem & Happee, 2016). The model reflects the findings from the literature by identifying the factors whose role has been investigated in current acceptance studies (e.g. perceived usefulness, social influence, trust) and factors that have been proven to be relevant for the acceptance and use of information technology, but which still remain hypothetical in the context of automated driving (e.g. habit, price value). The 4P Acceptance Model assumes that the acceptance and use of driverless 4P’s is influenced by utilitarian-functional characteristics and symbolic-affective motives. The utilitarian-functional characteristics are captured by the Unified Theory of Acceptance and Use of Technology (UTAUT) as one of the most common technology acceptance models. The symbolic-affective motives are represented by the Pleasure-Arousal-Dominance (PAD) framework to address the emotional value of 3


driving or being driven, which has been one of the key reasons for the popularity of the private car. It was recently extended and validated (Nordhoff, Kyriakidis, Van Arem, Graff, Ruhrort & Happee, submitted) in the sense that both the UTAUT model and PAD framework are influenced by a range of different lower-level factors inherent to the individual (e.g. innovativeness) or the vehicle itself (e.g. vehicle speed) and environmental conditions that are external and hence beyond the direct control of the individual (e.g. accident with automated vehicle). The effects of the UTAUT constructs effort expectancy and social influence on the acceptance and use of driverless vehicles could not be investigated as their underlying items were excluded in the process of the exploratory factor analysis. Therefore, one of the research objectives of this study is to examine the effects of all UTAUT constructs, including performance expectancy, effort expectancy and social influence, on the acceptance and use of driverless vehicles in public transport. To tackle the gap in research involving “real” physical automated vehicles without steer and pedals that are tested by individuals under real environmental conditions, the purpose of this study is to capture the first perceptions of individuals who tested the driverless shuttle on the EUREF campus (Fig. 1). Questionnaires were distributed to respondents on tablet computers after they took a ride with the shuttle. This is the second study next to the study by Madigan et al. (2016) that we are aware of to the best of our knowledge that has studied the perceptions of real users of real vehicles on (semi-) public roads. For this reason, the research findings of current acceptance studies on automated vehicles need to be interpreted with care, taking into account the general lack of real and concrete experiences among respondents, which could bias and threaten the validity of results. Exposing the public to the technology at a very early stage on a small-scale under limited conditions (e.g. limited speed) makes it possible to gradually and slowly expose and familiarize the public with automated public transport technology. In this way the introduction of automated vehicles can be linked to the creation of realistic expectations, which have been defined as key driver of acceptance (Nees, 2016). People can experience these vehicles and also get to know their system limitations so that they develop a realistic idea of the current state of the art. Current system limitations include sharper braking in front of obstacles in close proximity to the automated shuttle as well as the transition to the manual mode in overtaking maneuvers or when the shuttle faces some localization problems. These technical difficulties are often not presented in the media, which is why people may develop overly unrealistic expectations that exceed the actual technological capabilities of the system.

4


Fig. 1. Driverless shuttle on EUREF campus in Berlin-Schöneberg

2. Methodology 2.1. Instrument development The 37-items-questionnaire reported in this study was administered together with a 15-items questionnaire to assess individual perceptions on certain technical features and the attractiveness of driverless shuttles as mobility offer in the city as well as in rural areas that was executed as part of a pilot project involving the testing of driverless shuttles on the EUREF campus in Berlin. The 37-items questionnaire asked respondents about their intended use of driverless shuttles in everyday conditions, the perceived usefulness (performance expectancy), ease of use (effort expectancy), social influence, level of trust, ecological norms, and perceived enjoyment. The original version of the questionnaire was in German, but was translated into English as the EUREF Campus attracts many international visitors and guests. Some of the items have been double checked before given their use in previous studies (e.g. Nordhoff et al., submitted; Madigan et al., 2016; Yap et al., 2016), which ensures their correct translation. 2.2. Instrument administration The operators onboard the driverless shuttle distributed the questionnaire to respondents by tablet computers, after they took a ride in the driverless shuttle. In this way, it was ensured that only respondents who tested the driverless shuttle filled out the questionnaire. Data collection took place in close proximity of the shuttle demonstration. All questionnaires were self-administered. The operators provided some assistance when respondents could not handle the technical functionality of the device. Data were collected between December 19th, 2016 and April 12th, 2017 where the vehicle provided daily shuttle operations from 9 to 17 hrs. The information was recorded 5


