Self-mapping radio maps for location fingerprinting

Abstract

Localisation of a mobile device in Wi-Fi based indoor environments has received much attention and is an important area of research, considering the uptake of devices with built-in Wi-Fi and often blanket coverage in enterprise. Most localisation techniques in this area require an off-line calibration phase where a radio map is manually built using the laborious efforts of administrators. This radio map is then subject to becoming inaccurate with time and requires refreshing as the equipment is upgraded and replaced. We propose a system that can automatically build and calibrate a radio map for use with location fingerprinting techniques to provide indoor and outdoor localisation with Wi-Fi. This system first uses a self-mapping technique that makes use of user roaming patterns to build a weighted undirected graph of Access Point locations. We then combine this initial map with a map built by users of smartphones with built-in Wi-Fi and GPS to generate anchor nodes. We show that the unique combination of these two techniques provides a zero-configuration calibration map for use with location fingerprinting techniques, which not only saves time and effort in the calibration phase, but provides a constantly fine-tuning and self-healing map.

Indoor Positioning on Smartphones Using Built-In Sensors and Visual Images

“…In this paper, the similarity between the query image and the candidate image is calculated by clustering the feature vectors to form a feature histogram. To improve the accuracy, this paper adopts the normalized term frequency (NTF) weighting algorithm [ 24 ] to process the feature histogram. Image I consists of a statistical histogram with k columns , where is represented as: where k is the number of visual words and is the total number of visual words in image I .…”