Loading…

Accurate Indoor Positioning Based on Learned Absolute and Relative Models

To improve the accuracy of indoor positioning systems it can be useful to combine different types of sensor data. This paper describes deep learning methods both for estimating absolute positions and for performing pedestrian dead reckoning, and then how to combine the resulting estimates using weig...

Full description

Saved in:

Bibliographic Details
Main Authors:	Kjellson, Christoffer, Larsson, Martin, Astrom, Kalle, Oskarsson, Magnus
Format:	Conference Proceeding
Language:	English
Subjects:	Buildings Deep learning fingerprinting indoor positioning Neural networks Optimization PDR radio beacons sensor fusion smartphone Task analysis Time measurement Training
Online Access:	Request full text
Tags:	Add Tag No Tags, Be the first to tag this record!

cited_by
cites
container_end_page	8
container_issue
container_start_page	1
container_title
container_volume
creator	Kjellson, Christoffer Larsson, Martin Astrom, Kalle Oskarsson, Magnus
description	To improve the accuracy of indoor positioning systems it can be useful to combine different types of sensor data. This paper describes deep learning methods both for estimating absolute positions and for performing pedestrian dead reckoning, and then how to combine the resulting estimates using weighted least squares optimization. The positioning model is based on a custom neural network which uses measurements of received signal strength indication from one instant of time as input. The model for estimating relative positions is on the other hand based on inertial sensors, the accelerometer, magnetometer and gyroscope. The position estimates are then combined using a least squares approach with weights based on the standard deviations of errors in predictions from the used models.
doi_str_mv	10.1109/IPIN51156.2021.9662534
format	conference_proceeding
fullrecord	<record><control><sourceid>ieee_CHZPO</sourceid><recordid>TN_cdi_ieee_primary_9662534</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><ieee_id>9662534</ieee_id><sourcerecordid>9662534</sourcerecordid><originalsourceid>FETCH-LOGICAL-i272t-9a36508e003c91e40fbf75511a2be5787b979d067f09877a298de3c1f83f7a123</originalsourceid><addsrcrecordid>eNotj9FKwzAUhqMgOOeeQJC8QOs5SZM0l3XoLFQdouDdSNtTidREmk7w7R24q--_-PjhY-waIUcEe1Nv6yeFqHQuQGButRZKFifsArVWBRQgzClbiMJgZtG8n7NVSp8AgBpRQ7lgddV1-8nNxOvQxzjxbUx-9jH48MFvXaKex8AbclM4zKpNcdwfZBd6_kKjm_0P8cfY05gu2dngxkSrI5fs7f7udf2QNc-bel01mRdGzJl1UisoCUB2FqmAoR2MOjQ40ZIypWmtsT1oM4AtjXHClj3JDodSDsahkEt29f_riWj3PfkvN_3ujuXyD6bcTMA</addsrcrecordid><sourcetype>Publisher</sourcetype><iscdi>true</iscdi><recordtype>conference_proceeding</recordtype></control><display><type>conference_proceeding</type><title>Accurate Indoor Positioning Based on Learned Absolute and Relative Models</title><source>IEEE Xplore All Conference Series</source><creator>Kjellson, Christoffer ; Larsson, Martin ; Astrom, Kalle ; Oskarsson, Magnus</creator><creatorcontrib>Kjellson, Christoffer ; Larsson, Martin ; Astrom, Kalle ; Oskarsson, Magnus</creatorcontrib><description>To improve the accuracy of indoor positioning systems it can be useful to combine different types of sensor data. This paper describes deep learning methods both for estimating absolute positions and for performing pedestrian dead reckoning, and then how to combine the resulting estimates using weighted least squares optimization. The positioning model is based on a custom neural network which uses measurements of received signal strength indication from one instant of time as input. The model for estimating relative positions is on the other hand based on inertial sensors, the accelerometer, magnetometer and gyroscope. The position estimates are then combined using a least squares approach with weights based on the standard deviations of errors in predictions from the used models.