| 期刊名称： |
Sensors |
| 全部作者： |
Tao Liu,Xing Zhang*,Qingquan Li,Zhixiang Fang |
| 出版年份： |
2017 |
| 卷 号： |
17 |
| 期 号： |
8 |
| 页 码： |
1790 |
| 查看全本： |
|
Localizationofusersinindoorspacesisacommonissueinmanyapplications.Amongvarioustechnologies,aWi-Fifingerprintingbasedlocalizationsolutionhasattractedmuchattention,sinceitcanbeeasilydeployedusingtheexistingoff-the-shelfmobiledevicesandwirelessnetworks.However,thecollectionoftheWi-Firadiomapisquitelabor-intensive,whichlimitsitspotentialforlarge-scaleapplication.Inthispaper,avisual-basedapproachisproposedfortheconstructionofaradiomapinanonymousindoorenvironments.Thisapproachcollectsmulti-sensordata,e.g.,Wi-Fisignals,videoframes,inertialreadings,whenpeoplearewalkinginindoorenvironmentswithsmartphonesintheirhands.Then,itspatiallyrecoversthetrajectoriesofpeoplebyusingbothvisualandinertialinformation.Finally,itestimatesthelocationoffingerprintsfromthetrajectoriesandconstructsaWi-Firadiomap.Experimentresultsshowthattheaveragelocationerrorofthefingerprintsisabout0.53m.Aweightedk-nearestneighbormethodisalsousedtoevaluatetheconstructedradiomap.Theaveragelocalizationerrorisabout3.2m,indicatingthatthequalityoftheconstructedradiomapisatthesamelevelasthoseconstructedbysitesurveying.However,thisapproachcangreatlyreducethehumanlaborcost,whichincreasesthepotentialforapplyingittolargeindoorenvironments.
