International Global Navigation Satellite Systems Association
IGNSS Symposium 2018
Colombo Theatres, Kensington Campus, UNSW Australia
7 – 9 February 2018
Quantized RSS Based Wi-Fi Indoor Localization with
Room Level Accuracy
Yan Li
University of Melbourne
liy19@student.unimelb.edu.au
Simon Williams
RMIT University
Phone: +61 3 9925 1631
Simon.williams2@rmit.edu.au
Bill Moran
University of Melbourne
Phone: +61 3 8344 4000
wmoran@unimelb.edu.au
Allison Kealy
RMIT University
Phone: +61 3 9925 1261
allison.kealy@rmit.edu.au
ABSTRACT
The prevalent deployment of Wi-Fi infrastructure provides a potentially low-
cost way to track Wi-Fi enabled devices in a building. Many indoor Location
Based Systems (LBS) aim to get sub-meter precision for grid position
estimation, however such accuracy is not necessary for some indoor location-
aware applications, such as conference room identification and elder-care alert
etc. Our system is designed to track the mobile user at room level granularity
with high accuracy and reliability.
In this paper, a probabilistic fingerprint approach is proposed based on
quantized Received Signal Strength (RSS) measurements. In the training
phase, a histogram based radio map is constructed for each room by storing
various levels of RSS. The motion dynamics of the user is modelled as a
Markov process and a Hidden Markov model (HMM) is applied to track the
mobile user, where the hidden states comprise the possible room locations and
the RSS measurements are taken as observations. In the positioning phase,
backtracking the trajectory of the user can be carried out by the Viterbi
Algorithm.
Our proposed system does not need the prior knowledge of the true scale of
the floor plan nor the true coordinate of each reference point (RP) within the
room. The experimental results show that the offline labour consumption can
be significantly reduced. The proposed approach can give the user location as
