ArticlePDF Available

Abstract and Figures

Technological advances have recently led to the development of urban air mobility (UAM), an alternative transportation mode with several concepts including vehicles operated by on-demand fully-automated vertical takeoff and landing aircraft (VTOL) for intra-city passenger transportation. However, despite a growing interest in UAM, understanding users' perceptions to it remains limited. This research aims to identify and quantify the factors affecting the adoption and use of UAM, based on relevant tools from the literature, such as recurring factors in studies on aerial vehicle concepts, ground autonomous vehicles, but also acceptance models, such as the Technology Acceptance Model by Davis et al. (1989). A stated-preference survey was developed to assess the perception of users in terms of adoption time horizon, including options such as the first six years of the service's implementation, "unsure", and "never". The obtained results were evaluated using exploratory factor analyses, and the specification and estimation of suitable discrete choice models, multinomial logit models (MNLs) and ordered logit models (OLMs), with adoption time horizon as dependent variable. Findings revealed the importance of safety and trust, affinity to automation, data concerns, social attitude, and socio-demographics for adoption. Factors, such as the value of time savings, the perception of automation costs, and service reliability, were also found to be highly influential. There was also an indication that skeptical respondents, i.e. answering "unsure", had a behavior similar to late and non-adopters, i.e. adoption time horizon higher than six years or answering "never". The summarized results were represented in an extended Technology Acceptance Model for urban air mobility, and provided insights for policymakers and industrial stakeholders.
Content may be subject to copyright.
Factors affecting the adoption and use of urban air mobility
Christelle Al Haddada,
, Emmanouil Chaniotakisb, Anna Straubingerc, Kay Pl¨otnerc, Constantinos
aChair of Transportation Systems Engineering, Technical University of Munich, Arcisstrasse 21, 80333 Munich, Germany
bMaaSLab, Energy Institute, University College London, 14 Upper Woburn Place, WC1H0NN London, UK
cBauhaus Luftfahrt e.V., Willy–Messerschmitt–Strasse 1, 82024 Taufkirchen, Germany
Technological advances have recently led to the development of urban air mobility (UAM), an alternative
transportation mode with several concepts including vehicles operated by on-demand fully-automated verti-
cal take-off and landing aircraft (VTOL) for intra–city passenger transportation. However, despite a growing
interest in UAM, understanding users’ perceptions to it remains limited. This research aims to identify and
quantify the factors affecting the adoption and use of UAM, based on relevant tools from the literature,
such as recurring factors in studies on aerial vehicle concepts, ground autonomous vehicles, but also accep-
tance models, such as the Technology Acceptance Model by Davis et al. (1989). A stated–preference survey
was developed to assess the perception of users in terms of adoption time horizon, including options such
as the first six years of the service’s implementation, “unsure”, and “never”. The obtained results were
evaluated using exploratory factor analyses, and the specification and estimation of suitable discrete choice
models, multinomial logit models (MNLs) and ordered logit models (OLMs), with adoption time horizon
as dependent variable. Findings revealed the importance of safety and trust, affinity to automation, data
concerns, social attitude, and socio–demographics for adoption. Factors, such as the value of time savings,
the perception of automation costs, and service reliability, were also found to be highly influential. There
was also an indication that skeptical respondents, i.e. answering “unsure”, had a behavior similar to late
and non-adopters, i.e. adoption time horizon higher than six years or answering “never”. The summarized
results were represented in an extended Technology Acceptance Model for urban air mobility, and provided
insights for policymakers and industrial stakeholders.
Keywords: urban air mobility, adoption, acceptance, technology acceptance model, perception, stated
preference, exploratory factor analysis, discrete choice modeling
1. Introduction
Shared mobility services are providing users with more efficient travel options, characterized by lower
demand for parking spaces, lower vehicle ownership, but also reduced environmental impacts resulting from
lower emissions (Baptista et al.,2014). At the same time, autonomous vehicles promise safe and comfortable
transportation, and most automobile manufacturers are likely to launch fully autonomous vehicles in the5
coming decade (Bimbraw,2015). This trend has led to a research interest in ground shared autonomous
mobility (Fagnant & Kockelman,2014), and possibly to the exploration of the third dimension: the skyscape.
This has been reflected in aerial vehicle concepts for passenger transportation, also called as “urban air
mobility” (UAM). The concepts for short–haul passenger air trips are facilitated by technological advances
Corresponding author
Email addresses: (Christelle Al Haddad), (Emmanouil
Chaniotakis), (Anna Straubinger), (Kay
Pl¨otner), (Constantinos Antoniou)
Preprint submitted to Transportation Research Part A; https: // doi. org/ 10. 1016/ j. tra. 2019. 12. 020 January 2, 2020
in terms of battery storage, electrical power transmission and distributed propulsion systems (Shamiyeh10
et al.,2017). The US National Aeronautics and Space Administration (NASA) is developing a framework for
the integration of urban air mobility airspace research with the efforts of different partners and stakeholders
(Thipphavong et al.,2018). For instance, Airbus (2018) introduces urban air mobility as the integration of
the third dimension to urban transport networks, including the on–demand sharing mobility operated by
vertical take–off and landing aircraft (VTOL) for intra–city passenger trips, with a long–term vision entailing15
electrical self–piloted VTOLs, such as Vahana and CityAirbus demonstrators, for one or more passengers
respectively. The introduction of urban air mobility, according to Airbus (2017), promises to offer a safer,
more reliable, and more environmental alternative to alleviate congestion on transport networks. Currently,
Voom (an Airbus company) is providing on–demand helicopter bookings in megacities like Mexico City and
ao Paulo, laying the ground for further developments in urban air mobility (Airbus,2018). Uber has also20
introduced air taxi models in their business plans, such as the economic model assuming a four–seat capacity
(including the pilot if the vehicle is not self–piloted), for which passengers have the possibility to share the
flight and thereby save on the ride cost (Uber Elevate,2016). Porsche Consulting (2018) defines the process
of this service as follows: starting from the first mile (access to the vertiport), followed by the boarding
on the vehicle, the flight time (including take–off and landing), the de–boarding, and finally the last mile25
transfer; the service can be booked online and the assumed boarding and de–boarding times are around
three minutes each. In this study, Porsche Consulting (2018) presents three e–VTOL concepts with different
certification times, travel speeds (from 70 to 300 km/h), routes, and purposes (intra– and/or intercity trips),
projecting for 2035 a passenger market of around 23 000 aircraft and worth $32 Billion.
Despite different model definitions, urban air mobility is constrained by many aspects related to regula-30
tions, infrastructure availability, air traffic control, environmental impacts, but also community acceptance
(Vascik,2017). In terms of infrastructure challenges, Cohen (1996) studied vertiport prototypes to optimize
land use, site selection, and community acceptance and more recently Fadhil (2018) focused on a GIS–based
analysis for ground infrastructure selection for urban air mobility. In modeling UAM impacts, the service
implications on the inhabitants and the changes on the city were evaluated (Straubinger & Verhoef,2018),35
but also an agent–based simulation and an initial analysis of the service’s performance were conducted
(Rothfeld et al.,2018).
To better understand the service demand and to be able to predict it, understanding consumer adoption
is crucial; in other words, if and when users are going to use urban air mobility. For instance, a study from
the Georgia Institute of Technology designed a survey to collect responses from 2,500 high–income workers40
in different areas of the US to predict their demand for eVTOL in urban areas, with flights ranging up to
90 km (Garrow et al.,2018). Passenger adoption in a UAM environment was also explored using choice
modeling in a case study based in Munich (Fu et al.,2019). Setting the context for the characteristics that
imply the use of VTOL, Antcliff et al. (2016) proposes Silicon Valley as an ideal region for early introduction
of on–demand civil VTOL operations, mostly due to good weather, high housing prices, high incomes, but45
also number of hyper commuters (those who commute two or more hours per day) who would greatly reduce
their time in congestion (Antcliff et al.,2016). In Germany, research on UAM has targeted the city of
Ingolstadt as part of the “Urban Air Mobility Initiative”, supported by the European Commission.
Vascik & Hansman (2018) emphasized the significance of noise and altitude levels as areas of concerns
for community acceptance of urban air mobility systems. To the best of the authors’ knowledge, however,50
the acceptance of urban air mobility has not yet been investigated in terms of users’ profiles (not related
to the service characteristics) outside of a mode choice context, and in relation to adoption time horizon.
This paper helps to fill this gap by developing a framework for UAM adoption and use, assuming an on-
demand intra–city passenger transportation, operated by fully-automated electric VTOL, integrated with
public transportation systems, with the possibility to ride-pool (share the ride with other passengers to55
save money). This is based on the development of a novel stated preference survey design with adoption
time horizon as a dependent variable, aiming to reveal factors affecting adoption of this service. Futuristic
scenarios were introduced to evaluate a hypothetical system starting in 2030, aiming to better understand
if and when users would adopt them, but also the barriers and opportunities from a behavioral perspective.
Accordingly, factors were identified by exploring those found to be significant in the perception of aerial60
vehicle concepts, in technology and automation acceptance models, and affecting automation acceptance in
transport. This research contributes to the existing body of literature on the acceptance and adoption of
new forms of mobility by presenting a novel methodological approach using a non–conventional dependent
variable for ordered logit models: time variable in a scale including a non–ordinal parameter, such as
“unsure”, as to the best of the authors’ knowledge these models are mostly used with ordinal scales, such as65
Likert scales, for attitudes and satisfaction. The obtained results are represented in a Technology Acceptance
Model for urban air mobility that could be generalized for disruptive transport technologies. The significantly
identified factors offer a good basis for the investigation of more realistic scenarios in a nearer future, as
they show an aptitude of users towards them.
The remainder of this paper is structured as follows. A literature review is first presented, followed70
by the methods used including the survey design, descriptive analysis and model development, and study
limitations. The data collected and main statistical results are then derived. After that, the developed and
estimated models are shown, namely the exploratory factor analyses and discrete choice models. Then, a
discussion of the main survey findings is presented along with the proposed Technology Acceptance Model
for Disruptive Transport Technologies, and the policy level insights and recommendations. Industrial stake-75
holders could thereby appropriately care for the needs of different classes of users, while policymakers could
benefit from these factors by setting guidelines or more stringent regulations if required. Finally, a conclusion
is drawn with some recommendations on future relevant work.
2. Literature review
To examine the adoption of a new service, pertinent literature suggests that it is necessary to understand80
the conditions for trusting this service and accepting its operational characteristics. In the case of UAM,
this inevitably passes through the investigation of technology and automation acceptance models, and of
factors commonly affecting the acceptance of automation in transport. As only a few studies have explicitly
focused on consumer adoption of urban air mobility, examining studies on community acceptance of other
aerial vehicles or services might be useful despite its limitations.85
2.1. Factors affecting aerial vehicles perception
Recent studies commissioned by NASA investigated market potentials for urban air mobility, focusing
on the different barriers to its implementation, including community acceptance and key factors for its
users and non–users’ adoption. The study commissioned to Crown Consulting studied the viability of three
UAM use cases in 15 US cities (NASA,2019): last mile delivery (for packages), air metro (an autonomous90
public transit style commuter system) and air taxi (an autonomous on-demand ridesharing system). This
study revealed that nearly half of all surveyed consumers were potentially comfortable with delivery and
UAM use cases. Major concerns presented were safety, privacy, job security, environmental threats and
noise and visual pollution. Moreover, cybersecurity, affordability and willingness to pay were perceived as
barriers to the UAM viability. The study commissioned to Booz Allen Hamilton (NASA,2018) looked at95
three potential UAM markets (airport shuttle, air taxi and air ambulance) in ten target urban areas, to
explore market size and barriers, understand the potential viability of the use cases, but also the societal
and environmental impacts of UAM; this study included a survey, focus groups, and stakeholder interviews.
Societal barriers were analyzed by looking at both user and non-user perspectives. Key factors were safety
(including passengers’ need for security screening), privacy and noise, preference for piloted aircraft, impact100
of having a flight attendant, cybersecurity, cost and convenience, but also socio–demographics; men (along
with millennials and younger respondents) were found to be more comfortable and willing to use UAM.
In a stated preference survey based in Munich, Fu et al. (2019) explored transportation mode preferences
in an Urban Air Mobility environment including four alternatives: private car, public transportation, au-
tonomous taxi, and autonomous flying taxi (or UAM). The study results indicated that travel time, travel105
cost, and safety might be critical for determining UAM adoption. Socio–demographics were also found as
highly influential in UAM use, as market segments showed that younger individuals as well as older ones
with higher household income are more likely to adopt UAM. Also, trip purpose proved to be significant,
with non–commuting being the respondents’ most preferred option.
