Privacy Preserved Spectral Analysis Using IoT mHealth Biomedical Data for Stress Estimation

In recent years, quantitative analysis of sleep quality and stress estimation during sleep have been important social issues due to sleep deprivation. Conventionally, sleep quality is mainly subjectively evaluated by pittsburgh questionnaire, while stress is estimated by power spectral analysis of electrocardiogram. However, measurement is difficult during sleep since restrictions on respiration rate and body motion. Sleep depth transition presumable by heart rate variability is achieved, however, the correlation between heart rate and sleep quality during sleep is not clarified. In this paper, heart rate and sleep depth data are collected by wearable IoT devices. Then, stress index during sleep is estimated by autonomic balance evaluation index and correlation is analyzed using the collected biomedical data. Furthermore, homomorphic cryptography is applied to analysis for privacy preserving approach.