anonymously and no compensation was offered to respondents. 2.3. Operationalization of study variables Pre-validated measures were used for all constructs using multiple-item, four-point and six-point scales. The UTAUT constructs (performance expectancy, effort expectancy, social influence) were measured using at least two items of each construct of Venkatesh et al.’s (2003) measurement scale as well as the scales adopted and refined by Madigan et al. (2016). 4P acceptance and use was measured by behavioral intention, using Venkatesh et al.’s behavioral intention scale and some of the indicators regarding the use of automated vehicles as last mile transport by Yap et al. (2016). The items for the construct of behavioral intention were formulated in such a way to assess respondents’ behavioral intentions of driverless shuttles as an integral element of their daily lives beyond the demonstration. The items used to measure the willingness to share driverless shuttles were constructed on the basis of two different use cases for automated vehicles presented in a study by Krueger et al. (2016). We see the willingness to share driverless vehicles with other passengers having the same destination as indicator for their acceptance and use of driverless vehicles in public transport. Also, in Nordhoff et al. (2016), it was argued that acceptance is a multidimensional measure. 2.4. Analyses at the individual level Descriptive statistics (e.g. means, medians, standard deviations, and frequencies) were calculated for each of the variables. In the second step, spearman correlation coefficients were calculated between age, gender, and certain items measuring performance expectancy, effort expectancy and social influence as well as behavioral intention. In a later version when a more sufficient sample size is achieved, hierarchical multiple regression analysis will be carried out to examine more closely the relations between our independent and dependent study variables. 3. Results 3.1. Demographic profile of respondents In total, 318 respondents participated in the questionnaire. Of our current sample, 62.9% of respondents are male and 35.4% are female. The mean and median age of respondents was 35.11 and 32.00 years, respectively (SD= 14.13). 77.1% of respondents did not work on the EUREF campus, whereas 21% stated to do so. When being asked about their common transport mode respondents used on the EUREF campus, 70.4% of respondents stated to walk, 17.1% cycle, 7.7% use a conventional vehicle and 3.6% an electric vehicle. 95.7% already used the shuttle, while 3.1% of respondents did not use it before. 3.2. Analyses at the individual level: responses 6


In this section, the descriptive statistics and frequencies for the respondents will be reported. Generally, our study respondents indicated that the driverless shuttle is useful (M=5.12, SD=1.053, on a scale from 1-6) and would be an important part of the public transport system (M=4.77, SD=1.218, on a scale from 1-6). However, respondents do not necessarily think that the driverless shuttle is more efficient/faster than their current form of travel (M=2.47, SD=1.575, on a scale from 1-6), or better and more convenient than their current form of travel (M=3.19, SD=1.615, on a scale from 1-6), which is not surprising given the limited speed of the shuttle and its operation under restricted conditions. Concerning the ease of use, respondents generally stated that using the driverless shuttle is easy for them (M=4.66, SD=1.426, on a scale from 1-6), and that it would not take a lot of time to learn how to use driverless shuttles (M=4.96, SD=1.60, on a scale from 1-6). Furthermore, respondents believe that the use of the driverless shuttle is similar to current transport systems, such as the bus or train (M=4.15, SD=1.413, on a scale from 1-6). However, when compared to their current form of transport, respondents generally did not believe that driverless shuttles will be easier to use than their existing form of transport (M=3.92, SD=1.758, on a scale from 1-6), which is probably due to its limited operation and its linkage to the ODD as mentioned before. Concerning the role of social influence, respondents may not base their adoption decisions on the views of other people who are important to them. Thus, respondents do not generally prefer that people who are important to them (e.g. family, friends) adopt the driverless shuttle before they do (M=2.69, SD=1.791, on a scale from 1-6). They agree stronger with the statement that people who are important to them would like it when they would use the driverless shuttle (M=3.81, SD=1.507, on a scale from 1-6). These findings may correspond with the findings of a recent study (König & Neumayr, 2017), which found that self-driving cars apparently may not give people social recognition. When it comes to the level of trust, respondents generally feel comfortable in a vehicle without steering wheel, gas or brake pedals (M=4.55, SD=1.284, on a scale from 1-6), and believe that the driverless shuttle will be safe and reliable in severe weather conditions, such as snow, heavy rain or fog (M=4.55, SD=1.284, on a scale from 1-6). Respondents are less convinced that the driverless shuttle can drive safely and reliably without their assistance (M=3.56, SD=1.722, on a scale from 1-6), but generally felt safe throughout the whole trip (M=4.63, SD=1.404, on a scale from 1-6). Considering people’s need to control the shuttle, our study respondents generally do not want to manually steer the shuttle whenever they want (M=3.29, SD=1.898, on a scale from 1-6), but are more willing to use a button, which can be pressed to stop the shuttle in cases of emergency (M=5.19, SD=1.211, on a scale from 1-6). These findings overlap with König and Neumayr (2017) who defined the possibility to take over vehicle control as effective strategic option to increase trust in automated vehicles, and Sanaullah et al. (2017) who refer to the psychological need of people to have at least some level of control, which may explain why people generally prefer some intermediate automation levels to full automation (Schoettle & Sivak, 2015, 2016). Furthermore, our respondents moderately like the idea that the driverless shuttle operates at a low speed (M=3.28, SD=1.533, on a scale from 7