Research on consumer acceptance of unmanned civilian drones identified common concerns among people,110
such as privacy (Wang et al.,2016;Clothier et al.,2015), risks associated with accidents, and security
regarding the recognition of drones used for emergency (Lidynia et al.,2017). Additionally, Chamata (2017)
identified social and economic concerns as factors delaying the adoption of civilian drones. According to
the trip purpose, the community acceptance of drones might differ (Boucher,2016). For unmanned aerial
vehicles (UAV), MacSween (2003) found that emotional and safety data increased users’ persuasions of UAV115
applications, including commercial, cargo, and passenger transportation. Whereas the above concerns apply
to community acceptance of drones, they could be very relevant to consumer adoption of UAM, as they all
discuss users’ perceptions of unmanned aviation.
A stated preference study by Peeta et al. (2008) investigating the travel propensity towards air taxi ser-
vice, defined as very light jet–based on–demand air service (ODAS), predicted the probability of individuals120
of switching from their usual mode of intercity transportation to an on-demand air service (ODAS) given a
set of scenarios of travel distance and cost. The adoption of ODAS was found to be highly dependent on
travel distance, service fare, and level of accessibility. Kreimeier & Stumpf (2017) found cost to be crucial
in determining consumer adoption in Germany. The market share for thin–haul on–demand air mobility
services has been estimated as 19% or 235 million trips per year, based on a choice model assuming 0.1125
AC /km costs gap between UAM services and cars; a sensitivity analysis showed that this volume would drop
to 24 million for a gap of 0.2 AC /km. This market would be potentially viable for distances above 100 km,
with 60% of the market expected for distances between 200 and 400 km (inter–city). Both studies by Peeta
et al. (2008) and Kreimeier & Stumpf (2017) presented inter–city applications or thin–haul flights, which
are arguably different than intra–city application of UAM; however, they still highlighted cost and time as130
significant factors for consumer adoption of on–demand aerial mobility.
2.2. Technology and automation acceptance models
Technology acceptance has long been an area of interest in research with the main purpose of modeling
how people accept and use a technology. Perhaps the most renown technology acceptance model is the
Technology Acceptance Model (TAM) by Davis et al. (1989), developed to investigate technology use of135
information systems, particularly computer technology, in which the correlation between the intention to
use and actual usage is measured. The main aim of the model is to present a framework for modeling the user
acceptance in terms of factors that influence his or her decision in using the technology. This model is based
on two main constructs: the perceived usefulness (PU) and the perceived ease of use (PEU), where PEU
reinforces PU. The former is the extent to which the user believes the technology use would enhance his or her140
job performance, whereas the latter is the degree to which using the technology requires effort. Both factors
determine the user attitude towards using the system, which in turn determine the behavioral intention (BI)
to use the system, and then the actual system use. This model also includes external variables, which affect
the defined constructs. Based on this model, the Technology Acceptance Model 2 (TAM2) was extended by
adding variables grouped into social influence and cognitive processes (Venkatesh & Davis,2000), impacting145
the perceived usefulness of the technology. Social influence includes subjective norms, voluntariness, image;
cognitive processes included job relevance, output quality, and result demonstrability. In TAM2, these main
factors directly influence user perception (PU and PEU). A further revision extended TAM into the Unified
Theory of Acceptance and Use of Technology (UTAUT), (Venkatesh et al.,2003), and a later revision into
TAM3 (Venkatesh & Bala,2008). According to Ghazizadeh et al. (2012), the UTAUT offers a more complete150
account of acceptance in comparison with TAM2. In UTAUT, three constructs directly determine BI:
performance expectancy, effort expectancy, and social influence (as defined in TAM2); facilitating conditions
is a fourth construct that directly determines the use behavior. Gender, age, experience, and voluntariness
moderate the impact of the four constructs. The role of moderating factors (factors impacting adoption and
reducing the limitations of the models’ explanatory power) has been found crucial (Sun & Zhang,2006) in155
technology acceptance models; for instance, the moderating effects of gender and self–efficacy in the context
of mobile payment adoption (Riad et al.,2014).
To combine constructs from TAM models with those from the Cognitive Engineering literature, Ghaz-
izadeh et al. (2012) proposed a framework for automation acceptance model (AAM), by adding two main
constructs: compatibility (namely task–techhnology compatibility) and trust. The former includes previous160
experience with automation, existing work practices, work style, and values; the latter (trust) influences BI
directly and indirectly through PU and PEU. The model is also reinforced by the feedback mechanisms of
the different constructs. Zhang et al. (2019) used structural equation modeling to test hypotheses on their
proposed autonomous vehicles acceptance model, in which initial trust is the most critical factor in promot-
ing a positive attitude towards using autonomous vehicles, which along with perceived usefulness determine165
the behavioral intention (to use autonomous vehicles). In this model, initial trust is affected by perceived
usefulness, perceived ease of use, but also two types of risk: the perceived safety risk and the perceived
privacy risk; initial trust can be improved by enhancing perceived usefulness or by reducing perceived safety
risk. More recently, a study at the University of St.Gallen in Switzerland developed an adapted TAM for
personal autonomous mobility, focusing on ground autonomous vehicles (Jenkins et al.,2018). The main170
additions in automation–related models could serve as a motivation for the extension or application of an
urban air mobility acceptance model.
2.3. Factors affecting automation acceptance in transport
Social acceptance is a widely researched area in the field of automation, particularly for ground au-
tonomous vehicles1. In a study on the trust and acceptance of shared autonomous vehicles (SAE2level 4),175
Merat et al. (2016) summarized the factors influencing the use of conventional ride–sharing vehicles, but also
those affecting the acceptance and trust of robotic systems, where trust includes reliability and safety. Choi
& Ji (2015) also noted the importance of trust in adopting autonomous vehicles; by conducting a survey
of 552 drivers and analyzing the data using partial least squares, they noted the importance of perceived
locus of control or situation management for gaining trust. Nees (2016) found that realistic expectations,180
and thereby perceived reliability, are influential in user acceptance of autonomous vehicles. In urban areas
in particular, passenger security as part of perceived safety has been noted as crucial for gaining trust (and
thereby the implementation and adoption of AVs); especially during the night (Piao et al.,2016). The user’s
feeling of safety could depend on the vehicle’s interior (Merat et al.,2016), but might also be hindered
by cyber–security concerns (Kyriakidis et al.,2015). In Germany, 90% of the respondents of a study on185
autonomous driving stated that they would feel safer, if they were able to intervene or control the vehicle
at any time, or at least in case of emergency (Deloitte Analytics Institute,2017). Control has also been
associated with feeling more independent (Gaggi,2017) and proved to positively impact the user’s perceived
ease of use (odel et al.,2014). Overall, users have to be aware and convinced of the benefits of the new
technology in order to be able to trust it and therefore use it. Manufacturers’ reputation could be a positive190
factor in gaining this trust (Deloitte Analytics Institute,2017). Real–life tests for autonomous vehicles and a
higher transparency in demonstrating the new modes could also lead to a higher trust through an increased
user awareness (Bjørner,2015;Choi & Ji,2015).
A higher perception of the benefits of automation could help users gain trust to use the technology, which
is reflected in a higher perceived usefulness. A decrease in congestion and improvement of road safety are195
associated with social benefits (Kaan,2017) resulting from the reduced number of road crashes. Improved
mobility for mobility–impaired users (Clements & Kockelman,2017) can be seen as another social benefit.
Moreover, the trip purpose could play a role in behavioral intention. Users might be more interested in
using autonomous vehicles for leisure trips, after alcohol consumption (Connected Automated Driving EU,
2017), or if the intended trip is unpleasant (Deloitte Analytics Institute,2017). The perceived advantages200
of automation may lead to a higher perceived usefulness, resulting in a higher user acceptance. Ease of
use in terms of effort expectancy in the case of autonomous vehicles, or ease of access for dependents or
mobility–impaired users (Merat et al.,2016) might influence the intention to use autonomous vehicles. For
UAM, however, the perceived ease of use (PEU) is not related to effort expectancy or driver intervention,
like in the case of ground AVs, as users are not expected at any time to take control of the air vehicle;205
instead PEU could be translated to simply the booking and boarding processes, since these are the only
tasks expected from the users.
1For practical reasons, the term “autonomous vehicles” will be used to refer to autonomous ground vehicles, and abbreviated
with AV. Otherwise, the term urban air mobility or on–demand air mobility will be used to refer to autonomous air vehicles.
2Society of Automotive Engineers
Attributes, such as time (Krueger et al.,2016) and costs (Merat et al.,2016), were found to be highly
influential in the acceptance of autonomous vehicles. Comfort (Rychel,2016), vehicle cleanliness (Merat
et al.,2016), and availability in different weather conditions (Merat et al.,2016) proved to impact the210
intention to use; the latter may also be influenced by social behavior, including the willingness to share
the ride with strangers (Merat et al.,2016) and the perceived fun of driving or driving enjoyment (Bjørner,
2015). Social attitudes can also include concerns from the loss of jobs induced by automation (Deloitte
Analytics Institute,2017), concerns regarding terrorism, crime, or cyber–security in general, data concerns,
such as data protection, data use, or privacy in broader terms (Kyriakidis et al.,2015), but also cultural215
values since acceptance and adoption vary globally. In industrialized countries, for example, automated
vehicles might face higher skepticism compared to emerging countries like India or China (Rychel,2016).
As in the former accident rates are lower, due to higher measures of safety, people might be more reluctant
towards automation, as they are not necessarily convinced of its benefits. Also, in more developed countries,
users are less likely to be comfortable with their data being shared or used (Kyriakidis et al.,2015).220
Socio–demographic factors were mentioned in several studies focusing on automation perception (odel
et al.,2014;Kyriakidis et al.,2015;Payre et al.,2014). Women were often found to have a lower intention to
use autonomous vehicles, compared to their male counterparts (Hohenberger et al.,2016); in contrast, young
individuals (Deloitte Analytics Institute,2017) and individuals with multimodal travel patterns (Krueger
et al.,2016) were found to be more likely to adopt shared autonomous vehicles. Similarly, technology225
awareness was a positive impact on autonomous vehicles acceptance; this includes, for instance, having
heard of Google cars (Bansal et al.,2016), autonomous vehicles (Schoettle & Sivak,2014), and/or previous
experience with advanced driver–assistance systems (ADAS).
3. Methods
3.1. Survey design230
As the service in question is not yet operational and can not be observed or used, a Stated Preference (SP)
survey was designed to identify the most influential factors in UAM’s adoption time horizon. The survey
was structured in four parts, with a total of 31 questions (or question groups3). The study particularly
focused on the Munich region, and therefore the survey was available in both languages: English (EN) and
German (DE).235
The first and last parts of the survey targeted respondents’ travel behavior and socio–demographics,
respectively. In the first part, respondents were asked about their commute behavior, their satisfaction with
the public transportation system of their region, and their attitudes towards automated systems (enjoyment,
trust, perceived usefulness, and previous experience with automation). As these attitudes are commonly
found in technology acceptance and use, but are rather difficult to measure directly, they were assessed240
through (five–point) Likert scale agreement statements (Likert,1932), with options ranging from “strongly
disagree” to “strongly agree”; answers were verbally labelled and midpoints (“neither disagree nor agree”
used for attitude measures) were used for attitude measures, as suggested by Dolnicar (2013). These
statements included: “I enjoy interacting with automated systems, such as Siri (Apple) or Alexa (Amazon)”,
“I trust such automated systems”, “I think driver assistance systems, such as adaptive cruise control, lane245
keeping assistance, or other advanced systems, are useful”, “I have used (in my own car or someone else’s)
such advanced driver-assistance systems”. Socio–demographic questions included age, gender, household
size, disability in household, education level, main occupation, household income, and current residence
location (city and country), with a “prefer not to answer” option for these. As stated preference studies
are susceptible to anchoring bias (McFadden,2001), socio–demographic questions were placed at the end of250
the survey, so that people would not be biased in answering them according to what they think would be
consistent with their choices. This can also overcome the stereotype threat (Steele,1995) in which people
try to avoid the confirmation of any stereotype they feel threatened by.
3A question consisted sometimes of a matrix including several agreement statements, focusing on one attribute for instance.
