Motion links for Audio Augmented Reality
Motion links for Audio Augmented Reality
Master’s research project
MEng Design Engineering
spacer
9 months: Individual project
C++, Python
E-textiles
Human Systems
Data Analysis
Most audio augmented reality (AAR) systems rely on environmental infrastructure or pre-set sound files, which limits them to calibrated spaces and prevents everyday, scalable use. This project asks whether generating audio directly from a user's own movement, rather than from the environment, can produce a more personalised, body-led AAR experience where the user's sense of control over the sound is built in from the start rather than assumed.
Building context to AAR systems for everyday environments
The hardware problem
Collecting reliable pressure data
Developing an E-textile solution that is undetectable to the user, such that it doesn’t hinder their walking behaviour, that reliably measures pressure data. The data should be transmitted in real time to generate procedural audio recreating the sound of users’ footsteps.
Design Process
Iterating on the design to determine the range of walking pace that functions reliably with the quickest latency for transparent augmentation. Evaluating calibration methods and audio generation algorithms which build the procedural audio to ensure that it produces distinct audio effects, with consistent quality between setup and individual characteristics.
Final Design
Research Question: Does procedural audio generated from an individual’s own movement influence gait behaviour while preserving the individual’s perceived sense of agency?
specifically by applying two independent character drifts to the sound of their footsteps, and measuring their behaviour, and awareness over the influence. Changing the perceived weight and tempo, while ensuring that the sound is still reactive to their behaviour.
Impact
The system confirmed that a fully inside-out AAR setup, generating audio entirely from a user's own movement, works reliably outside a controlled prototype. Procedurally generated sound produced a measurable effect on gait, significantly increasing peak heel strike force under the heavier-sound condition.
Participants' sense of agency stayed high even when their gait was influenced and even though most never noticed the manipulation, showing that AAR can shape behaviour without undermining the user's sense of control.
Learnings
Designing explicitly for agency from the outset, rather than treating it as a byproduct of the audio system, proved to matter as much as the gait manipulation itself. A fixed manipulation target doesn't suit everyone, and the next version of this system should calibrate to each user's own baseline sensitivity rather than applying a uniform drift.