| 期刊名称： |
Sensors |
| 全部作者： |
undefined |
| 出版年份： |
2017 |
| 卷 号： |
17 |
| 期 号： |
|
| 页 码： |
|
| 查看全本： |
|
Spatialrelationshipsarecrucialtospatialknowledgerepresentation,suchaspositioninglocalities.However,minimalattentionhasbeendevotedtopositioninglocalitiesindoorswithlocalitydescription.Distanceanddirectionrelationsaregenerallyusedwhenpositioninglocalities,namely,translatingdescriptivelocalitiesintospatiallyexplicitones.Weproposeajointprobabilityfunctiontomodellocalitydistributiontoaddresstheuncertaintyofpositioninglocalities.Thejointprobabilityfunctionconsistsofdistanceandrelativedirectionmembershipfunctions.Weproposedefinitionsandrestrictionsfortheuseofthejointprobabilityfunctiontomakethelocalitydistributionhighlypractical.Wealsoevaluatetheperformanceofourapproachthroughindoorexperiments.Testresultsdemonstratethatapositioningaccuracyof3.5mcanbeachievedwiththesemanticallyderivedspatialrelationships.
