Name
Title | ||
|---|---|---|
Enriched Training Database For Improving The Wifi Rssi-Based Indoor Fingerprinting Performance |
Abstract | ||
|---|---|---|
The interest for RF-based indoor localization, and in particular for WiFi RSSI-based fingerprinting, is growing at a rapid pace. This is despite the existence of a trade-off between the accuracy of location estimation and the density of a laborious and time consuming survey for collecting training fingerprints. A generally accepted concept of increasing the density of a training dataset, without an increase in the amount of physical labor and time needed for surveying an environment for additional fingerprints, is to leverage a propagation model for the generation of virtual training fingerprints. This process, however, burdens the user with an overhead in terms of implementing a propagation model, defining locations of virtual training fingerprints, generating virtual fingerprints, and storing the generated fingerprints in a training database. To address this issue, we propose the Enriched Training Database (ETD), a web-service that enables storage and management of training fingerprints, with an additional "enriching" functionality. The user can leverage the enriching functionality to automatically generate virtual training fingerprints based on propagation modeling in the virtual training points. We further propose a novel method for defining locations of virtual training fingerprints based on modified Voronoi diagrams, which removes the burden of defining virtual training points manually and which automatically "covers" the regions without sufficient density of training fingerprints. The evaluation in our testbed shows that the use of automated generation of virtual training fingerprints in ETD results in more than 25% increase in point accuracy and 15% in room-level accuracy of fingerprinting. |
Year | Venue | Field |
|---|---|---|
2016 | 2016 13TH IEEE ANNUAL CONSUMER COMMUNICATIONS & NETWORKING CONFERENCE (CCNC) | Data mining,Pace,Computer science,Testbed,Virtual training,Voronoi diagram,Database |
DocType | Citations | PageRank |
Conference | 0 | 0.34 |
References | Authors | |
14 | 5 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| filip lemie | 1 | 81 | 11.76 |
| vlado handziski | 2 | 570 | 50.64 |
| Giuseppe Caso | 3 | 47 | 8.98 |
| Luca De Nardis | 4 | 179 | 22.72 |
| Adam Wolisz | 5 | 2693 | 407.71 |