Show simple item record

dc.contributor.advisorCook, Diane J.
dc.creatorWilliams, Jennifer Ashleigh
dc.date.accessioned2018-05-08T16:53:33Z
dc.date.available2018-05-08T16:53:33Z
dc.date.issued2017
dc.identifier.urihttp://hdl.handle.net/2376/12957
dc.descriptionThesis (Ph.D.), Computer Science, Washington State Universityen_US
dc.description.abstractSmart home technologies offer a unique opportunity to monitor individuals in a non-invasive manner. Because of this, they provide an ideal opportunity for pervasive healthcare. Smart homes provide the resident the freedom of being at home while still providing monitoring and analysis of any health issues that may arise. The ultimate goal is to use smart home technology to enable people recovering from injuries or coping with disabilities (i.e., Post-Traumatic Stress Disorder, or dementia) to live independently. The immediate goal this dissertation is to monitor, analyze, and predict the overall behavior of the individual in a smart home environment for correlation with a range of chronic and sudden health issues. We accomplish this goal by designing computational approaches to analyze and forecast daily behavior. Additionally, we quantify daily behavior exhibited by an individual while they are both awake and asleep based on data collected in CASAS smart homes. Sleep is an important element in our everyday lives. In this dissertation we hypothesize that wake behavior has an impact on both current and future sleep and vice versa. To analyze the relationship between sleep and wake activity we consider sleep separately from other activities within the smart home. We examine the relationship based on prior studies and using our novel computational methods. To computationally evaluate the relationship between wake and sleep behavior, we explore three main elements: 1) behavior cycle detection, 2) behavior quantification methods, and 3) behavior forecasting. We provide validation for these approaches by using data collected in actual smart home test sites. In addition to analyzing the relationship between wake and sleep behaviors, we propose that a reliable sleep quality analysis can be achieved using the CASAS smart home system. We pursue this goal by performing a pilot study in which wearable sensors are used in conjunction with the CASAS smart home sensors. We use the data from this pilot study to validate the use of smart home sensors as a tool for the analysis of sleep patterns.en_US
dc.description.sponsorshipWashington State University, Computer Scienceen_US
dc.language.isoEnglishen_US
dc.rightsIn copyright
dc.rightsPublicly accessible
dc.rights.urihttp://rightsstatements.org/vocab/InC/1.0/
dc.subjectComputer scienceen_US
dc.subjectArtificial intelligenceen_US
dc.subjectBehavior forecastingen_US
dc.subjectMachine learningen_US
dc.subjectSleep analysisen_US
dc.subjectSmart environmenten_US
dc.titleInvestigating the Relationship between Sleep and Wake Behavior using Machine Learning and Smart Home Sensorsen_US
dc.typeText||Electronic Thesis or Dissertationen_US


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record