Content uploaded by Earl Huff
Author content
All content in this area was uploaded by Earl Huff on Aug 15, 2020
Content may be subject to copyright.
3
Exploring the Needs, Preferences, and Concerns of Persons
with Visual Impairments Regarding Autonomous Vehicles
JULIAN BRINKLEY and EARL W. HUFF JR., Clemson University
BRIANA POSADAS and JULIA WOODWARD, University of Florida
SHAUNDRA B. DAILY, Duke University
JUAN E. GILBERT, University of Florida
Fully autonomous or “self-driving” vehicles are an emerging technology that may hold tremendous mobil-
ity potential for blind or visually impaired persons who are currently unable to drive a conventional motor
vehicle. Despite the considerable potential of self-driving vehicle technology to address this mobility issue,
however, the needs and preferences of persons with visual disabilities regarding this technology have been
insuciently investigated. In this article, we present the results of two studies that are focused on exploring
the needs, preferences, and concerns of persons with visual impairments as it relates to self-driving vehicles.
Study one investigated user acceptance, concerns, and willingness to buy partially and fully automated vehi-
cles using a 39-question Internet-based survey distributed in the United States to visually impaired respon-
dents (n=516). Study two explores the opinions of 38 participants who are blind and low vision, using focus
group methodology, regarding emerging self-driving vehicle technology. Collectively our ndings suggest
that while persons with visual impairments may be optimistic regarding the potential for enhanced mobility
and independence that may result from the emergence of self-driving vehicles, concerns exist regarding the
implementation of this technology that have been largely unexplored and under investigated.
CCS Concepts: • Human-centered computing →Accessibility;Empirical studies in accessibility;
Additional Key Words and Phrases: Autonomous vehicles, self-driving vehicles, accessibility, visual impair-
ment
ACM Reference format:
Julian Brinkley, Earl W. Hu Jr., Briana Posadas, Julia Woodward, Shaundra B. Daily, and Juan E. Gilbert. 2020.
Exploring the Needs, Preferences, and Concerns of Persons with Visual Impairments Regarding Autonomous
Vehicles. ACM Trans. Access. Comput. 13, 1, Article 3 (April 2020), 34 pages.
https://doi.org/10.1145/3372280
1 INTRODUCTION
As autonomous vehicle technology transitions from the realm of science ction to scientic
reality, the potential benets of this technology have become ever more broadly discussed.
Developers of this technology imagine that the most advanced of such vehicles, referred to as
Authors addresses: J. Brinkley and E. W. Hu Jr., 226C McAdams Hall, Clemson, SC, 29634, USA; emails: julianbrink-
ley@clemson.edu, earlh@g.clemson.edu; B. Posadas, J. Woodward, and J. E. Gilbert, P.O. Box 116120, Gainesville, FL, 32611,
USA; emails: {bposadas, julia.woodward, juan}@u.edu; S. B. Daily, 100 Science Drive, Hudson Hall Suite 130, Durham, NC
27710, USA; email: shani.b@duke.edu.
Contact author: J. Brinkley, julianbrinkley@clemson.edu.
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee
provided that copies are not made or distributed for prot or commercial advantage and that copies bear this notice and
the full citation on the rst page. Copyrights for components of this work owned by others than ACM must be honored.
Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires
prior specic permission and/or a fee. Request permissions from permissions@acm.org.
© 2020 Association for Computing Machinery.
1936-7228/2020/04-ART3 $15.00
https://doi.org/10.1145/3372280
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
3:2 J. Brinkley et al.
Level 5 Automation or fully “self-driving” vehicles [35,51] will operate with minimal human
interaction and no direct human manipulation of safety-critical controls (e.g., accelerator, brake
pedals or steering apparatus). Removing error prone human beings from manual driving may
reduce by as much as 90% [45,61,64] the types of motor vehicle crashes that claim thousands of
lives annually and result in costly property damage and loss [35]. Self-driving vehicle technology
itself may render personal vehicle ownership obsolete as transportation is increasingly consumed
as a service [12]. This, in turn, may change the nancial models of a host of major industries and
the physical design of cities themselves [19,58]. To realize these benets consumer adoption of
such a paradigm shifting technology is critical given the potentially high costs of the technology
and the potential safety concerns of automated driving [13]. While discussions of the impact of
self-driving vehicle technology taking place in academia, regulatory bodies and in industry are
being accompanied by wide ranging consumer and user research, there are still large knowledge
gaps as it relates to the needs and preferences of specic groups of users. It has been suggested,
for instance, that most self-driving vehicle technology being developed is not in fact accessible to
individuals with visual impairments [44]. We argue that this may be at least partially attributable
to a knowledge gap as it relates to the needs and preferences of individuals with visual impair-
ments regarding self-driving vehicle technology. A knowledge gap driven by the reality that most
self-driving technologies are being designed around the driver of the present who in all cases is
sighted as opposed to the operator of the future who need not necessarily be.
We present the results of two studies conducted with an intent to explore the needs, prefer-
ences and concerns of persons with visual impairments regarding self-driving vehicles. In the
initial study, a 39-question Internet-based survey was distributed in the United States. The sur-
vey collected 516 useable responses from persons 18 years and older who self-identied as blind
or visually impaired. Respondents were asked questions about their familiarity with emerging
self-driving vehicle technology, their general opinion about such vehicles, the anticipated bene-
ts of the technology and opinions on relevant issues related to visual impairment and blindness.
In the subsequent study, using focus group methodology, 38 people who are blind and low vi-
sion participated in eight focus groups over a two-day period. Participants were asked to provide
their general opinions regarding self-driving vehicles, comment on their hopes for the technol-
ogy, reect on their concerns, and express their preferences regarding interaction mechanisms in
a semi-structured group discussion lasting approximately one hour. We believe that collectively
this research furthers our goal of contributing to the literature research that furthers the under-
standing of the needs and preferences of users with a range of visual impairments as it relates to
emerging self-driving vehicle technology. Research of this type will become increasingly critical if
universal access to this technology is to be realized. The present article combines and signicantly
extends two previously published works; a paper presented at the ACM SIGACCESS Conference
on Computers and Accessibility that is available within the conference proceedings [10], and an
extended abstract presented at the CSUN Assistive Technology Conference that is available in the
Journal on Technology and Persons with Disabilities [9].
2 RELATED WORK
The studies described in this manuscript were motivated by knowledge gaps in the related user
research on autonomous vehicles. To provide context to the studies described, we provide back-
ground on vehicle automation levels, existing consumer/user research on autonomous vehicles,
spatial cognition, and navigation for blind and visually impaired persons.
2.1 Levels of Vehicular Automation
The Society of Automotive Engineers (SAE) has described six levels of vehicular automation that
have been adopted by the National Highway Transportation and Safety Administration (NHTSA),
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
Visual Impairments and Self-driving Vehicles 3:3
the regulatory body responsible for vehicle automation in the United States [45]. The SAE Au-
tomation Levels are described in the SAE J3016 standard [3,51,52] and describe levels 0 through
5, which represent escalating levels of vehicular automation. Level 0, for instance, represents no
automation and completely conventional or manual driving. Level 3 vehicles, however, have tech-
nology for “conditional automation” meaning that automated driving is possible under certain
conditions. With Level 3 vehicles, such as Tesla’s equipped with Autopilot [41], the vehicle is ca-
pable of certain automated highway driving, but the driver must be prepared to assume manual
control when notied by the vehicle. Level 5 vehicles, however, can perform all driving functions
under all conditions though the driver may have the option of assuming manual control if desired.
At present only automation levels 0 through 3 are commercially available in the United States. We
primarily focus our discussion on Level 5 or fully “self-driving” vehicles given that this level of
automation holds the most promise for blind and visually impaired (BVI) users who cannot legally
operate a Level 0 through 3 vehicle with existing technology.
2.2 Consumer Opinions on Autonomous Vehicles
There have been a number of studies in recent years that have investigated public opinion re-
garding automated driving conceptually, have explored consumer preferences regarding specic
self-driving vehicle technologies and have examined related user needs.
In a 2013 survey conducted by Continental AG in Germany, China, Japan, and the U.S., 59% of
respondents considered automated driving to be a “useful advancement,” but 31% of respondents
in all three countries found the development of automated vehicles to be “unnerving” [59]. Sixty-
six percent of U.S. respondents, for instance, indicated that they were “scared” by the concept of
automated driving. Respondents were also skeptical of the reliability of the technology with 74%
of respondents in China and 50% of respondents in the U.S., indicating that they did not believe
that it would function reliably.
In 2013 professional services company KPMG conducted a focus group study with a total of
32 participants across California, Chicago and New Jersey [26]. Their results showed that women
(median =8.5 on a scale from 1 to 10) were more willing to use self-driving vehicles than men
(median 7.5). Safety was a dominant topic of discussion during the focus groups with many par-
ticipants expressing skepticism that the technology would work properly. Participants were near
unanimous in expressing a need to be able to take control of the vehicle at will for a variety of rea-
sons. Some participants expressed a lack of trust in the automated systems and expressed comfort
in having manual controls available. Participants also expressed joy in driving and appreciated
having manual controls as an option.
Howard and Dai in a 2013 survey explored the opinions of 107 respondents in Berkley, CA re-
garding self-driving vehicles [22]. Respondents indicated that safety (75%) and convenience (61%)
were the most attractive features of the technology. More than 40% of respondents expressed will-
ingness to either purchase a fully self-driving vehicle as their next vehicle or to retrot their ex-
isting vehicle with self-driving technology if such an option were made available. Respondents
indicated that liability concerns and the potential cost were the least attractive features of the
technology.
In a 2014 survey involving 1,533 respondents in the U.S., U.K., and Australia, Schoettle and Sivak
found that more than 60% of respondents in all three countries were aware of the technology and
more than 50% had positive expectations about its potential benets (e.g., less trac congestion,
shorter travel time) [54]. Respondents in all three countries expressed concerns about self-driving
vehicles, however. Across all three countries the most signicant concerns were expressed re-
garding system/equipment failure, followed by vehicle performance in unexpected situations. Ad-
ditionally, more than 90% of respondents expressed some degree of concern regarding the legal
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
3:4 J. Brinkley et al.
liability of drivers/owners of self-driving vehicles. A majority of respondents in all three countries
expressed some interest in having self-driving technology but were generally unwilling to pay
extra for it.
Payre, Cestac, and Delhomme conducted a public opinion survey in 2014 of French drivers to
investigate opinions on fully automated driving [46]. The study examined attitudes and acceptabil-
ity of fully automated driving technology among 421 drivers nding that men and those scoring
highly on the driving-related sensation seeking scale were more willing to use a fully automated
vehicle and were more inclined to purchase a self-driving vehicle. Older respondents were less
likely to indicate that they would purchase such a vehicle but showed higher acceptance of the
technology. Respondents expressed a preference for full automation on highways, in trac con-
gestion, for automatic parking, and when impaired by drug use or alcohol.
Ipsos MORI conducted a public opinion survey in 2014 involving 1,001 British respondents to
investigate attitudes related to cars and technological developments surrounding the automotive
industry [40]. Only 18% of those surveyed felt that it was important for car manufacturers to focus
on driverless technologies while 41% indicated that it was unimportant. Older people (55+)were
less likely to embrace the technology than the youngest group (16 to 24) and 50% of those aged
55+felt that the technology was unimportant compared to 30% of those in the 16 to 24 age group.
A similar study was conducted by Kyriakidis, Happee, and de Winter who conducted an interna-
tional public opinion survey in 2015 involving 5,000 respondents from 109 countries to investigate
user acceptance, concerns and willingness to buy partially, highly and fully automated vehicles
[28]. A plurality of respondents (22%) indicated that they were unwilling to pay any money for a
fully automated driving system, while 5% indicated a willingness to pay more than $30,000 for it.
Respondents were most concerned about software hacking/misuse, legal issues and safety.
In a 2016 survey involving 618 respondents in the U.S., Schoettle and Sivak found that among
those surveyed the most frequent preference for vehicle automation was, in fact, “no automation”
followed by a preference for partially self-driving vehicles [55]. Approximately 16% of respondents
indicated a preference for self-driving vehicles. Over 90% of respondents expressed some degree of
concern regarding self-driving vehicles when presented with a scenario where a self-driving vehi-
cle would be their only means of transportation. The vast majority of respondents also expressed a
desire for manual vehicle controls (e.g., steering wheel, gas pedal and brake) with 94.5% of respon-
dents indicating that they would like a self-driving car to have such controls to enable a human
driver to take control in the event of an emergency. In terms of entering a route or destination in
a self-driving vehicle, 38% of respondents preferred the use of a touchscreen compared to 34.5%
who preferred the use of a voice commands, 7.9% preferred the use of a personal portable device.
