--- title: "LyreFlute: Sensor-Based Musical Experience for COPD Rehabilitation" authors: - Andreia Valente - Claudio Cotto - Mark Billinghurst - Kunal Gupta year: 2026 venue: "Extended Abstracts of the 2026 CHI Conference on Human Factors in Computing Systems (CHI EA 2026)" doi: "10.1145/3772363.3799123" url: "https://andreia-valente.com/publications/valente2026lyreflute.html" pdf: "https://andreia-valente.com/pdfs/valente2026lyreflute.pdf" topics: - COPD Rehabilitation - Breathing - Biofeedback - Tangible Interaction - Musical Interaction --- # LyreFlute: Sensor-Based Musical Experience for COPD Rehabilitation ## Citation Metadata - Authors: Andreia Valente, Claudio Cotto, Mark Billinghurst, Kunal Gupta - Venue: CHI EA 2026 - Year: 2026 - DOI: https://doi.org/10.1145/3772363.3799123 - HTML: https://andreia-valente.com/publications/valente2026lyreflute.html - PDF: https://andreia-valente.com/pdfs/valente2026lyreflute.pdf ## Plain-Language Summary LyreFlute is a sensor-based musical rehabilitation experience for people with chronic obstructive pulmonary disease. It turns breathing exercises into an interactive flute-like rhythm game. The prototype uses separate airflow sensors for nose and mouth breathing, plus capacitive finger sensors, so it can track whether users are performing target respiratory patterns rather than simply blowing into an object. The experience is framed through a culturally grounded narrative involving an Australian lyrebird and environmental restoration after bushfire damage. Correct breathing and finger actions trigger musical and visual progress, turning rehabilitation from repetitive clinical exercise into meaningful interaction. ## System Design The system consists of a custom flute-like device and a 3D interactive experience. The flute tracks nasal inhalation, oral exhalation, breath timing, and finger placement. Users follow rhythm-game prompts and learn songs from a lyrebird guide. Successful performance restores parts of the virtual environment. The CHI demonstration presents a condensed three-day progression, moving from fire suppression to smoke clearing and habitat restoration. This structure shows how a longer rehabilitation program can be represented as staged musical and narrative progress. ## Contributions - Introduces a tangible breath-sensing musical interface for COPD rehabilitation. - Separates nasal and oral airflow to support technique-aware breathing feedback. - Combines rhythm-game mechanics, respiratory training, and environmental storytelling. - Demonstrates how at-home rehabilitation can be made more engaging without losing measurement fidelity. ## Design Implications Rehabilitation systems should measure correct technique, not just motivate repetition. LyreFlute shows how sensing, game structure, and narrative can work together: sensing verifies performance, rhythm mechanics make invisible breath goals legible, and story gives each breath a visible consequence. ## Breathing Protocol and Narrative Progression LyreFlute targets two established COPD rehabilitation techniques: pursed-lip breathing and diaphragmatic breathing. The interaction emphasizes inhaling through the nose and exhaling slowly through the mouth, with exhalation lasting at least twice as long as inhalation. This supports COPD breathing practice by encouraging controlled exhalation, positive back-pressure, and more efficient breathing patterns. The CHI demonstration condenses a longer rehabilitation trajectory into a three-day musical progression. Day 1 uses a single continuous phrase of 17 notes over 21 seconds, focusing on sustained exhalation and breath control. Day 2 uses three breath groups totaling 35 notes, adding inhale-exhale cycling and recovery breaths of about five to six seconds between groups. Day 3 uses four breath groups with 32 notes, shorter recovery times of about three to five seconds, and denser note timing, creating the highest respiratory demand. The narrative connects personal lung rehabilitation to ecological restoration after the 2019-2020 Australian bushfires. A lyrebird guide teaches the songs, and successful breathing restores the forest: rain suppresses fire, wind clears smoke and ash, and plants and wildlife return. ## Hardware and Software Implementation The custom flute is 3D printed and includes separate airflow sensors for nasal inhalation and oral exhalation. One FS3000 airflow sensor sits inside the flute to measure exhalation, while another is angled toward the nose to capture inhalation. Four capacitive finger holes detect finger placement using copper rings and an MPR121 sensor. An XIAO ESP32S3 microcontroller processes the sensor data, with a BNO08X IMU available for orientation tracking, and sends data over Bluetooth to the Unity experience. The rhythm game maps four finger holes to musical input tracks. Users must coordinate nasal inhalation timing, oral exhalation intensity and duration, and finger placement. The desktop system compares breath timing and duration against the target pattern for each song section and provides immediate feedback. Unity converts sensor input to MIDI and uses hosted virtual instruments to produce flute sounds while the forest responds visually to performance. ## Future Work The paper proposes clinical validation with a COPD cohort over a 15-day period, comparing LyreFlute with traditional training while measuring adherence, physiological effects, and completion rates. Future versions may add octave changes controlled by exhalation velocity, low-cost Android integration, and culturally specific flute designs and narratives for different patient communities. ## Why This Paper Matters This work is relevant to health technology, tangible interaction, biofeedback, COPD rehabilitation, musical interfaces, and serious games. It is useful for researchers interested in turning physiological training into meaningful, measurable, and motivating interaction. ## Recommended Citation Valente, A., Cotto, C., Billinghurst, M., & Gupta, K. (2026). LyreFlute: Sensor-based musical experience for COPD rehabilitation. In Extended Abstracts of the 2026 CHI Conference on Human Factors in Computing Systems. Association for Computing Machinery. https://doi.org/10.1145/3772363.3799123 ```bibtex @inproceedings{valente2026lyreflute, author = {Valente, Andreia and Cotto, Claudio and Billinghurst, Mark and Gupta, Kunal}, title = {LyreFlute: Sensor-Based Musical Experience for {COPD} Rehabilitation}, year = {2026}, booktitle = {Extended Abstracts of the 2026 CHI Conference on Human Factors in Computing Systems}, publisher = {Association for Computing Machinery}, doi = {10.1145/3772363.3799123}, url = {https://doi.org/10.1145/3772363.3799123} } ```