Exploring Aural and Haptic Feedback for Visually Impaired People on a Track: A Wizard of Oz Study
Kyle Rector, Rachel Bartlett, Sean Mullan · 2018 · Proceedings of the 20th International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS 2018) · doi:10.1145/3234695.3236345
Summary
This University of Iowa study investigates how to make outdoor jogging tracks accessible to people who are blind or have low vision, addressing a gap in exercise accessibility research. Standard 400-meter tracks rely entirely on visual lane markings, and the overwhelming majority are inaccessible to visually impaired people. The researchers conducted an ecologically valid Wizard of Oz study at real outdoor tracks comparing four feedback conditions for lane-keeping: a human guide (control), verbal commands via bone-conduction headphones (e.g., "Correct left," "Right 45"), wrist vibration via smartwatches on each wrist (vibration patterns on the side the person is veering toward), and head beat feedback via bone-conduction headphones (heartbeat sounds in the ear corresponding to the veering direction, with tempo indicating severity). The study involved 14 participants (8 blind, 5 low vision, 1 with no central vision) aged 24-72, who walked approximately one mile each across the four conditions on real tracks at schools and universities. The Wizard of Oz approach had the experimenter walking behind participants, observing via video, and triggering feedback remotely through a custom Android app that communicated with the participant's belt-mounted phone, which relayed signals to headphones or smartwatches. Feedback was designed to complement rather than replace existing navigation strategies — bone-conduction headphones kept ears open for environmental awareness, and the metaphor of "bumping into a wall" was used to frame corrective feedback.
Key findings
The technology conditions did not significantly affect accuracy — participants stayed in their lane and made forward progress at similar rates across verbal, wrist vibration, and head beat conditions. However, the human guide was significantly better than all three technology conditions for time elapsed and forward progress. The clear preference order was: human guide first, then verbal, wrist vibration, and head beat last. This order was consistent across rank, comfort, and use-again ratings. Several critical real-world factors emerged that would not have been apparent in laboratory studies. Cane users (the majority of participants) reported that holding a cane interfered with perceiving wrist vibrations — the constant tactile input from the cane running along the track surface competed with the watch vibrations. Seven participants suggested stronger vibrations, and the outdoor track texture (polyurethane pebbles) made this worse than it would be on smooth indoor surfaces. Wind and loud ambient sounds (physical education classes, football pre-game music) degraded the head beat and verbal conditions, with participants noting that low-frequency head beat sounds were particularly vulnerable to being masked. Wearing two smartwatches raised social acceptability concerns. People with low vision had higher accuracy than blind participants across all conditions, presumably because they could use residual vision to detect track lines. The verbal condition was perceived as most personal and GPS-like, but participants found the verbal statements too long and worried that conversation with others would cause them to miss cues.
Relevance
This study makes important contributions to both exercise accessibility and real-world navigation feedback design. For accessibility practitioners, the most significant finding is the gap between laboratory navigation research and real-world outdoor settings — cane interference with haptic feedback, wind degradation of audio cues, and social acceptability of wearing multiple devices are problems that only surface in ecologically valid studies. The recommendation that vibrations should be "pronounced or only on one wrist" so users can still hold their cane is directly actionable for wearable designers. The study also demonstrates that exercise space accessibility is an underserved area — while much navigation research focuses on getting from point A to B, the task of staying within a lane while exercising has different requirements (hands-free, compatible with existing mobility aids, works with ambient noise). The finding that human guides remain superior but are difficult to recruit highlights an opportunity for remote human guidance services, where trained guides could provide verbal navigation via video call, combining the reliability of human feedback with broader availability.
Tags: blindness · low vision · exercise accessibility · navigation · haptic technology · audio feedback · vibration feedback · Wizard of Oz · physical activity