Bansai, Kockelman, and Sing in a 2016 study involving 347 respondents in Austin, Texas found
that respondents on average were willing to pay $7,253 to add Level 4 automation or full self-
driving capabilities to their vehicles [4]. This stands in contrast to the ndings of much of the
previously described research but is largely consistent with the ndings of Daziano, Sarria, and
Leard [13], who found in a 2017 study involving 1,260 respondents that the average household was
willing to pay $4,900 for self-driving capabilities.
Brewer and Kameswaran in a 2018 design-based focus group study involving 15 BVI partici-
pants explored preferences for varying levels of automation and the desire for control [8]. They
found that despite being legally unable to drive, several participants indicated that they continued
to drive with the assistance of telescopic devices. The biggest perceived benet of autonomous
vehicles for participants was in improved independence and quality of life. This factor, while of
signicance to participants of the design-based focus groups, was largely unexplored by prior
research given the de facto focus on consumers with sight. Excitement regarding this potential
was tempered, however, by concerns regarding control and the human-machine interface. Control
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
Visual Impairments and Self-driving Vehicles 3:5
was explored within the focus groups using a design activity wherein participants conceptualized
assistive technologies that would be benecial within the autonomous vehicle context. Brewer
and Kameswaran found that participants broadly were interested in a means of exercising direct
control of the vehicle, though the means of such control often diered between participants. Par-
ticipants expressed concerns specically that the voice and tactile systems used for control may
malfunction or misinterpret a user’s actions.
Similar in design to the study of Brewer and Kameswaran, Hu et al. conducted focus groups
involving 39 African American older adults (55+), exploring participant perceptions regarding self-
driving vehicles generally, the technology’s development and its potential use [23]. They found
that participants felt condent in their ability to operate a self-driving vehicle and take control in
the event that a situation deemed that necessary, ndings that largely mirror those of Schoettle
and Sivak [54,55], and were optimistic regarding the potential mobility benets of the technology.
Participants questioned, however, whether the needs of older adults and persons with disabilities
were being properly considered in the design of automated vehicles broadly and were concerned
that the absence of such consideration might render the technology ultimately inaccessible. These
sentiments were followed by concerns regarding what was anticipated to be high initial costs of
ownership and a general sentiment that such vehicles might be prone to errors and malfunctions.
Each of the previously described studies had specic foci but contributed generally to the un-
derstanding of public opinion and consumer preferences as it pertains to self-driving vehicles.
Viewed collectively, while these studies advance the understanding or consumer needs and user
preferences regarding vehicle automation, none but the study of Brewer and Kameswaran directly
and specically incorporated respondents and participants with visual disabilities. We argue that
the present report is unique in its focus on the opinions of blind and low-vision persons whose
preferences and concerns, we argue, have not been adequately explored in recent research.
2.3 Spatial Cognition of Blind and Visually Impaired Persons
To contextualize our studies on autonomous vehicle preferences of visually impaired persons, we
provide background on the mobility, spatial cognition and navigation challenges faced by BVI
individuals as well as mobility-related assistive technologies. This discussion is germane to the
larger discussion of self-driving vehicles and BVI persons in the sense that self-driving vehicles
may be viewed as an especially unique mobility technology. As such, there are both potential
mobility benets for visually impaired persons and potentially unique challenges with respect to
spatial knowledge acquisition, orientation and navigation.
2.3.1 Spatial Navigation. Orientation and mobility are common terms found in the literature
related to visual impairment and spatial navigation. Montello [43] distinguishes between both
terms and provides an explanation of their use in the related literature. Mobility (locomotion) [43],
which is inherently egocentric, pertains to the immediate response to environmental features such
as stepping onto a sidewalk, avoiding an obstacle or gaining awareness of a nearby landmark [17].
Orientation (waynding) [43] involves reasoning about immediate [21] and remote [67] environ-
ments and accordingly may involve short term or long-term mental representations as well as
egocentric or allocentric perspectives. Orientation is dependent upon mobility skills and requires
both an awareness of a person’s current position and heading in the environment with respect to
their desired goal as well as the ability to update this information during travel. Golledge [18]has
described orientation as being critical for planning and determining routes through an environ-
ment, especially if these routes have not been previously traveled.
Waynding may prompt a navigator to at times adopt either an egocentric or allocentric per-
spective. In an egocentric (self-based) perspective or frame of reference, information such as
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
3:6 J. Brinkley et al.
direction and distance is presented in relation to the position and orientation of the observer. In an
allocentric or absolute frame of reference, information is presented independent of the observer
but with respect to a xed point of reference. Both frames of reference can be used to represent
spaces at spatial scales that are described somewhat dierently in the recent literature but with
some degree of overlap. Schinazi et al. [53], for instance, identify three categorizations of scales:
micro-, meso-, and macroscales. Giudice [17], however, presents four categorizations: gural, vista,
environmental, and geographical space. Microscale and vista space refer to space that does not re-
quire fully-body movement, such as objects on a tabletop. Mesoscale and vista space describe space
that is larger than the observer but can be seen from a single perspective such as an indoor room.
Macroscale and geographical space describe space that is larger than the observer but can only be
apprehended from multiple viewpoints.
2.3.2 Spatial Knowledge Acquisition. Researchers have described two frameworks for the ac-
quisition of spatial knowledge in the discrete [42] and continuous [11] frameworks. In the discrete
framework spatial knowledge is acquired via three distinct stages: landmark, route and survey
[57]. In the landmark stage knowledge is gained regarding features of the environment that may
be used for establishing a frame of reference. In the route stage, landmarks become connected
through routes that may enable a navigator to draw inferences regarding the straight-line dis-
tance between two points. In the survey stage these mini-maps are integrated using an objective
frame of reference into a global representation that is at times referred to as a cognitive map [56].
2.3.3 Models of Spatial Development. Fletcher [15] proposed the Dierence, Deciency, and
Ineciency theories of spatial development for BVI persons that have been subsequently extended
to the Convergent model, Cumulative model, and Persistent model, respectively. These theories
posit that vision provides sighted individuals with an initial advantage relative to blind individuals
though, within the literature, there are cases in which the blind outperform the sighted [5].
The Dierence/Convergent model [53]suggeststhatblindpeoplebeginatadisadvantagerela-
tive to sighted individuals but that the disparity decreases with experience until similar levels of
performance are realized. The Deciency/Cumulative model [53] suggests that vision is critical
for the development of spatial representation. While blind people may be able to acquire spatial
knowledge, the theory suggests, in the absence of vision blind persons are incapable of forming
spatial representations. The Ineciency/Persistent model [53] suggests that the absence of vision
results in an initial disadvantage that remains constant with experience, because auditory and pro-
prioceptive cues are less eective for spatial knowledge acquisition than vision. Of the three, Giu-
dice [17] argues that literature on the Dierence/Convergent model is mixed and therefore hard to
interpret and the Deciency/Cumulative model is not born out in the related literature. The Ine-
ciency/Persistent model, however, is supported by some studies in which early blind persons have
worse performance on spatial tasks such as updating and cognitive map development compared
to people with late-onset blindness or low vision on the same tasks [47,48].
2.3.4 Spatial Processing and Representation. Research has suggested that spatial representa-
tions can be abstracted from dierent perceptual modalities [29,30,62]. These modalities dier
with respect to simultaneous versus sequential information acquisition. Prior research has sug-
gested that vision enables simultaneous perception [16,38], whereas audition [7] and haptics [49]
allow for sequential perception. Vision conveys an advantage, it is argued [53], in the speed with
which the eyes can move compared to the speed of head and body movements. Vision provides
rapid access to highly precise distance and direction information about positioning of various
local and distant landmarks [17]. Vision also provides geometric information about spatial struc-
tures, aords object recognition over a large eld of view and aords access to precise motion
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
Visual Impairments and Self-driving Vehicles 3:7
Table 1. Breakdown of Survey
Respondents by Gender
Gender Percent (n=516)
Female 54.41
Male 45.40
cues for changing self-to-object and object-to-object relations that occur during navigation [17].
While other senses may provide similar cues, they are more limited and less precise than what
vision can produce. However, recent evidence has suggested that BVI persons can achieve similar
spatial behavior and performance when learning from a combination of dierent spatial inputs,
which is referred to as functional equivalence [17,53]. Specically, amodality or the Amodal Hy-
pothesis claims that spatial representations can be abstracted from dierent perceptual modalities,
thus building an amodal or sensory-independent representation that is not tied to any particular
sensory or cognitive source [17,32,33].
Navigating large, unknown environments is arguably the most dicult task faced by blind and
visually impaired travelers [17]. When coupled with the limited training typically experienced by
BVI persons on complex spatial skills, self-driving vehicles represent a potential mobility solution
that may also introduce new challenges with respect to BVI spatial cognition given the inherently
large and unknown nature of the operating environment.
3 STUDY 1: AN ONLINE SURVEY REGARDING SELF-DRIVING VEHICLES
3.1 Method
3.1.1 Online Survey. The initial study was conducted as an online survey using the Qualtrics
[65] survey platform. The questionnaire was adapted from a public opinion survey regarding self-
driving vehicles in the U.S., U.K., and Australia conducted by Schoettle and Sivak [54], with format
modications designed to enable screen reader accessibility, scale adjustments and content mod-
ications intended to address topics related to visual impairment. The order of the questions was
the same for all respondents. The topics addressed by the survey are described below:
•Respondent familiarity with self-driving vehicles (1 Question: Q1)
•General opinions about self-driving vehicles (2 Questions: Q2, Q31)
•Concerns regarding self-driving vehicles (18 Questions: Q3–Q5, Q14–Q28)
•Anticipated benets of self-driving vehicles (8 Questions: Q6–13)
•Willingness to pay for self-driving vehicle technology (2 Questions: Q29, Q30)
•Issues related to visual impairment and blindness (3 Questions: Q37–Q39)
•Demographic data (5 Questions: Q32–Q36)
Responses were gathered from January 4, 2017 through April 12, 2017.
3.1.2 Respondents. Participants were recruited through email notications distributed by
16 state agencies for the blind and by the American Council of the Blind [2]. Participation was
restricted to individuals 18 years of age and older whom self-identied as blind or visually im-
paired. Participants were entered into a drawing for a $300 prepaid gift card as compensation.
This recruitment strategy resulted in 556 replies from potential respondents with completed sur-
veys received from 516 respondents. The nal response rate of the survey was 92.8%. The margin
of error at the 95% condence level for the results is +/−4.0%. Demographic breakdowns for the
respondents are provided in Tables 1through 5. Approximately 54% of respondents were female
and approximately 45% were male (Table 1). More than half of respondents were 45 years of age
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
3:8 J. Brinkley et al.
Table 2. Breakdown of Survey
Respondents by Age Group
Age group Percent (n=516)
18–24 7.69
25–34 18.65
35–44 17.11
45–54 18.26
55–64 24.23
65–69 8.00
70+5.96
Table 3. Breakdown of Survey Respondents
by Level of Education
Education Percent (n=516)
Some High School 0.96
High School 8.25
Some College 20.54
Two-year Degree/Associate Degree 11.32
Bachelor’s Degree 31.86
Graduate Degree 27.06
Table 4. Breakdown of Survey Respondents by
Level of Employment
Education Percent (n=516)
Employed Full-time 35.12
Employed Part-time 13.82
Not Currently Employed 19.77
Retired 21.31
Full-time Student 7.68
Part-time Student 2.30
or older, while those in the 18–44 age range made up 43.45% of those participating in the survey
(Table 2). Nearly 60% of respondents held at least a bachelor’s degree (58.92%), while fewer than 1%
had less than a high school education (Table 3). Those employed full-time (35.12%) exceeded the
combined number of respondents who were full-time students, part-time students and those em-
ployed part-time (23.8%) as illustrated in Table 4. More than half of respondents (55.34%) indicated
that they had been blind or visually impaired all of their lives.
3.2 Results
3.2.1 General Opinion of Self-Driving Vehicles. The survey began with a paragraph-length
overview of vehicle automation and a description of self-driving vehicles specically. A major-
ity of survey respondents had heard of self-driving vehicles prior to the survey (95.96%) with most
respondents having a positive impression of the technology (50.18% extremely positive, 30.44%
moderately positive, and 7.75% slightly positive). Fewer than 10% had a negative impression of the
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
Visual Impairments and Self-driving Vehicles 3:9
Fig. 1. Summary of responses to Q6–Q13: “Regarding self-driving vehicles, how likely do you think the following
benefits will occur...?” All variations of “likely” and “unlikely” have been tallied.
technology with 2.03% of respondents indicating that they held an “extremely negative” impression
of self-driving vehicle technology.
