Embodied Exploration: Facilitating Remote Accessibility Assessment for Wheelchair Users with Virtual Reality
Siyou Pei, Alexander Chen, Chen Chen, Franklin Mingzhe Li, Megan Fozzard, Hao-Yun Chi, Nadir Weibel, Patrick Carrington, Yang Zhang · 2023 · Proceedings of the 25th International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS) · doi:10.1145/3597638.3608410
Summary
This paper presents Embodied Exploration, a Virtual Reality system that allows wheelchair users to remotely assess the accessibility of unfamiliar physical environments by exploring high-fidelity digital replicas while embodied by personalized avatars. The system addresses a critical gap: wheelchair users need detailed, individualized accessibility information before visiting new places, but current methods (phone calls, photos, videos, 360-degree virtual tours, ADA labels) fail to provide the granularity needed. A photo cannot reveal whether legroom under a table is sufficient; an ADA label saying a room is "accessible" may not account for a wider power wheelchair. The researchers conducted four iterations of user-centered design with wheelchair users, beginning with a YouTube content analysis of wheelchair accessibility challenges, followed by needs-finding interviews, co-design sessions with a wheelchair user who is also a VR content creator, and pilot testing. This process identified three key task categories for accessibility assessment: visibility (seeing things from a seated position), locomotion (navigating spaces in a wheelchair), and manipulation (reaching and operating objects). The final system uses Matterport 3D scans recreated as true-to-scale digital replicas in SketchUp, with personalized avatars generated via Meta Quest 2 that encode three critical biometric parameters: wheelchair maximum width, armrest height, and seated eye height. Interaction techniques include first-person head rotation for visibility, manual wheel-rolling or joystick locomotion, and direct hand manipulation of virtual objects.
Key findings
A user study with six wheelchair users (ages 28-55, from five US states, 2-26 years wheelchair experience) compared Embodied Exploration (EE) against Photo Gallery (PG) and Virtual Tour (VT) baselines across 13 tasks in four environments (office, lab, house, apartment). EE achieved the highest average confidence score (6.68/7, SD=0.67) compared to PG (5.95, SD=1.48) and VT (6.35, SD=1.01). All participants preferred EE for manipulation tasks, and most preferred it for visibility and locomotion tasks. Participants described EE as "the most truthful replica of real-life experience" and noted they could assess accessibility "instantly without thinking." Key advantages emerged: for locomotion, participants could feel their virtual wheelchair scraping against furniture and check whether armrests would clear under tables; for manipulation, they could verify whether countertops were too high or objects too far to reach; for visibility, the correct seated eye height revealed what could actually be seen from their position. Photos and virtual tours consistently led to inaccurate depth perception and false confidence. P2 noted "I am confident that my legs would not fit under the table" when his virtual armrest visibly collided with the table edge. However, motion sickness during manual wheelchair locomotion was a significant challenge, with participants preferring the powered wheelchair joystick mode despite five of six being manual wheelchair users in daily life. The learning curve for in-air manual wheel-rolling gestures was high, and lack of force feedback reduced realism.
Relevance
This research opens a promising new direction for physical accessibility by demonstrating that VR embodiment can deliver assessment fidelity approaching a physical visit at a fraction of the effort. For the approximately 75 million wheelchair users worldwide, advance accessibility assessment is not a convenience but a necessity to avoid dangerous, frustrating, or impossible situations. The personalization aspect is particularly significant: accessibility is not binary, and what works for one wheelchair user may not work for another with a different chair width, armrest height, or range of motion. Current approaches like ADA labels and Airbnb accessibility tags collapse this individual variation into coarse categories that users reported finding unreliable. For practitioners in hospitality, real estate, education, and public spaces, the implication is clear: providing 3D-scannable environments could give wheelchair users agency to assess accessibility on their own terms. The practical barriers remain significant — creating digital replicas currently requires Matterport scanning and manual SketchUp post-processing — but the researchers note that automated 3D reconstruction from smartphones is advancing rapidly. The design guidelines around safety (avoiding movements that could cause real injury), embodiment granularity (different tasks need different avatar detail), and practicality (usability sometimes trumps realism) are valuable for any VR accessibility application.
Tags: wheelchair accessibility · virtual reality · embodied interaction · accessibility assessment · digital twin · 3D scanning · user-centered design · mobility disability · remote assessment
Standards referenced: ADA