1-6), but agree stronger to the supervision of the shuttle by an operator throughout the whole trip (M=4.14, SD=1.609, on a scale from 1-6). Respondents believe that the use of driverless shuttles in public transport (M=5.25, SD=1.098, on a scale from 1-6) is the most attractive application scenario, followed by their use in rural areas (M=5.10, SD=1.222, on a scale from 1-6) and in the city (M=5.04, SD=1.234, on a scale from 1-6). As the automated shuttle is 100% electric, respondents were asked about the role of ecological norms for their choice of transport. A majority of respondents would like it to use a fully-electric driverless shuttle from the train station to their final destination (M=5.05, SD=1.177, on a scale from 1-6) and indicated that the protection of the environment would be crucial for their choice of the driverless shuttle (M=4.72, SD=1.366, on a scale from 1-6). Respondents are willing to use driverless shuttles when they are available on the market (M=5.15, SD=1.140, on a scale from 1-6), and on their daily mobility trips (M=4.25, SD=1.568, on a scale from 1-6), but are more hesitant to replace their current form of transport by an automated shuttle (M=3.55, SD=1.664, on a scale from 1-6). As regards the frequency of use, respondents can imagine using the driverless shuttle on average on 1-3 days a week (M=2.24, SD=1.218, on a scale from 1=daily or almost daily, 2=on 1-3 days a week, 3=on 1-3 days a month, 4= less than monthly, 5=never or almost never). They display a high willingness to use shared driverless shuttles and indicated that they like the idea to share the driverless shuttle together with a few unknown travelers having the same destination like them (M=4.74, SD=0.1.383, on a scale from 1-6), and to use it with other 6-8 passengers at the same time (M=5.35, SD=0.940, on a scale from 1-6). 3.3. Correlational analyses at the individual level The calculation of Spearman correlation coefficients found no statistical significant effects between gender and performance expectancy, effort expectancy, social influence on the one hand and between the willingness to use driverless shuttles in public transport on the other (behavioral intention). This is in line with prior research (Krueger et al., 2016; Zmud et al., 2016). A negative statistical significant correlation was found between age and the extent to which driverless shuttles are perceived to be easy to use (r=-0.147, p=0.01). Furthermore, the correlation between age and social influence is negative and statistical significant, indicating that for elderly people the influence of other people important to them matters less for their adoption decisions of automated vehicles (r=-0.146, p=0.01) The extent to which respondents consider the driverless shuttle to be useful and an important part of the traffic system is positively correlated to individuals’ behavioral intentions to use a driverless shuttle on their daily trips (correlation coefficient r=0.440, p=0.01; r=0.510, p=0.01). Additionally, individuals who consider driverless shuttles to be useful are more likely to use driverless shuttles 8