3.2.2 Expected Benefits of Self-Driving Vehicles. Respondents were asked eight questions re-
lated to the anticipated benets that might occur through the use of self-driving vehicle technol-
ogy. With each question they were asked to select “extremely likely,” “moderately likely,” “slightly
likely,” “neither likely nor unlikely,” “slightly unlikely,” “moderately unlikely” or “extremely un-
likely.” Figure 1illustrates respondent perception of potential benets accounting for all variations
of “likely” (“extremely,” “moderately,” and “slightly”), “neither likely nor unlikely” and all varia-
tions of “unlikely” (“extremely,” “moderately,” and “slightly”). The majority of respondents felt
that each of the expected benets were likely to occur with self-driving vehicles with respondents
expressing the most condence in the likelihood of fewer automobile crashes (79.96% when all
variations of “likely” combined), reduced severity of automobile crashes (79.21%) and better fuel
economy (75.76%). Lower insurance rates were viewed as least likely (27.52% when all variations
of “unlikely” combined).
3.2.3 Self-Driving Vehicle Operational Concerns. Respondents were asked how concerned they
would be about riding in a fully autonomous or self-driving vehicle as the primary operator. A
denition describing a fully autonomous or self-driving vehicle accompanied the question. The
most frequently selected response was “slightly concerned” (38.96%), followed by “moderately
concerned” (22.82%), “very concerned” (16.70%), and “not at all concerned” (21.52%). Subsequently,
respondents were asked how concerned they would be about riding in a partially autonomous ve-
hicle as the primary operator. A denition describing a partially autonomous vehicle accompanied
the question. The most frequently selected response was “slightly concerned” (30.91%), followed by
“very concerned” (27.56%), “not at all concerned” (23.84%), and “moderately concerned” (17.69%).
A majority of respondents expressed some degree of concern regarding their ability to operate a
self-driving vehicle if one was made available to them (32.16% slightly concerned, 15.80% mod-
erately concerned, 17.66% very concerned). The most frequently selected response, however, was
“not at all concerned” (34.49%).
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
3:10 J. Brinkley et al.
Fig. 2. Summary of responses to Q19–Q28; “not at all concerned” is not displayed: “Regarding self-driving
vehicles, how concerned are you about...”
3.2.4 Self-Driving Vehicle Issue-based Concerns. Respondents were asked 10 questions related
to self-driving vehicle related issues; Figure 2provides a summary of their responses. For each
question respondents were asked to select “very concerned,” “moderately concerned,” “slightly
concerned,” or “not at all concerned.” Respondents expressed the most concern (when all varia-
tions of concern are accounted for) about equipment failure or system failure (93.18%), followed
by vehicles getting confused in unexpected situations (92.69%) and the interaction between self-
driving vehicles and pedestrians and bicycles (87.55%). The least concern was expressed about
learning to use self-driving vehicles (44.09%).
3.2.5 Self-Driving Vehicle Scenario-based Concerns. Respondents were presented with ve po-
tential scenarios involving self-driving vehicles; Figure 3provides a summary of their responses.
For each scenario they were asked to select “very concerned,” moderately concerned,” “slightly
concerned,” or “not at all concerned.” Respondents expressed the most concern (when all varia-
tions of concern are accounted for) about self-driving commercial vehicles (85.07%) followed by
self-driving public transportation (e.g., buses, 80.5%). Respondents were least concerned about the
prospect of self-driving vehicles moving by themselves from one location to another (39.01% “not
at all concerned”).
3.2.6 Ownership Interest and Willingness to Pay. More than 90% of respondents expressed some
interest in owning self-driving vehicle technology with 93.31% indicating that they were “ex-
tremely/very/moderately/slightly interested.” Respondents on average indicated that they were
willing to pay $6,346 US extra for this technology with those at the 50th percentile indicating that
they would pay $1,000 extra and those at the 90th percentile indicating that they would pay $10,000
extra. About a third (n=171) of respondents (33.11%) indicated that they would not be willing to
pay extra for self-driving vehicle technology.
3.2.7 Self-Driving Vehicle Travel Time. Respondents were asked how they would occupy their
time were they to travel in a self-driving vehicle and were presented with a list of nine options
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
Visual Impairments and Self-driving Vehicles 3:11
Fig. 3. Summary of responses to Q30–Q34; “not at all concerned” is not displayed: “Regarding self-driving
vehicles, how concerned are you about...”
Table 5. Percentage of Responses to Q37: “If you Were to Ride in a
Completely Self-driving Vehicle What Do You Think You Would Use the
Extra Time Doing Instead of Monitoring the Roadway?”
Response Percent
Text or Talk with Friends or Family 14.94
Read 20.31
Sleep 2.87
Watch movies/TV 1.92
Play Games 1.34
Work 12.07
Monitor the Road Even Though I Would Not Be Driving 31.99
Other 9.39
I Would Not Ride in a Completely Self-driving Vehicle 5.17
from which to choose from. Table 5provides a summary of responses. The most frequently chosen
response was “monitor the road even though I would not be driving” (31.99%), with “read” (20.31%)
being the second most frequently chosen response.
3.2.8 Issues Related to Visual Impairment and Blindness. Each respondent was asked if they
agreed that their needs, as a person with a visual disability, were being considered in the
development of self-driving vehicles. More than half of respondents, 62.36%, said that they
“strongly/somewhat/agreed” while 20.47% indicating that they “strongly/somewhat/disagreed.”
When asked about their concern regarding laws being put in place to prevent people who are
blind from operating self-driving vehicles, 94.37% stated that they were “very/moderately/slightly
concerned”; 60.58% of whom indicated that they were “very concerned.”
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
3:12 J. Brinkley et al.
Table 6. Statistically Significant Eects of Demographic Groupings on Responses to Individual estions
as Presented Through Results from a Series of One-way ANOVAs
Question
number
Ever
heard of
General
opinion
Length of
visual disability
Employment
status Education Transportation
type Age Gender
Q3 ***
Q4 ** *
Q5 *** * *
Q6 *** *
Q7 *** **
Q8 *** ** **
Q9 *** ***
Q10 *** *
Q11 *** **
Q12 *** * *
Q13 *** ** **
Q14 *** ** * *
Q15 ***
Q16 *** * ** *
Q17 ** *** * * *
Q18 ***
Q19 ***
Q20 *** *
Q21 *** * *
Q22 *** * **
Q23 * *** * ** **
Q24 *** * ***
Q25 * *** * * * **
Q26 * *** * * ***
Q27 * *** ***
Q28 *** ***
Q29 *** ** ***
Q38 * * ** *
Q39 *** *
*=p≤0.05.
**=p≤0.01.
*** =p≤0.001.
3.3 Statistically Significant Demographic Eects
Multiple One-way Analyses of Variance (ANOVA) were used to compare responses to survey ques-
tions for each individual demographic variable. Using the D’Agostino-Pearson normality test it
was determined that the collected data was non-normal [20]. Given the greater statistical power
of parametric tests [20], coupled with research that suggests one-way ANOVA is acceptable even
with non-normal data depending on sample size and grouping [39], the use of multiple one-way
ANOVA was chosen over the use of the Kruskal-Wallis test [20]. Table 6presents a summary ma-
trix from the series of ANOVAs, indicating statistically signicant eects of demographic grouping
on individual questions, either at p≤0.05, p≤0.01, or p≤0.001.
3.3.1 Prior Awareness of Self-Driving Vehicles (Q1). Respondents who had not previously heard
of self-driving vehicles expressed greater concerns about the technology than those who had.
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
Visual Impairments and Self-driving Vehicles 3:13
These respondents were more concerned about data privacy, destination tracking, self-driving
vehicles not driving as well as human beings, self-driving vehicles moving by themselves from
location to location, self-driving public transportation like buses and self-driving heavy trucks
and semis. Respondents who had not previously heard of self-driving vehicles expressed greater
agreement with the contention that their needs were being considered in the development of the
technology than those who had heard of it prior to the survey.
3.3.2 Initial Opinion of Self-Driving Vehicles (Q2). Respondents’ initial opinion of self-driving
vehicles had a signicant eect on every response. Respondents who began the survey with an
initial favorable opinion of the technology were more likely to expect the occurrence of its benets
(i.e., a reduction in vehicle crashes) and less likely to express concerns. Respondents who held
initial unfavorable views of the technology were more likely to express concerns, viewing the
technology as unsafe and unreliable for instance, and less likely to expect the occurrence of its
benets. This general principle holds for every question in the survey.
3.3.3 Gender (Q32). Female respondents were found to be generally more concerned than male
respondents about the self-driving vehicle issues and topics investigated during the survey. With
respect to respondent concerns about their ability to operate a self-driving vehicle if one was
made available to them (Q5), a signicant eect of gender was observed at the p<0.05 level
for the two conditions [F(1,517)=6.18,p=0.0132]. Post hoc comparisons using the Tukey HSD
test indicated that the mean score for the female condition (M=2.26,SD =1.08) and the male
condition (M=2.02,SD =1.06) were signicantly dierent. Females were thus shown to be more
concerned about their ability to operate a self-driving vehicle if one was made available to them.
Females continued to express this relatively greater concern in response to questions about riding
in a self-driving vehicle with no driver controls available (e.g., gas pedal, brake pedal or steering
wheel), self-driving vehicles moving by themselves from one location to another while unoccupied,
self-driving commercial vehicles like heavy trucks, self-driving public transportation like buses
and self-driving taxis.
Males were more likely to express a belief that the majority of the benets of self-driving vehicles
were likely to occur. With respect to respondent belief in the likelihood that the use of self-driving
vehicles would result in less trac congestion for instance (Q9), a signicant eect of gender was
observed at the p<0.001 level for the two conditions [F(1,516)=14.09,p=0.0002]. Post hoc
comparisons using the Tukey HSD test indicated that the mean score for the female condition
(M=4.94,SD =1.78) and the male condition (M=5.52,SD =1.70) were signicantly dierent.
Males were thus shown to have expressed a greater belief in the likelihood that self-driving vehicles
would result in less trac congestion. Males continued to express this relatively greater belief in
response to questions about the likelihood of fewer automobile crashes, the reduced severity of
crashes, an improved emergency response to crashes, shorter travel time, better fuel economy and
lower insurance rates.
With respect to respondent interest in buying or leasing a self-driving vehicle (Q29), a signicant
eect of gender was observed at the p<0.001 level for the two conditions [F(1,518)=11.52,p=
0.0007]. Post hoc comparisons using the Tukey HSD test indicated that the mean score for the
female condition (M=3.94,SD =1.37) and the male condition (M=4.32,SD =1.11) were signif-
icantly dierent. Mirroring their increased concerns and relative skepticism regarding their bene-
ts, females were shown to be less interested in buying or leasing a self-driving vehicle than males.
3.3.4 Age (Q33). Older respondents (55+) were the most likely to express concerns about their
ability to operate a self-driving vehicle, while those in the 35 to 44 age group were the least
likely (Q5). A signicant eect of age was observed at the p<0.05 level for the ve conditions
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
3:14 J. Brinkley et al.
(18–24,25–34,35–44), 45–54,55+)[F(4,513)=3.03,p=0.0172]. Post hoc comparisons using the
Tukey HSD test indicated that the mean score for the 35–44 condition (M=1.95,SD =1.06) was
signicantly dierent than the 55+condition (M=2.35,SD =1.14). Younger respondents were
more concerned about the potential for self-driving vehicles to be compromised by hackers than
older respondents (Q16). A signicant eect of age was observed at the p<0.05 level for the
ve conditions [F(4,513)=2.55,p=0.0384]. Post hoc comparisons using the Tukey HSD test
indicated that the mean score for the 25–34 condition (M=2.84,SD =1.11) was signicantly
dierent than the 45–54 condition (M=2.62,SD =1.08). The youngest respondents were also
more concerned about commercial self-driving vehicles like heavy trucks and semis (Q26). A
signicant eect of age was observed at the p<0.05 level for the ve conditions [F(4,514)=
2.46,p=0.0445]. Post hoc comparisons using the Tukey HSD test indicated that the mean score
for the 18–24 condition (M=3.25, SD =0.96) was signicantly dierent than the 25–34 condition
(M=2.62,SD =1.20). Younger respondents were more interested in owning or leasing a self-
driving vehicle and older respondents (55+) were the least interested (Q29). A signicant eect of
age was observed at the p<0.01 level for the ve conditions [F(4,514)=4.68,p=0.0010]. Post
hoc comparisons using the Tukey HSD test indicated that the mean score for the 25–34 condition
(M=4.39,SD =1.08) and the mean score for 35–44 condition (M=4.37, SD =1.20) were both
signicantly dierent from the 55+condition (M=3.86,SD =1.38). Older respondents also were
the least concerned about the prospect of laws being enacted that would prevent people who are
blind from operating self-driving vehicles whereas those in the 25 to 34 age group were the most
concerned (Q39). A signicant eect of age was observed at the p<0.05 level for the ve con-
ditions [F(4,509)=2.75,p=0.0275]. Post hoc comparisons using the Tukey HSD test indicated
that the mean score for the 25–34 condition (M=3.53,SD =0.78) was signicantly dierent than
the 55+condition (M=3.19,SD =1.00).
