Exhale to Execute: Transforming Commercial Microphone Arrays into Directional Airflow Sensors for Ubiquitous Interaction

Abstract

Ubiquitous computing environments frequently require hands-free input methods when users are physically occupied. Current exhalation detection approaches rely on dedicated airflow sensors or utilize standard microphones with basic volume thresholds. We introduce a software framework designed to repurpose commercial microphone arrays for directional airflow sensing. To address the limitations of simple amplitude thresholds, our system employs an absolute-energy gate, a cross-channel energy-uniformity ratio, and non-linear energy-weighted mapping. This acoustic pipeline enables standard microphones to estimate airflow direction under controlled conditions. Technical evaluations with a calibrated mechanical airflow source show that the framework distinguishes localized airflow impacts from the tested far-field background noise and separates coarse directional targets. This work includes no human-subject evaluation; the interactive demonstrations illustrate proof-of-concept mappings rather than validated interaction performance.