Headlock: A Wearable Navigation Aid that Helps Blind Cane Users Traverse Large Open Spaces
Alexander Fiannaca, Ilias Apostolopoulous, Eelke Folmer · 2014 · ASSETS '14: Proceedings of the 16th International ACM SIGACCESS Conference on Computers & Accessibility · doi:10.1145/2661334.2661453
Summary
This paper presents HEADLOCK, a navigation aid for optical head-mounted displays (OHMDs) like Google Glass that helps blind cane users traverse large open spaces—environments that are particularly challenging because they lack the tactile features (walls, corridors, floor transitions) that cane users rely on for orientation. Without these features, blind users tend to veer from their intended path. HEADLOCK works in two phases: a discovery mode where the user scans the space by moving their head and the system uses computer vision (OpenCV blob detection) to detect salient landmarks such as doorways, and a guidance mode where audio feedback directs the user toward the selected landmark. The system tracks the landmark's position relative to the camera's field of view, providing veering correction and distance-to-landmark feedback. Two audio feedback modes were implemented: sonification (varying beep pulse delay for veering direction and tone frequency for distance) and text-to-speech (spoken directional cues like "Left" or "Right" plus distance updates).
Key findings
A user study with eight blind participants (five totally blind, three legally blind) in a 10x12 meter conference room showed HEADLOCK to be feasible and effective. TTS feedback was significantly faster than sonification for both discovery (12.83s vs 21.34s, 39.9% faster) and guidance (20.37s vs 31.09s, 34.5% faster). However, there was no significant difference in veering between the two modes. Qualitative analysis revealed that TTS allowed users to continue navigating while correcting course, whereas sonification often caused users to pause to interpret the feedback. Users agreed that HEADLOCK would help them navigate open spaces more independently (M=3.88/5) and liked the system overall (M=4.00/5). Without HEADLOCK, all participants used wall-following strategies to find the doorway—walking to a wall first, then following it. Key suggestions for improvement included adding obstacle detection and making feedback adjustable. Users were enthusiastic about the OHMD form factor, finding it lightweight and hands-free, though the touchpad placement on the right side was problematic since all participants held their cane in their right hand.
Relevance
This research addresses a specific gap in blind navigation: crossing large open spaces where traditional cane techniques are insufficient. For accessibility practitioners and assistive technology developers, HEADLOCK demonstrates the potential of OHMDs as navigation aids—their head-mounted camera provides natural alignment with the user's orientation, they are lightweight and hands-free, and bone conduction speakers work well in noisy environments. The finding that TTS outperformed sonification for navigation efficiency suggests that familiar, intuitive feedback is preferable for real-time guidance tasks, even though sonification can convey more data simultaneously. The veering problem identified here is relevant to any indoor navigation system for blind users. The work also highlights practical design considerations: touchpad placement must account for cane use, and obstacle detection is an essential complement to landmark-based navigation.
Tags: visual impairment · blindness · navigation · wayfinding · wearable technology · computer vision · sonification · Google Glass · veering