3.3.5 Education (Q34). For four of the proposed benets (i.e., likelihood of fewer crashes, re-
duced severity of crashes, reduced trac congestion, and shorter travel time) of self-driving vehi-
cles discussed within the survey higher education levels were not associated with higher or lower
expectation regarding those benets. For three of the four remaining proposed benets, improved
emergency response to crashes, lower vehicle emissions and better fuel economy, expectations in-
creased with education level up to “Some College” but decreased as education increased further. For
the potential benet “lower insurance rates,” expectations rose through the “Associates Degree”
education level before decreasing. Respondents who stated their education level as “Some High
School” were most likely to express concerns regarding system failure and data privacy and ex-
pressed the most concern about self-driving vehicles driving as well as human drivers. Higher ed-
ucation levels were associated with decreased concern regarding self-driving vehicles not driving
as well as human drivers (Q23). A signicant eect of education was observed at the p<0.01 level
for the six conditions [F(5,512)=3.53,p=0.0038]. Post hoc comparisons using the Tukey HSD
test indicated that the mean score for the “Some High School” condition (M=3.40,SD =0.89)
was signicantly dierent than the “Some College” (M =1.98, SD =1.17), “Associate’s Degree”
(M=1.86,SD =0.91), and “Graduate Degree” conditions (M=1.84,SD =1.00). Lower education
levels were associated with an increased belief in a respondent’s agreement that this/her needs
were being considered in the development of self-driving car technology (Q38). Respondents who
indicated that they possessed a graduate degree expressed the least agreement that their needs
were being considered. A signicant eect of education was observed at the p<0.05 level for
the six conditions [F(5,512)=2.95,p=0.0122]. Post hoc comparisons using the Tukey HSD test
indicated that the mean score for the “Some College” condition (M=5.10,SD =1.66) was signif-
icantly dierent than the “Graduate Degree” condition (M=4.43,SD =1.75).
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
Visual Impairments and Self-driving Vehicles 3:15
3.3.6 Employment Status (Q35). Respondents who were part-time students were more con-
cerned about the safety consequences of equipment failures, self-driving vehicles getting confused
by unexpected situations, riding in a self-driving vehicle with no manual controls available, and
unoccupied self-driving vehicles moving by themselves from one location to another. Regarding
the potential for self-driving vehicles to be compromised by hackers (Q16), a signicant eect of
employment status was observed at the p<0.01 level for the six conditions [F(5,512)=3.29,p=
0.0061]. Post hoc comparisons using the Tukey HSD test indicated that the mean score for the
“employed full-time” (M=2.68,SD =1.04) and “retired” (M=2.69,SD =1.02) conditions were
signicantly dierent than the ‘employed part-time’ condition (M=3.66,SD =0.49).
With respect to respondent concerns about self-driving vehicles driving as well as human dri-
vers (Q23), a signicant eect of employment status was observed at the p<0.01 level for the six
conditions [F(5,512)=3.61,p=0.0032]. Post hoc comparisons using the Tukey HSD test indi-
cated that the mean score for the “employed full-time” condition (M=1.81,SD =1.00) and “not
currently employed” condition (M=2.24,SD =1.08) were signicantly dierent. Those not cur-
rently employed were thus shown to be more concerned about self-driving vehicles driving as well
as humans.
3.3.7 Modes of Transportation (Q36). A respondent’s mode of transportation was not found
to be associated with a particular perspective relative to concerns regarding self-driving vehicles
or the potential benets that may occur as a result of self-driving vehicles. Regarding respon-
dents’ concern about riding in a partially autonomous vehicle as the primary operator (Q4), a
signicant eect of mode of transportation was observed at the p<0.05 level for the six con-
ditions [F(5,512)=2.34,p=0.0407]. Post hoc comparisons using the Tukey HSD test, how-
ever, indicated that the mean scores for the “passenger car” (M=2.39,SD =1.11), “minivan/van”
(M=2.47,SD =1.16), “pickup truck” (M=2.38,SD =1.12), “SUV” (M=2.13,SD =1.15), “mo-
torcycle” (M=1.60,SD =0.54), and “public transportation” conditions (M=2.62, SD =1.12) were
not signicantly dierent from one another.
3.3.8 Length of Time of Visual Disability (Q37). Being blind or visually impaired for all of a
respondent’s life was associated with higher levels of concern regarding equipment failure, data
privacy, learning to use self-driving vehicles and self-driving vehicles not driving as well as hu-
mans. Respondents who indicated that they had been blind or visually impaired for a shorter
period of time, some of their life, were generally less concerned about system performance in poor
weather and about self-driving vehicles getting confused in unexpected situations. Respondents
who selected “I recently became blind” were associated with greater concern regarding unoccupied
self-driving vehicles moving by themselves from one location to another but with lower concern
about learning to use self-driving vehicles. With respect to respondent concerns about self-driving
vehicles not driving as well as human drivers (Q23), a signicant eect of length of visual disability
was observed at the p <0.05 level for the six conditions [F(3,509)=2.89,p=0.0349]. Post hoc
comparisons using the Tukey HSD test indicated that the mean score for the “some of my life”
condition (M=1.76, SD =0.91) and “I recently became blind” condition (M=2.37,SD =1.11)
were signicantly dierent. Those who indicated that their visual disability was recent were thus
shown to be more concerned about self-driving vehicles not driving as well as human drivers.
3.4 Discussion
3.4.1 General Opinions Regarding Self-Driving Vehicles. The vast majority of respondents had
heard of self-driving vehicles prior to participating in the study (95.96%), the majority held a gen-
erally positive view of the technology and a majority of respondents had an optimistic view re-
garding the potential benets. Our ndings, in this regard, are similar to those of Schoettle and
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
3:16 J. Brinkley et al.
Sivak [54], though respondents in the present study were most condent regarding the likelihood
of fewer automobile crashes and least condent regarding lower vehicle emissions. In the Schoettle
and Sivak study, respondents were most condent regarding better fuel economy and least con-
dent regarding shorter travel time. In the present study, men were more condent regarding the
likelihood of potential benets whereas women were relatively more skeptical.
Signicant concerns were raised regarding all issues addressed within the study with respon-
dents most concerned about equipment and system failure, self-driving vehicles getting confused
by unexpected situations, and interactions between self-driving vehicles, bicycles, and pedestrians.
While these ndings are consistent with the literature in that public opinion surveys have gener-
ally suggested that consumers have signicant concerns regarding self-driving vehicle technology
[13,54,55], our ndings suggest that the concerns of blind and visually impaired consumers may
be somewhat dierent than consumers generally. While concerns regarding legal liability for own-
ers and drivers has been a primary concern of respondents in studies by Howard and Dai [22]and
Schoettle and Sivak [54], it placed fth on the list of concerns in the present study. The same is true
for concerns regarding self-driving vehicles not driving as well as human drivers, which placed sev-
enth on the list of concerns in the present study but was thrid in the Schoettle and Sivak study [54].
These dierences continued through the ve scenarios (Q30–Q34) of potential concern that
were presented to respondents. While much of the literature has suggested that consumers are
especially concerned about the potential absence of manual controls in self-driving vehicles [26,54,
55], this was viewed as one of the least concerning scenarios for respondents of the present study.
Respondents were most concerned regarding self-driving commercial vehicles (e.g., heavy trucks)
and self-driving public transportation vehicles like buses. Overall, regarding the eight issues and
ve scenarios of potential concern, women expressed more concern than men and not surprisingly
expressed the least interest in owning or leasing a self-driving vehicle.
3.4.2 The Ability to Operate Self-Driving Vehicles. While learning to use a self-driving vehicle
was one of the least concerning issues for respondents of the eight presented, the ability to operate
the technology was an area of concern for older respondents (55+). Respondents in both groups
were more likely to express concerns regarding their ability to operate a self-driving vehicle while
those in the 35–44 age group were least likely. The issue of the control of self-driving vehicles
has typically been investigated only from the perspective of the inclusion or exclusion of manual
controls [8] or the preferred type of route input device [55].
3.4.3 Interest in Ownership and Willingness to Pay. More than 90% of respondents expressed
an interest in owning self-driving vehicle technology; high interest in ownership that is con-
sistent with the literature [28,54,55], which has indicated that the majority of those surveyed
would be interested in ownership. A majority of respondents, however, indicated a willingness
to pay extra for self-driving vehicle technology. Much of the literature in this regard, which
has presumably focused on sighted consumers, has indicated that consumers generally are un-
willing to pay extra for self-driving vehicle technology [28,54,55]. Our ndings indicate that
blind and visually impaired consumers, on average, may be willing to pay more than $6,000 ex-
tra for self-driving vehicle technology, a sum higher than the $4900 found by Daziano et al. [13],
presumably with sighted consumers and approaching the recent ndings of Bansai et al. [4]of
$7253.
3.4.4 Issues Related to Visual Impairment and Blindness. A majority of respondents indicated
that they believed their needs were being considered in the development of self-driving vehicle
technology, however, higher education levels were associated with a decrease in a respondent’s be-
lief that their needs were being considered. This nding is signicant in that it has been suggested
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
Visual Impairments and Self-driving Vehicles 3:17
that most self-driving vehicle technology being developed is not in fact accessible to individu-
als with visual impairments [44]. There appears to be a mismatch between respondents’ belief
regarding the consideration of their needs and the actual accessibility of the self-driving vehicle
technology being brought to market.
While additional research is needed to conclusively verify both sides of this question, the mis-
match may be the result of inuence of marketing eorts and the media. Google’s Waymo, one of
the leaders in the development of self-driving vehicle technology, has prominently featured a blind
user operating of one of its self-driving cars in self-produced promotional videos for years [63].
The operation of its self-driving car by a blind person has also been covered with much fanfare
by the media, being written about with titles like, “Blind man sets out alone in Google’s Driver-
less Car” [6]. Given these types of marketing eorts and coverage in the media it is perhaps not
surprising that many blind and visually impaired individuals believe that their needs are being
considered in the development of self-driving vehicle technology. It is, of course, entirely possi-
ble that most self-driving vehicle technology being developed is, in fact, accessible to individuals
with visual impairments and reports to the contrary are mistaken. The uncertainty in this regard
underscores the need for additional research both as it pertains to consumer opinions on the topic
as well as the actual accessibility of the technology.
Over the past ve years, a number of laws and regulations have been proposed that, if enacted,
would potentially prevent blind persons from operating self-driving vehicles. Laws that would
require a licensed driver as the primary vehicle operator or require that the primary operator
have the ability to take over control in the event of an emergency [14] would eectively prevent
blind and many visually impaired individuals from operating self-driving vehicles using current
technology. Over 90% of respondents indicated that they were concerned regarding the prospect
of laws being put in place to prevent people who are blind from operating self-driving vehicles.
3.5 Limitations
In adapting the survey instrument of Schoettle and Sivak, we made a number of design choices re-
lated to content, accessibility and structure while attempting to preserve the general themes of the
original survey to facilitate direct comparison. In the process, we omitted a question related to re-
spondents medically diagnosed visual acuity that could have been used as a demographic variable
during data analysis. The point of this exclusion was to view the survey and its data collectively
as representing the perspective of blind and visually impaired consumers. An argument can be
made, however, that an opportunity was missed to dig deeper into the data to identify dierences
between blind and low-vision consumers. A similar question (Q37) that explored the length of
time of each respondent’s visual impairment, used arbitrary time periods (e.g., “Recently,” “Some
of your life”) and should have instead collected the actual length of time for grouping and analysis.
Questions may also be raised regarding the generalizability of our ndings given that our sam-
ple is questionably representative of the target population. In terms of educational attainment, for
instance, our sample is skewed toward those with college education. Within our study those hold-
ing bachelor’s degrees or more education comprised approximately 59% of respondents. According
to 2015 census gures, 45% of the population generally reported bachelor’s degrees or advanced
degrees such as master’s or doctorates [50] and the National Federation of the Blind has indicated
that for persons with visual impairments this gure was 14.9% in 2015 [60].
4 STUDY 2: FOCUS GROUP DISCUSSIONS
4.1 Method
While the data collected from the online survey added much to the literature on the opinions,
preferences, and concerns of visually impaired persons regarding self-driving vehicles, our analysis
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
3:18 J. Brinkley et al.
Table 7. Breakdown of Focus Group Participants
by Degree of Vision Loss
Degree of vision loss % Female (n=20) % Male (n=18)
Blind 45.0 72.2
Low Vision 55.0 27.7
Table 8. Breakdown of Focus Group
Participants by age Range
Age Range % Female (n=20) % Male (n=18)
18–24 10.0 22.2
25–34 0.0 16.6
35–44 5.0 11.1
45–54 15.0 11.1
55–64 30.0 27.7
65–74 30.0 5.5
75+10.0 5.5
generated additional questions that were worthy of further investigation. Following the online
survey, a qualitative study was planned to investigate many of the issues that became apparent in
our analysis of the survey data; focus groups were ultimately undertaken to explore the identied
issues in a more open-ended format. The use of focus group methodology [25,68] specically was
chosen over other methods, because it provided the research team with an opportunity to elicit
subjective perspectives regarding the research topics while allowing for a signicant amount of
exibility to pursue themes that emerged during the course of discussion.
