Exploring the role of artificial intelligence and inclusive technologies during navigation-based tasks for individuals who are blind or who have low vision: Future directions and priorities.

doi:10.21203/rs.3.rs-3715501/v1

Download PDF

Research Article

Exploring the role of artificial intelligence and inclusive technologies during navigation-based tasks for individuals who are blind or who have low vision: Future directions and priorities.

https://doi.org/10.21203/rs.3.rs-3715501/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background

Mainstream smartphone applications are increasingly replacing the use of traditional visual aids (such as hand-held telescopes) to facilitate independent travel for individuals who are blind or who have low vision.

Objective

The goal of this study was to explore the navigation-based apps used by individuals who are blind or low vision, the factors influencing these decisions, and perceptions about gaps to address future needs.

Methods

An international online survey was conducted with 139 participants who self-identified as blind or low vision (78 women, 52 men) between the ages of 18 and 76.

Results

Findings indicate that the decision to use an app based on artificial intelligence versus live video assistance is related to whether the task is dynamic or static in nature. Younger participants and those who are congenitally blind are significantly more likely to employ apps during independent travel. Although a majority of participants rely on apps only during unfamiliar routes (60.91%), apps are shown to supplement rather than replace traditional tools such as the white cane and dog guide. Participants underscore the need for future apps to better assist with indoor navigation and to provide more precise information about points of interest.

Conclusions

These results provide vital insights for rehabilitation professionals who support the growing population of clients with acquired and age-related vision loss, by clarifying the factors to consider when selecting apps for navigation-based needs. As additional technology-based solutions are developed, it is essential that blind and low vision individuals, including rehabilitation professionals, are meaningfully included within design.

Physical Medicine & Rehab

Visual impairment

blind

low vision

orientation and mobility

accessibility

inclusive technologies

artificial intelligence

quality of life

universal design

According to the World Health Organization, at least 2.2 billion people have a visual impairment (blindness or low vision) worldwide [1]. This number is projected to at least double by 2050 across Global North countries, as individuals continue to age [2]. This poses an important challenge for the future, as visual impairment may significantly impact quality of life, even when compared to some other chronic conditions [3–7]. Indeed, individuals with visual impairments, regardless of their age group, may be more likely to experience anxiety [8] and/or depressive symptoms and may experience lower health related quality of life due to limited physical activity and sedentary behaviors [9–11]. Furthermore, visual impairment may affect safe and independent travel and is associated with an increased risk of falls. Consequently, expanding access to rehabilitation services and assistive technology is vital for enhancing overall quality of life. Forward-facing solutions must also take into account broader aspects of universal accessibility and inclusion to encourage and facilitate independent travel. This is especially important since the ability to safely perform daily activities and independent travel is correlated with higher levels of education, employment, social inclusion [12] and can increase the likelihood of aging in place at home .

Two interrelated navigation-based competencies that individuals who are blind or have low vision employ are orientation (the ability to determine one’s position in space) and mobility (the ability to safely navigate within one’s environment). Mobility tools that facilitate independent travel include optical aids (i.e., magnifier, telescope), a long white cane for assistance with obstacle detection and the use of a dog guide trained for obstacle avoidance while following route directions provided by the handler [13]. Blind and low vision travelers also draw on a scope of compensatory strategies including the use of auditory and tactile cues and echolocation within indoor and outdoor environments [14]. While these existing tools and techniques significantly facilitate independent travel, a number of external barriers persist. Travellers with disabilities must negotiate increasingly complex and dynamic spaces, including congested pedestrian traffic, complex multi-lane intersections and silent electric vehicles that cannot be discerned by sound, and challenging environments such as unpredictable construction zones and climate conditions. Members of the blind and low vision community underscore the need to consider innovative approaches to address these persisting barriers, including environmental accessibility and the use of inclusive technologies [15].

Mainstream smartphones and tablets (such as Apple iPhone and Google Android products) incorporate built-in accessibility features (including screen magnification, electronic braille connectivity and text-to-speech software) that can be activated by all users [16]. Based on the principles of universal design, these accessibility features enable users to access applications that support daily living activities, including those to read email, manage calendars and provide GPS support during travel [17]. Few prior studies have explored the extent to which smartphone applications are addressing different facets of navigation-based needs within the blind and low vision population. However, a recent study found that mainstream devices are increasingly replacing the use of traditional visual aids (such as specialized stand-alone GPS devices for blind users) for navigation-based tasks [17]. Previous investigations have centered heavily on the design and evaluation of sight substitution prototypes to address perceived challenges, including assisting with localization, multiple object tracking and obstacle avoidance, and the ability to follow a moving target, such as a line in a coffee shop [18–20]. Other investigations have explored which navigation-based applications are generally used by individuals with visual impairments, but have not examined the extent to which current solutions address specific navigation-based needs [21, 22]. This is particularly relevant given that barriers that impede orientation and mobility may arise at any point on the travel continuum (see Table 1), yet knowledge on the role of inclusive technologies at each of these points remains limited.

