Understanding User-perception of Sleep to Inform Sensing and Provide Actionable Feedback

Abstract

It is becoming increasingly clear of the importance of sleep and its impact on our daily lives. Despite the pervasiveness of sleep issues, people struggle to assess and improve their sleep. Sleep is an unconscious, passive activity and accurate, manual, self-tracking of sleep is often unattainable. A majority of the general population remains unaware of long-term patterns in their sleep duration and consistency, unless the individual suffers from a chronic sleep related disorder that affects day-time functioning or causes other related health issues. The growing popularity of commercial sleep sensors for use at home shows that these technologies have the potential to provide long-term, low-cost, and accurate representations of people’s daily sleep patterns in the comfort of their home environment. However, current sleep tracking devices are limited by hardware and algorithmic constraints. Un-actionable sleep feedback, in the context of long term sleep tracking for individuals without sleep disorders, leads to mental models that are in tension with recommendations for good sleep health from sleep medicine professionals. Through a comprehensive survey of commercial sleep sensing devices, I identify gaps between user perception of sleep data and best practices in sleep health. I contribute to the design, implementation and evaluation of a wireless sleep sensing system that enables non-contact monitoring of sleep related physiological signals throughout the night. I also explore, through the design and implementation of a smart phone app, how a personalized sleep model can be used to provide meaningful and actionable feedback that can provide users with the intent to make changes to their behavior to improve sleep health.