4.1.1 Participant Recruitment. Interested individuals were invited to participate if they fullled
the following criteria: age 18 or above and currently considered themselves to be a visually im-
paired person based on an accompanying denition that dened that as blindness or with limited
vision not correctable by glasses or contact lenses. Advertisements indicated that interested in-
dividuals should also have transportation to one of two focus group locations in north central
Florida. Participants were recruited by email, newsletter and social media announcements dis-
tributed by organizations serving individuals with visual impairments in north central Florida and
by vocational rehabilitation organizations. Those interested in participating were asked to call or
email for additional information and scheduling. The Institutional Review Board of the authors’
university approved this study and each participant provided written informed consent the day
of his or her focus group session. Participants were compensated with a $20 prepaid gift card for
their participation, which they were informed about prior to participating.
4.1.2 Description of Participants. Eight focus groups were conducted over a two-day period in
two separate locations in north central Florida. In total, 38 participants were involved in the study
in groups of between 4 and 6 people. Participant demographics are provided in Tables 7through
9. Study participants had a mean age of 51.5 (range =18 to 90 years old) and a household annual
income that ranged from under $15,000 to over $99,000. Twenty-two participants self-identied as
blind based on an accompanying denition that described blindness as visual acuity in the better
seeing eye of 20/200 or worse or a visual eld limited to 20 degrees with conventional correction
(e.g., glasses or contact lenses). Sixteen participants self-identied as low vision based on an ac-
companying denition that described low vision as visual acuity of between 20/70 to below 20/200
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
Visual Impairments and Self-driving Vehicles 3:19
Table 9. Breakdown of Focus Group Participants by Ethnicity
Ethnicity % Female (n=20) % Male (n=18)
White/Caucasian 90.0 61.1
Black/African American 10.0 16.6
Hispanic 0.0 16.6
Mixed Race 0.0 5.5
in the better seeing eye with conventional correction. Krueger and Casey have argued that focus
group participants should share similar characteristics (e.g., gender group, age-range, social class
background) to encourage open dialogue within the group but with enough variation to allow
for contrasting opinions [27]. An attempt was made to group participants near their preferred lo-
cation in accordance with this principle with a primary factor of similarity being a participant’s
characterization of their vision loss and a secondary factor being their age. No other factors were
considered for the purpose of constructing the focus groups (e.g., race, gender, education). During
the screening and scheduling process participants were briey presented with functional deni-
tions of blindness and low vision [34]. They were then asked to choose the denition that best
characterized their degree of vision loss. As a result of this approach, when feasible, participants
were placed in a group with other individuals who similarly characterized their vision loss as either
blind or low vision. Where placement in such a group was not possible, typically due to logistical
or scheduling constraints, an attempt was made to place the participant in a group with at least one
individual of similar age irrespective of whether or not the majority of the group’s participants had
expressed a dierent degree of vision loss. The approach resulted in one group that was predomi-
nantly composed of blind persons, one group predominantly composed of low-vision persons, one
group evenly split by degree of vision loss, and three groups that were relatively homogeneous in
terms of age and predominantly composed of blind persons, each consisting of older participants
(55+). The remaining groups were each slightly weighted toward blind participants in terms of
composition though each group contained at least two low-vision participants.
4.1.3 Procedure. Each focus group lasted approximately one hour with a procedure that was
identical for each session. After each participant was seated in the meeting space, the informed
consent document, which had been emailed to each participant in an accessible format the day
before each focus group, was read aloud by the study moderator. Participants were then provided
with assistance, as needed, signing the informed consent document. After being provided light
refreshments, a brief ice breaking exercise was led by the focus group moderator to introduce
participants to one another and encourage interaction between participants. A semi-structured
interview followed. Between three and ve days following the focus group session a telephone in-
terview was conducted with each participant to provide an opportunity to ask follow-up questions
after a period of reection and to gather additional demographic information. This article reports
on the results of the semi-structured interview as well as relevant demographic data.
4.1.4 Focus Group Guide (Wrien Script). We created a written semi-structured script to elicit
information from participants about their understanding and awareness of current developments
regarding self-driving vehicles, hopes and aspirations for future self-driving vehicle technologies,
concerns related to the accessibility of self-driving vehicles and opinions regarding the legal envi-
ronment for the use of this technology by individuals with visual impairments. The written script
or focus group guide was pilot tested with three participants in a group setting prior to beginning
data collection to ensure that it was comprehensible and comprehensive. Pilot test participants
were not visually impaired, and the data was not included for analysis.
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
3:20 J. Brinkley et al.
4.1.5 Data Capture and Transcription. Each focus group session was video recorded and tran-
scribed verbatim by a professional transcriptionist prior to analysis. The completed transcript was
then veried by a member of the research team against the original recordings.
4.1.6 Analysis. In preparation for analysis all transcripts were entered into MAXQDA [36], a
computer program for qualitative data analysis. After initially familiarizing ourselves with the
data two investigators independently coded all quotations from participants. For each researcher
this hybrid process began with a small set of a priori codes agreed upon by the research team in
advance then continued with codes inductively identied within the data. Each coding was then
categorized and rened by each researcher independently. Both independent analyses were then
merged into a single denitive version by a third researcher with any disagreements in coding and
categorization settled by this third researcher and agreed upon by the research team collectively.
4.2 Results
Results of the analyses of focus group data are organized around six major thematic ndings:
(1) self-driving vehicle concerns, (2) potential benets of self-driving vehicles, (3) licensing and
training, (4) the human-machine interface of self-driving vehicles, (5) purchase considerations, and
(6) risk and trust. Across the eight focus groups self-driving vehicle concerns were raised 223 times
followed by 151 comments related to the potential benets of self-driving vehicles. Comments
associated with licensing and training occurred 121 times and were followed by comments related
to the human-machine interface of self-driving vehicles (88) and vehicle purchase considerations
(74). Issues related to risk and trust were the least discussed, occurring 33 times. In many instances
more than one theme was addressed in a single conversational turn.
4.2.1 Self-Driving Vehicle Concerns. Are the needs of individuals with visual impairments being
adequately considered in the development of autonomous vehicle technologies? While opinions
varied, most participants expressed the view that the needs of individuals with visual impairments
were not being adequately considered in the development of autonomous vehicle technologies
(57%). Negative experiences with past technologies were frequently cited as examples of why they
believed these needs were likely being ignored:
No, I don’t/I don’t think they do. And they didn’t when they invented the quiet
car. And that’s, you know, that was an invention that’s plagued me. I mean, I love
hybrid vehicles and what they do for the environment, but I hate hybrid vehicles
and how dangerous they are. They’re like sharks in the water now. You don’t know
when it’s coming up behind you. It may strike. (P38)
I know that in the past with all the new technology that’s come out so far, blind
people have been kind of an afterthought. (P37)
Many participants (37%) expressed a belief that the technology presently exists to solve most
autonomous vehicle accessibility challenges but that manufacturers need to be made aware of the
importance of the issue:
Like, on the side like Tesla and Uber and whatnot, I’m sure there’s someone some-
where thinking about that, but I don’t think it’s in, like, the forefront. (P34)
If you KNOW of all these obstacles, then they can be overcome. And the technol-
ogy, it pretty much exists now. It’s just a matter of marrying that technology, you
know? (P17)
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
Visual Impairments and Self-driving Vehicles 3:21
Parking, Orientation, and Vehicle Location. A variety of issues were raised related to the parking
of an autonomous vehicle, orientation to a destination upon parking and the location and identi-
cation of a visually impaired operator’s vehicle at the conclusion of a trip.
Many participants expressed concerns regarding their ability to provide parking guidance to
an autonomous vehicle without the ability to see the surrounding area. Some of the concerns
expressed related to safety:
How do you try to set a car to park where you want it to park if you can’t see
where to show it to park or tell it to park or whatever? You might be parking in
the/across a railroad track. (P12)
Some expressed concerns related to convenience and practicality, with travel to shopping malls
being a frequently provided example of the need to tell a self-driving vehicle specically where to
park:
When you pull into a shopping mall, you know, you got like this one out here is,
you got Penney’s on one end, Sears in the back and Belk’s on the other end. Well,
you know, you need to be able to tell it. . . (P11)
The ability to orient oneself to the desired destination (e.g., a building’s front door) upon arrival
to a location was a widely discussed topic. Many participants felt that this was an unanswered
question that would need to be addressed by the technology:
Now, if I drive up in a car and get out of the car, how do I know where the door is?
In other words,there’s a...thecar has todoa lot besidesjustdrive. (P24)
One participant raised questions about how the technology would enable her to orient herself
to her desired destination in the event that she were dropped o in a parking lot:
And there’s also, you know, if it takes you to a location and it like drops you o
in the middle of like a parking lot, you know, which building or which direction
that your actual destination is? You know, something that would also, like, help
you gure that out and know where that is. (P23)
In response, another participant (P21) suggested dedicated autonomous vehicle parking to aide
in general orientation.
Twenty-eight percent of participants raised questions about the technology neces-
sary to enable an autonomous vehicle operator with visual impairments to locate
his or her vehicle in a large crowded space like a shopping center parking lot. A
number of solutions were expressed across focus groups to include programming
the vehicle to return at a specied time and location (P30, P32), to key chains de-
signed to vibrate based on proximity to the vehicle (P34).
Location Verication. The desire for a vehicle feature that would enable an operator with visual
impairments to verify his or her arrival at a desired location was a topic of great interest. A majority
of participants across focus groups (53%) commented that such a feature was important to some
degree and that they would be concerned about inadvertently getting out of the vehicle in the
wrong place. A commonly expressed view was that the technology should provide some type of
backup system to prevent such an occurrence:
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
3:22 J. Brinkley et al.
Ithinkmyoperatingsystem shouldtake careofit.Youknow?It’slike she’d...she’d
take care of everything. So, I think the operating system should say, “Don’t get out
of the car. You’re inthe wrong...you’re ina eld.”(P32)
In many cases the expressed desire for this feature closely related to negative experiences with
automotive navigation systems where the participant had observed these systems lead others in
the wrong direction, make errors or otherwise fail to operate properly. One participant’s comment
was representative of many regarding this topic:
...you know, I’ve experienced on MV, you know, our public transportation that,
you know, the GPS saying “You have arrived at your destination” and, you know,
and the driver says, “No, we haven’t.” So, you know, it’s not all it seems. Like GPS
is fallible, you know, at times. (P20)
Situational Awareness. The desire for a vehicle feature that would, at any given time, provide
information on an operator vehicle’s relationship to other vehicles, landmarks, pedestrians, obsta-
cles and the nal destination was frequently discussed. Forty-ve percent of participants across
focus groups expressed a view that the self-driving vehicle’s technology should be able to provide
such information in real time to mirror the type of information that a sighted operator would have
available through the use of cameras, mirrors and sight:
I think it would make the person sitting in the car feel a little bit more at ease. You
know, that you know, since you can’t see where you’re going or where you’re at,
that it would tell you like, “Well, you know, we are like half a mile away from your
destination.” (P2)
...youdon’t necessarily know,like, where youare inrelationship to whereyou’re
actually going. Something that might help you, like place, like where you are. (P23)
Interaction with Non-Autonomous Vehicles. Participants generally expressed skepticism that au-
tonomous vehicles would be able to successfully operate alongside non-autonomous vehicles due
to the unpredictability of human drivers (63%):
I think that the fact that they have to interact with people driving is a downfall too,
because, you know, people are really bad drivers. I ride with my daughter all the
time and the people constantly running red lights, cutting us o and doing things,
you know. And I would...you know, you would wonder howthe car would react
to things like that. (P33)
My downside is...To me, my–this is my opinion–I just don’t see an autonomous
car being able to react to human error that quickly. (P25)
Many participants (39%) commented that as autonomous vehicles became more common and
displaced non-autonomous vehicles on the road, they believed the interaction problem would
lessen as captured in this exchange between three participants in group six:
Well, I think the autonomous vehicles are like the u shot. The more people that
get the u shot, (P25 laughs) the less chance of an epidemic. The more autonomous
vehicles that are on the road. . . (P27)
The more of them that are on the road, the safer the roads will be. (P26)
Help and Roadside Assistance. Howtoandwhotoaskforhelpintheeventofavehiclebreakdown
or unforeseen event was a topic that was brought up repeatedly in the study with 39% of study
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
Visual Impairments and Self-driving Vehicles 3:23
participants raising questions in this regard. While the overwhelming majority of participants
stated that they had never owned or operated a vehicle, the suggested approaches took the form
of a type of monitoring and assistance service similar to General Motors’ OnStar [31], which was
specically named several times:
Yeah, that would be good, if they had some kind of like, like a home center (P2:
Yeah. Yeah), you know, that paid attention. Like a security system for cars that
keeps track of where your car is going. So, if you’re lost or something, you’re able
to get help. (P4)
Well,I...Ithink about somethingalong thelines of—and I’ve never usedit before,
but, you know, something along the lines of OnStar. In other words, you know,
if you get into some trouble, if there’s vehicle trouble, you know, with the actual
vehicle itself, being able to call, you know, somebody to come and help you. ’Cause,
obviously, if I was riding it by myself and the car had a problem and was maybe
automatically pulled over, I wouldn’t know where I was. (P20)
Legal Liability. A small number of participants raised questions regarding legal liability in the
event of an accident (13%). Some participants expressed concerns about the seemingly unsettled
liability questions in the event of an autonomous vehicle accident while several blind participants
expressed concerns that they would be automatically assumed to be liable as captured in this
exchange between two participants in group two:
You can’t even be your own witness? (laughs) (P6)
(laughs) Yeah, you’re just getting blamed for it for sure. (P5)
4.2.2 Potential Benefits of Self-Driving Vehicles. The majority of participant comments in the
study that focused on the potential benets of self-driving vehicles centered on the potential for
increased independence, personal mobility, and the potential for time savings versus participants’
existing means of transportation.