In the second part, UAM was introduced by presenting some of its properties found in the pertinent
literature (presented in Section 1). The assumptions for the vehicle design were mostly based on the economic255
model by Uber Elevate (2018) who proposed a vehicle with four seats. The process was defined as in Porsche
Consulting (2018), starting with the first mile, followed by boarding, in–vehicle flight including take–off and
landing, de–boarding, and finally the last mile; in this study, both boarding and de–boarding processes were
estimated at around three minutes, with flight ranges of 20 to 50 km (for intra–city). In particular, for our
study, an example of 30 km was given for the transfer from Munich’s airport to its city center (Marienplatz).260
Based on the above studies, the system was defined as a future mobility service provided by fully–
automated electrically powered vertical take–off and landing aircraft (eVTOL) with four–seat capacity (in-
cluding one wheelchair seat), operating at a speed of around 150 km/h. The service would be on–demand,
booked online (through the service’s app or website), and passengers would have the possibility to book for
up to four passengers or to ride–pool with other passengers and save money. The required boarding time was265
set to be 5 minutes prior to take–off, with access and egress through vertiports (helipad–like infrastructures)
that would be integrated with the existing public transportation system.
Two scenarios were given to present realistic attributes of the service (trip duration and fare), starting
hypothetically in 2030, with taxi as a benchmark. These targeted the region of Munich, by presenting two
trips where public transportation would need more than an hour with a necessary transfer in the center,270
with intra–city urban ranges of 20 and 30 km (for the second and first examples respectively). Other modes
were not presented as the aim of the survey was to understand the user adoption and not to conduct a mode
choice study; scenarios therefore aimed to familiarize respondents with the system. For UAM, the access
time was given as the combination of travel time to the nearest vertiport and an average waiting time of
five minutes (assuming a service integrated with public transportation and a 10–min frequency4). In–vehicle275
travel times (including vehicle take–off and landing) were based on a travel speed approximation of 150 km/h
and the direct distance between the origins and destinations. Boarding time was approximated as the sum of
both boarding and de–boarding times, assumed to be 5 minutes in total. On the other hand, taxi attributes
were based on traffic conditions and a waiting time of five minutes was assumed for taxi; service scenarios
are presented in Figure 1. Following this introduction, respondents were first asked to rank the given factors280
they believed to be the most important for adopting and using UAM. These were “booking experience”,
“on-time performance (reliability)”, “operation characteristics (availability/frequency of service)”, “process
of boarding”, “safety”, “trip cost”, “trip purpose”, “trip duration”, “vehicle characteristics (comfort and
cleanliness)”. Then, respondents’ perceived usefulness of the service was assessed in a Likert-scale question
with options ranging from “not at all useful” to “extremely useful”.285
In this second part of the survey, latent constructs affecting automation acceptance were investigated,
in the form of five–point Likert scale agreement statements (ranging from “strongly disagree” to “strongly
agree”), such as data concerns, trust and safety, cost perception, and the value of travel time savings.
Statements on data concerns were: “I am worried that my data goes to a third party”, “My fear of cyber–
security could prevent me from using UAM”, “I am concerned about the loss of jobs induced by automation”;290
on trust and safety through UAM’s operation characteristics: “Service reliability (on-time performance) is
a very important feature for trusting UAM”, “In order for me to feel safe, I would expect UAM’s vehicles
to be equipped with surveillance cameras”, “I should be able to talk to an operator on the ground at any
time”, “The operator should be able to override the system and remotely control the UAM’s vehicles in case
of emergency”, “The service provider’s reputation is very important to gain trust to use UAM”. Statements295
on cost perceptions included the following: “I would be willing to use the service, as long as its price is
in the same range as that of a taxi”, “I would first think about cost when deciding whether or not to use
UAM”, “I think UAM’s service costs provided in the scenarios are reasonable”. Finally, statements on time
savings were as follows: “Travel time saving is a key factor in deciding whether or not to use UAM”, “5-min
travel time saving is important”, “10-min travel time saving is important”, “20-min travel time saving is300
Latent statements were followed by the stated preference question on adoption time horizon: “When
are you most likely going to use UAM?”. Alternatives included options ranking from the first year of
4Both access and egress time were based on assumptions of a 10–min frequency of the service.
Figure 1: Survey scenarios
operation (Y1), the second or third years of operation (Y2–Y3), the fourth or fifth years of operation (Y4–
Y5), starting the sixth year of its operation (Y6+), to “never” (non–adopters), and “unsure” (uncertain305
adopters). Respondents were also asked about the most probable trip purpose they would use UAM for,
including “daily commute”, “business travel”, and “leisure”.
The third part of the survey also aimed at revealing latent variables on the respondents’ social atti-
tudes, through similar Likert-scale constructs. The familiarity with various on–demand services (Airbnb,
DriveNow/Car2Go, Uber, BlablaCar: “I am not familiar with the service” to “I use it frequently”), the use310
of social media platforms (Facebook, WhatsApp, Instagram, Twitter: “I don’t use it” to “I use it several
times a day”), the comfort with online services (online booking, banking, shopping: “very uncomfortable”
to “very comfortable”), with flying (including “Prefer not to answer”), the willingness to share a ride (such
as in a taxi or BlaBlaCar: “very unwilling” to “very willing”) with strangers, the enjoyment of driving a
car (“don’t enjoy it at all” to “very much enjoy it”), and environmental awareness were investigated. The315
latter included five–point Likert scale agreement statements: “I am concerned about global warming”, “I
do not change my behavior based on environmental concerns”, “I am willing to spend a bit more to buy a
product that is more environmentally friendly”. Respondents were also asked if they had been involved in a
car crash and its severity level (if yes, between “no injuries”, “minor injuries”, and “major injuries”). The
survey ended with a free comment field where respondents could optionally express their further concerns320
or suggestions.
First, a pilot survey was conducted to gain useful insights on potential biases in the survey design. The
insights gained from this pilot study (in the form of comments and by evaluating preliminary models) were
incorporated in the final survey, which was then published online using Limesurvey Pro (
The survey was publicly available for two months starting from the 18th of July 2018 and disseminated325
through various channels, including Facebook, Instagram,and mailing lists (such as student lists at the
Technical University of Munich).
3.2. Descriptive analysis and model development
A preliminary descriptive analysis was performed to understand the sample distribution and the stated
adoption of different demographics. Then, models were built using exploratory factor analysis and discrete330
choice modeling (namely multinomial logit models and ordered logit models).
The exploratory factor analysis (EFA) was applied to the second and third parts of the survey pertaining
to the respondents’ perceptions (mostly of UAM automation) and social behaviors, respectively. This
method was used to primarily identify latent constructs behind the variables and secondarily to reduce the
data dimensionality. The variables mainly used were those with the same scale, such as five–point Likert335
scale agreement statements. In this study, EFA was applied using the maximum likelihood estimation (MLE)
as a factor extraction method: factanal in the R statistical computing software (R Core Team,2019). The
suggested number of factors was obtained using the Kaiser–Guttman method (Yeomans & Golder,1982).
Also, after trying different rotations, varimax orthogonal rotation was used (Kaiser,1958). Factor scores
were computed using factor.scores and the “component” method, a weighted sum of the factor loads.340
As the answer options of the survey are discrete categories of adoption time horizon, discrete choice
modeling was applied to identify and analyze the factors significantly affecting adoption (independent vari-
ables). Using pythonbiogeme (Bierlaire,2003), both multinomial and ordered logit models were specified
and estimated. As the outcomes for “Y6+” and “never” represented very small percentages of the sample
size (less than 5%), both categories were ultimately merged. Starting from multinomial logit models and345
based on the inputs from the factor analysis, models were developed in a stepwise fashion, first backward
(from saturated models), where only variables of high significance (95% or 90 % confidence interval) were
kept, then forward (from empty models), where significant variables were added one after the other, using
similar confidence levels. Both generic and alternative specific models were developed, where independent
variables were specific to the different outcomes. Hypothesis testing was used in case some variables had350
similar estimates in order to test if merging them into generic variables would improve the overall model.
Also, models were compared against each other by performing a log–likelihood test, and by assessing the
values of the statistical parameters AIC and BIC. After assessing the results of the multinomial logit mod-
els (MNLs), ordered logit models (OLMs) were proposed by initially specifying models with variables that
showed patterns across the outcomes (adoption time horizon), then testing them for significance like in the355
MNLs. After testing the model performance, the late adopters (Y6) and non–adopters (never) were merged
in the proposed OLM. Despite the limitations of OLMs for the non–ordinal scale, these were used for testing
the possible order of the outcomes, mostly for “unsure” respondents.
3.3. Study limitations
The convenience sample resulting from the online dissemination of the survey is a limitation of the360
sampling approach, might influence the results, and may potentially lead to a lower representativeness of
the population. To overcome this, the authors conducted statistical tests to test the significance of attitudes
across demographics; for instance gender, or language. These were then taken into consideration for the
model development. Moreover, the assumption that factors used for the factor analysis are uncorrelated is
subject to limitations; Fabrigar et al. (1999) advise instead the use of oblique correlations, and Mokhtarian365
et al. (2009) use it in their study to reveal correlations among factors.
Since this is a Stated Preference (SP) survey, some statements might be biased. Particularly, attitude
statements may induce some prompting due to their formulation, as they lead to respondents’ tendency to
agree, also known as the acquiescence bias. According to Baumgartner & Steenkamp (2001), this response
bias is a threat to the validity of the questionnaire results. To overcome this, these statements were controlled370
for their quality, as discussed in Section 4.
Finally, most factors in our study are perception variables, only related to respondents’ attitudes and
not to the service attributes, which might affect the dataset variability.
4. Data collection and results
The survey generated 221 responses, with most respondents residing in Europe (181 respondents), par-375
ticularly in Germany (138 or about 60% of the sample size), and specifically the region of Munich (97
respondents). The rest were scattered outside Europe between the US, Latin America, and the Middle
East. Due to its high concentration and significance, the group of Munich was considered as a significant
subsample. It should be noted that the location only represents the place of residence, and not the place of
origin. The average completion time was 13.6 min with a standard deviation of 6.5 min. Overall, a good380
representation of gender, occupation, household size, and household incomes was observed, as well as an
overrepresentation of the age category 25–34 and of higher levels of education, possibly due to the online
method of survey distribution. Also, public transportation as a main commute mode was overrepresented
in the Munich subsample. In all cases, missing values (representing less than 5% of the sample size) were
recoded using mean or median values (Tabachnick et al.,2007). The summary statistics of the entire sam-385
ple and the major subsample (Munich) are presented (before further processing) in Table 1, benchmarked
against the latest Munich Census for reference (Statistische ¨
Amter des Bundes und der L¨ander,2014).
Among the 221 respondents, 36.65% stated that they would adopt UAM in the second or third year of
its implementation, followed by 22.17% claiming an adoption during its first year, 14.03% during its fourth
and fifth year, 2.71% starting its sixth year, and 3.17% stating that they would never adopt the service.390
Moreover, 21.27% of the respondents expressed uncertainty (“unsure”) on their adoption time horizon of
UAM. The analysis of the results highlighted the importance of safety for UAM adoption as the majority
of respondents (more than 50 %) ranked it as the most important factor in their intention to adopt UAM.
Also, a strong indication towards the importance of UAM costs (second factor), trip duration (third factor),
on–time reliability (fourth factor), and operation characteristics (fifth factor) was given. The suggested395
ranking was obtained by extracting for each rank (e.g.: first rank, second rank, etc.) the factor with the
highest percentage of respondents. For the second to fifth ranks, the mentioned factors had a share of about
20 % of the respondents. On the other hand, vehicle characteristics, boarding process, booking experience,
and trip purpose were ranked as sixth to ninth factor by most of the respondents, gathering each also about
20 % of the respondents.400
Table 1: Summary of sample and subsample characteristics
Total sample
Munich subsample
Munich Census
Gender Female 43.0% 51.6% 48.6%
Male 56.1% 47.4% 51.4%
Prefer not to an-
0.9% 1.0% -
Agea0-17 0.5% - -
18-24 19.5% 23.7% 9.2%
25-34 45.7% 56.7% 21.7%
35-44 19.0% 16.5% 22.4%
45-54 9.5% 16.5% 22.2%
55-64 5.0% - 16.8%
65+ 0.9% - 7.7%
Main occupation Full–time employed 57.9% 46% 87.1%
Part-time employed 9.1% 8%
Student 28.1% 42% 2.9%
Unemployed 0.5% 1% 2.2%
Self-employed 2.3% 1% 7.8%
Retired 0.9% -
Prefer not to an-
1.4% 1% -
Education High School 8.6% 6% 34.1%
Apprenticeship 2.7% 1% 40.7%
Bachelor 26.7% 26% 22.7%
Master 47.5% 52%
Doctorate 13.1% 13% 2.5%
Prefer not to an-
1.4% 2% -
Household incomeb<500 AC 7.2% 13.4 %
500-1000 AC 8.6% 13.4%
1000-2000 AC 11.3% 10.3%
2000-3000 AC 14.0% 16.5%
3000-4000 AC 10.9% 6.2%
4000-5000 AC 10.9% 7.2%
5000-6000 AC 4.9% 3.1%
6000-7000 AC 5.4% 30.9%
>7000 AC 6.3% 21.7%
Prefer not to an-
20.4% 3.1%
Main commute cCar as a driver 33.9% 14.0% 31.0%
mode Car as a passenger 1.8% -
Public transporta-
40.7% 63.0% 28.8%
Bike 15.4% 20.0% 14.7%
Walk 5.9% 1.0% 25.5%
Other 2.3% 2.0% -
aCensus values for age are based on different age categories and are therefore best-fit values for the survey classes.
bThere are no census values for income levels; however the average disposable household income for the year 2016 is
4220 AC (Euromonitor International,2017).
cThe given values are trip distribution percentages (MVG 2011).