Current navigation-based applications can be classified into three broad categories, based on the audience that is targeted (mainstream apps designed for the general public or specialized options for blind and low vision users), the task that is addressed, and the underlying mechanism used to drive the application. Applications tend to focus on specific facets of independent travel. For instance, there are applications that assist with route planning, provide information about street names and other points of interest, vocalize stop announcements on public transportation, or those which enable users to geolocate services (such as restaurants) within a given geographic range [13]. Several of these applications have been designed to meet unique travel needs for blind and low vision users. For example, blind and low vision smartphone users can take a picture of signage or other environmental text, and use artificial intelligence (AI) to perform optical character recognition (whereby the text is interpreted and read aloud) [23]. Additionally, these applications also differ based on the ways in which environmental information is gathered and transmitted. While some applications draw on AI to identify objects, locate points of interest or perform optical character recognition, others enable users to call upon remote video assistance, provided by sighted volunteers, family and friends or trained visual interpreters [15, 24].

Objectives

To date, research on factors influencing the decision to use specific navigation-based apps remains scant, including whether the use of AI or live remote sighted assistance is preferrable depending on the nature of the task. Therefore, we conducted an international survey to collect information from the blind and low vision community about their travel habits and smartphone application usage in navigation-based tasks. The threefold objectives of this study were to explore the applications used for different navigation-based tasks among blind and low vision users, identify the factors that correlate with application usage, and explore the extent to which current solutions are meeting navigation-based needs. These findings will better inform technology-based recommendations made by rehabilitation practitioners who support clients with vision loss and will direct developers engaged in inclusive technology design.

Data were collected through an accessible anonymous online survey hosted on the Université de Montréal’s Lime Survey platform between June 2022 and February 2023. Ethics approval consistent with the Declaration of Helsinki [25] was obtained through the Université de Montréal (CERC 2022 − 1465). Prior to commencing the survey, participants read the Information and Consent form on the initial page of the platform. Informed consent was obtained by the decision to proceed to the online survey.

Eligibility and recruitment

Participation was open internationally to individuals aged 18 years or older, who had been using a smartphone for at least three months, who could communicate in English or French and who self-identified as blind or as having low vision. To provide additional context about level of vision, data about reading and writing methods used (e.g. large print, text-to-speech software and braille) and visual diagnosis were also collected. The invitation to participate was posted to over 150 social media groups geared towards members of the blind and low vision community. Snowball sampling (whereby participants were invited to share the invitation with others) provided additional reach beyond these means [26].

Sample size

Given the exploratory nature of this study, no a priori sample size was computed. A sample size of 100 + participants was expected, given previous success with similar studies aimed at blind and low vision technology users [17].

Materials and procedure

The survey included between 45 and 50 questions (depending on participant responses). Data were collected through primarily closed-ended (multiple choice) questions, with the option of providing additional open-ended comments (see supplementary file 1 for the full instrument). The survey was comprised of three distinct sections. Section one of the survey (between 22 and 27 questions) gathered data about demographic characteristics. Section two included 14 questions about navigation-based tasks and the type of smartphone applications employed to perform these tasks, if any. Table 1* includes the categories of navigation-based tasks explored in this section and provides a brief definition for each. The different navigation-based tasks investigated reflect the typical trajectory for completing a route independently, from route planning to reaching a point of interest [13]. Participants were asked to indicate whether they use an application for each task, and to identify the application either by selecting from the provided list or by supplying an additional response of their own. If a participant did not use any applications for a given task, they were asked to specify the reason either by selecting among the provided choices or by inserting their own open-ended response. The final section, section three, of the survey (consisting of 9 questions) asked participants to comment further on the factors which influence their decision to use a smartphone application for different navigation-based tasks, and the advantages and disadvantages of currently available solutions.