Independence and Mobility. Forty-seven percent of participants specically referenced “indepen-
dence” when referring to the potential benets of self-driving vehicles with many participants ex-
pressing a hope that self-driving vehicles would enable them to no longer rely on friends, relatives
or public transportation for their transportation needs:
I would denitely say the benet would be to not have rely on people to take
you (P6 nods) around. That you could just get up and go whenever you want to,
whenever you want to. (P7)
Independence and the ability to go from point A to point B without having to ask
somebody to drop what they’re doing to do that for you. (P16)
Within the context of the discussion regarding independence several participants specically
discussed a desire for basic transportation capabilities that would enable them to return to work
as expressed by one blind participant:
You know, I think one of the things that’s important to understand is, you know,
everybody is looking at these high-end abilities, you know, things like luxury, com-
fort, you know, safety about, you know, an ability to go anywhere. I don’t think
that we have to start there. You know, if I could purchase an AV [autonomous ve-
hicle] that...Iwasinhome healthcaresalesrep, but, you know, two years ago. You
know, I...Everyweek, Icalledonthesamedoctor’soces,the samehospitals,the
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
3:24 J. Brinkley et al.
same nursing homes. You know, I ran the same route every day, on time. You know.
And if I could have a vehicle that I could program to do nothing but that route and
how to get here and maybe the grocery store, I would be totally 100% happy. (P16:
Denitely, yeah.) You know? And I think that’s one of the things that these engi-
neers are missing is: they’re trying to make the all-inclusive, you know, panacea,
you know. And that’s not really what people with disabilities are concerned with.
(P14: Right.) Yeah, it would be nice to have. But we’re just looking for the basic
mobility and independence that most people totally take for granted.
Participants in the study commented on the impact that the inability to operate a personal ve-
hicle had on their daily lives. Participants indicated that it impacted where they lived, their diet,
their shopping habits, their medical care, their employment and their ability to visit friends and
relatives. Nearly all participants responded positively to the potential independence and mobility
benets of self-driving vehicles (97%):
Well, the idea of being able to get in a car at the time you wanted to and go wherever
you want it to and get in it and come back home when you wanted to, it would
mean an awful lot, I think. (P12)
One blind participant indicated that she did not see any benets, however (P29).
Time Savings. Of participants commenting on the topic (18%), opinions were overwhelmingly
positive about the potential time savings of the use of a self-driving vehicle over their current
forms of transportation; in most cases was public transportation, friends or family. Many shared
personal stories about their daily challenges or the challenges of friends with visual impairments
and how these challenges could potentially be overcome with the use of a self-driving vehicle:
I know a guy who goes to work, and he spends three hours to get there and like
two hours coming home. And that’s for 25 years he’s been doing that. (P25: Oh my
God!) So, can you imagine how valuable that time is that he is losing? (P26)
And...And what you...Whatpeopledon’t realizewhatyou couldin30to45min-
utes, it takes on a bus two hours to four hours. [P28: Or more.] Like getting here.
(P25)
4.2.3 Licensing and Training. The discussion of legal and regulatory issues related to self-
driving vehicles centered on laws that could potentially restrict access to individuals with visual
impairments. Broadly within the study these discussions generally focused on the potential re-
quirement that self-driving vehicle operators possess a valid driver’s license or on potential train-
ing requirements.
More than half of participants (55%) expressed concerns about the prospect of laws being put
in place that would eectively prevent individuals with visual impairments from operating self-
driving vehicles. The most concerning prospect for many was the requirement that a valid dri-
ver’s license be required to operate a self-driving. Several participants expressed the view that this
amounted to a form of discrimination.
So,Ithink they would have tomakesome laws,likehe wassaying, to...for people
not to do that. ’Cause we have rights too. (P3)
But regardless,I’d really...I personally hate...That’s whatADA is allabout. (P25:
Yeah.) (P24)
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
Visual Impairments and Self-driving Vehicles 3:25
The topic elicited a variety of comments, however, and some degree of disagreement with several
participants suggesting that a driver’s or operator’s license of some kind should be required to, at
a minimum, promote safe operation and prevent children or individuals with limited cognitive
abilities from operating self-driving vehicles alone:
I think, the rst thing that comes (P19: Safety.) to my mind would be, you know:
Would...I mean, it seems logical that if we’re gonna operate something like that
we’dhavetogetadriver’s license.AndI...I’mjustnotsurehow that’sgonnawork.
(laughs) (P20)
I would say, you don’t put a 16-year-old behind the wheel. (P38)
In many of these discussions within the study a compromise of sorts that seemed acceptable to
the group that emerged was the concept of an “operator’s license” that could be earned through a
training program and testing:
When you drive a normal car you gotta go and get your driver’s license and take a
class, so there should probably be a class to take for this, too. I think. I mean, that
would make sense. Maybe not that extensive or whatever, but, you know, just like
all the safety issues and stu like that. It would make sense to have that. (P5)
And I think to be on the safe side and keep you from having some issues, I think
it’s better that you have to take a test. (P31)
4.2.4 The Human-Machine Interface of Self-Driving Vehicles. The discussion of the human ma-
chine interface (HMI) of self-driving vehicles focused on the manner in which participants antici-
pated that they would interact with a self-driving vehicle and their preferences in this regard.
The vast majority of participants (71%) anticipated that self-driving vehicles would implement
some type of speech interaction capability. For example:
I hope it’s so that you can just get in the car and say, “I want to go to “such and
such church” and it’ll take you there. And I hope I can get in and say, “I want to
go home,” when it’s over. (P11: Well) I hope they can simplify it somewhat. (P12)
Concerns were expressed about the accuracy of speech input, however, with personal experi-
ences with voice recognition errors in Apple’s Siri [24] being cited as examples by several partici-
pants.
Participants expressed a variety of opinions, however, regarding their preferred primary means
of interacting with a self-driving vehicle. While some expressed a preference for speech input, sev-
eral expressed a desire to interact with the vehicle primarily through the use of their smartphone
as captured in this exchange in group ve:
Because like most visually impaired or blind individuals, like we have smart
phones. So, is there a way to put it in through our phone that it will sync with
the car? Well, I’m not saying “Is there a way?” but “Could there be a way?” I guess,
that’s what I’m asking. I guess, that would probably be the number one thing. (P21)
I think that would probably be a good idea ’cause a phone is something I can, like,
hold up closer to me so I can see it easier. And also, it will read it back to me, like,
that’s already accessible in that way versus like having to look at something else.
And I’d also say to make like buttons or dierent things maybe more distinguish-
able, so you know which one is which. (P23)
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
3:26 J. Brinkley et al.
The use of a touchscreen to interact with an autonomous vehicle was described by many partic-
ipants as a backup means of interaction. One participant with low vision who favored the use of a
smart phone as a primary means of interaction preferred the use of a touchscreen as a backup form
of interaction in the event that her smartphone failed albeit with an iPhone like “voiceover” feature
that would indicate what icon was being touched. Several blind participants disagreed with this
comment, however, citing past problems with voiceover and the inaccessibility of standard touch
screen displays.
4.2.5 Purchase Considerations. Vehicle purchase considerations as it related to self-driving ve-
hicles largely revolved around cost, vehicle maintenance and the presence of backup systems.
Most participants who commented on the topic (24%) expressed a view that while the costs are at
present unknown, self-driving vehicles would likely be costly, if not prohibitively so. For example:
Let’s face it, most of these vehicles to begin with, they’re gonna be economically
prohibitive for most of us to be able to purchase and maintain. (P36)
Several participants also commented about the uncertainty of repair costs or even where or how
to get the vehicle repaired or maintained. For example:
And since it’s run mainly by technology, you know, where do you take it for its
50,000 overhaul or whatever? (P19)
A small number of participants (8%) expressed concerns about the reliability of primary self-
driving vehicle systems. These participants questioned what would happen if primary systems
failed during vehicle operation and considered the presence of a backup system to be a key pur-
chase consideration. For instance:
It must have. What if the main system malfunctions, you know, and it can’t literally
drive you? Well, if you’re on the interstate and it malfunctions, would they have
like a backup system? (P29)
4.2.6 Risk and Trust. Participants expressed a broad range of views as it related to whether or
not they would trust in the safety and reliability of self-driving vehicle technology.
Some participants (5%) rejected the idea outright as too risky as did one blind participant:
I just would not feel full trust at turning myself totally loose with any type of
a/Even if it’s a computer. And I’ve worked with computers too long that I know
they can make/get a glitch in them. And I think there is still a possibility of risk.
(P1)
Sixteen percent of study participants raised concerns regarding the possibility of malfunctions
due to computer viruses and hacking. One participant stated for example:
Computers aren’t infallible. You know, they get viruses, they get/other people get
in and take control. You know? So, you have to think about all that stu, too. And
that would make me very nervous. (P10)
Of the 76% of participants who commented on the topic, many indicated that they viewed riding
in a self-driving vehicle as being no more risky than riding with a human driver:
The risk involved. For me, it’s no, even more than getting in the car of my wife.
(P11)
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
Visual Impairments and Self-driving Vehicles 3:27
Some participants, however, indicated that they would actually feel safer in a self-driving
vehicle:
But, we almost daily risk our lives by getting into a vehicle (P25: Oh yeah, I agree!)
with a stranger who may be the worst driver. (P29: That’s true, yes.) (P28: Yeah.)
They may be drunk, they may be on drugs, they may have stayed up all night ght-
ing with their girlfriend. Whereas the autonomous vehicle does not get distracted,
does not get tired. I think it’s just way, way, way safer. (P28: I agree.) (P27)
A blind participant who had recently lost his sight described to the group being driven by his
wife who had never driven before and was in the process of learning to drive:
She’s got to learn it. So now I’m riding in a computer that’s learning, you know,
a self-taught computer. So instead of having, you know, the option of maybe a
program glitch happening, I’m riding a computer that doesn’t have a clue what
it’s doing yet. You know? So, I’d feel safer in an AV. (P17)
For some participants the potential independence and mobility benets simply outweighed any
risks. One participant who was progressively losing her sight and no longer had the ability to drive
stated:
Pitfalls? None. I don’t care if they crash me into a wall, I’ll wear a helmet. I just
want to get out of my house. I don’t care. I’ll be a test driver. I just want out. I just
want my freedom. They outweigh any/any problems. I’ll sign a liability waiver.
(P32)
4.3 Discussion
4.3.1 General Opinions Regarding Self-Driving Vehicles. Similar to Study 1, the vast majority of
participants expressed an awareness of self-driving vehicle technology with all but one participant
within the study indicating a belief in their general understanding of the technology when asked at
the outset of their respective focus group. Participants expressed views that, while somewhat skep-
tical, were generally more positive than the related literature [54,55], however, as it pertained to
the benets of self-driving vehicle technology. The overwhelming majority of participants were
resolute regarding their belief that self-driving vehicles could potentially be life-changing tech-
nology for them personally. While the majority expressed a view that there are considerable un-
knowns regarding the technology that need to be addressed, they were generally optimistic about
the potential personal benets. Participants also expressed optimism for society at large believing
that self-driving vehicles could potentially make driving generally safer, while reducing the impact
on the environment.
Concerns regarding the potential legal liability of owners/drivers/operators of self-driving ve-
hicles have been raised as a substantial concern in the related literature [22,54]. Participants ex-
pressed similar concerns in the present study but generally related these concerns to their visual
impairment and not the absence of an active driver, as is typically the case in the related literature.