A preliminary analysis of the attitudes of different demographics showed the importance of these factors
for the adoption intention. This was mostly observable for gender, main occupation, education, survey
language, and previous crashes, as observed in Figure 2. Females were found to have a lower tendency of
being immediate adopters (Y1), and a higher of being “unsure” about their adoption time; as expressed by
the 12% of female respondents for Y1 compared to 30% for their male counterparts. For the main occupation,405
respondents working full–time expressed a higher immediate adoption intention (30%) compared with part–
time employees (0%), or students (14%), which can be due to higher income levels and/or higher value of
time. The survey language also seemed to impact stated adoption. Respondents who filled the survey in
English showed a higher interest in early adoption (31%) and less skepticism (12%) compared to those filling
the survey in German. Respondents with previous crashes and no injuries expressed a higher intention410
of immediate adoption (26%) compared to those who had injuries. Respondents with different education
levels also stated different preferences: those with lower levels of education (bachelor or lower) expressed
less uncertainty (19%) compared to the ones with higher levels of education (doctorate). On the other hand,
the latter group showed a higher interest in immediate adoption (26%) compared to the former (18%). Of
course some of these comparisons are limited to their statistical significance, since the compared outcomes415
were not well represented for all categories, such as occupation, crash history, and education levels. On the
other hand, gender and survey language were rather well distributed.
Previous crashes
or lower
(N=105) Doctorate
(N=29) Female
(N=97) Male
(N=124) English
(N=128) German
(N=93) Student
(N=62) Part−time
(N=20) Full−time
(N=128) Yes
(no injuries)
Percentage adoption (%)
Stated time adoption Y1 Y2−Y3 Y4−Y5 Y6+ Never Unsure
Figure 2: UAM adoption by different demographics
The exploration of respondents’ attitudes, based on the Likert scale statements, showed that gender and
survey language were the demographics where most differences were observed. As presented in Figure 3,
females seemed to give a higher level of importance to the loss of jobs due to automation as well as to safety;420
the latter is demonstrated by higher safety camera expectations and requirements for an operator on the
ground and to override the vehicle in case of emergency. On the other hand, male respondents seemed to
have more flying comfort than their female counterparts. These findings are statistically significant (error
bars in Figure 3drawn based on the standard error) and compatible with those from the literature; in
particular the study by Hohenberger et al. (2016), where women were found to have a lower intention to use425
autonomous vehicles, possibly due to the effect of gender on anxiety, rather than on pleasure.
Different attitudes were observed based on the survey language. Notably, the respondents filling the
survey in English expressed a higher trust in automation, enjoyment of automation, perceived usefulness of
ADAS, and of UAM, compared to those filling the survey in German. A similar, less remarkable, pattern was
observed for respondents residing in Munich as opposed to those who don’t. This finding parallels the results430
of the study on Autonomous Vehicles Readiness, where Germany was ranked 12th in consumer acceptance,
despite being third in technology and innovation (KPMG,2018): it can be assumed that language reflects
better the culture than the place of residence.
Expect in−vehicle
safety cameras
Flying comfort
Loss of job concerns
Operator on ground
Operator overrides
in emergency
Mean value on a 5−point Likert scale
All significant to at least 97% confidence level
Gender Female Male
Enjoy automation
Fear of cyber−
Perceived ADAS
Perceived UAM
Trust automation
0 1 2 3 4
Mean value on a 5−point Likert scale
All significant to at least 99.9% confidence level,
except for cyber−security with a 94% confidence
Survey language English German
Figure 3: Attitudes towards different statements: (a) by gender; and (b) by language
Attitude statements from the second and third parts of the survey (see Section 3.1) were also analyzed
to assess the quality of the responses. After a careful examination of the distribution of the answers across435
the Likert–scale (“strongly disagree” to “strongly agree”), two statements were identified as prompted. This
was observed for the agreement statements on the necessity of an operator to override the vehicle in case
of emergency, and on the importance of the service provider’s reputation to gain trust to use UAM. For
both statements, the respondents’ tendency to agree, also known as the acquiescence bias, was noted, as
mentioned in the study limitations (see Section 3.3). Accordingly, both statements were removed from the440
subsequent analysis and model specifications.
Finally, a qualitative analysis of the comments left at the end of the survey showed additional concerns
of the respondents. The biggest concern was expressed by eight respondents on environmental impacts,
including concerns from noise and visual pollution (four respondents for each). For the latter, the respondents
expressed the disturbance they would feel from being watched from above (“Not in My Backyard” effect,445
as described by one respondent). Two respondents expressed their concerns about privacy (concerning data
sharing) and safety (one of them stated the desire to wait for an incident–free operation of UAM). Moreover,
three respondents expressed their concern of economic impacts of UAM; one of them was concerned that the
service would target a niche market. Also, two respondents expressed their desire for a system integration of
UAM (with the existing and future transportation systems), one respondent was interested in the information450
sharing inside the vehicle, another for a high frequency service with on–demand availability, and finally one
more about the use case or purpose of UAM.
5. Model development and estimation
This section describes the model development, starting from exploratory factor analysis and continuing
with the developed discrete choice models.455
5.1. Exploratory factor analysis
The outcome of the exploratory factor analysis for respondents’ perceptions on automation is presented
in Table 2. Four factors were eventually extracted, explaining a cumulative variance of 52% of the total
variance. As observed in Table 2, each factor explains more than 10% of the total variance (with the exception
of one that explains 9%5), which is considered acceptable according to Costello & Osborne (2005). This is460
compatible with the work of several researchers who used factor analysis to understand users’ perceptions
of transportation systems (Tyrinopoulos & Antoniou,2008;Efthymiou et al.,2013), reducing the initial
indicators to fewer factors each explaining more than 10% of the total variance (with one exception at
most), with a cumulative total variance ranging roughly from an average of 46% to 50%. Loadings lower
than 0.4 are not shown to simplify the table.465
By looking at the latent meaning that these variables might have, the factors are interpreted as follows.
The value of time savings is a cluster of the three variables on the perception of saving 5, 10, and 20 min
of travel time. The enjoyment of automation6, trust of automation, and perceived usefulness of UAM are
interpreted as the affinity to automation. The fear of cyber–security, the concern of data being shared to a
third party, and the concern of loss of jobs due to automation are grouped under data and ethical concerns.470
Finally, the need for an operator on the ground, and of in–vehicle safety cameras are interpreted as safety
Table 2: Factor analysis on respondents’ perceptions
Loadings Factor 1 Factor 2 Factor 3 Factor 4
Travel time savings 5min 0.78
Travel time savings 10min 0.98
Travel time savings 20min 0.59
Enjoy automation 0.80
Trust automation 0.75
UAM is useful 0.51
Fear of cyber–security 0.70
Fear that data goes to a third party 0.53
Loss of job concerns 0.49
Operator on the ground 0.67
In–vehicle safety cameras 0.50
Sum of square of loadings 2.01 1.66 1.14 0.96
Proportion variance 0.18 0.15 0.10 0.09
Cumulative variance 0.18 0.33 0.44 0.52
Factor interpretation Value of
time savings
Affinity to
Data and
ethical concerns
For respondents’ social attitudes, four factors were extracted, explaining a cumulative variance of 55% of
the total variance, presented in Table 3. Variables pertaining to social attitudes were grouped under affinity
to online services, environmental awareness, affinity to social media, and affinity to sharing, with all factors475
explaining at least 10% of the a variance and therefore also considered acceptable (Costello & Osborne,
5This factor was retained due to its high explanatory power and interpretability.
6Enjoyment of automated systems like Alexa or Siri
Table 3: Factor analysis on respondents’ social attitudes and behaviors
Loadings Factor 1 Factor 2 Factor 3 Factor 4
Online booking 0.81
Online banking 0.86
Online shopping 0.65
Concerned about global warming 0.99
Spend on environmental products 0.5
Instagram 0.99
Facebook 0.43
BlaBlaCar 0.67
Airbnb 0.5
Willingness to share 0.5
Sum of square of loadings 1.93 1.28 1.24 1.06
Proportion variance 0.19 0.13 0.12 0.11
Cumulative variance 0.19 0.32 0.45 0.55
Factor interpretation Affinity to
online services
Affinity to
social media
Affinity to
The factor analysis interpretation (Tables 1 and 2) is similar to the one on shoppers’ attitudes by
Mokhtarian et al. (2009), where a factor analysis was applied to 42 attitude agreement statements (five-
point Likert Scale), reducing them to 13: shopping enjoyment, time consciousness, trust, trendsetting,480
pro-environment, technology; these are comparable to the presented factors: affinity to automation, time
savings, safety concerns, affinity to online services and social media, environmental awareness, and affinity
to automation and social media, respectively.
5.2. Multinomial logit model
Aiming at a better understanding with regards to the factors which affect adoption, a multinomial logit485
model was estimated; the model specification and estimation results are shown in Table 4. The coefficient
estimates were in general reasonable in sign and magnitude and consistent with prior expectations; most of
them significant at least to the 95% confidence level with a t–value higher than 1.96, with a couple being
significant to the 90% with a t–value higher than 1.65. For instance, estimates for the affinity to automation
were highly significant and positive mostly for early adoption (estimate of 1.15 for Y1), but also (to a lesser490
extent) for later years (estimate of 0.79 for Y2–Y5). For most of the respondents (Y1–Y5 and unsure),
highly significant and positive impacts were also noted for the affinity to social media and to WhatsApp
(estimate coefficients of 0.50 and 0.53, respectively); the same impact was observed for full–time employment
(estimate of 0.63 for Y1–Y3). Cost factors also had a significant positive impact on adoption. Scaling UAM
prices to taxis’ highly contributed to immediate adoption (estimate of 0.32 for Y1). On the other hand,495
higher data and ethical concerns (including fear of cyber–security, concerns of data being passed to third
parties, and loss of job concerns) had influential negative impacts on early adoption (-0.24 for Y1). Previous
crash experiences also had a negative impact on adoption in general, with lesser impacts on later adoption
(estimate of -2.42 for Y1, compared to -1.89 for Y2–Y5). Safety concerns (including the importance of having
an operator on the ground at any time and of in–vehicle safety cameras) also negatively affected adoption500
(estimate of -0.34) with a high influence on immediate adoption (Y1) and later adoption (Y4–Y5). Female
respondents were in general less likely than their male counterparts to adopt UAM (estimate of -2.94 for all
years) and highly educated respondents (doctorate level or higher) less likely to adopt it in its early years
(estimate of -1.43 for Y2–Y3). Moreover, German as a survey language (compared to English) was highly
correlated with uncertainty (estimate of 1.42). Furthermore, higher income level respondents were less likely505
to be late adopters of UAM (estimate of -1.42 for Y4–Y5). The perception of travel time importance was
also decisive in UAM adoption for the second and third years (estimate of 0.28 for Y2–Y3) and the value
of time savings (including the perception of each of 5,10, and 20 minutes of time savings) for later adoption
(estimate of 0.33 for Y4–Y5). Finally, public transport commuters were more likely to adopt UAM during
its later years (estimate of 0.85 for Y4–Y5) and the alternative–specific constant for respondents stating a510
very late or non–adoption (Y6+, Never) was positive and highly significant (estimate of 6.73).