</description><identifier>EISSN: 2471-917X</identifier><identifier>EISBN: 1665404027</identifier><identifier>EISBN: 9781665404020</identifier><identifier>DOI: 10.1109/IPIN51156.2021.9662534</identifier><language>eng</language><publisher>IEEE</publisher><subject>Buildings ; Deep learning ; fingerprinting ; indoor positioning ; Neural networks ; Optimization ; PDR ; radio beacons ; sensor fusion ; smartphone ; Task analysis ; Time measurement ; Training</subject><ispartof>2021 International Conference on Indoor Positioning and Indoor Navigation (IPIN), 2021, p.1-8</ispartof><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://ieeexplore.ieee.org/document/9662534$$EHTML$$P50$$Gieee$$H</linktohtml><link.rule.ids>309,310,776,780,785,786,23908,23909,25117,27901,54529,54906</link.rule.ids><linktorsrc>$$Uhttps://ieeexplore.ieee.org/document/9662534$$EView_record_in_IEEE$$FView_record_in_$$GIEEE</linktorsrc></links><search><creatorcontrib>Kjellson, Christoffer</creatorcontrib><creatorcontrib>Larsson, Martin</creatorcontrib><creatorcontrib>Astrom, Kalle</creatorcontrib><creatorcontrib>Oskarsson, Magnus</creatorcontrib><title>Accurate Indoor Positioning Based on Learned Absolute and Relative Models</title><title>2021 International Conference on Indoor Positioning and Indoor Navigation (IPIN)</title><addtitle>IPIN</addtitle><description>To improve the accuracy of indoor positioning systems it can be useful to combine different types of sensor data. This paper describes deep learning methods both for estimating absolute positions and for performing pedestrian dead reckoning, and then how to combine the resulting estimates using weighted least squares optimization. The positioning model is based on a custom neural network which uses measurements of received signal strength indication from one instant of time as input. The model for estimating relative positions is on the other hand based on inertial sensors, the accelerometer, magnetometer and gyroscope. The position estimates are then combined using a least squares approach with weights based on the standard deviations of errors in predictions from the used models.</description><subject>Buildings</subject><subject>Deep learning</subject><subject>fingerprinting</subject><subject>indoor positioning</subject><subject>Neural networks</subject><subject>Optimization</subject><subject>PDR</subject><subject>radio beacons</subject><subject>sensor fusion</subject><subject>smartphone</subject><subject>Task analysis</subject><subject>Time measurement</subject><subject>Training</subject><issn>2471-917X</issn><isbn>1665404027</isbn><isbn>9781665404020</isbn><fulltext>true</fulltext><rsrctype>conference_proceeding</rsrctype><creationdate>2021</creationdate><recordtype>conference_proceeding</recordtype><sourceid>6IE</sourceid><recordid>eNotj9FKwzAUhqMgOOeeQJC8QOs5SZM0l3XoLFQdouDdSNtTidREmk7w7R24q--_-PjhY-waIUcEe1Nv6yeFqHQuQGButRZKFifsArVWBRQgzClbiMJgZtG8n7NVSp8AgBpRQ7lgddV1-8nNxOvQxzjxbUx-9jH48MFvXaKex8AbclM4zKpNcdwfZBd6_kKjm_0P8cfY05gu2dngxkSrI5fs7f7udf2QNc-bel01mRdGzJl1UisoCUB2FqmAoR2MOjQ40ZIypWmtsT1oM4AtjXHClj3JDodSDsahkEt29f_riWj3PfkvN_3ujuXyD6bcTMA</recordid><startdate>20211129</startdate><enddate>20211129</enddate><creator>Kjellson, Christoffer</creator><creator>Larsson, Martin</creator><creator>Astrom, Kalle</creator><creator>Oskarsson, Magnus</creator><general>IEEE</general><scope>6IE</scope><scope>6IL</scope><scope>CBEJK</scope><scope>RIE</scope><scope>RIL</scope></search><sort><creationdate>20211129</creationdate><title>Accurate Indoor Positioning Based on Learned Absolute and Relative Models</title><author>Kjellson, Christoffer ; Larsson, Martin ; Astrom, Kalle ; Oskarsson, Magnus</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-i272t-9a36508e003c91e40fbf75511a2be5787b979d067f09877a298de3c1f83f7a123</frbrgroupid><rsrctype>conference_proceedings</rsrctype><prefilter>conference_proceedings</prefilter><language>eng</language><creationdate>2021</creationdate><topic>Buildings</topic><topic>Deep