iNav : a hybrid approach to WiFi localization and tracking of mobile devices

Abstract

This thesis presents iNav - a hybrid system for 802.11-based localization targeted at low-power mobile devices. WiFi localization enables numerous location-based services and applications without requiring a separate GPS module, thus offering device cost and power consumption savings. iNav is a WiFi localization system targeted at low-power mobile devices, capable of utilizing multiple data sources to produce location estimates with accuracy higher than that of pure WiFi estimates. iNav uses a stochastic location estimation algorithm based on particle filters to integrate streams of WiFi access point observations and 3-axis accelerometer data. The system is tailored towards localization of vehicles and relies on a road network map to increase localization accuracy. iNav is designed with low-power devices in mind, and is capable of computing real-time location estimates on embedded devices like the iPhone.