Table 4: Multinomial logit model results (N=221)
Parameters estimate std. error t–stat
ASC Y6+ or never 6.73 2.20 3.06
Affinity to automation Y1 1.15 0.32 3.59
Affinity to automation Y2–Y5 and unsure 0.79 0.31 2.60
Cost as taxi Y1 0.32 0.18 1.83
Previous crash with injuries Y1 -2.42 0.97 -2.49
Previous crash with injuries Y2–Y5, unsure -1.89 0.86 -2.20
Starting language German unsure 1.42 0.34 4.13
Data concerns Y1 -0.24 0.12 -1.98
Doctorate level of education Y2–Y3 -1.43 0.53 -2.67
Female Y1–Y5 and unsure -2.94 1.14 -2.58
Full–time employment Y1-Y3 0.63 0.31 2.09
High income: 3000–7000 Euro Y4–Y5 -1.42 0.65 -2.20
PT commute Y4–Y5 0.85 0.40 2.11
Safety concerns Y1 and Y4–Y5 -0.34 0.10 -3.46
Affinity to social media Y1–Y5 and unsure 0.50 0.29 1.71
TT important for UAM Y2–Y3 0.28 0.07 3.81
Value of time savings Y4–Y5 0.33 0.08 4.07
Affinity to WhatsApp Y1–Y5 and unsure 0.53 0.26 2.07
Summary statistics
AIC 546.42
BIC 607.587
5.3. Ordered logit model
Ordered logit models were also built with adoption time horizon as a dependent variable. Although
time frame can theoretically be ordered from the first to the sixth or more years (even never), the category
“unsure” cannot be ranked in that scale. Therefore, based on previously observed patterns in the MNL515
model, OLMs were proposed . For instance, the affinity to automation estimate from the MNL (Table 4)
was positive and strongly significant (more than 95%) for both immediate adoption (Y1) and later adopters
(Y2 to Y5) and unsure respondents; however, the impact was higher for immediate adopters (Y1), expressed
by the higher coefficient estimate (1.15 compared to 0.79). Similarly, previous crashes with injuries negatively
impacted adoption, expressed by strongly significant (more than 95%) and negative coefficient estimates;520
higher even for Y1 (-2.42 compared to -1.89 for the rest). These for instance guided the specification of
an ordered logit model, where such attributes would be ranked, as suggested by the patterns in the MNL.
Accordingly, two cases were proposed. In the first case, the ordered categories were as follows: Y1; Y2–Y3;
Y4–Y5; Unsure; Y6+/Never. The second case was built from the first by merging the “unsure” category
with the one of late and non–adopters. The corresponding categories were therefore: Y1; Y2–Y3; Y4–Y5;525
Unsure, Y6+ and Never.
Table 5represents the final OLM model for both cases. The highly significant cut values for case 1
indicate that adoption is indeed ordered and people who are unsure display a behavior that is ranked
between late (Y4–Y5) and extremely late or non–adopters (Y6+ or Never). This was rather expected from
the patterns observed in MNL models. The significant parameters are the same for both cases: affinity to530
automation, full–time employment, and cost as taxi (meaning the willingness to use the service for prices
in the range of taxis) are associated with an early adoption. On the other hand, data and ethical concerns
and starting language German are strongly correlated with a later adoption.
Table 5: Ordered logit models results (N=221)
Case 1 Case 2
Parameters estimate std. error t–stat estimate std. error t–stat
Affinity to automation -0.26 0.07 -3.54 -0.23 0.07 -3.11
Cost as taxi -0.37 0.13 -2.82 -0.38 0.14 -2.70
Starting language German 0.61 0.30 2.05 0.70 0.30 2.32
Data and ethical concerns 0.27 0.08 3.32 0.26 0.08 3.19
Full–time employment -0.96 0.26 -3.68 -1.03 0.27 -3.84
Case 1
Y1—Y2–Y3 -4.22 0.85 -4.97
Y2–Y3—Y4–Y5 -2.16 0.21 9.96
Y4–Y5—Unsure -1.36 0.14 5.93
Unsure—Y6+/Never 0.87 0.31 7.10
Case 2
Y1—Y2–Y3 -4.02 0.88 -4.57
Y2–Y3—Y4–Y5 -1.96 0.21 9.92
Y4–Y5—Unsure/Y6+/Never -1.15 0.14 5.93
Summary statistics
¯ρ20.25 0.26
AIC 585.10 526.04
BIC 615.69 553.23
Compared to Case 1, Case 2 presents similar values for the relevant parameters with close estimates
in terms of value and significance. This model also presents significant cutoff values between the ordered535
categories and indicates that uncertain respondents can be merged with very late (Y6+) and non–adopters.
In other terms, respondents showing some skepticism regarding their adoption time horizon of UAM are more
likely to use it at a later stage, or not to use it at all. In both cases, uncertain adopters are associated with
rather late adoption; the only difference is that one case considers uncertainty as part of non–adoption and
the other as one degree less. However, due to the difference in the nature of the data between these models540
(since the first case provides five outcomes for the dependent variable compared to four in the second), a
likelihood ratio test is not possible. Therefore as both models have more or less the same meaning, deciding
on the number of ordered categories depends on personal judgment and preference: parsimony vs. richness.
6. Discussion
6.1. Summary findings545
Despite high stated adoption rates for UAM (about 37% for Y2–Y3 and 22 % for Y1), a non-negligible
number of respondents expressed their uncertainty about their intention to use the service, as shown by the
percentage of “unsure” respondents (more than 22 %). The analysis of the responses showed that safety
was perceived as the factor of highest importance in using UAM (ranked by the majority of respondents
as the first factor in UAM adoption). A high indication for the importance of trip cost, trip duration,550
service reliability and operation characteristics was also observed, as they were mostly ranked as second,
third, fourth, and fifth most important factors for UAM adoption. The survey statistics indicated a high
impact of socio–demographic factors on adoption and attitudes. Females expressed a much lower interest in
immediate UAM adoption, expressing overall lower trust and perceived usefulness of automation, and greater
security and safety concerns and expectations, expressing a higher desire for operators on the ground, and555
of in–vehicle safety cameras, as reflected in the mean values of the attitude Likert–scale questions (partially
presented in Figure 3). On the other hand, both fully-employed and higher income respondents expressed
a greater interest in early adoption, as discussed in the survey statistics (Section 4.1). Cultural impact was
observed through the survey language, which seemed to also have an impact on automation; respondents
filling the survey in German expressed a lower interest in early adoption and a higher skepticism observed560
through a higher degree of uncertainty. This is interesting when compared to the impact of the place of
residence (Munich or Germany), which was less influential than the survey language, as only the latter was
a significant variable in the adoption models (MNL and OLM).
The discussed findings from the statistics were confirmed by the specified models. For instance, MNLs
demonstrated the high impact of gender (female), employment (full–time), language (German language),565
income levels on adoption. The MNLs also showed that public transportation users were more likely to be
late UAM adopters, as expressed by the positive and significant coefficient estimate for PT commuters (Table
4). Moreover, respondents with previous crash experiences with injuries were less likely to be immediate
adopters, as revealed by the model finding estimates (Table 4). Safety concerns were also found to play
an inhibiting role in early and late adoption, expressed by the negative and highly significant (more than570
99.9 %) coefficient estimate in the MNL. These included the need for in–vehicle surveillance cameras, and
for an operator on the ground at any time (resulting from the factor analysis in Table 2). Lack of safety
may hinder UAM, and lead to late UAM adoption. Moreover, the affinity to automation was found crucial
for UAM adoption (as demonstrated by the positive coefficients for the MNL, higher still for Y1 than for
Y4-Y5, and negative for the OLM, proving as well the higher effect on immediate adoption of the affinity575
to automation) and included the enjoyment and trust of automation, and the perceived usefulness of UAM.
Furthermore, the value of time savings and perceived costs were highly influential for late UAM adoption.
As some of these findings might be counter–intuitive, their occurrence might be due to respondents’ prior
expectations and judgments of service properties. For instance, respondents may be skeptical about actual
time savings and therefore as they value this factor, they might decide to adopt UAM later on, waiting for580
the service to improve its performance. This was also observed for respondents who are highly sensitive
to costs as they might believe that costs would decrease a few years after the service implementation to
be in the range of taxis. Moreover, higher safety concerns were negatively associated with late adoption,
possibly due to the users’ beliefs that safety would only need a few years to be demonstrated; waiting for too
long would therefore not be needed. Also, data and ethical concerns were found to be hindering immediate585
adoption; this could refer to the respondents who are rather worried about their data being shared to third
parties, and who tend to be uncertain about the future as they are concerned from cyber–security threats
and/or the loss of jobs due to automation. Finally, the impact of social media and WhatsApp was found to
be positive on adoption.
OLMs confirmed some of these factors found in the MNLs, notably the cost range importance (comparable590
to the taxi range for early adoption), the cultural impact (correlation between German language and later
adoption), the data and ethical concerns (associated with later adoption and including the fear of cyber–
security, data sharing, and the loss of job concerns), the affinity to automation, and the employment impact.
Finally, the ordered models gave meaningful insights regarding “unsure” respondents, in which these were
found to have similar or close (one degree less) behavioral intentions to late adopters.595
When comparing the results of this study with the reports commissioned by NASA and presented in
Section 2.1, parallel findings were observed. The study by Crown Consulting (NASA,2019) mentioned
that nearly half of the consumers were potentially comfortable with delivery and UAM use cases, which
is quite comparable with our results (more than 50% stated time adoption within the first three years of
implementation). Moreover, common factors were observed such as safety, privacy, loss of jobs concerns,600
environmental concerns, and noise and visual pollution that were extracted from our qualitative analysis.
Common results were also shared with the study by Booz Allen Hamilton (NASA,2018), particularly for
user concerns like safety, privacy and noise, piloted aircraft and flight attendant, cybersecurity, cost and
convenience, and gender impact with men being more comfortable and willing to use UAM; in our study
however, the system was assumed to be fully automated, and therefore a flight attendant was not mentioned.605
Instead, the importance of having an operator (available on the ground and/or remotely control the vehicle
in case of emergency) was investigated.
6.2. Technology Acceptance Model for Disruptive Transport Technologies
The model findings pertaining to acceptance were incorporated to extend the Technology Acceptance
Model for the application of urban air mobility, and more generally disruptive transport technologies (Fig-610
ure 4). Compared to the initial TAM by Davis et al. (1989), the model retains only one of the two main
constructs: the perceived usefulness (PU), which directly impacts the behavioral intention (BI). The per-
ceived ease of use (PEU; in the case of urban air mobility: service booking or boarding) was not found
influential and was therefore removed from the model. More generally, for fully-automated systems, the
user is not expected to do any effort. Actual system use is also omitted, as there is no way to measure it615
for non–existing systems. The attitude towards using the system is removed; instead the following factors
(including PU) are considered as attitudes directly impacting the behavioral intention: perceived usefulness,
social behavior (social attitudes related to social media use, environmental concerns, etc.), value of time
(the value given to time savings), perceived costs (cost perceptions: cost range), data and ethical concerns
(fear of data sharing to a third party, fear of cyber–security, and the loss of job concerns). These factors620
result from the estimated multinomial logit models (MNLs); data and ethical concerns and cost perceptions
also result from the ordered logit models (OLMs).
The construct of trust is added as a main construct in this model, as suggested by the Automation
Acceptance Model (AAM) by Ghazizadeh et al. (2012). Additionally, the proposed model introduces factors
directly and positively impacting trust. These are the perceived reliability of automation, the perceived625
vehicle’s safety, the perceived locus of control, and the previous experience with automation. The vehicle’s
safety and perceived locus of control are supported by the MNLs and refer to the safety from in–vehicle
safety cameras and the locus of control confirmed by the importance given to the human factor (operator
on the ground at any time). Moreover, both service reliability and automation experience were found from
analyzing respondents’ attitudes. Finally, external variables, similarly to the AAM directly impact trust,630
but not PU. These are given as socio–demographics and affinity to automation. The former includes gender,
education, occupation, income, cultural impact (demonstrated by the survey language). In the proposed
model, these external variables directly impact behavioral intention as well.
Trust/ Value of
Value of Time
Perceived Usefulness
Social Behavior
Behavioral Intention
Affinity to Automation
Perceived Costs
Data and Ethical
Vehicle‘s Safety
Locus of Control
Perceived Reliability
of Automation
Automation Experience
External variables
Figure 4: Technology Acceptance Model for Disruptive Transport Technologies, adapted from the original TAM by Davis et al.
(1989) and the Automation Acceptance Model (AAM) by Ghazizadeh et al. (2012)
6.3. Insights for policymakers and industrial stakeholders
The findings of this research provide insights for both responsible policymaking and industrial stake-635
holder engagement, to ensure a smooth urban air mobility implementation and integration with the existing
transport systems. Recommendations on such areas can be summarized in the following points.