Data analysis

Results from the survey were exported into excel where incomplete submissions were removed, resulting in a total of 139 respondents of which 125 completed section two, and 110 completed section three. Following data cleaning, all applications used by participants were classified into four distinct categories (Human assistance, Computer Vision/AI, Camera magnification/Image rendering, and GPS), such that participants could be classified as either using or not using apps within that category for each navigation-based task (see Table 1 and 2)*.

Table 1

Categorization of data OR descriptions
Navigation-based tasks across the independent travel trajectory
Task		Definition
Establishing a route itinerary		Determining route instructions prior to leaving, preparing a travel plan
Obstacle detection		The ability to detect and avoid obstacles in one’s line of travel, such as other pedestrians, poles, and signposts
Visual identification		The ability to identify aspects of the visual environment, such as an environmental object (e.g. product at a store) or text (e.g. public signage and street names)
Street crossings		The ability to determine when it is time to safely cross the street, read traffic signals and cross while maintaining alignment
Using public transport		Determining schedules (e.g. bus times), locating bus stops and tracking bus stop announcements (e.g. to determine one’s location while on route)
Identifying points of interest in close proximity		The ability to confirm a nearby point of interest/destination (e.g. restaurant, pharmacy)
Geolocation		Identifying one’s position while on route (e.g. where am I?) such as a specific street, intersection or address (GPS).
Categories of apps
Category		Definition
Human assistance		Interacting with a sighted person virtually through video and /or audio (i.e.,AIRA)
Computer vision/AI		Automated Visual interpretation using artificial intelligence (AI) of an image or text from the smartphone camera (i.e., Seeing AI)
Camera magnification/image rendering		Altering of an image to increase visibility / readability to the user (i.e., Magnifying Glass)
GPS		Use of the smartphone geolocation and possible integration of other geolocation applications (i.e., Google Maps)
App audience
Audience		Definition
Specialized apps		Apps purposefully designed to be used by individuals with visual impairments (ex: BeMyEyes)
Mainstream apps		Apps meant to be used by the general public, with or without considerations of disability (ex: google maps)
Categories of unmet needs
Task		Definition
Indoor navigation		Difficulty navigating large indoor spaces (i.e., finding store in mall)
Accessibility		Some or all the functions of the app/website were inaccessible
POI/geolocation		Difficult finding specific locations when nearby (i.e., entrance to a store, gate at airport, etc.)
Street crossing		Information about street crossing unavailable (i.e., crossing time, crossing signal)
Route planning		Lack of information to accurately plan route (i.e., map of bus terminal not available, information not available while traveling in car)
App precision		Errors or unreliability of information (i.e., GPS in mountainous area, bus stops not properly labeled/updated)
Versatility		Prefer to have multiple functions in one app
Object ID		Identify objects and traffic sounds
Obstacle detection		Finding obstacles that the white cane does not

The data were then analysed using three main statistical tests. First, Cochran’s Q tests were performed to assess if there were differences in matched sets of proportions (binary variables: yes or no). When Cochran’s Q tests revealed that the proportions of yes and no responses varied in a certain set, pairwise Post-hoc Dunn’s Multiple Comparison Test (with Bonferroni corrections) were performed to locate what proportions significantly differed. Second, Chi-square tests of independence were performed to examine the relation between different pairs of nominal or binary variables. Third, Spearman Rho tests were used to investigate the relation between variables when at least one of these was continuous or ordinal.

Table 2

Dependent and independent variables
Independent variables
Variable	Type	Levels
Age	Continuous	≥ 18
Age group*	Binary	< 50 or > 50
Gender	Nominal	M, F, non-binary, or other
Visual impairment	Nominal	Blind, Low vision, or other
Age at onset of visual impairment	Continuous	≥ 0
Age at onset group*	Ordinal	1 = During early childhood: [0,3[ 2 = During childhood: [3,12[ 3 = During adolescence: [12,20[ 4 = During adulthood [20,50[ 5 = During late adulthood: 50+
Visual impairment duration*	Continuous	= Age - Age at onset of visual impairment
Type of mobility aid	Nominal	Long cane, dog guide, or no aids/non-traditional aids
Type of phone	Nominal	iPhone or Android
Time using phone	Ordinal	1 = > 3h/day 2 = 1-3h/day 3 = > 1h/week
Confidence with technology	Ordinal	1 = Beginner 2 = intermediary 3 = advanced
Frequency of travel	Ordinal	1 = daily 2 = several times a week 3 = once a week 4 = sometimes per month
Confidence during independent travel	Ordinal	1 = Beginner 2 = intermediary 3 = advanced
Dependent variables
Why Variable	Type	Levels
Use of app	Binary (set)	Yes or No (for each navigation task)
Use of app*	Binary (set)	Yes or No (for each app category, one set per navigation task)
Reported reasons why not using apps for a navigation task	Binary (set)	Yes or No (for each reason, one set per navigation task)
Are travels possible without the use of apps	Binary	Yes or No
Reason for choosing an app	Binary (set)	Yes or No (for each reason)
How apps are learned	Binary (set)	Yes or No (for each way to learn)
Conditions in which apps do not work properly	Binary (set)	Yes or No (for each condition)
Are there needs that are not met by apps	Binary	Yes or No
Unmet needs*	Binary (set)	Yes or No (for each category of need)
*Variable derived from the data. O&M, Orientation and mobility

