Resilience to Disruption: Accessible Navigation for People with Visual Impairment
Trevor Cross, Ishani Pandey, Sophia S Jit, Robert Soden, Priyank Chandra · 2026 · Proceedings of the 2026 CHI Conference on Human Factors in Computing Systems (CHI '26) · doi:10.1145/3772318.3791617
Summary
This paper investigates how people with visual impairments (PwVI) actually navigate urban environments, especially during severe weather events such as snowstorms, high winds, and icy conditions, and how digital navigation technologies do or do not support those practices. The authors, based in Toronto, ran a 3.5-hour remote journey-mapping design workshop with 11 self-identified blind and low-vision adults, recruited through a local blind services organization. Participants were asked to walk through recent and typical trips step-by-step across three phases - pre-trip, in-trip, and post-trip - identifying the tools they used (weather apps, Google Maps, Apple Maps, Voice Vista, Soundscape, Be My Eyes), the cues they relied on (auditory, tactile, spatial memory), the points at which they changed plans or asked for help, and the anticipatory redundancies they prepared. The researchers analyzed audio and video recordings through reflexive inductive thematic analysis, generating four overarching themes: flexible mode-switching, anticipatory planning, the "braiding" of digital tools with Orientation and Mobility (O&M) skills and human assistance, and routine adaptation to disruptions. The paper draws on prior HCI work on wayfinding, crisis informatics, temporality in HCI, and disability studies' concept of crip time, and situates accessible navigation not as a technical routing problem but as relational work that assembles and re-assembles heterogeneous supports to produce a viable "navigation space" in the face of recurrent breakdowns.
Key findings
Four central findings emerge. First, participants continuously switched between walking, public transit, paratransit, taxis, and rideshare depending on weather, construction, fatigue, and time pressure - often chaining modes in a single trip, and prioritizing predictability, safety, and confidence over speed. For example, a snowstorm could convert a normally-walked route into a paratransit or Uber trip, while construction around a destination could force an advance paratransit booking days earlier. Second, anticipatory planning was extensive: participants called ahead to confirm building layouts, placed personal beacons (e.g., a 10-metre offset for a pharmacy entrance), downloaded offline maps, charged spare battery packs, and identified safe indoor waiting areas before leaving home. Third, participants routinely braided digital tools with O&M skills (cane use, guide dog handling, listening for traffic and concierge voices) and human assistance (friends, passersby, Be My Eyes volunteers). Digital tools failed in predictable ways - weather apps lacked wind information, transit apps lacked construction alerts, GPS drifted in "the last few metres" - and O&M plus human help filled the gaps. Fourth, disruptions (snow-blocked bus stops, construction detours, wind that masked traffic cues, sewer-cleaning machine noise) were routine rather than exceptional, and participants absorbed them through "situated repair" - in-situ reconfiguration of tools, routes, and helpers. The authors introduce the distinction between functional interoperability (sharing state across apps, O&M, and people) and temporal interoperability (alignment with external rhythms like paratransit windows and personal pacing).
Relevance
For accessibility practitioners, this paper reframes what accessible navigation software should optimize for. Most turn-by-turn apps privilege shortest-time routing and seamless linear guidance; these findings argue that predictability, safety, and confidence matter more to PwVI, and that beauty, quietness, and low-walking options are legitimate routing criteria. Three design directions follow directly: (1) design for routine breakdown - apps should preserve trip state and intent across mode-switches, allow handoffs between apps, O&M skills, and remote helpers, and integrate construction and real-time weather alerts with the granularity needed to decide whether a sidewalk is passable; (2) design for multiple temporalities - make pauses, detours, and "crip time" first-class rather than treating them as user error, and surface uncertainty in schedule estimates (e.g., paratransit return windows); (3) treat interoperability as a first-class mechanism of resilience, not just a technical nicety. The paper is most relevant for teams building mainstream mapping apps, transit apps, dedicated accessibility tools like Soundscape and Voice Vista, paratransit booking systems, and remote-assistance services. Limitations include a small (N=11) urban Canadian sample, retrospective rather than in-situ methods, no intersectional analysis of race, gender, or socioeconomic status, and no direct evaluation of proposed design changes; future work should test these recommendations in deployed systems and across different transit ecologies.
Tags: blind and low vision · navigation · wayfinding · orientation and mobility · crisis informatics · resilience · journey mapping · assistive technology · crip time