pages_top_content_img_9.jpg
 Logo_JS18_Copie_3.jpg

Géolocalisation et navigation

28-29 mars, 2018
Observatoire de Paris

Site de Meudon

Rechercher > Par auteur > Elgui Kevin

A semi parametric model for RSSI-based localization
Kevin Elgui  1@  , Pascal Bianchi  1, *@  
1 : TELECOM & Management SudParis (Institut Mines-Télécom)
Télécom ParisTech
9, rue Charles Fourier, F-91011 Evry -  France
* : Auteur correspondant

Accurate and reliable geolocation of an object in the context of an Internet of Things (IoT) network must deal with
a lack of available information. The approach followed in this paper is based on Received Signal Strength Indicators
(RSSI) measured at all base stations (BS) as a useful information to infer the object's position. The proposed technique is
based on a maximum a posteriori (MAP) estimator derived within a flexible semi-parametric model. The non-parametric
components of the model are estimated with the help of Nadaraya-Watson type estimators. Our proposal has been tested
and compared with an other method found in the literature: a simple model consisting of taking a weighted barycenter
of the BS as the estimator of the real position. Our approach, although computationally less efficient, has revealed
to outperform other methods and to come up with some very interesting information on the behaviour of the studied
variables.


Personnes connectées : 1