Demographics

A total of 139 participants completed the survey (78 women, 52 men, mean age = 48.9 ± 15.5 years) from a variety of countries, mainly Canada (n = 85), the USA (n = 18), and countries within Europe (n = 27) Table 3 contains additional demographic information.

Table 3

Participants Demographics
Response	n respondents
Gender
Female	78
Male	54
Did not answer	6
Country
Canada	85
USA	18
Europe	27
Others	5
Did not answer	3
Age
18–19	3
20–29	13
30–39	29
40–49	19
50–59	26
60–69	34
70–76	10
VI category
Blind	86
Low vision	43
Other	6
Did not answer	3
Age at VI onset
0–2	60
3–11	17
12–20	12
21–50	23
50–68	8
Auditory impairment
Yes	24
No	114
Type of phone
Iphone	120
Android	23
Time phone use
> 3 h/day	78
1–3 h/day	44
< 1h/day	12
Did not answer	3
Competence with technology
Beginner	10
Intermediary	47
Advanced	77
Did not answer	4
Aids used
None or non-traditional aids	20
White cane	92
Guide dog	26
Area of travel
Urban	105
Suburban	66
Rural	23
Frequence travel
Everyday	58
Several times a week	54
Once a week	12
A few times a month	6
Other	4
Did not answer	3
Confidence in independent navigation
Beginner	20
Intermediary	51
Advanced	63
Did not answer	4

Q1: What types of apps are used for navigation-based tasks?

In total, 125 participants completed the section about navigation-based tasks. Findings indicate that most participants (n = 120, 96%) use applications during independent travel, with only 4% (n = 5) of participants indicating that they do not use apps for navigation-based tasks at all. However, the proportions of participants using apps significantly differed between the navigation-based tasks (Q(6, n = 125) = 253.003, p < .001). More precisely, participants are less likely to use apps for street crossings and obstacle detection compared to other navigation-based tasks (p < .001). More detailed information can be found in Fig. 2 and supplementary file 2.

Apps to access visual information

Overall, the navigation-based task for which apps are most commonly adopted (76% of respondents) relates to visual interpretation (e.g., locating/identifying objects or reading environmental text). For such visual interpretation, respondents significantly preferred using specialized apps (n = 87, 91.6%) compared to mainstream apps (n = 50, 52.63%), and predominantly used computer vision/AI and human assistance apps for this purpose. For those who indicated not using apps for visual interpretation, the most common reason reported was the lack of awareness of apps to assist with this task (n = 13, 48.15%).

Apps for planning and following routes

For navigation-based tasks related to general orientation and routes (e.g., planning routes, finding POIs, using public transport, and geolocation), GPS apps were predominantly used. Mainstream apps were significantly preferred to specialized apps for planning routes (92.5% vs 43.62%) and taking public transportation (77.03% vs 52.70%), but there was no apparent preference between mainstream and specialized apps for geolocation (74.07% vs 72.84%) and finding POIs (68.60% vs 82.56%). For participants who indicated not using apps for these tasks, the main reasons identified were that respondents did not feel the need to use apps for this purpose or did not know apps that could help them.

Apps for dynamic tasks related to perceived risk

Participants were significantly less likely to use apps for dynamic navigation-based tasks (such as street crossings and obstacle detection while in movement) because most respondents: 1) already use other aids for these tasks (street crossings, n = 49; obstacle detection, n = 68); 2) don’t know apps that could help them (street crossings, n = 49; obstacle detection, n = 47); and 3) don’t feel the need to use apps for this purpose (street crossings, n = 35; obstacle detection, n = 41). Other reported reasons included the lack of trust in applications or technology in general (e.g., incomplete maps, battery life, processing delays, or smartphone processing capacities, n = 9). However, among those who do use apps for dynamic tasks (street crossings, n = 23; obstacle detection, n = 15), human assistance apps are more commonly used than those based on AI, with a preference for specialized apps over mainstream apps (street crossings, 86.95% vs 47.82%; obstacle detection, 72.22% vs 53.33%), a pattern that was only significant for street crossing.

