Jolley Hall, Room 309
"Towards Large-Scale Indoor Positioning"
Advisers: Tao Ju and Yasutaka Furukawa
The emerging of the Global Position System (GSP) has opened up a a new area of location-based service, such as navigation and location-aware advertising. However, such system does not work for indoor environments, where building roofs block the satellite signals. Indoor positioning systems typically rely on building-dependent infrastructures such as Bluetooth and WiFi, in which the laborious calibration process prevents such systems from large-scale deployment. In this project, we propose a pipeline that use data-driven methods on crowd-sourced data to build an indoor positioning system that can be deployed at scale. Towards the goal we will propose an algorithm for inertial-only motion tracking, a method that constructs the WiFi radio map automatically from crowd-sourced data and finally a system that combines all information to provide reliable indoor positions.