6533b82efe1ef96bd1293094
RESEARCH PRODUCT
A Flexible 4G/5G Control Platform for Fingerprint-based Indoor Localization
Ilenia TinnirelloGian Michele Dell'aeraMichele GucciardoMarco Carettisubject
Artificial neural networkSmart objectsbusiness.industryComputer scienceReal-time computing020206 networking & telecommunications02 engineering and technologyBase stationSoftware0202 electrical engineering electronic engineering information engineeringCellular network020201 artificial intelligence & image processingbusinessClassifier (UML)5Gdescription
In this paper we propose a centralized SDN platform devised to control indoor femto-cells for supporting multiple network-wide optimizations and applications. In particular, we focus on an example localization application in order to enlighten the main functionalities and potentialities of the approach. First, we demonstrate that the platform can be exploited for reconfiguring some operational procedures, based on standard signalling mechanisms, at the programmable femto-cells; these procedures enable customized logics for collecting measurements reports from mobile terminals. Second, assuming that high-density devices such as smart objects are disseminated in the controlled indoor space, we experimentally validate the measurement collection mechanism and the possibility to build indoor radio-maps. Finally, we propose a comparison between a k-nearest neighbors classifier and a trained neural network for supporting fingerprint-based localization. Experimental tests show that the approach can lead to accurate positioning results if the fingerprint maps are updated at regular time intervals (able to capture the dynamic variations of the environment).
| year | journal | country | edition | language |
|---|---|---|---|---|
| 2019-04-01 | IEEE INFOCOM 2019 - IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS) |