Sleep disorders affect a large percentage of the population; however, most disorders can only be diagnosed through the use of polysomnography. In this overnight sleep study, patients are asked to sleep in a clinical lab attached to many different machines. In addition, patients might be on a waiting list for up to a year before an available lab time comes up. Current techniques are expensive, uncomfortable, and resource-consuming. In an effort to meet the need for an affordable, at-home sleep monitoring system, SleepSmart was developed. SleepSmart is a mattress sensor topper that leverages arrays of accelerometers and thermistors to acquire dynamic data profiles and temperature distribution. Using the data from the sensors, various information such as sleeping postures, biosignals, movement events can be predicted with machine learning methods. The frequency and magnitude of posture changes can be used to determine sleep quality by calculating the restlessness index. These sensors are all incorporated into a mattress topper which would replace a fitted mattress sheet, allowing for unobtrusive monitoring.
A prototype for SleepSmart has been fabricated, and software implementations are currently being implemented to allow for the characterization of biosignals (heart rate and respiratory rate) as well as improve on the existing temperature sensors, and the classification of movement events. The project is being funded by Kids Brain Health Network, a trans-Canada initiative dedicated to studying children’s brain development, and studies are currently underway in collaboration with the BC Children’s Hospital. The final product will be used to assist in the diagnosis of sleep disorders of children with neurodevelopmental disorders. In addition, SleepSmart will allow clinicians to gather a repository of physiological data that can be used in future analysis to further the understanding of sleep in the pediatric population.
If you’re interested in participating, more details can be found here.
Supervisor: Machiel Van der Loos
Researcher: Yi Jui Lee