Espace membre

Cet espace est dédié aux étudiants, aux enseignants et au personnel administratif de l'école

Valider

Mot de passe oublié?

Agenda

18/03/2017

Doctorate thesis defense of Leïla Gazzah Ben Abdeljelil




Doctorate thesis defense on March 18th 2017 at 10H00 AM ,in Amphi I, Sup’Com.


Entitled :Cooperative and Uncooperative Hybrid Localization in NLOS environment

Presented by :  Leïla Gazzah Ben Abdeljelil 





Committee


President

Mr. Ridha Bouallegue

Professor at Sup'Com, Tunisia

 

 

 

Reviewers

Mrs. Monia Turki

Professor at ENIT, Tunisia

 

Mr. Ali Gharsallah

Professor at the Faculty of Sciences of Tunis

 

Examiners

Mr. Hatem Boujemâa

Professor at SUP’Com, Tunisia

 

Supervisor:

Mrs. Leïla Najjar Atallah

Associate Professor at SUP’Com, Tunisia

 

Co-supervisor:

Mr. Hichem Besbes

Professor at SUP’Com, Tunisia

 

Abstract


The problem of source localization is considered, which is of growing interest in several applications in cellular networks as well as in wireless sensor networks. A major source of error in location estimation from available AOA and/or RSS measurements is NLOS propagation which induces both range bias and angular deviation.


This thesis proposes different approaches to mitigate location error caused by NLOS. The first approach is based on node selection technique and can be processed in two different manners : either choose the best higher three RSS or choose all nodes whose RSS exceeds a given threshold. It was proved to be an effective approach to mitigate the estimation bias induced by NLOS propagation. The second approach provides a selective hybrid weighting approach when both angle and range measurements are possibly corrupted due to NLOS propagation. In this approach, LOS are not identified from NLOS measurements. Rather, range bias is mitigated by computing scale factors or weights that adjust the NLOS-corrupted measurements, at the three best nodes, to approximately their LOS values and uncertainty on AOA is accounted for. In the third approach, a discriminative method is composed of LOS links identification step using a hybrid hypothesis test (HHT) followed by an elimination phase of the links identified as NLOS while using the hybrid combination RSS/AOA.


In the case of limited connectivity and dense users, the cooperative localization, involving measurements from different nodes in the network is an interesting alternative to conventional localization. This thesis has several contributions in the cooperative localization framework. First, we propose a cooperative algorithm based on reference node selection technique with a selective weighting iterative LS, involving the best nodes in terms of higher RSS measurements. Secondly, a cooperative MSs localization technique is proposed for the uncooperative emitters case which corresponds to the case where transmitted power is unknown. The MS position is derived in two steps. In a first step, a Selective Hybrid Weighting Least Squares is applied for all MSs in order to obtain a coarse estimation of their positions. In a second step, the MSs final location, considering anchors and other MSs, is obtained by minimizing a function where weights, selected according to the range measurements (between each MS and targeted MS) reliability, ponderate the error function terms. Finally, a cooperative scheme based on P2P communications in limited connectivity context is proposed. It operates by only two nodes after LOS BS identification using HHT.


Keywords


RSS, reliable node selection, AOA, hybrid localization, Cooperation, NLOS mitigation, range weighting, NLOS identification, hybrid hypothesis test, reference node selection, uncooperative emitters.