An accessibility mode which makes NCR's Kiosks more accessible for wheelchair users

Problem Statement
Increasing accessibility for NCR Kiosks for wheelchair users

Solution Statement
Our solution aims to augment the uNav interface by creating an accessibility mode which changes the layout to make it more accessible using uNav control panel. The uNav Compatibility Mode targets user’s pain points on a community level for users who are using the uNav interface.

I conducted user experience research to finalize the design direction and performed usability tests on the prototype.

Kevin Key, Varnit Jain, Benton Humphreys, Yunfei Wang

Aug 2019 - Nov 2019


1. Market Research

Analyzing the current NCR kiosk products for accessibility features reveals many intentional design choices. For NCR SelfServe checkout hardware, wheelchair height and reach have been considered for easier accessibility. These checkout hardware products also include UNav keypads, integrated headphone jacks, volume control, and tactile keys to assist users with physical, hearing, and visual impairments. Specific accessibility considerations have been implemented for NCR ATMs, such as universal tactile features for ATM function keys and a touchscreen designed for users with visual impairment. NCR ATMs also include interactive teller features, which allow users to communicate with a remote teller for assistance.

NCR has approached accessibility design with the inclusion of uNav, and it is important to understand how this fits into the context of the kiosk manufacturing market. Out of NCR, Rockwell Collins , Toshiba, Fujitsu, IER SAS, and SITA, only NCR and IER SAS include physical hardware button controls as part of their accessibility design solution. All other manufacturers mentioned either do not include accessibility design or it is implemented within the software, requiring users to interact with a touch screen. Since many manufacturers allow for custom kiosk software often times accessibility features and design may be overlooked entirely, especially when no physical buttons are implemented in the kiosk.

Though uNav has been found to improve kiosk interaction, the lack of uniform accessibility design between kiosk manufacturers can be and has been problematic for users. It has been suggested that kiosk manufacturers work together on kiosk accessibility to mitigate these issues. This is an interesting proposition considering that NCR’s patented uNav can be bought as a standalone unit to be implemented in other brands of kiosks.

2. Literature Review

There is an ever-growing field of research to suggest best practices for accessibility within digital design, yet there seems to be a lesser focus on integrating accessibility into kiosks. We have sourced, reviewed, and synthesized several prior research papers that either touch on kiosk accessibility or an applicable field (e.g. web design).

Some of the key takeaways from literature review are:
1. While we will not be focusing on the idea of mechanical adjustments, but the idea that a system can recognize and adapt to its user is something to be considered.
2. Many models exist that can be good reference points if we need to do an overall accessibility test of kiosks, or understanding how other kiosks have improved physical accessibility.
3. A good contextual enquiry can be very insightful - they are not only able to correct the user flow but are also able to add additional features.

3. Task Analysis

‍As a starting point, we decided to first complete a task analysis of the XK32 kiosk. This allowed us to familiarize ourselves with the overall feel of the system, understand interactions from a high level, and prompted us to understand the limitations of the system for those with physical disabilities. The task analysis begins from the moment that a user approaches the kiosk with the intent to order a meal and ends with the completion of the interaction, when they pay for and receive a receipt for their order. Our contact at NCR was unable to provide wireframes to us due to confidentiality concerns, but were able to provide screen grabs of every screen.

It is important to note that the specific information displayed on the kiosk depends entirely on the context in which it is used. For instance, if the kiosk is located in a certain restaurant it will only display information pertinent to that restaurant. If it is located in a food court, however, it will display options for several restaurants. Therefore, the task analysis we created represents a general overview of the interactions a user will experience, separated by what is experienced in a food court-type of context from a dedicated-restaurant context.

Infromation Goal
We conducted a task analysis to primarily learn about the user flow and the decision-making process of a user as they interact with the XK32 kiosk. The ordering process on the XK32 kiosk can involve many steps from start to completion, so it was necessary to understand how these steps connected to one another and how a user typically navigates from the start screen to order completion, and follows them during the entire ordering process. By understanding how the system currently operates, we were able to form a baseline on which we could develop further research efforts.

4. Observation

‍In order to better understand how users interact with the current system, our team decided to observe users interacting with the kiosk during lunch hour at NCR. Though our target user group consists of those with disabilities, we found it important to observe all users who might interact with our system. Any design that targets those with physical disabilities will also affect how others interact with the system.

Infromation Goal

Our goal with observation was mostly exploratory and built off of our task analysis. The task analysis we completed served as the base to form this observation study. All though the task analysis informed us of the overall flow a user must take to complete a transaction, it did not highlight any pain points or inform any subjective feelings the users might have about the system. We wanted to determine if users experienced difficulty with any point of the ordering process due to usability issues.
Additionally, we aimed to explore baseline data on the learnability of the kiosk interface. In short, we wanted to know how a novice user interacts with the system compared to a user that has been familiar with the system for some time. We felt this was an important step in deciding how users of different experience levels interacted with the system, and we rationalized that we could use this information to inform design decisions in subsequent sections.