from the train station or some other public transport stop to the final destination and vice versa (r=0.443, p=0.01), or as mobility offer in the city (r=0.257, p=0.01). The perceived usefulness of driverless shuttles does not predict its use as mobility offer in rural areas, which will be further discussed in the remaining section. Individuals who believe that the use of the driverless shuttle is easier than their current form of transport are more likely to consider the choice of driverless shuttles on their daily trips (r=0.312, p=0.01). Individuals who would like it that people who are important to them adopt the driverless shuttle before they do it (social influence) are more likely to use driverless shuttles in rural areas (r=0.244, p=0.01), whilst social influence seems to be irrelevant for the use of driverless shuttles in the city. 4. Conclusions and discussion This is the second study to the best of our knowledge next to the study of Madigan et al. (2016) that investigates the acceptance of driverless vehicles in public transport that were tested by real users on semi-public roads. Our study respondents were generally very positive towards the use of driverless shuttles in public transport and can also imagine their use in the city as well as in rural areas. Also, the willingness to share driverless shuttles with other fellow travelers is very high among our respondents, which is seen as additional indicator of acceptance and use. Moreover, this study confirms the importance of the UTAUT constructs performance, effort expectancy and social influence for individuals’ behavioral intentions. Our study respondents believed that driverless shuttles will be useful and easy to use, but when comparing them to their current mode of transport, the perceived usefulness and ease of use declines. This means that our research findings should also be interpreted with the current restrictions under which driverless shuttles currently operate (e.g. limited speed). Future research is needed to re-visit individuals’ perceptions as regards automated shuttles when their use is more widespread, and when they operate under less restrictive conditions on public roads. The desire of individuals to take over control from the vehicle needs to be considered with care as the human intervention in the vehicle’s trajectory may seriously jeopardize traffic safety and nullify the effects of vehicle automation altogether. Therefore, a safety button inside the driverless vehicle may be a good way to increase trust among respondents and their acceptance, but only if it’s safe and proper use can be warranted under all conditions. Its misuse can have serious counter effects and may hamper acceptance in the end. Thus, the present study sheds light on the design implications to make driverless vehicles a success. Furthermore, the use of driverless shuttles in rural areas was not predicted by perceived usefulness, which may be due to the higher car dependency in these areas. These findings are in line with a recent study (König & Neumayr, 2017), which investigated the attitudes of 489 respondents from 33 countries by an online survey and found that urban residents tended to be more open towards 9


automated driving than rural residents. Therefore, to attract rural residents to driverless shuttles, it would be important to increase the attractiveness of public transport systems and reduce the car dependency in these areas. The popularity of new mobility services (e.g. carsharing, scooter sharing) can be improved so that rural residents are also aware of other possible mobility options next to their private car. Driverless shuttles can improve the accessibility of transport stops, providing seamless, reliable, frequent and around-the-clock transport so that individuals in rural areas rely less on motorized forms of travel. Generally, no statistical significant gender-specific effects were found, which mirrors current literature. A negative statistical significant correlation was found between age and the extent to which driverless vehicles are perceived to be easy to use, which shows that elderly people find automated shuttles less easy to use. More sophisticated analyses will be performed in a consecutive study to model the correlations of the key constructs in this study by applying a more powerful multivariate statistical analysis. The limitations of the study need to be taken into account when analyzing the results. First, the vehicle currently operates under restricted conditions (e.g. limited speed), which are likely to influence individual perceptions of the vehicle itself and may make it more difficult for our study respondents to envision its use as daily transport mode. Second, the sample of respondents may be skewed towards individuals that are more innovative and open towards sustainable mobility innovations as the EUREF campus assembles many companies whose business model is centered on the conceptualization and implementation of sustainable mobility innovations. Third, the views of rural residents are likely to be underrepresented in this study sample as the pilot project takes place in Berlin.

10


References 1.

AAA.

(2016).

Fact

Sheet:

Vehicle

Technology

Survey.

http://publicaffairsresources.aaa.biz/wpcontent/uploads/2016/02/Automotive-Engineering-ADAS-Survey-Fact-Sh eet-FINAL-3.pdf. Accessed April 11, 2016. 2. Abraham, Lee, Brady, Fitzgerald, Mehler, Reimer, and Coughlin. (2016). Autonomous Vehicles, Trust, and Driving

Alternatives:

A

survey

of

consumer

preferences.

http://agelab.mit.edu/files/publications/2016_6_Autonomous_Vehicles_Consumer_Preferences.pdf. Accessed September 12, 2016. 3. Bansal, P., Kockelman, K.M., & Singh, A. (2016). Assessing public opinions of and interest in new vehicle technologies: An Austin perspective. Transportation Research Part C, 67, 1-14. 4. Bazilinskyy, P., Kyriakidis, M., and De Winter, J. (2015). An internal crowdsourcing study into people’s statements on fully automated driving. Procedia Manufacturing, 3, 2534-2542. 5.

Continental

(2013).

Mobility

Study

2013.

http://www.continentalcorporation.com/www/download/pressportal_jp_jp/general/ov_automated_driving_jp /ov_mobility_ study_jp/download_channel/pres_mobility_study_jp.pdf. Accessed September 22, 2016. 6.

Continental

(2015).