Several participants questioned whether there would be a presumption of their liability if a self-
driving vehicle were to be operated by a person with visual impairments and how an individual
would defend his or herself from spurious legal claims without the ability to provide visual testi-
mony regarding an accident. These concerns, while similar, are somewhat dierent from the legal
liability issues raised in the related literature that often explore whether legal liability should be
assigned to the vehicle manufacturer or operator.
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
3:28 J. Brinkley et al.
4.3.2 Human-Machine Interface. Our ndings suggest that the Human Machine Interface (HMI)
of a self-driving vehicle for a person who his blind or low vision will likely need to do more than
simply allow an operator to input a route or destination. The HMI for an operator who is blind
or low vision will likely need to serve this purpose while also satisfying the need for situational
awareness, location verication and other, as yet unspecied needs.
Study participants’ desires for a self-driving vehicle HMI closely aligned with their degree of
vision loss and their age. Blind participants generally expressed preferences for both speech in-
put capabilities and the use of a smartphone application. Many participants anticipated that self-
driving vehicles would utilize speech input as a default means of interaction but were wary of the
system inaccurately interpreting their utterances or of generally poor performance based on past
experiences with Apple’s Siri [24], Microsoft Cortana [37], and Amazon Alexa [1]. Many partici-
pants indicated that they found their personal smartphone to be very accessible and would prefer
to control a self-driving vehicle using a smartphone application if possible though they acknowl-
edged that a backup system would be necessary in the event that they did not have a smartphone,
or their battery died. Blind participants were extremely resistant to a standard touchscreen, view-
ing them as being entirely inaccessible with several indicating some acceptance of a touchscreen
with Apple “voiceover” [66] capabilities.
Participants with low vision also expressed preferences for both speech input capabilities and
the use of a smartphone application but also expressed an interest in touchscreen interaction. Like
blind participants, many participants with low vision expressed a belief that self-driving vehicles
would utilize speech input as a default means of interaction but expressed somewhat less concern
about the potential for voice recognition problems. Participants with low vision also found their
personal smartphones to be very accessible and expressed a desire to control a self-driving vehicle
using a smartphone application if possible. Participants with low vision expressed interest in the
use of a touchscreen, largely as a backup to the use of a smartphone application, but suggested
the incorporation of contrasting colors, enlarged buttons and Apple “voiceover” [66] capabilities.
Our ndings stand in contrast to those of Schoettle and Shivak [55] who, in a survey presumably
of sighted respondents, found that a plurality of respondents preferred touchscreen interaction.
It is important to note that relatively younger (under 30) blind and low-vision participants in
the present study were the most emphatic about the desire to use their personal smartphone or
a smartphone application as a control device for a self-driving vehicle with older participants
accepting of the concept but open to other approaches.
Studies by Schoettle and Shivak [55] and KPMG [26] have suggested that a majority of potential
operators of self-driving vehicles have a preference for manual controls to allow the vehicle to be
controlled by a human operator in an emergency situation or for enjoyment. A very small number
of the participants of the present study expressed a similar desire. All but one of these participants
were people with low vision who felt that they should have some ability to control the vehicle in
an emergency situation if they observed a potential hazard.
The need for the self-driving vehicle’s technology to provide some means to satisfy the need
for situational awareness while in transit and to aid the operator in verifying their arrival at the
correct location was expressed by the majority of participants in the study. Most participants ex-
pressed some concerns about traveling without having some awareness of the direction that they
were traveling in and their immediate surroundings (e.g., current street, surrounding vehicles,
buildings, pedestrians). Participants also expressed hesitance about getting out of the vehicle upon
arrival, especially in the case of a self-driving taxi that would perhaps immediately leave, without
assurances that they had arrived in the correct location. These concerns were closely related to
skepticism regarding the reliability of GPS technologies and negative assumptions regarding the
technologies that would be used to input desired destinations (e.g., unreliable speech systems).
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
Visual Impairments and Self-driving Vehicles 3:29
Older participants (55+) were especially concerned about the ability of the HMI to relay informa-
tion regarding parking options and building locations. Many of these older participants expressed
a belief that the challenge of getting from the vehicle to their desired building or nal destination
and back would be nearly as complicated a technical challenge as the development of the self-
driving vehicle itself. While the majority of participants felt that these challenges could be solved
with existing technology, they were unsure how. Findings in this regard are perhaps not surprising
when viewed in the context of related research on BVI spatial orientation and navigation, which
posits that navigating large unknown environments is one of the more challenging navigational
tasks for visually impaired persons [17].
4.3.3 Interest in Ownership and Willingness to Pay. Much of the self-driving vehicle literature
of the past decade has suggested that in cases where people would consider owning self-driving
vehicle technology these same individuals in most instances would not consider paying extra for it
[28,54]. Recent studies [4,13] have suggested a break from this trend, nding that consumers may
be increasingly willing to pay for self-driving vehicle technology depending upon the cost. In the
present study, only a few participants took the position that they were unwilling to pay anything
extra for autonomous vehicle technology with very few being steadfast in this view. While not
every participant verbalized a position on the issue, only one participant indicated during the
discussion that they would not be willing to pay for a self-driving vehicle.
In the study the question for many participants was not whether they were willing to pay or pay
a premium for this technology but rather whether they would be able to aord what many assumed
would be very costly technology. This question was exacerbated by the fact that the overwhelm-
ing majority of participants had never owned or purchased a vehicle before and therefore many
indicated that they had little direct frame of reference for vehicle costs, self-driving or otherwise.
Several participants in the study expressed an interest in the Tesla brand and the “Autopilot” fea-
ture specically [41], indicating that they had conducted research on related features and pricing.
These participants, coupled with those who had exposure to vehicle pricing from past experience,
friends and relatives, ultimately helped to frame the discussion around vehicle costs in several
of the groups. The majority of groups arrived at an estimate at or near $100,000 that seemed to
make reference to the believed cost of a near-autonomous Tesla that was often discussed in the
same context. This cost for most participants was viewed as unaordable though quite a few par-
ticipants indicated that they were willing to pay that or more if that was what it would cost for
a self-driving vehicle. Many of the participants espousing this latter view were older participants
(55+) who questioned whether the technology would become commercially available soon enough
for them to benet.
4.3.4 Legal and Regulatory Issues, Licensing, and Training. Discussions related to the legal envi-
ronment for self-driving vehicles, the licensing requirements for their operation and the need for
training were cordial but generally contentious throughout the study. In some cases, participants
indicated that their concerns were admittedly based on second-hand information or rumors. The
majority of the study’s participants, however, indicated that they had been following national and
local discussions around these issues with many participants indicating that they believed, based
on these discussions, that it was possible that individuals with visual impairments would eec-
tively be blocked from owning or operating self-driving vehicles. Of particular concern were two
types of laws: (1) laws requiring that an individual operating a self-driving vehicle have the ability
to take over control in the event of an emergency and (2) laws that would require a licensed driver
to operate a self-driving vehicle.
Participants were split regarding laws that would require an individual operating a self-driving
vehicle to have the ability to take over as necessary. While some participants found this premise to
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
3:30 J. Brinkley et al.
be unreasonable given that such a vehicle would be designed to be fully automated, others argued
that present or near-future technology would exist that would enable a blind or low-vision driver
to safely bring a vehicle to a stop in the event of an emergency. These participants, while in a
minority, did not believe that such a law would therefore prevent the operation of a self-driving
vehicle by individuals with visual impairments.
While some participants were overtly against the idea of requiring a conventional driver’s li-
cense, some participants felt that a license of some sort should be a requirement given the serious
nature of vehicle operation, self-driving or otherwise. These participants repeatedly raised con-
cerns about the prospect of children or people with cognitive disabilities operating self-driving
vehicles. Those against licensing raised a number of arguments against it but the majority viewed
it as an attack on their civil liberties and as overt discrimination against persons with disabilities.
In all but two of the focus groups the discussion gradually trended toward the concept of an “op-
erator’s license” that would require a certain amount of training on the operation of self-driving
vehicles generally but would not exclude individuals with physical disabilities from ownership or
operation.
4.3.5 Risk, Trust, and Safety. Much of the recent literature on self-driving vehicles and reli-
ability, risk, trust and safety nd that most people are concerned that self-driving vehicles will
not function reliably [26,54]. While the present study included participants who held this view,
the majority expressed the view that self-driving vehicles are at least as safe as human drivers.
Some participants expressed a view that self-driving vehicles would be safer than humans and for
some the potential independence and mobility benets simply outweighed what they viewed as
minimal safety risks. Overall, the attitude expressed toward the technology was generally more
trusting than that expressed in much of the related literature. A small number of participants raised
questions about the potential for software hacking, a topic found to be of importance to those sur-
veyed by Kyriakidis et al. [28], but it was a topic of limited discussion within the study and the
issue generally did not appear to undermine trust in the technology.
4.3.6 Consideration of User Needs. Many participants shared impassioned personal stories
while discussing whether or not they felt the needs of people who are blind or low vision were be-
ing adequately considered in the development of self-driving vehicles. While opinions varied, most
participants expressed the view that their needs were being largely ignored. In all cases, evidence
for this was not any experience with an inaccessible self-driving vehicle, as none of the participants
had rst-hand experience with the technology but was instead negative experiences in other areas
of their lives related to other technologies. Many participants expressed a view that technology
was often designed and developed without adequate input from people with disabilities generally
and was often forced upon them without adequate training or explanation. A number of partic-
ipants commented that major companies routinely moved to new technologies without properly
considering the needs of people with disabilities and potential concerns regarding inaccessibility.
The movement toward touchscreen soda machines was cited several times by participants as a
problematic technology that was being broadly introduced without proper consideration of ac-
cessibility concerns. Several participants described challenges that they had faced in having their
accessibility concerns addressed on a local and state level and questioned their ability to have
their concerns addressed by, arguably, much more dicult to reach international businesses. The
majority of participants specically cited awareness as a key issue. Many expressed a belief that
the technology and technical capability currently exists to make self-driving vehicles universally
accessible but questioned whether manufacturers and decision makers were generally aware of
the importance of doing so. At a minimum these ndings suggest that more needs to be done to
reach out to organizations that serve people who are blind and low vision and to people directly
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
Visual Impairments and Self-driving Vehicles 3:31
to better inform them about what is being done to develop accessible self-driving technology. At
the same time, more, in fact, may need to be done to make self-driving vehicle technology more
accessible while including more stakeholders with a range of visual impairments in the design and
development process.
4.3.7 Limitations. Despite our attempts to structure this research in manner that would hold
up to scrutiny it is not without limitations. While the use of focus group methodology has its
merits there are also weakness with this approach. Perhaps one of the major weaknesses may be
found in the composition of the focus groups themselves. While we have and do argue the merits
of our approach, we recognize that in implementing Krueger and Casey’s [27] guidelines in the
composition of our focus groups, our design choices may be called into question. Given the topic
of study, we argue that our chosen variables, degree of vision loss and age, were rather logical
choices. Disagreement can be found, however, in our manner of implementation. While our goal
in implementing the guidelines was to structure each group according to degree of vision loss
and participant age range, due largely to recruitment, scheduling and logistical constraints several
of the focus groups ended somewhat mixed in terms of degree of vision loss and participant age
range. Generally, low-vision persons were underrepresented in most groups as most groups were
slightly weighted toward blind persons. Our use of a functional denition of vision loss as opposed
to the use of a medical denition of visual acuity may have also practically limited our ability to
precisely group participants by their degree of vision loss. Given the active participation of nearly
all participants in the focus groups, it is our opinion that the aforementioned limitations had little
impact on the discussions themselves, however.
5 CONCLUSION
Viewed collectively, both studies further the process of illuminating the under explored perspec-
tive of blind and visually impaired persons regarding self-driving vehicles. In both studies, visually
impaired persons responded favorably to the concept of self-driving vehicles. The general consen-
sus, from both the initial survey and subsequent focus groups, being that self-driving vehicles have
considerable potential to aid visually impaired persons from the perspective of mobility and in-
dependence. This optimism, for some participants, even superseded concerns regarding risk and
safety. Participants of both activities were also optimistic about societal benets though focus
group participants tended to view this prospect more optimistically than the survey respondents.
Collectively, these ndings suggest that blind and visually impaired persons may have an op-
timism regarding self-driving vehicle technology that surpasses that of the presumably sighted
participants and respondents in the related literature.
While this initial optimism regarding the conceptual prospects of self-driving technology is
noteworthy there are considerable warning signs for manufacturers and the research community
that the described studies highlight. Older adults (55+) were especially concerned regarding their
ability to operate emerging self-driving vehicle technologies. When viewed relative to the aging
population this paints a troubling picture absent additional research in this regard. Visually im-
paired survey respondents expressed some concerns regarding legal liability in the event of an
accident, that when furthered examined within the focus groups was associated with concerns of
spurious claims tied to visual disability. Our ndings also suggest that those with higher levels of
educational attainment (e.g., bachelor’s degree or more) have concerns about whether their needs
are being adequately considered in the design of this emerging technology.