Q2: What factors are correlated with app usage?

Among the 110 responses, 36 (32.72%) of participants indicated that they do not rely on apps for travel; 67 (60.91%) rely on apps only in unfamiliar areas, and 7 (6.36%) rely on apps for all travel. This was found to be true for both blind and low vision participants, with a significant correlation with the frequency at which participants travel. Specifically, participants who travel less often are more likely to rely on apps during travel (rho = 0.228, p = .021). Furthermore, we investigated the correlation between several factors and apps usage for the different navigation-based tasks, with the following significant correlations observed:

Age. Younger respondents were more likely to use apps for obstacle detection (rho=-0.179, p = .046) and visual identification (rho=-0.202, p = .025).

Frequency of travel. The more respondents travel, the more likely they are to use apps to prepare routes (rho = -0.263, p = .004).

Level of vision. Blind respondents were more likely to use apps for visual interpretation (X²(1, n = 116) = 12.14, p < .001), finding points of interest (X²(1, n = 116) = 5.02, p = .025), and geolocation (X²(1, n = 116) = 6.001, p = .014). In general, blind and low vision respondents equally used apps of the different categories except for visual interpretation where low-vision respondents are more likely to use image rendering apps (13/23 vs 22/68, stats), and blind respondents, to use computer vision/AI apps (66/68 vs 15/23, stats).

VI onset. Participants with earlier VI are more likely to use apps for street crossing (rho=-0.193, p = .044), finding POI (rho=-0.349, p < .001), using public transport (rho=-0.396, p < .001), and geolocation (rho=-0.282, p = .003). Similarly, participants with a longer VI duration – found to be linked with earlier VI onset (rho=-0.605, p < .001) – are also more likely to use apps for street crossing (rho = 0.323, p < .001), finding POI (rho = 0.204, p = .037), and public transport (rho = 0.254, p = .009).

Type of navigational aid. As the type of navigational aids used was closely related to level of vision (X²(2, n = 129) = 23.842, p < .001), the analysis revealed the same pattern of results. Participants using non-traditional aids, or no aids (30% of low-vision participants, 3% of blind participants) were less likely to use apps for visual interpretation (X²(2, n = 119) = 12.14, p = .002), finding POI (X²(2, n = 119) = 11.19, p = .004) and geolocation (X²(2, n = 119) = 6.02, p = .049). However, it was found that individuals using non-traditional aids, or no aids, were less likely to use apps to take public transport (X²(2, n = 119) = 8.31, p = .016).

Overall, the most reported condition that prevents the use of an app is the presence of loud ambient sounds (n = 60, 54.55%), and the least reported was high luminosity (n = 9, 8.18%). Low-vision participants are more likely to be bothered by high luminosity (16.67% vs 1.45%, X²(1, n = 105) = 8.804, p = .003). As for the factors participants take into account when selecting apps, the results show that the importance of the different factors significantly varied (Q(3,n = 110) = 121, p < .001). The most important factors were 1) the ease of use/accessibility of the app (N = 90, 81.82%); and 2) that the app responds to a certain need (n = 79, 71.82%). Moreover, the price (reported by 42 respondents, 38.18%) was perceived to be more important than the amount of data used by the app (see Fig. 3). Blind respondents were more likely to choose apps based on whether the app responds to a specific need (54/69 vs 21/36, X²(1, n = 105) = 4.603, p = .032), and also tended to choose their apps according to their ease of use/accessibility” (60/69 vs 26/36, X²(1, n = 105) = 4.5, p = .063).

Q3: Are current apps addressing navigation-based needs?

