Open Collections

A system for the monitoring of sleep-related parameters based on inertial measurement units

University of British Columbia

Sleep is an essential process needed by the body; as a result, the effects of sleep deprivation are severe and include increased risk of heart diseases and dementia. It is estimated that 40% of Canadians suffer from sleep disorders; nevertheless, most of them are unaware of their condition. Approximately 75–80% of cases of sleep apnea (SA), the most common sleep disorder, are still undiagnosed and thus lacking treatment. However, demand for sleep studies is unmet by the available spaces—high operating costs and expensive equipment limits their availability and the access to diagnosis. Thus, the necessity of inexpensive, yet adequate alternatives for the monitoring of sleep parameters is evident. We present the development of a respiratory effort (RE), body position (BP), and heart rate (HR) monitoring systems for sleep using inertial measurement units (IMU) comprising a 3D accelerometer, gyroscope, and magnetometer. These parameters are required by the American Academy of Sleep Medicine (AASM), and necessary for diagnosis of SA. The chest’s angle variations due to breathing are tracked using all three sensors in the IMU and an extended Kalman filter (EKF) as a data fusion method, thus obtaining the RE in supine and lateral recumbent positions, as well as monitoring BP by tracking gravity vector orientation relative to the body. The system is self-contained, wireless, battery operated and real time. Thus, we improve over previous works that only measure RE limited to a supine position of the body. The HR is estimated by detecting cardiac induced vibrations from a single axis angular rate measuring channel. Using a simple algorithm with an adaptive threshold for peak detection, we obtain good instantaneous HR readings. Although its simplicity sacrifices some accuracy, it is amenable for real-time implementation in a low cost, wireless microcontroller, such as the one utilized for RE/BP. Both tests show promising results, with a high correlation value for RE (e.g. 96.26%) and low error values for HR (e.g. mean 0.35bpm) with respect the reference signals. Furthermore, preliminary tests show the possibility of obtaining all three signals using a single IMU device, suggesting a viable alternative to current technologies.

Text

2018-12-05

Attribution-NonCommercial-NoDerivatives 4.0 International

http://hdl.handle.net/2429/67971

Electrical and Computer Engineering

University of British Columbia

UBCV

http://creativecommons.org/licenses/by-nc-nd/4.0/