Continental

Mobility

Study

2015.

https://www.contionline.com/www/download/pressportal_com_en/themes/initiatives/chanel_mobility_study _en/ov_mobility_study2015_en/dow. Accessed September 22, 2016. 7. Danise, A. (2015). Women Say No Thanks to Driverless Cars, Survey Finds; Men Say Tell MeMore. https://www.nerdwallet.com/blog/insurance/survey-consumer-fears-self-driving-cars/. Accessed April 11, 2016. 8. Detecon International GmbH. (2016). Autonomes Fahren: Wenn das Lenkrad zur Sonderausstattung wird: Eine

empirische

Untersuchung

der

Akzeptanz

autonom

fahrender

Fahrzeuge.

https://www.detecon.com/sites/default/files/gladbach_-_studie_autonomes_fahren_final.pdf. Accessed April 21, 2016. 9. Fraedrich, E., Cynganski, R., Wolf, I., and Lenz, B. (2016). User Perspectives on Autonomous Driving: A Use-Case-Driven

Study

in

Germany.

https://www.geographie.huberlin.de/de/institut/publikationsreihen/arbeitsberichte/download/Arbeitsberichte _Heft_187.pdf. Accessed August 02, 2016. 10. Hoedemaeker, M., & Brookhuis, K. A. (1998). Behavioural adaptation to driving with an adaptive cruise control (ACC). Transportation Research Part F: Traffic Psychology and Behaviour, 1(2), 95-106. 11. Hohenberger, C., Sporrle, M., Welpe, I.M. (2016). How and why do men and women differ in their

11

User Acceptance of Driverless Shuttles Running in an Open and Mixed Traffic Environment willingness to use automated cars? The influence of emotions across different age groups. Transportation Research Part A, 94, 374-385. 12. Krueger, R., Rashidi, T.H., and Rose, J.M. (2016). Preferences for shared autonomous vehicles. Transportation Research Part C, 69, 343- 355. 13. Kyriakidis, M., Happee, R., and De Winter, J.C.F. (2015). Public opinion on automated driving: Results of an international questionnaire among 5.000 respondents. Transportation Research Part F: Traffic Psychology and Behavior, 32, 127-140. 14. König, M. & Neumayr, L. (2017). Users’ resistance towards radical innovations: The case of the self-driving car. Transportation Research Part F, 44, 47-52. 15. Louw, T., & Merat, N. (2017). Are you in the loop? Using gaze dispersion to understand driver visual attention during vehicle automation. Transportation Research Part C: Emerging Technologies, 76, 35-50. 16. Madigan, R., Louw, T., Dziennus, M., Graindorge, T., Ortega, E., Graindorge, M., and Merat, N. (2016). Acceptance of Automated Road Transport Systems (ARTS): An adaptation of the UTAUT model. Proceedings of the 6th Transport Research Arena, April 18-21, Warsaw, Poland. 17. Merat, N., & Jamson, A. H. (2009). How do drivers behave in a highly automated car. In Proceedings of the 5th international driving symposium on human factors in driver assessment, Training and Vehicle Design (pp. 514-521). 18. Merat, N., Jamson, A. H., Lai, F. C., & Carsten, O. (2012). Highly automated driving, secondary task performance, and driver state. Human factors, 54(5), 762-771. 19. Milakis, D., Snelder, M., Van Arem, B., Van Wee, B., Correira, G.H.D.A. (2017). Development and transport implications of automated vehicles in the Netherlands: scenarios for 2030 and 2050. EJTIR 17, 1, 63-85. 20. Nordhoff, S., Van Arem, B., and Happee, R. (2017). A Conceptual Model to Explain, Predict, and Improve User Acceptance of Driverless 4P Vehicles. Transportation Research Records, 2602-2608. 21. Nordhoff, S., Kyriakidis, M., Van Arem, B., Ruhrort, L., Graff, A., & Happee, R. (not yet submitted). Modelling Acceptance of Driverless Vehicles in Public Transport: Results of an International Survey with 10.001 Respondents Worldwide. 22. Nees, M.A. (2016). Acceptance of Self-Driving Cars: An Examination of Idealized versus Realistic Portrayals with a Self-Driving Car Acceptance Scale. Proceedings of the Human Factors and Ergonomics Society 2016 Annual Meeting, pp. 1449-1453. 23. Rudin-Brown, C. M., & Parker, H. A. (2004). Behavioural adaptation to adaptive cruise control (ACC): implications for preventive strategies. Transportation Research Part F: Traffic Psychology and Behaviour, 7(2), 12