In total, 55 respondents (50%) reported having navigation-based needs that are currently not met by applications available to them. Among these, blind respondents were more likely to have unmet needs than those with low vision (38/69 vs 12/36, X²(1, n = 105) = 4.482, p = .034). Respondents were then invited to elaborate on what needs were currently not met. The 49 responses received were classified into nine categories (see Table 1 and 2*). Results showed that some categories were more frequently reported than others (Q(8, n = 49) = 41.096, p < .001). The three most reported categories of unmet needs were: “POI/geolocation” (n = 17, 34.70%), “indoor navigation” (n = 14, 28.57%), and “route planning” (N = 15, 30.61%), while the least reported categories were: “Obstacle detection” (n = 1, 2.41%), “Object Identification” (n = 2, 4.82%) and “versatility” (n = 1, 2.41%). More detailed information can be found in Fig. 3 and supplementary file 2. Additionally, respondents traveling in semi-urban/suburbs were more likely to have unmet needs for POI/geolocation, than those not traveling in such areas (11/19 vs 6/30, X²(1, N = 49) = 7.373, p = .007).

The goals of this study were to explore 1) the navigation-based apps used by blind and low vision individuals, 2) the factors that correlate with the decision to use an application, and 3) the perceptions on the extent to which applications are meeting current navigation-based needs. Our findings raise important insights to strengthen the potential for future inclusive technologies to better support the health, safety, and independent travel needs of blind and low vision adults.

Navigation-based apps used and influential factors

While prior studies have found that mainstream solutions such as smartphones are increasingly replacing the use of traditional aids (such as hand-held telescopes) during independent travel [17, 27], the findings of the current study underscore an additional distinction between mainstream and specialized solutions. Participants are adopting mainstream applications (i.e., Google Maps) available to the general public for planning and following routes, while selecting more specialized solutions (i.e., AIRA, BlindSquare) when additional visual information about a route is required. In fact, a majority of navigation-based tasks performed by blind and low vision users mirror those of the general population. Like sighted individuals, persons with visual impairments may draw on technologies to plan a route to an unfamiliar destination or to obtain new route instructions. However, they may also draw on additional visual interpretation support (such as information about specific landmarks, intersections, and points of interest) that move beyond the level of detail provided by mainstream applications.

This study also suggests that AI is more frequently employed when the user remains stationary (i.e., reading signage, confirming a point of interest). Conversely, participants are more likely to draw on live sighted assistance during dynamic situations, such as locating a front door while walking or crossing a street. These findings point to the inherent limitations of AI, which is not yet able to provide the same timely and precise information afforded by a live visual interpreter. Time sensitive and precise information about a physical environment is especially vital in dynamic situations, where users must negotiate constantly moving targets, such as people and traffic, highlighting the benefit of asking precise and timely questions.

Separate from technology competencies, users must also have opportunities to learn about the benefits and limitations of these solutions and to assess the best option, based on their safety and needs [28]. Such insights underscore the value of interdisciplinary collaboration between both assistive technology and orientation and mobility specialists, but also the potential benefit of multifunctional devices that incorporate both AI and human-based assistance for greater flexibility.

Findings also accentuate the importance of proactively incorporating accessibility within mainstream solutions [15]. While accessibility standards for websites and applications exist, these are not consistently applied across platforms, leading to persisting access barriers for users with diverse needs, including older adults and second language speakers [29]. The accessibility of a mainstream app may also change when software updates are implemented, if designers do not consider these factors prior to launch [30]. Additionally, a recent study has shown that O&M specialists tend to adopt independent travel technologies slower than the clients they serve [31], highlighting the need for professionals to remain informed about application developments and changes.

In line with previous studies [32], participants with congenital, or early onset, blindness or low vision are more likely to draw on applications during independent travel. Likewise, participants who are blind are more likely to use applications for a variety of navigation-based tasks, including object identification and the reading of text during independent travel. These findings raise the importance of supporting the early introduction of these tools for users who have progressive vision loss. Ample research indicates that the early adoption of tools facilitates cortical plasticity and the transition to non-visual methods once additional vision loss occurs [33].

Unmet needs and future directions

Participants highlighted that although technologies are beneficial in a number of contexts, these tools supplement rather than replace the use of traditional tools and techniques, such as the use of a dog guide or white cane. This is evidenced by the fact that participants may not use navigation-based apps in familiar locations. However, respondents highlight an expressed need for further support during indoor navigation and for soliciting more precise information about points of interest, such as the exact location of an entry door. Importantly, previous research has centered on the development of solutions to assist with other aspects of independent travel not identified as significant barriers here, such as the ability to avoid obstacles or maintain one’s position in line [34, 35]. It is evident that while helpful for some end-users, not all applications necessarily reflect the needs and priorities of a majority of blind and low vision users. Experts with disabilities have increasingly emphasized the need to meaningfully include end-users with disabilities as core members of the team from inception, to ensure that technology designs are based upon priorities from within the community [36]. Beyond meaningful partnerships with blind and low vision users, assistive technology specialists and orientation and mobility professionals (who may also be members of the blind and low vision community) can also provide professional insights to further clarify issues pertaining to environmental accessibility and safety to inform the development of new tools.