While we argue that the described studies have involved foundational research on self-driving
vehicles that had not been conducted previously, the topic of self-driving vehicle accessibility is
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
3:32 J. Brinkley et al.
broad and requires much additional research. Specically needed are studies that explore the use
of this technology in a practical or naturalistic context.
ONLINE APPENDIX
We have provided an online Appendix, available within the ACM Digital Library, that contains
the instrument used for Study 1, adapted from Schoettle and Sivak [54], as well as the focus group
guide and nal codes from Study 2.
REFERENCES
[1] Alexa. 2017. Retrieved from https://developer.amazon.com/alexa.
[2] American Council of the Blind. 2013. Retrieved from http://www.acb.org/.
[3] Automated Vehicles for Safety. 2017. Retrieved from https://www.nhtsa.gov/technology-innovation/automated-
vehicles-safety.
[4] P. Bansal et al. 2016. Assessing public opinions of and interest in new vehicle technologies: An Austin perspective.
Trans. Res. Part C: Emerg. Technol. 67, (2016), 1–14. DOI:https://doi.org/10.1016/j.trc.2016.01.019
[5] F. Blanco and D. Travieso. 2003. Haptic exploration and mental estimation of distances on a ctitious island: From
mind’s eye to mind’s hand. J. Visual Impair. Blind. 97, 5 (2003), 298–300.
[6] Blind man sets out alone in Google’s driverless car. 2016. Retrieved from https://www.washingtonpost.com/
local/tracandcommuting/blind-man- sets-out- alone-in-googles- driverless-car/2016/12/13/f523ef42-c13d-11e6-
8422-eac61c0ef74d_story.html.
[7] A. S. Bregman and S. Pinker. 1978. Auditory streaming and the building of timbre. Can. J. Psychol./Revue canadienne
de psychologie 32, 1 (1978), 19.
[8] R. N. Brewer and V. Kameswaran. 2018. Understanding the power of control in autonomous vehicles for people with
vision impairment. In Proceedings of the 20th International ACM SIGACCESS Conference on Computers and Accessibility.
185–197.
[9] J. Brinkley et al. 2018. A survey of visually impaired consumers about self-driving vehicles. J. Technol. Persons Disabil.
6, (2018), 274–283.
[10] J. Brinkley et al. 2017. Opinions and preferences of blind and low-vision consumers regarding self-driving vehicles:
Results of focus group discussions. In Proceedings of the 19th International SIG ACCESS Conference on Computers and
Accessibility. 290–299.
[11] S. M. Casey. 1978. Cognitive mapping by the blind. J. Visual Impair. Blind. (1978).
[12] M. A. Cusumano. 2014. How traditional rms must compete in the sharing economy. Commun. ACM 58, 1 (2014),
32–34. DOI:https://doi.org/10.1145/2688487
[13] R. A. Daziano et al. 2017. Are consumers willing to pay to let cars drive for them? Analyzing response to autonomous
vehicles. Transport. Res. Part C: Emerg. Technol. 78, (2017), 150–164. DOI:https://doi.org/10.1016/j.trc.2017.03.003
[14] Deployment of Autonomous Vehicles for Public Operation. 2017. Retrieved from https://www.dmv.ca.gov/portal/
dmv/detail/vr/autonomous/auto.
[15] J. F. Fletcher. 1980. Spatial representation in blind children. 1: Development compared to sighted children. J. Visual
Impair. Blind. 74, 10 (1980), 381–85.
[16] E. Foulke. 1982. Perception, cognition and the mobility of blind pedestrians. Spatial Abil.: Dev. Physiol. Found. (1982),
55–76.
[17] N. A. Giudice. 2018. Navigating without vision: Principles of blind spatial cognition. Handbook of Behavioral and
Cognitive Geography. Edward Elgar Publishing.
[18] R. G. Golledge. 2003. Human waynding and cognitive maps. The Colonization of Unfamiliar Landscapes. Routledge.
49–54.
[19] J. B. Greenblatt and S. Saxena. 2015. Autonomous taxis could greatly reduce greenhouse-gas emissions of U.S. light-
duty vehicles. Nature Climate Change 5, 9 (2015), 860–863. DOI:https://doi.org/10.1038/nclimate2685
[20] J. F. Hair et al. 2006. Multivariate data Analysis (Vol. 6). Pearson Prentice Hall, Upper Saddle River, NJ.
[21] H. Heft. 1996. The ecological approach to navigation: A Gibsonian perspective. The Construction of Cognitive Maps.
Springer. 105–132.
[22] D. Howard and D. Dai. 2014. Public perceptions of self-driving cars: The case of Berkeley, California. In Proceedings
of the Transportation Research Board 93rd Annual Meeting.
[23] E. W. Hu et al. 2019. Am I too old to drive?: Opinions of older adults on self-driving vehicles. In Proceedings of the
21st International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS’19). 500–509.
[24] iOS - Siri. 2017. Retrieved from http://www.apple.com/ios/siri/.
[25] J. Kitzinger. 1995. Qualitative research: Introducing focus groups. BMJ 311, 7000 (1995), 299–302.
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
Visual Impairments and Self-driving Vehicles 3:33
[26] G. Silberg et al. 2013. Self-Driving Cars: Are We Ready. Kpmg Llp. 1–36.
[27] R. A. Krueger and M. C. Casey. 2009. Focus Groups a Practical Guide for Applied Research.SagePublications.
[28] M. Kyriakidis et al. 2015. Public opinion on automated driving: Results of an international questionnaire among 5000
respondents. Transport. Res. Part F: Trac Psychol. Behav. 32, (2015), 127–140. DOI:https://doi.org/10.1016/j.trf.2015.
04.014
[29] S. Lacey et al. 2014. Spatial imagery in haptic shape perception. Neuropsychologia 60, (2014), 144–158.
[30] S. Lacey and K. Sathian. 2012. Representation of object form in vision and touch. The Neural Bases of Multisensory
Processes. CRC Press/Taylor & Francis.
[31] Learn About OnStart. 2017. Retrieved from https://www.onstar.com/us/en/home.html.
[32] J. Loomis and R. Klatzky. Functional equivalence of spatial representations from vision, touch, and hearing: Relevance
for sensory substitution. In Blindness and Brain Plasticity in Navigation and Object Perception. Taylor & Francis, 167–
196.
[33] J. M. Loomis et al. 2013. Representing 3D space in working memory: Spatial images from vision, hearing, touch, and
language. Multisensory Imagery. Springer, 131–155.
[34] Low Vision and Legal Blindness Terms and Descriptions—Visionaware. 2015. Retrieved from http://www.
visionaware.org/info/your-eye-condition/eye- health/low-vision/low- vision-terms-and- descriptions/1235.
[35] A. Marinik et al. 2014. Human Factors Evaluation of Level 2 and Level 3 Automated Driving Concepts: Concepts of
Operation. Report No. DOT HS 812 044. National Highway Trac Safety Administration.
[36] MAXQDA: Qualitative Data Analysis Software | Windows & Mac. 2017. Retrieved from http://www.maxqda.com/.
[37] Microsoft Cortana Dev Center. 2017. Retrieved from https://developer.microsoft.com/en-us/cortana.
[38] S. Millar. 1981. Crossmodal and intersensory perception and the blind. Intersensory Perception and Sensory Integration.
Springer. 281–314.
[39] Laerd Statistics. One-way ANOVA using Minitab. Retrieved on 25 March, 2020 from https://statistics.laerd.com/
minitab-tutorials/one- way-anova- using-minitab.php.
[40] J. Missel. 2014. Ipsos MORI Loyalty Automotive Survey. Retrieved from https://www.ipsos.com/ipsos-mori/en-uk/
only-18- cent-britons- believe-driverless-cars-be-important-development-car-industry- focus.
[41] Model S. Autopilot Press Kit | Tesla Motors. 2016. Retrieved from https://www.teslamotors.com/presskit/autopilot.
[42] D. R. Montello. 1998. A new framework for understanding the acquisition of spatial knowledge in large-scale envi-
ronments. In Spatial and Temporal Reasoning in Geographic Information Systems. Oxford University Press, 143–154.
[43] D. R. Montello. 2005. Navigation. Cambridge University Press.
[44] National Federation of the Blind Resolutions for 2016 | National Federation of the Blind. 2016. Retrieved from
https://nfb.org/2016-resolutions.
[45] National Highway Transportation Safety Administration 2017. Automated Driving System: A Vision for Safety 2.0.
Technical Report #DOT HS 812 442.
[46] W. Payre et al. 2014. Intention to use a fully automated car: Attitudes and a priori acceptability. Transport. Res. Part
F: Trac Psychol. Behav. 27 (2014), 252–263.
[47] J. J. Rieser. 1989. Access to knowledge of spatial structure at novel points of observation. J. Exper. Psychol.: Learn.
Memory Cogn. 15, 6 (1989), 1157.
[48] J. J. Rieser et al. 1986. Sensitivity to perspective structure while walking without vision. Perception 15, 2 (1986), 173–
188.
[49] G. Ruggiero et al. 2012. Egocentric/allocentric and coordinate/categorical haptic encoding in blind people. Cogn.
Process. 13, 1 (2012), 313–317.
[50] C. L. Ryan and K. Bauman. 2016. Educational attainment in the united states: 2015. United States Census Bureau.
[51] SAE International. 2016. Automated Driving: Levels of Driving Automation Are Dened in New SAE International Stan-
dard J#016. SAE International.
[52] SAE J3016 automated-driving graphic. 2019. https://www.sae.org/news/2019/01/sae-updates-j3016- automated-
driving-graphic.
[53] V. R. Schinazi et al. 2016. Spatial navigation by congenitally blind individuals: Spatial navigation bycongenitally blind
individuals. Wiley Interdisciplinary Reviews: Cognitive Science. 7, 1 (2016), 37–58. DOI:https://doi.org/10.1002/wcs.1375
[54] B. Schoettle and M. Sivak. 2014. A Survey of Public Opinion About Autonomous and Self-Driving Vehicles in the US, the
UK, and Australia. Report No. UMTRI-2014-21. University of Michigan, Ann Arbor, Transportation Research Institute.
[55] B. Schoettle and M. Sivak. 2016. Motorists’ Preferences for Dierent Levels of Vehicle Automation: 2016. Report No.
UMTRI-2015-22. University of Michigan Sustainable Worldwide Transportation.
[56] S. J. Shettleworth. 2010. Cognition, Evolution, and Behavior. Oxford University Press.
[57] A. W. Siegel and S. H. White. 1975. The development of spatial representations of large-scale environments. Advances
in Child Development and Behavior. Elsevier. 9–55.
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
3:34 J. Brinkley et al.
[58] R. Snyder. 2016. Implications of autonomous vehicles: A planner’s perspective. Inst. Transport. Eng. J. 86, 12 (2016),
25.
[59] K. Sommer. 2013. Continental Mobility Study 2011. Continental AG. (2013), 19–22.
[60] Statistical facts about blindness in the united states. 2018. Retrieved from https://nfb.org/blindness-statistics.
[61] N. Strand et al. 2014. Semi-automated versus highly automated driving in critical situations caused by automation
failures. Transport. Res. Part F: Trac Psychol. Behav. 27 (2014), 218–228. DOI:https://doi.org/10.1016/j.trf.2014.04.005
[62] L. Tcheang et al. 2011. Visual inuence on path integration in darkness indicates a multimodal representation of
large-scale space. Proc. Natl. Acad. Sci. U.S.A. 108, 3 (2011), 1152–1157.
[63] Technology. 2017. Retrieved from https://waymo.com/tech/.
[64] Ten ways autonomous driving could redene the automotive world | McKinsey & Company. 2015. Retrieved
from http://www.mckinsey.com/industries/automotive-and- assembly/our-insights/ten-ways- autonomous-driving-
could-redene- the-automotive- world.
[65] The Leading Research & Experience Software. Retrieved from https://www.qualtrics.com/.
[66] Vision Accessibility-Mac. 2018. Retrieved from https://www.apple.com/accessibility/mac/vision/.
[67] R. F. Wang and J. R. Brockmole. 2003. Simultaneous spatial updating in nested environments. Psychonom. Bull. Rev.
10, 4 (2003), 981–986.
[68] S. Wilkinson. 1998. Focus group methodology: A review. Int. J. Soc. Res. Methodol. 1, 3 (1998), 181–203.
Received April 2018; revised July 2019; accepted January 2020
ACM Transactions on Accessible Computing, Vol. 13, No. 1, Article 3. Publication date: April 2020.
