Hacking Blind Navigation
João Guerreiro, Hernisa Kacorri, Jeffrey P. Bigham, Edward Cutrell, Daisuke Sato, Dragan Ahmetovic, Chieko Asakawa · 2019 · CHI Conference on Human Factors in Computing Systems Extended Abstracts · doi:10.1145/3290607.3299015
Summary
This workshop summary describes a one-day CHI 2019 workshop designed to bring together researchers from multiple disciplines working on assistive navigation technologies for blind people. The organizers — experts from Carnegie Mellon University, University of Maryland, Microsoft Research, IBM Research Tokyo, and University of Turin — identified that blind navigation research is fragmented across communities including HCI, accessible computing, cognitive sciences, computer vision, and ubiquitous computing. The paper surveys the current landscape of blind navigation research, covering in-situ turn-by-turn navigation systems, environmental information tools like Microsoft Soundscape and Footnotes (which provide functional, visual, historical, and social descriptions of surroundings), computer vision for detecting obstacles and points of interest, tactile map exploration, virtual navigation for gaining prior knowledge of environments through audio-haptic and 3D audio interfaces, and the cognitive science foundations of how blind users build spatial mental representations. The workshop structure included an interactive panel on open challenges, minute-madness presentations of position papers, and a hands-on hacking session where participants designed or prototyped new navigation solutions.
Key findings
The workshop identified several key open challenges in blind navigation research: sharing large-scale data across research communities while respecting privacy; easing the transfer from research prototypes to real-world deployable products; exploring non-visual augmented and virtual reality for helping blind users gain environmental knowledge; establishing common metrics and benchmarks for comparing navigation systems across studies; and the need for adaptive interfaces that respond to different users and situations rather than presenting static navigation instructions. The organizers noted that most current solutions fail to adapt to individual users or changing contexts, and that computer vision — while increasingly powerful for detecting crosswalks, obstacles, and objects — still faces challenges in real-time deployment on mobile and wearable devices. The workshop also highlighted the importance of cognitive science contributions, particularly research on spatial knowledge acquisition and mental representation of environments, which should inform the design of navigation interfaces. Crowdsourcing was identified as a promising approach for installing and maintaining indoor localization infrastructure at scale.
Relevance
This workshop summary serves as a valuable snapshot of the blind navigation research landscape as of 2019, identifying the key research threads, open problems, and disciplinary gaps. For accessibility practitioners, it highlights that effective blind navigation requires more than just turn-by-turn directions — users need rich environmental information, prior knowledge of spaces, and adaptive interfaces that account for individual differences in navigation strategies and spatial cognition. The emphasis on interdisciplinary collaboration underscores that no single field can solve blind navigation alone: computer vision researchers need interface design expertise, cognitive scientists need engineering partners, and all need input from blind users themselves. The identified challenge of transferring research to real-world deployment remains highly relevant, as many promising navigation prototypes never reach the people who need them. The workshop also points toward the growing role of non-visual AR/VR as a tool for pre-trip exploration and spatial learning.
Tags: blind navigation · orientation and mobility · assistive technology · visual impairment · computer vision · indoor localization · virtual navigation · workshop