Limitations and Future Research

Participants were primarily recruited through English-speaking social media platforms. It is likely that the sample does not represent the views of those with lower technology competencies and those who are less connected to online services, including those located in developing countries. To supplement this data, members of this research team are pursuing a number of related investigations, including those that focus on the navigation-based needs of those newer to blindness or low vision and with less self-identified technology and independent travel competencies.

As technology developers work towards the design of new and innovative solutions, it is vital that these tools respond to the increasing diversity of individuals. For those who are blind or low vision, navigation-based applications provide essential insights about the visual environment to supplement the knowledge they gain through traditional methods. For these tools to be most effective, blind and low vision users including assistive technology, vision rehabilitation and orientation and mobility specialists, should be meaningfully included in the consultation and design process from inception [37]. Through meaningful inclusion, developers can gain vital insights about the decisions underpinning the use of specific navigation-based apps, alongside future development priorities to respond to unmet needs.

Conflicts of interest

The authors declare having no conflict of interest.

World Health Organization. World report on vision. 2019; Available from: https://www.who.int/publications/i/item/9789241516570.
Varma, R., T. Vajaranant, and B. Burkemper, Visual impairment and blindness in adults in the United States: Demographic and geographic variations from 2015 to 2050. Journal of the American Medical Association: Opthalmology, 2016. 134(7): p. 802-809.
Wiener, W.R., R.L. Welsh, and B.B. Blasch, Foundations of orientation and mobility. Vol. 1. 2010: American Foundation for the Blind.
Marron, J.A. and I.L. Bailey, Visual factors and orientation-mobility performance. Optometry and Vision Science, 1982. 59(5): p. 413-426.
Popescu, M.L., et al., Age-related eye disease and mobility limitations in older adults. Investigative ophthalmology & visual science, 2011. 52(10): p. 7168-7174.
Turano, K.A., et al., Association of visual field loss and mobility performance in older adults: Salisbury Eye Evaluation Study. Optometry and vision science, 2004. 81(5): p. 298-307.
Loprinzi, P.D., B.K. Swenor, and P.Y. Ramulu, Age-related macular degeneration is associated with less physical activity among US adults: cross-sectional study. PLoS One, 2015. 10(5): p. e0125394.
Langelaan, M., et al., Impact of visual impairment on quality of life: a comparison with quality of life in the general population and with other chronic conditions. Ophthalmic epidemiology, 2007. 14(3): p. 119-126.
Ahmad Bahathig, A., et al., Relationship between physical activity, sedentary behavior, and anthropometric measurements among Saudi female adolescents: a cross-sectional study. International journal of environmental research and public health, 2021. 18(16): p. 8461.
Haegele, J.A., R. Famelia, and J. Lee, Health-related quality of life, physical activity, and sedentary behavior of adults with visual impairments. Disability and rehabilitation, 2017. 39(22): p. 2269-2276.
Smith, L., et al., Visual impairment and objectively measured physical activity and sedentary behaviour in US adolescents and adults: a cross-sectional study. BMJ open, 2019. 9(4): p. e027267.
Martiniello, N. and W. Wittich, Employment and visual impairment: Issues in adulthood, in Routledge Handbook of Visual Impairment, J. Ravenscraft, Editor. 2019, Routledge. p. 415-437.
Wiener, W.R., R.L. Welsh, and B.B. Blasch, Foundations of Orientation and Mobility, 3rd Edition: Volume 2, Instructional Strategies and Practical Applications. 2010: AFB Press.
Kreidy, C., et al., How Face Masks Affect the Use of Echolocation by Individuals With Visual Impairments During COVID-19: International Cross-sectional Online Survey. Interactive Journal of Medical Research, 2022. 11(2): p. e39366.
Siu, Y.-T. and I. Presley, Access technology for blind and low vision accessibility. 2020: APH Press, American Printing House for the Blind.
Khan, A. and S. Khusro, An insight into smartphone-based assistive solutions for visually impaired and blind people: issues, challenges and opportunities. Universal Access in the Information Society, 2021. 20: p. 265-298.
