Stream beat-to-beat signals from many Garmin wearables to a single iPad — live.
UeduPAD is a research-grade iOS app built by the Uedu team at National Central University, Taiwan. Through the Garmin Health SDK it streams beat-to-beat intervals (BBI) in real time — time-aligned, with the raw sequence fully preserved — giving multimodal learning analytics (MMLA) a cohort-level source of physiological data.
Illustrative readout · dummy data
The real open problem is how to scale it to whole classes and cohorts, aligned with classroom events, so teachers and researchers get multimodal signals they can actually use. UeduPAD is built for exactly that — not another single-user health app, but a classroom-grade instrument for physiological signal capture.
一台 iPad 同時監看多支手錶的逐拍訊號，跨個體同一時間軸比較——群體層級的視角，而非個人健康追蹤。
訊號在課堂當下從手錶經 Garmin Health SDK 直達 iPad，不必等課後才從雲端下載彙整。
每一段生理訊號都與課堂事件日誌時間對齊，支援生理—行為的時序延遲分析。
Watch → SDK → iPad → aligned stream → Uedu platform.
手錶端逐拍偵測心搏間期（RR/BBI）。
透過 BLE 把逐拍訊號即時送到 iPad，非事後雲端匯出。
同時聚合多支手錶，對齊課堂事件時間軸，即時計算 HRV。
寫入 Uedu Fit / Uedu Mind 與 Educational Omics 資料湖供研究分析。
Each tile is one wearer’s live heart rate (BPM) and beat-to-beat interval (BBI). Devices going on- and offline and signals fluctuating are shown transparently — this is what multi-person synchronized streaming actually looks like.
Purpose-built for research: synchronized capture, faithful raw data, and an ethical workflow — by design.
一台 iPad 同時監看多支手錶的逐拍訊號，跨個體共用同一時間軸——群體層級的比較，而非單人追蹤。
逐拍 BBI 不做破壞性壓縮，完整落地，任何 HRV 指標都可在課後重新計算與重製。
生理訊號與課堂事件日誌對齊時間軸，支援生理—行為的時序延遲（temporal lag）分析。
從整班同步比較到個別回溯，隨時在群體視角與個人序列間切換。
RMSSD、pNN50、lnHF、LF/HF 等標準指標以 5 分鐘滑動視窗即時計算。
知情同意流程、去識別化、倫理資料流內建——在 IRB 核准的框架下收集與保存。
Beat-to-beat intervals and derived HRV metrics are objective proxies for cardiac autonomic activity. Absent validated, context-specific algorithms, UeduPAD does not treat these signals as direct measures of stress, emotion, or learning; every physiology–behavior correspondence must be established through research. Garmin wearables are consumer wellness products, not medical devices.
大學 Python 等課堂的多模態學習分析示範：讓學生親手體驗生理訊號如何被採集、對齊與解讀。
MMLA 與 PALM（生理感知 AI 輔導）的第一手資料來源，串接 Educational Omics 資料湖供後續分析。
運動科學、認知、學習分析、臨床試驗——同一套採集與同步機制，換場景不換工具。
UeduPAD’s data collection is covered by an umbrella IRB approved by the National Taiwan University Research Ethics Committee (No. 202507EM058). With the built-in informed-consent flow and de-identification SOP, data collected under the protocol falls within IRB coverage.
From Wearables to Classrooms: A Person-Centered Feasibility Study of HRV-Based Physiological Monitoring for Learning Analytics
PALM: Scaling Physiologically-Aware AI Tutoring Through Consumer Wearables and Large Language Models
Temporal Lag Effects in Multimodal Learning Analytics: Physiological–Behavioral Characterization
C-GRASP: Clinically-Grounded Reasoning for Affective Signal Processing
Designing an Educational Omics Data Lake: A Multimodal Infrastructure for Technology-Enhanced Learning
UeduPAD currently runs as a research instrument for academic and educational collaboration. If you have an idea for classroom multimodal research or wearable integration, we’d love to talk.