Experiments with RouteNav, A Wayfinding App for Blind Travelers in a Transit Hub
Peng Ren, Jonathan Lam, Roberto Manduchi, Fatemeh Mirzaei · 2023 · Proceedings of the 25th International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS '23) · doi:10.1145/3597638.3608428
Summary
This paper presents RouteNav, an iOS wayfinding app designed to help blind travelers navigate indoor/outdoor transit hubs without requiring any external infrastructure such as Bluetooth Low Energy (BLE) beacons or Wi-Fi fingerprinting. The system achieves localization by fusing GPS data (when available) with inertial dead reckoning from the phone's accelerometers and gyroscopes, processed through RoNIN—a machine learning algorithm for pedestrian tracking—and combined via particle filtering. A key innovation is the "tile-based navigation" paradigm, where routes are represented as sequences of discrete spatial units called tiles (typically 10-25 meters in size). This approach deliberately works with localization imprecision rather than against it: because tiles are larger than the expected localization error radius, the system avoids giving misleadingly precise directions. Each tile contains descriptive metadata about its type (crosswalk, ramp, tunnel), points of interest, and obstacles. The user interface employs redundant modalities—speech, haptic feedback, and sound/vibration backdrops—to communicate direction, progress, and contextual information. A "switching goalposts" strategy ensures directional guidance remains accurate even when the user's estimated position has significant error. The app was iteratively refined through participant feedback over the course of the study, with features like sound/vibration backdrops, chime alerts before notifications, and contextual tile descriptions all added based on user suggestions.
Key findings
Seven blind participants (ages 59-74, using long canes or dog guides) tested RouteNav across three challenging routes at the Palo Alto Transit Center, totaling up to 495 meters and involving underground tunnels, crosswalks, ramps, traffic islands, and fence openings. All participants completed all routes except for two localization/tracking failures. Median traversal times ranged from 9 to 20 minutes depending on route complexity. However, experimenter intervention was needed at various "challenging spots" along each route—particularly at tunnel entrances, open spaces with poor GPS, and transitions between parallel paths like walkways and ramps. The tile navigation concept received mixed but generally positive feedback; participants valued knowing their progress along the route. The sound/vibration backdrop indicating correct vs. wrong direction was one of the most appreciated features. A "user as a sensor" mechanism—asking users to confirm whether they had entered a tunnel based on their own auditory perception—proved very effective for resolving ambiguous localization near tunnel entrances. The system's infrastructure-free approach achieved localization accurate enough for practical navigation in most situations, though open spaces and underground passages remained challenging.
Relevance
RouteNav addresses a critical gap in accessible transit: the last-meter navigation problem within transit hubs where mainstream GPS-based apps fail. The infrastructure-free approach is particularly significant because it removes the dependency on transit authorities installing and maintaining beacons—a barrier that has stalled many indoor navigation projects. For accessibility practitioners, the tile-based navigation paradigm offers a transferable design pattern for handling localization uncertainty gracefully rather than pretending it does not exist. The iterative, participant-driven design process demonstrated how blind travelers actively integrate app guidance with their own environmental perception, challenging the assumption that users are passive followers of navigation instructions. The finding that redundant modalities (speech plus haptics plus sound) significantly improved the experience reinforces multimodal design principles for assistive technology.
Tags: wayfinding · indoor navigation · blind and low vision · transit accessibility · mobile application · localization · dead reckoning · particle filtering
Standards referenced: ITU-T F.921