Martiniello, N., et al., Exploring the use of smartphones and tablets among people with visual impairments: Are mainstream devices replacing the use of traditional visual aids? Assistive Technology, 2022. 34(1): p. 34-45.
Kuribayashi, M., et al. Linechaser: A smartphone-based navigation system for blind people to stand in lines. in Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems. 2021.
Murata, M., et al., Smartphone-based localization for blind navigation in building-scale indoor environments. Pervasive and Mobile Computing, 2019. 57: p. 14-32.
Faubert, J. and L. Sidebottom, Perceptual-cognitive training of athletes. Journal of Clinical Sports Psychology, 2012. 6: p. 85-102.
Kuriakose, B., R. Shrestha, and F.E. Sandnes. Smartphone navigation support for blind and visually impaired people-a comprehensive analysis of potentials and opportunities. in Universal Access in Human-Computer Interaction. Applications and Practice: 14th International Conference, UAHCI 2020, Held as Part of the 22nd HCI International Conference, HCII 2020, Copenhagen, Denmark, July 19–24, 2020, Proceedings, Part II. 2020. Springer.
Kuriakose, B., R. Shrestha, and F.E. Sandnes, Tools and technologies for blind and visually impaired navigation support: a review. IETE Technical Review, 2022. 39(1): p. 3-18.
Kelley, S., Access to information: Electronic listening, recording, and reading devices, in Foundations of Vision Rehabilitation Therapy, H. Lee and J. Ottowitz, Editors. 2020, APH Press.
Martiniello, N., et al., Artificial intelligence for students in postsecondary education: a world of opportunity. AI Matters, 2021. 6(3): p. 17-29.
World Medical Association, Declaration of Helsinki: ethical principles for medical research involving human subjects. Journal of the American Medical Association, 2013. 310(20): p. 2191-4.
Goodman, L.A., Snowball Sampling. Ann. Math. Statist., 1961. 32(1): p. 148-170.
Wittich, W., et al., Device abandonment in deafblindness: a scoping review of the intersection of functionality and usability through the International Classification of Functioning, Disability and Health lens. BMJ open, 2021. 11(1): p. e044873.
Mino, N.M., Problem solving in structured discovery cane travel. Journal of Blindness Innovation and Research, 2011. 1(3).
W3C. Web Content Accessibility Guidelines (WCAG) 2.1. 2018 5 June 2018 [cited 2021 June 9]; Available from: https://www.w3.org/TR/WCAG21/.
Lee, H., Keyboarding and Access Technology, in Foundations of Vision Rehabilitation Therapy, H. Lee and J. Ottowitz, Editors. 2020, APH Press: Louisville, KY. p. 313-355.
Deverell, L., et al., Use of technology by orientation and mobility professionals in Australia and Malaysia before COVID-19. Disability and Rehabilitation: Assistive Technology, 2022. 17(3): p. 260-267.
Martiniello, N., et al., Exploring the use of smartphones and tablets among people with visual impairments: Are mainstream devices replacing the use of traditional visual aids? Assistive Technology, 2019: p. 1-12.
Sadato, N., et al., Tactile discrimination activates the visual cortex of the recently blind naive to Braille: a functional magnetic resonance imaging study in humans. Neuroscience Letters, 2004. 359: p. 49-52.
Ptito, M., et al., Brain-machine interfaces to assist the blind. Frontiers in Human Neuroscience, 2021. 15: p. 638887.
Xu, P., et al., Wearable obstacle avoidance electronic travel aids for blind and visually impaired individuals: A systematic review. IEEE Access, 2023.
Mankoff, J., G.R. Hayes, and D. Kasnitz, Disability studies as a source of critical inquiry for the field of assistive technology, in Proceedings of the 12th international ACM SIGACCESS conference on Computers and accessibility. 2010, Association for Computing Machinery: Orlando, Florida, USA. p. 3–10.
Nemargut, J.P., et al. Next Gen Health Powered by The Open Grid And Edge AI [Moderated Panel Discussion]. in IEEE Future Networks World Forum. 2022. Montreal, QC, Canada.

Supplementary Files are not available with this version

Download PDF

Version 1

posted

You are reading this latest preprint version

Exploring the role of artificial intelligence and inclusive technologies during navigation-based tasks for individuals who are blind or who have low vision: Future directions and priorities.

Status:

Version 1

Abstract

Background

Objective

Methods

Results

Conclusions

Figures

Introduction

Methods

Data analysis

Results

Discussion

Conclusion

Declarations

Conflicts of interest

References

Supplementary Files

Status:

Version 1