Supporting the Design and Selection of Accessible Interaction Techniques for Virtual Reality

Over the last 10 years, virtual reality (VR) has become a popular consumer technology because of its diminishing hardware size and increasing affordability. Yet, like the accessibility of past consumer technologies, the accessibility of VR has not been a priority, and as a result, people with limited mobility have been excluded from adopting VR. One of the benefits of VR is the numerous ways of viewing and travelling in virtual environments with a variety of gestures, input types, and parts of the body. Over the last 30 years, hundreds of scene-viewing and locomotion techniques have been developed, some of which could be accessible given a user’s individual abilities, and more techniques continue to be developed to this day. Despite the potential for improving accessibility that the scene-viewing and locomotion design spaces afford, there is no guidance for designers to select or design accessible scene-viewing or locomotion techniques.

To address this gap, I developed two design frameworks that facilitate the classification and invention of scene-viewing techniques that do not require head movement. I also demonstrated how a novel scene-viewing technique could potentially improve the accessibility of VR for people with color vision deficiencies. To support researchers and designers in evaluating VR interaction techniques, I created an extensible and customizable testbed in Unity. I then used the testbed to evaluate the accessibility of six locomotion techniques with participants with upper-body impairments. My proposed work will compare the performance and perceptions of people with and without upper-body impairments when they use six locomotion techniques. I will also investigate whether there are any differences in how people with and without upper-body impairments interact with locomotion techniques and encode these differences in a set of locomotion metrics. Finally, I will create and evaluate a machine learning model that predicts a new VR user’s performance with a locomotion technique as a step towards developing a recommender system that personalizes VR interaction based a user’s individual abilities. The thesis this work will demonstrate is: The accessibility of VR can be improved by the design and selection of scene-viewing and locomotion techniques using design frameworks, testbed evaluations, and movement analysis and modeling.

Supervisory Committee
Chair: Jacob O. Wobbrock, Professor, The Information School
GSR: James Fogarty, Professor, Computer Science & Engineering
Member: Hrvoje Benko, Director of Research Science, Meta
Member: Leah Findlater, Associate Professor, Human Centered Design & Engineering
Member: Anind Dey, Dean and Professor, The Information School