Echolocation as an Accessible Navigation Tool in a Virtual 3D Environment
Mai Ricaplaza Thøgersen, Rasmus Jens Frølich Kjeldsen · 2024 · Assets '24: Proceedings of the 26th International ACM SIGACCESS Conference on Computers and Accessibility · doi:10.1145/3663548.3688547
Summary
This paper investigates whether active echolocation — emitting a sound and interpreting its echo — can serve as a viable navigation tool for blind and low-vision (BLV) individuals in virtual 3D environments. Virtual spaces remain overwhelmingly visual, lacking the physical and auditory cues (ambient noise, sound shadows, surface textures) that BLV people rely on for real-world navigation. The researchers built a web-based 3D environment using Three.js and the Web Audio API, implementing a simplified echolocation mechanic: users emit a click sound that travels forward from their position and reflects back with volume proportional to distance, functioning more like an active sonar beam than true omnidirectional echolocation. The system also simulates footstep sounds that change dynamically based on room size, providing passive spatial information. Thirteen BLV participants (all blind, recruited from the audiogames.net community) navigated a tutorial level and six progressively more complex maze-like environments. The study collected real-time telemetry data (paths taken, distances traveled) alongside qualitative feedback through questionnaires. The implementation deliberately omitted wall collision sounds and environmental ambience (sonification) to isolate the echolocation mechanic for evaluation, though this design choice proved to be a significant limitation that participants identified.
Key findings
The study identified three distinct navigation strategies among participants. "Shorelining" — walking along walls to find exits, mirroring a common real-world BLV navigation technique — was used consistently by two participants who demonstrated strong spatial cognitive approaches. "Ping-pong" — using echolocation to find the far wall, walking to it, then repeating — was used by three participants and proved effective for building mental maps. The remaining participants used random or inconsistent approaches. Of the 13 participants, four demonstrated spatial cognitive approaches (building mental maps), three took pragmatic approaches (just finding the exit), and six were inconclusive. The echolocation was effective for gauging straight-line distance to walls and locating the exit (identified by a pitch change), but struggled significantly with detecting openings in the middle of walls — a critical limitation for real navigation. Six participants found the echolocation useful, four saw potential but found the current implementation lacking, and three gave no clear feedback. Multiple participants emphasised that echolocation without ambient sonification of the environment felt unrealistic, noting that real-world echolocation always operates alongside environmental sounds. The clicking sound also became harsh and fatiguing during extended use.
Relevance
This study is valuable for several reasons. First, it directly tests accessibility in the sonic design space — a domain where much theoretical work exists (including on design spaces and self-adapting interfaces) but relatively little empirical user research with BLV participants. The finding that BLV participants transferred real-world navigation strategies (shorelining) to virtual environments is significant, suggesting that accessible virtual environments should be designed to support, not replace, existing navigation skills. The study also reveals a fundamental design tension: isolating a single accessibility feature for evaluation strips away the holistic sensory context that makes it effective in practice. The participants' insistence on sonification — ambient sound, reverberation, spatial audio cues — reinforces that accessible interfaces in any design space require rich, layered feedback rather than single-channel solutions. For practitioners working on accessible gaming, VR, or spatial computing, the paper provides concrete design lessons: echolocation needs omnidirectional feedback (not just forward-facing), environmental sonification is essential context, and sound design must account for extended-use fatigue.
Tags: virtual environments · echolocation · blind and low vision · spatial navigation · sonic design space · user study
Standards referenced: WCAG 2.2