User Acceptance of Driverless Shuttles Running in an Open and Mixed Traffic Environment 59-76. 24. Payre, W., Cestac, J., and Delhomme, P. (in press). Intention to use a fully automated car: Attitudes and a priori acceptability. Transportation Research Part F: Traffic Psychology and Behavior, 27, 252-263. 25. Payre, W., Cestac, J., and Delhomme, P. (2016). Fully Automated Driving: Impact of Trust and Practice on Manual Control Recovery. Human Factors, 58 (2), 229-241. 26. Sanaullah, I., Hussain, A., Chaudhry, A., Case, K., & Enoch, M. Autonomous Vehicles in Developing Countries: A Case Study on User’s View Point in Pakistan. In: Advances in Human Aspects of Transportation, Advances in Intelligent Systems and Computing 484. Springer International Publishing Switzerland, 2017, pp. 561-569. 27. Schoettle, B., & Sivak, M. (2014). A Survey of Public Opinion about 26 Autonomous and Self-Driving Vehicles in

the

U.S.,

the

U.K.,

and

Australia.

27

Michigan,

USA.

http://deepblue.lib.umich.edu/bitstream/handle/2027.42/108384/103024.pdf?sequence=1&isAllowed=y. Accessed January 2, 2015. 28. Schoettle, B., and Sivak, M. (2015). Motorists’ Preferences for Different Levels of Vehicle Automation. https://deepblue.lib.umich.edu/bitstream/handle/2027.42/114386/103217.pdf?sequence=1&isAllowed=y. Accessed July 22, 2015. 29. Schoettle, B., and Sivak, M. (2016). Motorists’ Preferences for Different Levels of Vehicle Automation: 2016. http://www.umich.edu/~umtriswt/PDF/SWT-2016-8.pdf. Accessed July 22, 2015. 30. Schaefer, K.E., and Straub, E.R. (2016). Will Passengers Trust Driverless Vehicles? Removing the steering wheel and pedals. 2016 IEEE International Multidisciplinary Conference on Cognitive Methods in Situation Awareness and Decision Support (CogSIMA). 31. Souders, D., and Charness, N. Challenges of Older Drivers’ Adoption of Advanced Driver Assistance Systems and Autonomous Vehicles. In: Human Aspects of IT for the Aged Population. Healthy and Active Aging. Lecture Notes in Computer Science. Springer International Publishing Switzerland, 2016, 9755, pp. 428-440. 32. Tennant, C., Howard, S., Franks, B. Bauer, M.W., Stares, S., Pansegrau, P., Stysko-Kunkowska, M. and Cuevas-Badallo, A. (2016). Autonomous Vehicles – Negotiating a Place on the Road: A study on how drivers feel about

Interacting

with

Autonomous

Vehicles

on

the

Road.

Executive

Summary.

www.lse.ac.uk/newsAndMedia/PDF/AVs-negotiating-a-place-on-the-road-1110.pdf. Accessed October, 31, 2016. 33.

TNS

Opinion

&

Social.

(2015).

Special

Eurobarometer

427

„Autonomous

Systems”.

http://ec.europa.eu/public_opinion/archives/ebs/ebs_427_en.pdf. Accessed September 30, 2016 34. Wintersberger, P., Frison, A.-K., Riener, A. & Boyle, L.N. (2016). Towards a Personalized Trust Model for Highly Automated Driving. In: Weyers, B. & Dittmar, A., Mensch und Computer 2016 –Workshopband. Aachen:

13

User Acceptance of Driverless Shuttles Running in an Open and Mixed Traffic Environment Gesellschaft für Informatik e.V. 35. Van den Beukel, A. P., & Van der Voort, M. C. (2017). How to assess driver's interaction with partially automated driving systems–A framework for early concept assessment. Applied ergonomics, 59, 302-312. 36. Yap, M.D., Correira, G., and Van Arem, B. (2015). Valuation of travel attributes for using automated vehicles as egress transport of multimodal train trips. Transportation Research Part A: Policy and Practice, 94, 1-16. 37. Zmud, J., Sener, I., and Wagner, J. (2016). Consumer Acceptance and Travel Behaviour: Impacts of Automated Vehicles. http://d2dtl5nnlpfr0r.cloudfront.net/tti.tamu.edu/documents/PRC-15-49-F.pdf. Accessed April 13, 2016.4

14