6.3.1. Insights for policymakers
Safety concerns
As highlighted in the proposed Technology Acceptance Model, trust or the value of safety is a crucial640
component in users’ adoption of urban air mobility and is determined by several components. Among
these, the perception of automation reliability, vehicle’s safety and locus of control are areas where
stringent regulations could improve users’ trust and reduce skepticism resulting from missing or erro-
neous information. Accordingly, safety standards enforced by the responsible authorities could ensure
users’ trust, alleviating negative perceptions on automation reliability and service performance. En-645
forcing regulations acting upon in-vehicle safety, such as the presence of in-vehicle surveillance cameras,
might help increasing trust and feeling of safety. Finally, regulations focusing on the human factor in
the operation of urban air mobility could take into consideration the importance of the availability (at
any time) of an operator on the ground, to increase the users’ perceived locus of control, and thereby
trust and safety towards the service.650
Data concerns
The proposed TAM also highlights the importance of data and ethical concerns for UAM adoption,
and their negative impacts on safety perception. These include concerns from information sharing
to third parties, from cyber-security, and from the loss of jobs due to automation, as shown by the
exploratory factor analysis results. Responsible policymaking in the area of data sharing and protection655
would probably ensure a higher sense of privacy for the users. For ground Autonomous Vehicles,
Fagnant & Kockelman (2015) recommend the creation of nationally recognized licensing framework
for autonomous vehicles (ground), determining appropriate standards for liability, security, and data
privacy. Such frameworks could be appropriately developed and transferred to urban air mobility.
Other concerns660
Not directly stemming from the models, environmental concerns expressed by respondents demonstrate
the need for policy–making in the area of noise and visual impact. Regulating the allowed noise levels
and flying altitudes of the vehicles could address community acceptance of both users, and non-users
of urban air mobility. Moreover, an inclusive and integrated system, as defined by our study should
ensure a proper integration of UAM services with existing transportation systems.665
6.3.2. Insights for industrial stakeholders
Service attributes
The proposed TAM revealed the importance of service attributes, i.e. cost and time, for the adoption of
urban air mobility. If the system aims to attract a wider public, user requirements must be prioritized,
notably pricing schemes must be defined, and regulated by the authorities, so that the service would670
not only be a niche market. These for instance need to be within a certain reasonable range, i.e. taxi
prices, allowing competition with existing ground vehicles. Similarly, the time from access to egress
(to and from the vertiports) must be optimized: a smoother process must enable an efficient process
for the service to be accepted; again, if the service aims to be integrated with the existing systems, this
can only be done through a proper stakeholder engagement with the responsible authorities. Finally,675
a higher transparency on both time and costs of the service would ensure a higher awareness on the
service attributes, which is crucial for users’ perceptions towards it.
For an inclusive service, industrial stakeholders should consider the importance of socio-demographics
in the perception of the service. User categories with different gender or cultural background could be680
targeted according to their specific needs. Incentives might as well help some respondents to overcome
their fears of automation.
7. Conclusions
In this paper, users’ acceptance and adoption of urban air mobility was assessed by analyzing an online
stated preference study that gathered 221 respondents from all over the world, with a subsample of 97 Munich685
residents. Using mostly factors from the literature on the acceptance of both technology and automation
(notably autonomous vehicles), and by projecting them to UAM, the survey aimed at uncovering many
hypotheses on this new mobility service and thereby also validated an extended technology acceptance
model for an urban air mobility context. The analysis of the survey data highlighted the importance
of socio–demographic parameters and their attitudes in adoption. Using the exploratory factor analysis,690
many variables were clustered into groups having a high explanatory power and thus reducing the dataset
dimensionality. These included the affinity to automation, safety concerns, data and ethical concerns, the
value of time savings, in addition to social attitudes such as environmental awareness, the affinity to social
media, online services, and sharing. Using the inputs from the factor analysis, significant multinomial and
ordered logit models were developed with adoption time horizon as a dependent variable, which can be a695
considered as a theoretical contribution to the use case of ordered logit models with a non–conventional
ordinal dependent time variable (including non-ordinal parameters such as “unsure” and “never”). Mostly,
the influence of socio–demographics, cultural impact, affinity to automation (including the enjoyment and
trust of automation) were shown. Trust and safety were found to be key components for UAM adoption; in
particular, the presence of in–vehicle cameras and operators, as well as performance expectancy in terms of700
service reliability and on–time performance were noted. Data and ethical concerns, the value of time savings
and costs, social attitudes including a high affinity to social media were also found as highly influential for
UAM adoption. Finally, public transportation as a commute mode was found to be rather related to late
Overall, the study extracted relevant and highly significant factors, extending accordingly the TAM705
model for UAM use, and contributing in filling the gap in research on this new service. The findings of this
work have strong insights for policymakers. Regulatory bodies should have departments targeting UAM
regulations and work closely with other countries to have some unified standards; for instance, a unified
European framework on suggested safety standards and allowed noise levels.
Most factors in the developed models are related to the perception and attitudes of respondents, and710
not to the service attributes; newer models could therefore be developed with alternative-specific variables.
Scenarios could for instance present travel times or costs for different time frames. Future work could
therefore focus on uncertainties pertaining to the service attributes and business models. Another research
motivation could be to build a nested ordered logit model with one or more ordered nests; one of the nests
could comprise uncertain or non–adopters and the other the certain adopters including the ordered (time)715
adoption categories. Other studies could test the developed UAM TAM by applying a confirmatory factor
analysis (CFA) and developing a hybrid latent class model, in which both analysis methods (factor analysis
and choice models) would be combined.
This work has been partly funded by the Bavarian Ministry of Economic Affairs, Regional Development and720
Energy through the OBUAM project.
Airbus (2017). Rethinking urban air mobility. URL:
rethinking-urban- air-mobility.html.
Airbus (2018). Urban air mobility - the sky is yours. URL:
Urban-air- mobility-the- sky-is-yours.html.
Antcliff, K. R., Moore, M. D., & Goodrich, K. H. (2016). Silicon valley as an early adopter for on-demand civil VTOL
operations. In 16th AIAA Aviation Technology, Integration, and Operations Conference (p. 3466).
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: Emerging Technologies,67 , 1–14.730
Baptista, P., Melo, S., & Rolim, C. (2014). Energy, environmental and mobility impacts of car-sharing systems. empirical
results from lisbon, portugal. Procedia-Social and Behavioral Sciences,111 , 28–37.
Baumgartner, H., & Steenkamp, J.-B. E. (2001). Response styles in marketing research: A cross-national investigation. Journal
of Marketing Research,38 , 143–156. URL: doi:10.1509/jmkr.38.2.143.
Bierlaire, M. (2003). Biogeme: a free package for the estimation of discrete choice models. In Swiss Transport Research
Conference TRANSP-OR-CONF-2006-048.
Bimbraw, K. (2015). Autonomous cars: Past, present and future a review of the developments in the last century, the present
scenario and the expected future of autonomous vehicle technology. In Informatics in Control, Automation and Robotics
(ICINCO), 2015 12th International Conference on (pp. 191–198). IEEE volume 1.740
Bjørner, T. (2015). A priori user acceptance and the perceived driving pleasure in semi-autonomous and autonomous vehicles.
In European Transport Conference 2015.
Boucher, P. (2016). ‘You wouldn’t have your granny using them’: Drawing boundaries between acceptable and unac-
ceptable applications of civil drones. Science and Engineering Ethics,22 , 1391–1418. URL:
s11948-015- 9720-7.745
Chamata, J. (2017). Factors delaying the adoption of civil drones: A primitive framework. The International Technology
Management Review,6, 125–132.
Choi, J. K., & Ji, Y. G. (2015). Investigating the importance of trust on adopting an autonomous vehicle. International
Journal of Human-Computer Interaction,31 , 692–702.
Clements, L. M., & Kockelman, K. M. (2017). Economic effects of automated vehicles. Transportation Research Record:750
Journal of the Transportation Research Board, (pp. 106–114).
Clothier, R. A., Greer, D. A., Greer, D. G., & Mehta, A. M. (2015). Risk perception and the public acceptance of drones. Risk
Analysis,35 , 1167–1183.
Cohen, M. M. (1996). The Vertiport as an Urban Design Problem. Technical Report SAE Technical Paper.
Connected Automated Driving EU (2017). New study on user acceptance of driverless cars released. URL: https:755
// user-acceptance- driverless-cars-released/.
Costello, A. B., & Osborne, J. W. (2005). Best practices in exploratory factor analysis: Four recommendations for getting the
most from your analysis. Practical Assessment, Research & Evaluation,10 , 1–9.
Davis, F. D., Bagozzi, R. P., & Warshaw, P. R. (1989). User acceptance of computer technology: a comparison of two theoretical
models. Management Science,35 , 982–1003.760
Deloitte Analytics Institute (2017). Autonomous driving in Germany how to convince customers. URL: https://www2. products/Autonomous-driving- in-Germany_
Dolnicar, S. (2013). Asking good survey questions. Journal of Travel Research ,52 , 551–574.
Efthymiou, D., Antoniou, C., & Waddell, P. (2013). Factors affecting the adoption of vehicle sharing systems by young drivers.765
Transport Policy,29 , 64–73.
Euromonitor International (2017). Munich city review. URL:
Fabrigar, L. R., Wegener, D. T., MacCallum, R. C., & Strahan, E. J. (1999). Evaluating the use of exploratory factor analysis
in psychological research. Psychological methods,4, 272.
Fadhil, D. N. (2018). A GIS-Based Analysis for Selecting Ground Infrastructure Locations for Urban Air Mobility. Master’s770
Thesis Technical University of Munich.
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.
Fagnant, D. J., & Kockelman, K. M. (2014). The travel and environmental implications of shared autonomous vehicles, using
agent-based model scenarios. Transportation Research Part C: Emerging Technologies,40 , 1–13.775
Fu, M., Rothfeld, R., & Antoniou, C. (2019). Exploring preferences for transportation modes in an urban air mobility
environment: a Munich case study. Transportation Research Record: Journal of the Transportation Research Board,In
Gaggi, S. (2017). Driverless cars and user acceptance. URL:
driverless-cars- and-user- acceptance/.780
Garrow, L. A., German, B., Mokhtarian, P., Daskilewicz, M., Douthat, T. H., & Binder, R. (2018). If you fly it, will commuters
come? A survey to model demand for e–VTOL urban air trips. In 2018 Aviation Technology, Integration, and Operations
Conference (p. 2882).
Ghazizadeh, M., Lee, J. D., & Boyle, L. N. (2012). Extending the technology acceptance model to assess automation. Cognition,
Technology & Work,14 , 39–49.785
Hohenberger, C., Sp¨orrle, M., & Welpe, I. M. (2016). How and why do men and women differ in their willingness to use
automated cars? the influence of emotions across different age groups. Transportation Research Part A: Policy and Practice,
94 , 374–385.
Jenkins, S., Linden, E., & Wittmer, A. (2018). Autonomous Mobility - An Application of the Technology Acceptance Model.
URL: consulting/cfac-working- paper-series.790
Kaan, J. (2017). User acceptance of autonomous vehicles: Factors & implications, . URL:
Kaiser, H. F. (1958). The varimax criterion for analytic rotation in factor analysis. Psychometrika ,23 , 187–200.
KPMG (2018). Autonomous vehicles readiness index. URL:
KPMG-Autonomous- Vehicle-Readiness- Index.pdf.
Kreimeier, M., & Stumpf, E. (2017). Market volume estimation of thin-haul on-demand air mobility services in Germany. In795
17th AIAA Aviation Technology, Integration, and Operations Conference (p. 3282).
Krueger, R., Rashidi, T. H., & Rose, J. M. (2016). Preferences for shared autonomous vehicles. Transportation Research Part C:
Emerging Technologies ,69 , 343 – 355. URL:
Kyriakidis, M., Happee, R., & de Winter, J. C. (2015). Public opinion on automated driving: Results of an international
questionnaire among 5000 respondents. Transportation Research Part F: Traffic Psychology and Behaviour,32 , 127–140.800
Lidynia, C., Philipsen, R., & Ziefle, M. (2017). Droning on about dronesacceptance of and perceived barriers to drones in civil
usage contexts. In Advances in Human Factors in Robots and Unmanned Systems (pp. 317–329). Springer.
Likert, R. (1932). A technique for the measurement of attitudes. Archives of psychology ,20 , 5–55.
MacSween, S. (2003). A public opinion survey-unmanned aerial vehicles for cargo, commercial, and passenger transportation.
In 2nd AIAA” Unmanned Unlimited” Conf. and Workshop & Exhibit (p. 6519).805
McFadden, D. (2001). Economic choices. American economic review,91 , 351–378.