4. Semi-Structure Interview

‍Our team decided to conduct a semi-structured one-on-one interview as our first research method involving a physically disabled user within the problem space. A one-on-one interview was logistically easier to setup than a focus group and allowed for a very personal conversation compared to other research methods. Using a semi-structured interview allowed our team to answer specific questions we had, but still explore the problem space by having an open discussion. We held our interview in the Usability Lab in Georgia Tech and it lasted about an hour.

Infromation Goal

Our main information goal from this research method was to gain insight into our user group’s wants and needs regarding digital technology accessibility. After conducting our observation it was clear that the kiosk has general usability issues, but we needed to focus on select issues affecting physically disabled users. Another goal was to understand the limitations of wheelchair users during point of sale transactions due to their seated position. Our understanding prior to the interview was that not all kiosks were designed with wheelchair height in mind for usability. Beyond this, we wanted to determine specific pain points that our interviewee had with kiosk accessibility. We also wanted to know what public technologies they currently found helpful, whether or not it was implemented in point of sale kiosks. We also went into the interview with the idea that we could potentially narrow down our user group if given the opportunity.

5. Survey

‍We chose an online survey as a research method to collect data largely because of its ability to collect data from a substantial set of users in a short time. Our preference was to do an in depth cognitive walkthrough of NCR’s system, but due to the lack of sourced participants this was unable to be completed before the research needed to be completed. Unfortunately, NCR was not able to fund in person interviews which provides less incentive for our user group. Also, we hoped to have our IRB protocol approved in time to source participants from Georgia Tech labs, but this was not the case. Ultimately, we used an online survey as an alternative due to lack of participants willing to meet in person.

Infromation Goal

A main information goal of using a survey was to learn what users with disabilities like and dislike about point of sale kiosks. Another goal was to learn how users’ disabilities hinder specific interactions with a kiosk. Based on our previous research methods we decided another goal was to learn about common frustrations at each stage of the process of completing a checkout. A secondary information goal was to refine our user group by recognizing common problem areas and patterns. Another secondary information goal was to determine if certain users prefer tactile interaction versus touch screen and for completing what tasks.


Our research methods provided valuable insight into who our user group is, what common points of frustration and difficulty they with interactive kiosks (and more generally, digital technology), and, perhaps most importantly, resulted in an informed design direction for the next part of this project. After every research method was completed our team met to discuss the findings and how we would want to pursue further research methods.


From the task analysis, observation study, interview, and survey, we had a general idea of where users (both with and without physical impairments) we're likely to experience problems. A few of the most straight-forward problems, and likely the most easily remedied, were:

  1. Users did not notice that alternate methods of interacting with the kiosk were available.
    • Nearly all of the users from the observation did not notice that the uNav existed at all. This is likely the result of the physical device being out of the line of sight of the user as they approach the kiosk
    • The initial welcome screen indicates that the user should “Tap anywhere to begin,” but does not mention the uNav system.
    • Though many users will never have any reason, or interest, in using the uNav, the lack of recognition of the feature might dissuade those with physical disabilities from approaching the kiosk in the first place.

  2. Users did not notice that there were more options to certain menu items due to the scroll bar being too light in contrast compared to the background.
    • Another takeaway from the observation study revealed that users would skip over certain menu options if they weren’t displayed on the screen immediately. Many items weren’t immediately visible, and instead were further down the page and had to be scrolled to.
    • They very light grey of the scroll bar likely went unnoticed when paired with the white background.

  3. Certain areas of the screen are challenging for users with physical impairments to reach or see from where they are seated.
    • According to information gathered from our survey, a common problem that users faced was reaching areas of the screen.
    • Further, from their seated position, many were unable to effectively see the upper areas of the screen to know what information was being displayed.

  4. Users became frustrated with the slow movement and interaction controls of the uNav device.
    • Users noted that they felt the uNav was cumbersome to use - both in how slowly it tracked across the screen and how some interactions required significant dexterity to successfully perform. For instance, in order to scroll down the screen, the user must place the cursor on the scroll bar and hold down two buttons at the same time (the enter button and up or down) to facilitate scrolling.

  5. Users were often unaware of how to add an item to cart before checkout.
    • Many times, users would go to check out only to find that their cart was empty. They would then be presented with the option to tap “x” to return to their order, but many were unsure if that would cancel their order outright.

  6. Other insights were less direct in their relevance to the XK32 kiosk, but were of enough significance to be considered for future design directions. Namely:

  7. Users want to remain independent, and not made to feel “special.”
    • This was explicitly stated during the interview and was an underlying subject from many survey respondents. In developing design solutions, it will be important to address how an accessibility feature will be perceived and, if possible, how it might be used universally.

