Espace membre

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


Mot de passe oublié?

Actualités de l'école


Soutenance de doctorat de Kaouther Hedhly

Soutenance de thèse de doctorat de Kaouther Hedhly le 14/09/2019 à 09H00 , à l'amphi Ibn Khaldoun de Sup'Com.

Intitulé : Contributions on Cooperative Localization and Tracking in Wireless Networks

Présentée par : Kaouther Hedhly



M. Ridha Bouallegue          

Professeur, SUP’COM, Université de Carthage, Tunisie





M. Hichem Snoussi            

Professeur, Université de Troyes

  Mme Noura Sellami Professeur, ENIS, Université se Sfax, Tunisie
Examinateur Mme Layla Najjar Atallah       Professeur, SUP’COM, Université de Carthage, Tunisie

Directeur de thèse

Mme Fatma Abelkefi             

Professeur, SUP’COM, Université de Carthage, Tunisie




Co Directeur de thèse

M. Mohamed Siala

M.Mohamed Laaraiedh  

Professeur, SUP’COM, Université de Carthage, Tunisie
Maitre Assistant, ISI


The evolution of information technology and communication systems (ITCS) empowered many location-based applications. Indeed, position finding and tracking have become a key feature and a challenging issue in several fundamental fields in our lives. In this thesis, we investigate the issue of cooperative localization based on a statistical signal processing perspective. We focus on three axes. The first axis is message passing-based cooperative localization within WSNs. Due to the constraint of limited measurements, several ambiguities may produce position estimation errors. Consequently, in this work, two novel techniques for cooperative localization are proposed which are the Evolved Variational Message Passing algorithm, namely E-VMP algorithm, and the Cooperative Robust Geometric Positioning Algorithm, namely C-RGPA algorithm. These two techniques offer more accurate location estimates with lower complexity compared to other cooperative localization techniques

In addition to the position estimate accuracy, the above-mentioned proposed algorithms must take into account the computation complexity in order to further reduce energy consumption for longer battery life. In order to achieve this objective, the second axis proposes novel approaches for cooperative localization which exploit E-VMP and C-RGPA algorithms. These approaches are based on a subset of nodes which belong to the considered WSN. The study of this subset, instead of the whole set, may guarantee similar performance in terms of accuracy and lower complexity. In this framework, Game Theory (GT) gives efficient means for the selection of the adequate nodes subsets which will achieve the localization task. Game theory is based on many analytic tools that allow complex interactions among entities. Indeed, we exploit the cooperative GT for localization based on E-VMP and C-RGPA algorithms. We study the advantages of a novel proposed technique for coalitions formation before the implementation of these algorithms. Secondly, we propose an enhanced version of the Junction Tree (JT) technique in order to resolve localization ambiguities in loopy networks with lower energy consumption.  This proposed approach exploits the E-VMP algorithm and considers some assumptions which are adapted to the localization context. Thirdly, we propose a new technique which is based on E-VMP algorithm and an Edge Participation Probability (EPP) approach for cooperative localization in WSN. We consider distributions over all spanning trees based on a proposed Edge Participation Probabilities. Besides, we propose a specific function for the optimal EPP which results in lower average error rate.

The third axis investigates Kalman and Particle filtering (KF and PF) for cooperative localization and tracking purpose. We apply these techniques to E-VMP and C-RGPA algorithms. The main added value of distributed tracking filters is to guarantee dynamic versions of these two algorithms.

The proposed techniques are evaluated and compared by means of real heterogeneous measurements carried out using ZigBee and OFDM devices and where location-dependent parameters such as RSSI and RTD are exploited. Experiments and realistic simulations reveal that the proposed techniques exhibit good localization accuracy for very low complexity and cost. Moreover, the comparative study shows that distributed particle filter (DPF) provides better performance than KF in terms of positioning accuracy and root-mean square error.