Merat, N., Madigan, R., & Nordhoff, S. (2016). Human factors, user requirements, and user acceptance of ride-sharing
in automated vehicles. In International Transport Forum Roundtable on Cooperative Mobility Systems and Automated
Driving (pp. 6–7).
Mokhtarian, P. L., Ory, D. T., & Cao, X. (2009). Shopping-related attitudes: a factor and cluster analysis of Northern California810
shoppers. Environment and Planning B: Planning and Design,36 , 204–228.
NASA (2018). Urban Air Mobility (UAM) Market Study. Technical Report. URL:
NASA (2019). Urban Air Mobility (UAM) Market Study. URL:
Nees, M. A. (2016). Acceptance of self-driving cars: an examination of idealized versus realistic portrayals with a self-driving
car acceptance scale. In Proceedings of the Human Factors and Ergonomics Society Annual Meeting (pp. 1449–1453). SAGE
Publications Sage CA: Los Angeles, CA volume 60.
Payre, W., Cestac, J., & Delhomme, P. (2014). Intention to use a fully automated car: Attitudes and a priori acceptability.
Transportation Research Part F: Traffic Psychology and Behaviour ,27 , 252–263.820
Peeta, S., Paz, A., & DeLaurentis, D. (2008). Stated preference analysis of a new very light jet based on-demand air service.
Transportation Research Part A: Policy and Practice,42 , 629 – 645. URL:
Piao, J., McDonald, M., Hounsell, N., Graindorge, M., Graindorge, T., & Malhene, N. (2016). Public views towards imple-
mentation of automated vehicles in urban areas. Transportation Research Procedia,14 , 2168–2177.825
Porsche Consulting (2018). The future of vertical mobility: Sizing the market for passenger, inspection, and good services until
R Core Team (2019). R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing
Vienna, Austria. URL:
Riad, M.-I., Jaradat, m.-i., Fagih, K., & S Faqih, M. (2014). Investigating the moderating effects of gender and self-efficacy830
in the context of mobile payment adoption: A developing country perspective. International Journal of Business and
odel, C., Stadler, S., Meschtscherjakov, A., & Tscheligi, M. (2014). Towards autonomous cars: the effect of autonomy levels
on acceptance and user experience. In Proceedings of the 6th International Conference on Automotive User Interfaces and
Interactive Vehicular Applications (pp. 1–8). ACM.835
Rothfeld, R. L., Balac, M., Ploetner, K. O., & Antoniou, C. (2018). Initial Analysis of Urban Air Mobility’s Transport
Performance in Sioux Falls. In AIAA AVIATION, Modeling and Simulation for Unmanned and Personal Aerial Vehicle
Operations. Atlanta.
Rychel, A. (2016). Cultural divides: How acceptance of driverless cars varies globally. URL:
driverless-cars- acceptance/.840
Schoettle, B., & Sivak, M. (2014). A survey of public opinion about autonomous and self-driving vehicles in the US, the UK,
and Australia.
Shamiyeh, M., Bijewitz, J., & Hornung, M. (2017). A review of recent personal air vehicle concepts. In Aerospace Europe 6th
CEAS Conference.
Statistische ¨
Amter des Bundes und der L¨ander (2014). Ergebnisse des zensus 2011. URL:
Steele, C. M. (1995). Stereotype threat and the intellectual test performance. Journal of Personality and Social Psychology,
69 , 1.
Straubinger, A., & Verhoef, E. T. (2018). (Working Paper) Options for a Welfare Analysis of Urban Air Mobility. Hong Kong.
Sun, H., & Zhang, P. (2006). The role of moderating factors in user technology acceptance. International Journal of Human-850
Computer Studies,64 , 53–78.
Tabachnick, B. G., Fidell, L. S., & Ullman, J. B. (2007). Using multivariate statistics volume 5. Pearson Boston, MA.
Thipphavong, D. P., Apaza, R., Barmore, B., Battiste, V., Burian, B., Dao, Q., Feary, M., Go, S., Goodrich, K. H., Homola, J.
et al. (2018). Urban air mobility airspace integration concepts and considerations. In 2018 Aviation Technology, Integration,
and Operations Conference (p. 3676).855
Tyrinopoulos, Y., & Antoniou, C. (2008). Public transit user satisfaction: Variability and policy implications. Transport
Policy,15 , 260–272.
Uber Elevate (2016). Fast-forwarding to a future of on-demand urban air transportation. URL:
Uber Elevate (2018). Fast-forwarding to the future of on-demand, urban air transportation. URL:
Vascik, P. D. (2017). Systems-level analysis of On Demand Mobility for aviation. Ph.D. thesis Massachusetts Institute of
Vascik, P. D., & Hansman, R. J. (2018). Scaling constraints for urban air mobility operations: Air traffic control, ground
infrastructure, and noise. In 2018 Aviation Technology, Integration, and Operations Conference (p. 3849).865
Venkatesh, V., & Bala, H. (2008). Technology acceptance model 3 and a research agenda on interventions. Decision sciences,
39 , 273–315.
Venkatesh, V., & Davis, F. D. (2000). A theoretical extension of the technology acceptance model: Four longitudinal field
studies. Management Science,46 , 186–204.
Venkatesh, V., Morris, M. G., Davis, G. B., & Davis, F. D. (2003). User acceptance of information technology: Toward a870
unified view. MIS quarterly, (pp. 425–478).
Wang, Y., Xia, H., Yao, Y., & Huang, Y. (2016). Flying eyes and hidden controllers: A qualitative study of people’s privacy
perceptions of civilian drones in the US. Proceedings on Privacy Enhancing Technologies,2016 , 172–190.
Yeomans, K. A., & Golder, P. A. (1982). The guttman-kaiser criterion as a predictor of the number of common factors. The
Statistician, (pp. 221–229).875
Zhang, T., Tao, D., Qu, X., Zhang, X., Lin, R., & Zhang, W. (2019). The roles of initial trust and perceived risk in publics
acceptance of automated vehicles. Transportation Research Part C: Emerging Technologies,98 , 207–220.
... Interestingly, these concerns about GAV adoption reappear in early studies on UAM adoption. In a stated preference survey study in Germany, ref. [19] explored user perception on UAM. The study indicated that safety and trust are primary concerns for UAM adoption, and adopters are younger and have higher incomes. ...
... Urban air mobility holds the promise of becoming a greener, faster, and quieter m of aerial transportation in the near future [4]. While the adoption of UAM faces m social acceptance barriers [19,38], identifying feasible landing site locations in buil metropolitan areas remains a major physical barrier to the mass deployment of UAM. study provides an initial assessment of potential vertiports locations and associated t demand in southern California by employing a systematic scenario analysis. ...
... Urban air mobility holds the promise of becoming a greener, faster, and quieter mode of aerial transportation in the near future [4]. While the adoption of UAM faces many social acceptance barriers [19,38], identifying feasible landing site locations in built-out metropolitan areas remains a major physical barrier to the mass deployment of UAM. This study provides an initial assessment of potential vertiports locations and associated travel demand in southern California by employing a systematic scenario analysis. ...
Full-text available
The deployment of urban air mobility in built-out metropolitan regions is constrained by infrastructure opportunities, land use, and airspace zoning designations. Meanwhile, the availability and spatial distribution of infrastructure opportunities influence the travel demand that can be potentially captured by UAM services. The purpose of this study is to provide an initial assessment of the infrastructure opportunities of UAM in southern California with different mixes of spatial constraints, such as noise levels, school buffer zones, and airspace zones. The corresponding travel demand that can be potentially captured under each scenario is estimated with a home–workplace trip table. The results of the analyses indicate that supply-side infrastructure opportunities, such as heliports and elevated parking structures, are widely available to accommodate the regional deployment of UAM services. However, current spatial constraints can significantly limit the scope of vertiport location choices. Furthermore, the low-income population, blue-collar workers, and young people live farther away from supply-side opportunities than the general population. Moreover, this study proposes a network of UAM based on the top home-based and workplace-based stations for long-distance trips.
... In UAM systems and VTOL vehicles, vehicle control (piloted or automated) is at the forefront of consumers' the safety concerns. Apart from safety and security concerns commonly encountered in the adaptation of UAM system, the presence and social media impact of in-car cameras and operators, particularly in autonomous vehicles, are also among the prominent factors of UAM acceptance (Al Haddad et al., 2020). ...
... These conditions were considered, in the development of UAM-AUM to determine consumers' attitudes towards acceptance and use of UAM systems (Yavaş and Tez, 2021). Previous studies were referenced (Chen and Lu, 2016;Chancey, 2020;Davis, 1989;Davis et al., 1989;Zhang et al., 2019;Al Haddad, Chaniotakis, Straubinger, Plötner and Antoniou, 2020;Yuen et al., 2020;Yuen et al., 2021) when developing the model. In addition to the sample items and approaches, new items deemed suitable for the subject were also added. ...
This study investigates the factors affecting users' acceptance of urban air mobility (UAM) systems, constructing a proposed theoretical urban air mobility acceptance and usage model (UAM-AUM). In this context, we posit that intention to fly, UAM conceptual intention, environmental consciousness, UAM affordability, general reliability and perceived usefulness variables are key determinants for behavioural intention to use. Using the UAM-AUM scale we apply structural equation modelling analysis for model validation based on data collected from 348 participants and the results strongly support the proposed model. Perceived usefulness is found to be the strongest determinant of user behavioural intention and mediates the effects of UAM acceptance, influencing the intention to use the system along with other variables. Another critical factor for promoting consumers’ positive intentions towards UAM system use is general reliability. The findings provide practical guidance for designing strategic interventions to improve public acceptance of UAM systems.
... Die Forschung beschäftigt sich aktuell verstärkt mit dem Thema Akzeptanz. Studien zum Thema Passagierdrohnen zeigen, dass Informationen und Berührungspunkte sowie Vertrauen in die Technologie und deren Sicherheit zu einer besseren Akzeptanz beitragen (Planing und Pinar 2019, S. 3;Al Haddad et al. 2020 ...
Full-text available
Zusammenfassung Die Versorgung der nordfriesischen Inseln und Halligen ist stark durch ihre Tiden‑, Witterungs- und Saisonabhängigkeit geprägt. Die vorliegende Untersuchung prüft, ob eine Belieferung dieser schwierig zu erreichenden Gebiete mit Transportdrohnen das Potenzial hätte, mögliche Versorgungslücken zu schließen, um zur Angleichung der Lebensverhältnisse an die des Festlands beizutragen. Qualitative Interviews mit Stakeholdern der Region zeigen, dass die Saisonalität der Fähren zu Einschränkungen in der Versorgung, vor allem mit zeitkritischen Gütern, führt. Obwohl die Anwohner*innen Anpassungsstrategien wie Vorratshaltung entwickelt haben, gibt es immer wieder kurzfristig unerfüllten Bedarf an Spezialgütern. Ein anhand bestehender regionaler Versorgungsrouten erarbeitetes Drohnenkonzept zeigt, dass die Nutzung von Drohnen das Potenzial hat, durch Ergänzung des bestehenden Schiffsverkehrs die Versorgungssituation zu verbessern. Allerdings ist der Großteil der Befragten dem Einsatz von Drohnen gegenüber skeptisch. Die geäußerten Bedenken beziehen sich hauptsächlich auf Umweltschutz, Nachhaltigkeit, touristische Attraktivität und die Einsatzfähigkeit von Drohnen bei widrigem Wetter. Somit zeigt sich, dass die Akzeptanz neuer Technologien durch die Bevölkerung eine Kernherausforderung für die erfolgreiche Lösung von Versorgungsengpässen durch Transportdrohnen ist und neben der Untersuchung der technischen und ökonomischen Machbarkeit solcher Konzepte zentrale Beachtung finden muss.
... The consideration of important others toward using FAVs can be attributed to the coefficient of imitation (word-of-mouth effect) in the classic Bass Diffusion model that describes the process of how new products get adopted in a population (Stasinopoulos et al., 2021). Furthermore, Al Haddad et al. (2020), by assessing the factors influencing the adoption of urban air mobility with high automation level, stated that social behavior, defined as social attitudes regarding the use of social media, or environmental concerns, found to be an important factor in explaining the adoption of UAMs. Contradictory to many previous studies, the results of a field experiment from 117 respondents who had experienced a highly automated minibus in China revealed that SN was not a significant predictor of intention (Dai et al., 2021). ...
