A Mobile Indoor Positioning System Founded on Convolutional Extraction of Learned WLAN Fingerprints

Avi Bleiweiss

2016

Abstract

The proliferation of both wireless local area networks and mobile devices facilitated cost-effective indoor positioning systems that obviate the need for expensive infrastructure. We explore a floor-level, indoor localization system to predict the physical position of a mobile device holder in an office space by sensing a fingerprint of signal strength values, received from a plurality of wireless access points. In this work, we devise an instructive model that tailors elemental algorithms for unsupervised fingerprint learning, and resorts to only using a single-layer convolutional neural-network, succeeded by pooling. We applied our model to a fingerprint-based dataset that renders large multi-story buildings, and present a detailed analysis of the effect of changing setup parameters including the number of hidden nodes, the receptive field size, and the stride between extracted features. Our results surprisingly show that classification performance improves markedly with a sparser feature extraction, and affirms a more intuitive gain, yet milder, as any of the number of features or the tile size increases. Despite its simplicity, the positional accuracy we attained is sufficient to provide a useful tool for a location-aware mobile application, purposed to automate the mapping of building occupants.

Paper Citation

in Harvard Style

Bleiweiss A. (2016). A Mobile Indoor Positioning System Founded on Convolutional Extraction of Learned WLAN Fingerprints . In Proceedings of the 5th International Conference on Pattern Recognition Applications and Methods - Volume 1: ICPRAM, ISBN 978-989-758-173-1, pages 214-223. DOI: 10.5220/0005685702140223

in Bibtex Style

@conference{icpram16,
author={Avi Bleiweiss},
title={A Mobile Indoor Positioning System Founded on Convolutional Extraction of Learned WLAN Fingerprints},
booktitle={Proceedings of the 5th International Conference on Pattern Recognition Applications and Methods - Volume 1: ICPRAM,},
year={2016},
pages={214-223},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0005685702140223},
isbn={978-989-758-173-1},
}

in EndNote Style

TY - CONF
JO - Proceedings of the 5th International Conference on Pattern Recognition Applications and Methods - Volume 1: ICPRAM,
TI - A Mobile Indoor Positioning System Founded on Convolutional Extraction of Learned WLAN Fingerprints
SN - 978-989-758-173-1
AU - Bleiweiss A.
PY - 2016
SP - 214
EP - 223
DO - 10.5220/0005685702140223