Tampu: Mobile Application Using Wearable Technology for Real-Time Detection and Alert of Anxiety and Panic Attacks in University Students

Curi I.C.; Quispe E.Y.; Lopez N.B. · Proceedings of the 2025 IEEE 5th International Conference on Advanced Learning Technologies on Education and Research, ICALTER 2025 (2025) · DOI: 10.1109/icalter69698.2025.11355147

Resumen

This article presents the design, development, and validation of Tampu, a mobile-wearable system for the realtime detection and alerting of anxiety episodes in university students. The article includes methodological details, a refined system architecture, extended experimental results, and a more comprehensive statistical analysis. Due to the high prevalence of anxiety disorders in academic environments, Tampu integrates continuous monitoring of physiological signals-heart rate, heart rate variability (HRV), and oxygen saturation (SpO2)with a machine learning model executed locally using TensorFlow Lite. The system identifies physiological patterns associated with anxiety episodes and delivers immediate alerts along with emotional support tools such as breathing exercises, event logging, and emergency contacts. A 15-day pilot study with 30 students generated 2,574 valid physiological records, equivalent to an average of 85.8 records per participant per day. The model achieved an accuracy of 86.7%, a sensitivity of 87%, and specificity of 93%, in addition to a high usability score (SUS =87 / 100). The results demonstrate the feasibility of integrating wearable technology and machine learning into mobile devices to support early detection and self-management of anxiety in higher education contexts.

Instituciones: Universidad Peruana de Ciencias Aplicadas