Human centric sensing by Android phones - II

Abstract

Current indoor localization systems usually rely on external infrastructures and devices such as Wi-Fi hotspots. However, such method is inaccurate as well unstable by itself due to the nature of Wi-Fi and additional infrastructures.

The aim of this project is to create an Android based Indoor Localization Application, which does not rely on infrastructures or any additional components. In order to achieve this, the application will use the indoor geo-magnetism field coupled with the virtual map or floor plan as a reliable way to locate and navigate around. The core data will comprise of mainly sampled magnetic fingerprints and gyroscope data from human centric sensing to form the virtual map with magnetic fingerprints.

Additionally, to increase performance and accuracy of the application, Wi-Fi coordinates will be included to work together with magnetic fingerprints.

Research has been done to understand the theories behind building an indoor localization application. Android programming as well as Android platform and its components such as sensors have been picked up and studied in order to implement and improve the application. Furthermore, in order to understand more about indoor localization application, the famous Google Maps was studied and tested.

At the end of this project, the application has gone through rigorous testing and it is able to deliver satisfactory results.