learning</topic><topic>fingerprinting</topic><topic>indoor positioning</topic><topic>Neural networks</topic><topic>Optimization</topic><topic>PDR</topic><topic>radio beacons</topic><topic>sensor fusion</topic><topic>smartphone</topic><topic>Task analysis</topic><topic>Time measurement</topic><topic>Training</topic><toplevel>online_resources</toplevel><creatorcontrib>Kjellson, Christoffer</creatorcontrib><creatorcontrib>Larsson, Martin</creatorcontrib><creatorcontrib>Astrom, Kalle</creatorcontrib><creatorcontrib>Oskarsson, Magnus</creatorcontrib><collection>IEEE Electronic Library (IEL) Conference Proceedings</collection><collection>IEEE Proceedings Order Plan All Online (POP All Online) 1998-present by volume</collection><collection>IEEE Xplore All Conference Proceedings</collection><collection>IEEE Electronic Library (IEL)</collection><collection>IEEE Proceedings Order Plans (POP All) 1998-Present</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Kjellson, Christoffer</au><au>Larsson, Martin</au><au>Astrom, Kalle</au><au>Oskarsson, Magnus</au><format>book</format><genre>proceeding</genre><ristype>CONF</ristype><atitle>Accurate Indoor Positioning Based on Learned Absolute and Relative Models</atitle><btitle>2021 International Conference on Indoor Positioning and Indoor Navigation (IPIN)</btitle><stitle>IPIN</stitle><date>2021-11-29</date><risdate>2021</risdate><spage>1</spage><epage>8</epage><pages>1-8</pages><eissn>2471-917X</eissn><eisbn>1665404027</eisbn><eisbn>9781665404020</eisbn><abstract>To improve the accuracy of indoor positioning systems it can be useful to combine different types of sensor data. This paper describes deep learning methods both for estimating absolute positions and for performing pedestrian dead reckoning, and then how to combine the resulting estimates using weighted least squares optimization. The positioning model is based on a custom neural network which uses measurements of received signal strength indication from one instant of time as input. The model for estimating relative positions is on the other hand based on inertial sensors, the accelerometer, magnetometer and gyroscope. The position estimates are then combined using a least squares approach with weights based on the standard deviations of errors in predictions from the used models.</abstract><pub>IEEE</pub><doi>10.1109/IPIN51156.2021.9662534</doi><tpages>8</tpages></addata></record>
fulltext	fulltext_linktorsrc
identifier	EISSN: 2471-917X
ispartof	2021 International Conference on Indoor Positioning and Indoor Navigation (IPIN), 2021, p.1-8
issn	2471-917X
language	eng
recordid	cdi_ieee_primary_9662534
source	IEEE Xplore All Conference Series
subjects	Buildings Deep learning fingerprinting indoor positioning Neural networks Optimization PDR radio beacons sensor fusion smartphone Task analysis Time measurement Training
title	Accurate Indoor Positioning Based on Learned Absolute and Relative Models
url	http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-02-24T11%3A33%3A54IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-ieee_CHZPO&rft_val_fmt=info:ofi/fmt:kev:mtx:book&rft.genre=proceeding&rft.atitle=Accurate%20Indoor%20Positioning%20Based%20on%20Learned%20Absolute%20and%20Relative%20Models&rft.btitle=2021%20International%20Conference%20on%20Indoor%20Positioning%20and%20Indoor%20Navigation%20(IPIN)&rft.au=Kjellson,%20Christoffer&rft.date=2021-11-29&rft.spage=1&rft.epage=8&rft.pages=1-8&rft.eissn=2471-917X&rft_id=info:doi/10.1109/IPIN51156.2021.9662534&rft.eisbn=1665404027&rft.eisbn_list=9781665404020&rft_dat=%3Cieee_CHZPO%3E9662534%3C/ieee_CHZPO%3E%3Cgrp_id%3Ecdi_FETCH-LOGICAL-i272t-9a36508e003c91e40fbf75511a2be5787b979d067f09877a298de3c1f83f7a123%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_id=info:pmid/&rft_ieee_id=9662534&rfr_iscdi=true