The use of off-the-shelf wearable sensors to analyze daily-living activities and emotional state of a person at the Smart Condo

• Author / Creator

  Diaz Romero, Dillam Jossue

• The Smart Condo is a model condo embedded with a wireless sensor network, developed by an interdisciplinary team including researchers from Occupational Therapy, Industrial Design, Pharmacy, and Computing Science. The Smart Condo aims to support older adults, including those with physical and cognitive disabilities, to live independently longer. Older adults with complex needs are often limited in their ability to perform necessary daily activities and may require task-specific supports. Continuous health-monitoring systems have the potential to enhance one’s quality of life and help older adults live safely in their home. This thesis describes two studies in the use of off-the-shelf wearable sensors in order to investigate the feasibility of using wearable devices and usefulness in the Smart Condo.

  In the first study, twenty-six participants spent a single two-hour session in the one-bedroom living environment, either alone or in pairs, and performed a scripted protocol of activities of daily living. Twelve of these participants wore the commercial smart eyewear device JINS MEME, which collected electrooculography (EOG), accelerometer and gyroscope data throughout their sessions. This study used an offline classification method to predict the participants’ activities. The approach showed that this method yields equal or better results with a variety of activities compared to approaches that involve more restrictive wearable device setups. The results demonstrate the suitability of JINS MEME for recognition of activities of daily living and identify limitations associated with the current model.

  In the second study, twenty-one participants viewed a sequence of images from the International Affective Picture System (IAPS) database. Using wearable sensor devices, we collected electroencephalography (EEG), electrooculography (EOG), and kinematic motion data as participants viewed the images; the participants also characterized their own emotional responses to the images. Participants then played the serious game “Whack-a-Mole,” wearing the sensor devices, and played three levels of the game that required varying amounts of cognitive effort. This study describes the method for emotion recognition (\ec~ and \rf) in participants as they played the serious game. This approach showed that emotional state during a task can be determined accurately using data collected from wearable sensor devices, with and without self-reported measures.

• Subjects / Keywords

  • Machine Learning
  • Valence & Arousal recognition

• Graduation date

  Fall 2019

• Type of Item

  Thesis

• Degree

  Master of Science

• DOI

  https://doi.org/10.7939/r3-dvqa-yc82

• License

  Permission is hereby granted to the University of Alberta Libraries to reproduce single copies of this thesis and to lend or sell such copies for private, scholarly or scientific research purposes only. Where the thesis is converted to, or otherwise made available in digital form, the University of Alberta will advise potential users of the thesis of these terms. The author reserves all other publication and other rights in association with the copyright in the thesis and, except as herein before provided, neither the thesis nor any substantial portion thereof may be printed or otherwise reproduced in any material form whatsoever without the author's prior written permission.