Files in this item

FilesDescriptionFormat

application/pdf

application/pdfWANG-DISSERTATION-2016.pdf (21MB)
(no description provided)PDF

Description

Title:Applying multimodal sensing to human location estimation
Author(s):Wang, He
Director of Research:Roy Choudhury, Romit
Doctoral Committee Chair(s):Vaidya, Nitin
Doctoral Committee Member(s):Lymberopoulos, Dimitrios; Nahrstedt, Klara
Department / Program:Electrical & Computer Eng
Discipline:Electrical & Computer Engr
Degree Granting Institution:University of Illinois at Urbana-Champaign
Degree:Ph.D.
Genre:Dissertation
Subject(s):sensing
location
visual fingerprinting
motion leaks
side-channel attacks
security
Abstract:Mobile devices like smartphones and smartwatches are beginning to "stick" to the human body. Given that these devices are equipped with a variety of sensors, they are becoming a natural platform to understand various aspects of human behavior. This dissertation will focus on just one dimension of human behavior, namely "location". We will begin by discussing our research on localizing humans in indoor environments, a problem that requires precise tracking of human footsteps. We investigated the benefits of leveraging smartphone sensors (accelerometers, gyroscopes, magnetometers, etc.) into the indoor localization framework, which breaks away from pure radio frequency based localization (e.g., cellular, WiFi). Our research leveraged inherent properties of indoor environments to perform localization. We also designed additional solutions, where computer vision was integrated with sensor fusion to offer highly precise localization. We will close this thesis with micro-scale tracking of the human wrist and demonstrate how motion data processing is indeed a "double-edged sword", offering unprecedented utility on one hand while breaching privacy on the other.
Issue Date:2016-07-11
Type:Thesis
URI:http://hdl.handle.net/2142/92743
Rights Information:Copyright 2016 HeWang
Date Available in IDEALS:2016-11-10
Date Deposited:2016-08


This item appears in the following Collection(s)

Item Statistics