Date of Award
Master of Science
Ubiquitous applications collect contextual information, process it, and then use this derived data to deliver valuable services. Location is one these contexts, and has been significant in providing navigation and guidance services for GPS devices. However, GPS is designed for outdoor use and is not precise enough, in terms of location accuracy for indoor applications.
There are many indoor location systems that rely on a single technology, but these systems are either inaccurate in uncontrolled environments or require the installation of a dedicated infrastructure. This has led to the investigation of hybrid systems. This thesis examines the creation of a hybrid indoor positioning system combining different tech nologies and techniques; Wi-Fi access points and their associated signal strength, image analysis using machine learning to create location specific scene classifiers, and an altimeter sensor to determine the user's current floor. This system is meant to provide indoor positioning data to location-aware applications.
Lemieux, Nathan Paul, "Exploring Hybrid Indoor Positioning Systems" (2009). Digitized Theses. 4034.