Beyond the Cane: Describing Urban Scenes to Blind People for Mobility Tasks
Karin M. P. Hoogsteen, Florian Mathis, Garreth W. Tigwell, Alastair R. Beresford, Mohamed Khamis · 2022 · ACM Transactions on Accessible Computing · doi:10.1145/3522757
Summary
This research investigates what environmental information is most useful for blind pedestrians navigating urban spaces, moving beyond obstacle-avoidance to understand comprehensive scene description needs. The authors conducted two experiments with 13 cane-using blind participants (5 early blind, 8 late blind) in Glasgow, Scotland. Experiment 1 used a think-aloud protocol during real mobility tasks where participants could ask any questions about their surroundings, generating 337 unique questions. Experiment 2 had participants rate the usefulness of scene descriptions for 18 different objects using various description types (identity, location, appearance, dimension) and annotation formats (precise vs. imprecise, distance, angle, combined). The research addresses a significant gap in Electronic Travel Aid (ETA) development: while most systems focus on obstacle detection and avoidance, blind travelers actually need richer contextual information for independent mobility. The methodology combined real-world mobility tasks with structured rating exercises, providing both qualitative insights into information-seeking behavior and quantitative data on description preferences. The study carefully distinguished between early and late blind participants, finding notable differences in information needs and navigation confidence between these groups.
Key findings
The study established a three-level information hierarchy: low-level (immediate surroundings like objects on the path), intermediate (streets and intersections), and high-level (neighborhood and city context). Across three information types—orientation, navigation, and behavioral—participants prioritized low-level navigation information most heavily. Key quantitative findings include: intersection-related objects (curb cuts, crosswalks, traffic signals) received significantly higher usefulness ratings than potential collision hazards (trees, fire hydrants, poles). Location and identity descriptions were rated more useful than appearance descriptions. Precise location descriptions ("building at 9 o'clock, 10 feet away") were preferred over imprecise ones for orientation tasks, though imprecise descriptions worked for navigation instructions. Late blind participants asked significantly more questions than early blind participants (mean 28.9 vs. 17.8), requested more information overall, and prioritized location information differently. Early blind participants showed greater independence in spatial navigation and prioritized identity information that helps locate things indirectly (like addresses), while late blind participants wanted information to locate objects directly from their perspective.
Relevance
This research has direct implications for designing accessible navigation systems and Electronic Travel Aids. The travel information framework presented—distinguishing orientation, navigation, and behavioral information across three abstraction levels—provides a structured approach for developers building scene description systems. For practitioners, the finding that obstacle avoidance is not the primary concern challenges assumptions underlying many current ETAs. Blind travelers want environmental context that helps build cognitive maps, not just hazard warnings. The differences between early and late blind users highlight the importance of personalizable systems rather than one-size-fits-all solutions. The research also demonstrates methodological value: combining real-world mobility tasks with structured rating exercises yields richer insights than lab studies alone. However, limitations include focus on cane users only (not guide dog users) and point-to-point navigation (not leisure walking or public transit scenarios).
Tags: blindness · mobility · orientation and mobility · navigation · scene description · Electronic Travel Aids · wayfinding · spatial orientation