  8. Feedback is important.
    • Many users with physical impairments also suffer from lack of feeling difficult in the movement of their extremities. As such, means of feedback other than visual will be considered.


‍Our group started this phase of the project by brainstorming different design concepts with users’ limitations and needs and design implications we analyzed from previous section in mind.
We strove to incorporate design implication from previous research into the redesign of the kiosk ordering system. We did this through an informed brainstorm session where we generated ideas relevant to our paint points and design implications. In doing this we realized we had three distinct sets of ideas that fit into the following groups:

  • Universal design (applicable to users beyond our user group)
  • Relevant only to our user group
  • Personalized/customized to the user

Using this we created “How Might We” (HMW) questions to represent each one of these groups. This guided the rest of our session to generate more ideas relevant in these groups. The following are the HMW questions with the corresponding ideas:

We finalized various features and ideas from brainstorm session into three abstract design concepts that covers different ranges of potential users of kiosks mentioned above. Sketches and descriptions were created for each concepts for assessment by experts and users for feedback.

Concept 1: Custom Profile

‍This concept utilizes the existing XK32 interface, but allows users to create a custom profile based on their individual needs and preferences. This concept targets users’ pain points on an individual level and, as such, can most effectively mitigate frustrations compared to other more generalized approaches.

  • Extremely customizable experience that appeals to individual needs and preferences
  • Quick setup that can be accomplished on personal device or kiosk interface
  • Maintains current GUI for those already familiar with the system

  • Weaknesses
  • Requires several entirely new interface “filters”
  • The requirement for profile creation might dissuade some users from using it

Concept 2: uNav Compatibility Mode

‍This system aims to augment the uNav interface by creating an accessibility mode which changes the layout to make it more accessible using uNav. The uNav Compatibility Mode targets user’s pain points on a community level for users who are using the uNav interface.

  • No need for any new hardware additions
  • Faster Interaction through uNav shortcuts
  • Appeals to our target user group’s preference for tangible interfaces

  • Weaknesses
  • A bit of a learning curve because of a new interface
  • Designers will have to create two sets of UI for all restaurants
  • Pasting instructions affects the aesthetic of the system

Concept 3: Feedback

‍This concept provides users with instant feedback for certain actions to keep users updated throughout the ordering process and tool tips option for troubleshooting. This concept takes a universal approach to target general user's pain point of losing confidence about their action and confusion towards the system.

  • Only minor design changes added
  • Doesn’t interrupt current flow
  • Cover both our target users and general kiosk users

  • Weaknesses
  • No features that solve the problem of accessibility
  • The popups can be disruptive for users


‍The first feedback session was designed for the perspective of an expert in the field of assistive technology. The set of questions from the first feedback session are intended for accessibility experts to review and identify obvious issues of our design concepts based on their experience with working with people with accessibility needs. We planned on leveraging the feedback we obtained from first feedback round to help make decision on which concept and features to take forward.
From there, the final design concept can be developed into a detailed wireframe that can be tested by user tasks. The second round of feedback session was done with our stakeholder, NCR, in order to work towards our ultimate goal of achieving better accessibility of the NCR kiosk for wheelchair users. This round was designed to understand users’ performance and opinions towards the system and for us to resolve issues discovered early so the later prototyping phase can move more smoothly.

Feedback Session 1 Design - Sketches

After fully defining our concepts and developing sketches to represent them, we began developing our first feedback session. At this point we wanted to present our concepts at a relatively low fidelity to allow for feedback with original and creative ideas for improvement. The goal of this feedback session was to provide usability experts with a broad understanding of the three concepts we created in order to obtain feedback on the idea as a whole, and for certain features within the idea. For each concept we wanted to understand the following:

  • The value each feature adds to the overall design
  • Preferences for specific features within the concept and why
  • Potential changes or improvements that may strengthen the concept

Our initial plan was to conduct this first session in person with our user group. Due to limited access to our user group, we had to divert to accessibility design experts. Sourcing participants to meet in person has been a continuous issue for our team, as well as for NCR who was only able to find two wheelchair users in Atlanta through a paid sourcing company. We decided that accessibility design experts were the next best option because their experience and feedback would be relevant and valuable while our designs were still at concept level.

Due to timing and professors’ schedules, we decided to use qualtrics to gather feedback online for this session. We had success gathering concept feedback this way in our HCI foundations course, and were able to design this session in a similar fashion.
We structured the session the same way we would have for an in-person session. We planned for this session to take approximately 45 minutes to allow for thorough and well thought out feedback. Our session breakdown was as follows:

  • Describe our project goals, explain prior research and findings done to this point, and provide context (15 min)
  • Present the participant each concept individually (5 min per concept)
    • Provide low to mid fidelity sketches of the design
    • Explain and highlight important concept features for the design
    • Give 1 or 2 example scenarios explaining how and when the concept would be used
  • Present users with relevant questions for feedback (5 min per concept)

Concept Refinement
After analyzing our expert feedback results and speaking with NCR, we decided to pursue concept 2 (uNav Compatibility Mode) as a basis, and incorporate features from other concepts. From a user’s perspective, concept 2 most significantly addresses the accessibility issues presented during the initial observation. From a stakeholder’s perspective, the proposed concept is a companion to the already existing uNav control, and might be considered as an evolution of this system.

