Enhancing Accessibility for Future Transportation
Defining a solution to make autonomous vehicles more accessible for people with visual impairments
Above: This diagram depicts the high level themes and components of a successful trip with a focus on audio interaction.
Sponsor: Traffic21, Heinz College and Carnegie Mellon University
Role: Primary Investigator, Design Lead
Deliverables: Novel research method development and implementation, primary research, conversational user interface design
Project Time Frame: December, 2017 through July, 2018
Above: The core aspects of a successful trip that directly contribute to personal agency and often are presented to our target population through audio.
BACKGROUND
This project was sponsored by the Traffic21 Initiative, an interdisciplinary research and development initiative focused on exploring and defining the future of transportation in the Pittsburgh, PA region and beyond.
In early 2017, the Disability Rights Education & Defense Fund, National Center for Aging and Disability, and Independent Living Research Utilization, met to outline a set of focus areas that they believe the automotive industry needs to focus on, mainly: designing vehicles and hardware with adaptive uses in mind, and creating accessible user interfaces.
The primary goal of this project was to increase personal agency and autonomy for persons with visual impairments through an improved transportation experience. This required developing a greater understanding and empathy for this population and acquiring a deeper knowledge of the context around the technology and systems in development already in order to inform our design work.
Leveraging both secondary and primary research techniques, we developed an in-car conversational interface to provide passengers with critical information and system controls to increase access and practicality for passengers with visual impairments.
KEY RESEARCH FINDINGS
Through our research, we discovered four key themes that provide constraints and requirements for an in-vehicle solution. These themes were woven throughout our review of prior work in transportation, mobility, and visually impaired research as well as our primary research, interviewing and observing people with visual impairments
1. Safety
A clear and direct means of escalation needs to be readily available in the event of system error or failure or in the event of other emergencies. The availability of such a mechanism needs to take into account the preferred means of interacting with a system for this population, meaning audio controls and interfaces.
2. Trust
The interface needs to promote trust through information presentation and transparency as well as provide clear controls for users to access different features.
3. Efficiency
While often transportation efficiency means getting from point A to point B as quickly as possible, other considerations are involved when it comes to daily transportation needs—considerations like preferences, routes, drop off locations, and transparency in information presentation all contribute to a make a ride both efficient in terms of time as well as effort expended by passengers.
4. Relationships
Transportation for some in this community is a regular source of delightful (and sometimes not delightful) interaction. Any system that removes the opportunity to interact with a human driver needs to make up for the loss with a better means of getting around.
Above: The foundational themes that come from having the basic needs met during a successful trip.
Above: A mapping of a greeting dialog module with example utterances.
Above: A mapping of a trip conclusion dialog module with plans to include detailed instructions for passengers to navigate the path from the vehicle to their desired destination.
Above: To document the approach we took to creating the script for the conversational user interface, we created a spreadsheet that maps utterance type, examples, and other specifications.
solution
We’ve design a conversational user interface to be used within a private vehicle in a shared autonomous fleet.