Fully automated vehicles (FAVs) are predicted to enter the passenger vehicle market in the near future. Being a new technology, there is significant uncertainty on adoption by users. There is a need for studies on the user acceptance of autonomous vehicles. This study aims to assess user acceptance of FAVs in a middle-income country, Iran, by adopting and comparing three popular user acceptance models: Technology Acceptance Model (TAM), Theory of Planned Behavior (TPB), and Unified Theory of Acceptance and Use of Technology (UTAUT). The data were collected via an online survey of 1381 drivers (with valid driver’s license) from different cities in Iran. The results showed that while 49.4% of participants had never heard of FAVs, the majority of them showed a high intention to use FAVs in the future. All three models successfully explained the behavioral intention to use FAVs. The TPB model was the best performing model to represent acceptance, which explained 70.9% of the variance in behavioral intention with subjective norms as the strongest predictor of the model, followed by attitudes toward FAVs as a significant predictor of intention to use them. The UTAUT model constructs accounted for 69.6% of the variance in intention, followed by the TAM model, which explained 57.6% of the variance in intention to use FAVs. Considering that the study on public perceptions of FAVs in middle-income countries is almost non-existent in the literature, it is expected that the current study will be a benchmarking resource for a comparative study for countries with a similar socio-economic and cultural context as the current study.
... Safety is highlighted as the most important factor for the adoption of Air Mobility by a majority of people [1]. In order to foster market readiness of tilt wing eVTOL, safe flight has to be maintained at all times, including abnormal situations like in the event of technical failure. ...
Conference Paper
The development of fault-tolerant control concepts for tandem tilt wing aircraft in nominal and non-nominal conditions requires an aerodynamic modeling tool that captures relevant effects while maintaining modeling flexibility. In this work, a semi-empirical approach for aerodynamic modeling with focus on the propeller-wing interaction is presented. As a foundation, two-dimensional aerodynamics of an airfoil in interaction with propellers and flaps are considered. Combining the effects of numerous airfoil sections, the approach is extended to 3D in order to model a complete aircraft by applying strip theory. Finite wing effects on the lift distribution are taken into account. Without reference data for a full tandem tilt wing configuration, validation was performed on each subcomponent individually or combinations of components. The validation tests showed satisfactory agreement with the experimental data. Together with weight & balance and propulsion models, the tool allows performing trim analyses for the complete flight envelope, which is the foundation for finding flight corridor for stable transition flight. Failure cases, like loss of a motor or stuck actuators, and their impact on the flight envelope and transition corridor can be investigated in the future. Besides these analyses, the presented model can also be used for dynamic flight simulation or aerodynamic model identification.
... • Automation of UAS-enabled AAM control and its interaction with the TMA: The diverse environment created by the substantial increase of aircraft in the urban space [147] [148] [149] will require solutions to optimize the interoperability of the airspace. Solutions that assume the UAS are also required for safe and efficient future operations; • Priority assessment in AAM disruptions: AAM is expected to bring various aircraft to fly simultaneously in the urban environment. ...
Full-text available
Air transportation is essential for society, and it is increasing gradually due to its importance. To improve the airspace operation, new technologies are under development, such as Unmanned Aircraft Systems (UAS). In fact, in the past few years, there has been a growth in UAS numbers in segregated airspace. However, there is an interest in integrating these aircraft into the National Airspace System (NAS). The UAS is vital to different industries due to its advantages brought to the airspace (e.g., efficiency). Conversely, the relationship between UAS and Air Traffic Control (ATC) needs to be well-defined due to the impacts on ATC capacity these aircraft may present. Throughout the years, this impact may be lower than it is nowadays because the current lack of familiarity in this relationship contributes to higher workload levels. Thereupon, the primary goal of this research is to present a comprehensive review of the advancements in the integration of UAS in the National Airspace System (NAS) from different perspectives. We consider the challenges regarding simulation, final approach, and optimization of problems related to the interoperability of such systems in the airspace. Finally, we identify several open challenges in the field based on the existing state-of-the-art proposals.
Conference Paper
We address the topic of network design for innovative transport systems using advanced air mobility (AAM) as an example. For any network design, nodes must be placed and connected with edges in the most optimal way possible with respect to specific objectives. Existing approaches for AAM network design focus on pre-defining nodes or directly linking nodes to existing transport systems. With these methods, the targeted identification of weaknesses in existing transport systems is not yet sufficiently taken into account. However, such targeted identification is necessary if AAM is to effect improvement for the transportation system as a whole. As a remedy, we present a metric based on general utility functions that quantifies the benefit of AAM for origindestination (OD) connections. This metric takes into account the strengths and weaknesses of existing transportation systems when designing AAM networks from scratch. The strengths and weaknesses of the existing network are identified based on the requirements of various stakeholders, which are evaluated in different scenarios. Using Bavaria as an example, we show that the proposed metric is suitable for identifying AAM connections that specifically complement the existing transportation network. Our evaluations show that AAM is particularly suitable for very specific point-to-point connections that do not form a closed network. The evaluations allow individual connections to be prioritized in a targeted manner, e.g. through subsidies. The proposed metric can also be used when planning other innovative transport systems in order to improve existing transport systems.
Not less than one hundred thousand Unmanned Aerial Vehicles (UAVs) are expected to perform flights simultaneously in Russia by 2035. The UAV fleet capacity triggers the development of the systems for informational support, operating control and management of UAV flights (Unmanned Aircraft System Traffic Management (UTM) systems) similar to that one already operating in manned aviation. The challenges arising in the sphere of civil aviation cannot be solved without wireless communication. The goals of this article are as follows: 1) familiarization of communication experts with the latest scientific developments of unmanned aerial technologies 2) description of the telecommunication-related problems of extensive systems of UAV control encountered by development engineers. In this article a schematic architecture and main functions of UTM systems are described as well as the examples of their implementation. Special emphasis is put on enhancing flight safety by means of a rational choice of communication technologies to manage conflicts (Conflict Management) known as "collision avoidance". The article analyzes the application of a wide range of wireless technologies ranging from Wi-Fi and Automatic Dependent Surveillance Broadcast (ADS-B) to 5G cellular networks as well as cell-free networks contributing to the development of 6G communication networks. As a result of the analysis, a list of promising research trends at the intersection of the fields of wireless communication and UAVs for civil application is made.
This research paper presents a novel approach for cost-benefit analysis for the bi-directional functionality of eVTOL (Electric Vertical Take-Off and Landing) aircraft charging. eVTOL aircrafts represents an emerging aviation market that seeks to revolutionize mobility, both intra-city and inter-city, through an air transportation system for passengers based on safety, efficiency and accessibility on demand. In this research, an optimization based on genetic algorithms has been used to find the number of eVTOL vehicles which makes vehicle-to-grid services profitable for each of the electrical systems considered. This research discusses the impact that vehicle-to-grid eVTOL aircraft would cause, changing the current demand curve peaks and the economic benefits for eVTOL end users, power plant companies and system operator. The results show that there is an increase in the penetration of renewable energy sources, causing a reduction of approximately 4.6% in global emissions. In addition, the benefits obtained by the stakeholders considered in the system is greater than zero. The methodology considered in this research offers a new perspective for electrical systems to predict the demand of eVTOL aircraft, in specific territories, such as in the presented case study in the Archipelago of Canary Islands.
Urban Air Mobility (UAM) has lately emerged as a time-saving mode of air transportation in congested urban areas. However, several challenges to the commercialization and adoption of UAM vehicles exist, such as licensing and restrictions, security, and construction infrastructure. Besides these, public perception is one of the significant aspects of easing psychological stress and representing users’ needs. A survey study was carried out to understand the public perception of UAM and to comprehend how the public perceives and expects UAM aircraft in terms of user requirement analysis as part of human-centered design. To achieve this, a total of 2,847 valid data sets were analyzed. A preliminary data analysis revealed the general level of awareness of UAM machines, expected costs and values for specific scenarios, perceived advantages of UAM vehicles, and overall opinion, along with distinctions based on demographic information such as age groups, residential areas, and income levels. Future research topics might include in-depth data analysis and subsequent user surveys to discover underlying requirements.
Full-text available
The advancement of electric Vertical Take-Off and Landing (eVTOL) vehicles in the world of aviation leads to a novel way of travel. Suggested to navigate in a metropolitan area, the availability of take-off and landing sites for eVTOL within the boundaries of the city are sparse. This research conducts a different approach in Urban Air Mobility (UAM) ground infrastructure placement using a geographic information system (GIS)-based analysis. A suitability analysis using the weighted linear combination (WLC) method is implemented to locate suitable areas for UAM ground infrastructure in a metropolitan area. Los Angeles and Munich are the two metropolitan areas that are taken as a case study for applying the GIS-based UAM ground infrastructure placement. The results of case studies in two metropolitan cities and three different methods show that the city center (which has high office rent prices in Munich and Los Angeles), airports and intercity train stations are most suitable for the initial operation of UAM. Point of interests (POIs) should also be taken into account to catch the niche market of tourism trips.
Full-text available
The literature on the acceptance or adoption of unmanned aircraft, aka civil drones, is scarce and underdeveloped. The majority of literature discusses several social concerns, whereas only little research promotes economic concerns. Additionally, the literature does not consider existing technology acceptance theory, but arbitrarily investigates the acceptance of several applications. This research discusses both social and economic concerns and establishes for a conceptual framework for predicting the public acceptance of drones. The conceptual framework considers the contemporary issues (i.e. social and economic concerns) and is based on risk theory. It proposes that the perception of risk associated with the use of drones influence the public acceptance of the technology. Social and economic concerns, in turn, are proposed as constituents of perceived risk. The proposed model establishes for further theoretical development and empirical assessment in pursue of deeper explanation of the factors which influence public acceptance. The research concludes with stating limitations and proposing recommendations for future research.
Urban Air Mobility (UAM) is a recent mobility concept with the potential to reduce travel time and change travel patterns. When evaluating the introduction of UAM, understanding the potential users’ choice behavior regarding current available urban transportation modes and autonomous transportation services is essential to demand estimation. This preliminary research intends to gain insight into the travel behavior impacts of autonomous transportation modes, especially UAM, by deriving measures for transportation service attributes and identifying characteristics of potential users who might adopt autonomous transportation services, particularly the services of UAM. Thus, a stated preference questionnaire was designed and distributed in Munich metropolitan region. A main mode choice multinomial logit model and several sub-models, based on market segmentation, were estimated regarding four transportation alternatives: private car, public transportation, autonomous taxi, and autonomous flying taxi. The results indicate that travel time, travel cost, and safety may be critical determinants in autonomous transportation mode adoption. The potential consumers may be willing to pay more for using autonomous transportation modes, especially the service of UAM. Among different market segments, younger individuals, as well as older individuals with high household income, are more likely to adopt UAM. In addition, during the market entry stage, potential travelers may favor UAM particularly for performing non-commuting trips.
Conference Paper
Urban Air Mobility (UAM) is currently often being discussed as a disruptive new mode of transport. Due to its potential to change urban transport systems and its large need for dedicated infrastructure it is essential to understand the overall impact of UAM introduction on the city. Hereby it is crucial to not only take changes in the transport system into account but to also consider changes on the labor, land and goods market, meaning that households' utility, companies' profit and location choice are affected. An evaluation of these effects can be performed by applying an urban spatial computable general equilibrium (CGE) model. The model is utility based and therefore, can assess welfare changes within in the city. This paper gives a first overview over required model extensions for UAM integration. A preliminary model shows first results that follow the expectation we got from analytical considerations of a monocentric city. The basic model shows that rents in all city zones increase and inhabitants tend to move from the city center to the suburbs due to the new transport mode. Yet, in the current setting the household's average utility decreases.
Connected and fully automated or autonomous vehicles (CAVs) may soon dominate the automotive industry. Once CAVs are sufficiently reliable and affordable, they will penetrate markets and thereby generate economic ripple effects throughout industries. This paper synthesizes and expands on existing analyses of the economic effects of CAVs in the United States across 13 industries and the overall economy. CAVs will soon be central to the automotive industry, with software composing a greater share of vehicle value than previously. The number of vehicles purchased each year may fall because of vehicle sharing, but rising travel distances may increase vehicle sales. The opportunity for heavy-truck drivers to do other work or rest during long drives may lower freight costs and increase capacity. Personal transport may shift toward shared autonomous vehicle fleet use, reducing that of taxis, buses, and other forms of group travel. Fewer collisions and more law-abiding vehicles will lower demand for auto repair, traffic police, medical, insurance, and legal services. CAVs will also lead to new methods for managing travel demand and the repurposing of curbside and off-street parking and will generate major savings from productivity gains during hands-free travel and reduction of pain and suffering costs from crashes. If CAVs eventually capture a large share of the automotive market, they are estimated to have economic impacts of $1.2 trillion or $3,800 per American per year. This paper presents important considerations for CAVs’ overall effects and quantifies those impacts.