Compared to the other concepts, concept 2 presented significant advantages like:

  • Hardware solutions are difficult to implement, concept 1 focused heavily on a new hardware interface in the form of an NFC Chip and a backend database management system which stores all user profiles. Concept 2 does not require any additional hardware requirements and uses only user interface changes to augment the present interactions.
  • Through our initial requirements gathering sessions, we realized that users preferred tangible interactions. Concept 1 and Concept 3 did not account for tangible interactions but relied on touch screen for user input. Concept 2 leveraged the existing accessibility system - uNav to provide richer and faster interactions.
  • Concept 1 targeted users on an individual level and concept 3 was for a universal population. Concept 2 targeted out audience (wheelchair users) directly and appealed to their need of an interface designed taking with respect to their needs. Concept 2 also directly helps solve the reach and dexterity issues through uNav, improving overall accessibility.
  • Concept 2 is also compatible with the next generation of Kiosks by NCR. The next generation of kiosks have a handheld uNav which can be taken out of the Kiosks and can be used at a distance. This idea will solve the major limitation of uNav - limited viewing angle. Once the user is at a distance, a greater viewing angle with give the user access to the whole screen.

Feedback Session 2 Design - Wireframes

Following the feedback from the concepts, we chose to take a deeper look at the uNav compatibility feature by generating mid-fidelity wireframes. These wireframes were constructed with the intent to inform a user of how they might navigate through the ordering process from a high level, what the information structure would look like (e.g. information hierarchy and element size and placement), and how certain features might influence their experience (e.g. tab indexing and highlighting selected elements). From this session we wanted to understand the following:

  • Relevant opinions towards the overall design
  • Preferences for specific features within the concept and justification
  • Potential changes or improvements that can be considered for the system

Similar to the feedback session for sketches, we wanted to hold this feedback session in person with our user group. Again, this was not possible due to the access and time constraints of this assignment. Our next best option was to receive stakeholder feedback from NCR. Their feedback at this stage in the design would be helpful to understand the feasibility of the concept and how well it may fit into the existing system. From this we could develop relevant design changes and suggestions to make the concept with better within their current framework. They may also have feedback on similar features or elements they have tried in the past and why they did or did not work. Beyond this, having mid-fidelity wireframes would allow NCR to understand and visualize ideas we had previously discussed.

We designed this session to be conducted in person at NCR’s HQ with one or two NCR employees. Our meetings often fluctuate between one and two employees, so we wanted to prepare for both situations. We planned on having access to a meeting room with a large TV. This would allow us to project our wireframes to a large size, similar to the size they would be on an XK32 kiosk. We designed the session to take approximately 45 minutes with the following breakdown:

  • Describe the previous three concepts and explain why we chose the specific concept we did. Also explain how the concept has evolved since the initial idea (15 min)
  • Go through each wireframe and explain the interactions between each frame (15 min)
    • Cover relevant features
    • Get feedback on interactions
    • Take note on hesitations and confusion
  • Provide users with relevant questions for feedback (15 min)

Prototype Design
Based on our feedback sessions on the sketch concepts and wireframes, we can make certain improvements to the current design shown in section 06. Regarding accessibility considerations, the entire concept is designed to address usability concerns that those with physical disabilities might face when using the kiosk. For this reason, all of the usability refinements mentioned below are designed with accessibility in mind.

  • Information hierarchy by
    • Including adequate space between interactive elements
    • Separating elements on screen by “interactive” and “non-interactive” elements, so as to keep any touch point within a certain area
    • Presenting information (e.g. food items) in categories, so as to minimize the amount of information present at any moment
  • Input feedback by
    • Highlighting the element the user has currently selected
    • Visually distinguish between interactive and non-interactive elements
    • Increase contrast on certain existing elements (e.g. scroll bar) to improve recognizability
  • Navigation by
    • Arranging elements in a grid-pattern to promote navigating through interactive elements by tab indexing
    • Incorporating “placeholder” elements that facilitate indexing, but are non-interactive, to allow a user to more easily traverse the screen
    • Refining the navigation bar as a “tree” that better filters information
    • Displaying the current, previous, and next steps that a user will take to complete their order, to better
  • Predictability by
    • Displaying current, previous, and next steps in the navigation bar that will communicate where a user is in the ordering process, and how many more steps they need to take to complete their order


We are currently testing the prototype. Results will be out soon.