Detection of atypical patterns of occupancy and mobility in smart homes and offices with a network of motion detectors
Abstract/Contents
- Abstract
- Ambient sensors are a category of sensing technologies designed to be embedded in an environment for the purposes of monitoring changes to that environment from either internal or external sources, such as occupants or the weather, respectively. The goal of this dissertation is to study aspects of the behavior and wellbeing of occupants through unobtrusive and non-invasive ambient motion sensors. We focus on passive versions of motion sensors, such as Passive InfraRed (PIR) motion detectors, which have been utilized by researchers for the recognition of Activities of Daily Living (ADL) in homes. Motion sensors are also used in office settings, where they act as occupancy sensors to better control HVAC and lighting systems to save energy. We begin by discussing the types of information that can be extracted from a dense network of PIR sensors in a home, focusing on indicators of wellness. We then discuss methods for extracting occupancy based features from dense motion sensor deployments that represent an occupant's behavior. We demonstrate how these occupancy based features exhibit a strong correlation with features derived from human annotated activities for multiple long term smart home datasets. From this we reason that our derived features can be used as a proxy for human labeled activities for certain applications such as atypical behavior detection. We then present a method to use occupancy based features to find days that are atypical, and compare our results with a classifier based on human annotated activities. Lastly we discuss how low level mobility features can be extracted from a network of motion sensors to form a mobility graph, which can be used to summarize the aggregate mobility of the occupants of an environment. We show how this mobility graph can be visualized to represent how occupants move between sensors and how changes in the mobility graph over time can be quantified. In a home environment, we show how significant changes in this mobility graph correlate with changes in the occupant's underlying behaviors. In an office environment, we show how changes in the mobility graph can be related to global events, like fire alarms.
Description
| Type of resource | text |
|---|---|
| Form | electronic; electronic resource; remote |
| Extent | 1 online resource. |
| Publication date | 2015 |
| Issuance | monographic |
| Language | English |
Creators/Contributors
| Associated with | Wong, Kevin Bing-Yung |
|---|---|
| Associated with | Stanford University, Department of Electrical Engineering. |
| Primary advisor | Law, K. H. (Kincho H.) |
| Thesis advisor | Law, K. H. (Kincho H.) |
| Thesis advisor | Aghajan, Hamid K |
| Thesis advisor | Poon, Ada Shuk Yan |
| Thesis advisor | Rajagopal, Ram |
| Advisor | Aghajan, Hamid K |
| Advisor | Poon, Ada Shuk Yan |
| Advisor | Rajagopal, Ram |
Subjects
| Genre | Theses |
|---|
Bibliographic information
| Statement of responsibility | Kevin Bing-Yung Wong. |
|---|---|
| Note | Submitted to the Department of Electrical Engineering. |
| Thesis | Thesis (Ph.D.)--Stanford University, 2015. |
| Location | electronic resource |
Access conditions
- Copyright
- © 2015 by Kevin Bing-Yung Wong
- License
- This work is licensed under a Creative Commons Attribution Non Commercial 3.0 Unported license (CC BY-NC).
Also listed in
Loading usage metrics...