Exploring Interface Design for Independent Navigation by People with Visual Impairments
Erin L. Brady, Daisuke Sato, Chengxiong Ruan, Hironobu Takagi, Chieko Asakawa · 2015 · ASSETS 2015: Proceedings of the 17th International ACM SIGACCESS Conference on Computers & Accessibility · doi:10.1145/2700648.2811383
Summary
This short poster paper from ASSETS 2015 explores how people with visual impairments respond to different navigation interface designs during in situ walking tasks. The authors observe that prior research has been limited in two ways: user studies of navigation applications are typically constrained by the capabilities of existing localisation technologies, and information needs are usually gathered through interviews rather than real-world trials. To move past these limits, the researchers used a Wizard-of-Oz methodology in which a sighted researcher walked behind participants and manually triggered navigation instructions as the walker passed landmarks or needed to correct their path. This allowed the team to rapidly prototype and compare fourteen distinct interface variants without building a full localisation stack. The variants covered three core navigation components — walking in direct paths, turning, and correcting veering — plus a landmark-based interface, and they systematically varied output modality (speech, non-speech sonification, and haptic vibration), the interval between instructions, the precision of distance or angle information, and the setting (empty corridor, trafficked indoor space, stairwell, lobby, outdoor quad). Nine participants aged 22 to 73 with a range of vision levels (four totally blind, three legally blind, two low vision), including two with hearing impairments and one with cerebral palsy, ranked each interface on a ten-point ease-of-use scale and discussed their experiences in follow-up interviews. The authors used the data to surface open questions for future work on adaptive navigation interfaces.
Key findings
There was no single best interface — participants disagreed strongly on ideal instruction frequency, with some wanting continuous reassurance during a straight path and others preferring silence so they could hear ambient cues. Consensus did emerge at decision points: most participants wanted a fixed instruction 5 or 10 feet before a waypoint regardless of their general preference. Exact numerical precision was rarely useful: eight of nine preferred clockwise directions ("turn to 11 o'clock") or plain verbal cues ("turn slightly left") over instructions given in degrees, which took too long to parse. Long exact distances such as "1400 steps" were described as meaningless. Speech dominated for path navigation, rated more distinguishable and informational than sonification or haptic output; only two participants preferred vibration for veering correction, and both had hearing impairments. Landmark information was valued but could overwhelm users when combined with background noise from construction or crowds. Ease-of-use scores show speech interfaces with frequent updates scoring highest for direct paths (8.17) and veering (8.17), while degree-based turning scored lowest (5.28). Across all interfaces, participants wanted more information than current tools offer, even when rating a specific design poorly.
Relevance
This paper is a useful reminder that "one size fits all" navigation aids for blind and low-vision users are unlikely to succeed — personal preference varies widely and shifts with context (familiar vs. unfamiliar routes, quiet vs. noisy environments, obstacles and seasonal changes). For practitioners building navigation apps or indoor wayfinding systems, the findings argue for configurable instruction frequency, user-selectable output modalities, and plain-language directional cues rather than precise measurements. The study also validates Wizard-of-Oz prototyping as a practical way to test navigation interfaces without investing in localisation hardware first. Limitations include the small sample of nine participants, the university-campus setting, and the poster format's necessarily brief analysis. Still, the work directly foreshadows adaptive navigation research and remains relevant to anyone designing turn-by-turn guidance, orientation and mobility tools, or audio-haptic wearables for visually impaired users.
Tags: visual impairment · blind navigation · non-visual navigation · wayfinding · wizard of oz · interface design · sonification · haptic feedback · speech output · orientation and mobility