Séminaire d'Automatique du plateau de Saclay : A relaxation result for state constrained delay differential inclusion

Séminaire le 28 Février 2017, 10h00 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Ihab Haidar (Laboratoire Quartz, ENSEA-CERGY)


This talk is interested by delay differential inclusions in finite dimensional real space. The celebrated Filippov’s theorem is extended to this case. Then, this theorem is generalized to the case when the state variable is constrained to the closure of an open state subset. Under a new “inward pointing condition”, a relaxation result stating that the set of trajectories lying in the interior of the state constraint is dense in the set of constrained trajectories of the convexified inclusion is shown.
Bio. Ihab Haidar was born in Beirut, Lebanon, in 1983. He received the Master’s degree in mathematics from the University of Aix-Marseille 1, France, in 2008 and the Ph.D degree from the University of Montpellier 2, France, in 2011. Since then he has been post-doc in different places (Laboratoire des Signaux et Systèmes, Institut de Mathématiques de Jussieu, Laboratoire QUARTZ-ENSEA). His research interests include control theory, time delay systems and systems biology.

S³ seminar : FastText: A library for efficient learning of word representations and sentence classification

Séminaire le 24 Février 2017, 10h30 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Piotr Bojanowski, (Facebook AI Research)


In this talk, I will describe FastText, an open-source library that can be used to train word representations or text classifiers. This library is based on our generalization of the famous word2vec model, allowing to adapt it easily to various applications. I will go over the formulation of the skipgram and cbow models of word2vec and how these were extended to meet the needs of our model. I will describe in details the two applications of our model, namely document classification and building morphologically-rich word representations. In both applications, our model achieves very competitive performance while being very simple and fast.

Séminaire d’Automatique du plateau de Saclay : Cooperative Control of Multi-Agents: On a Sphere Manifold and in the Euclidean Space

Séminaire le 13 Février 2017, 11h00 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Wei Li (Department of Control and Systems Engineering, Nanjing University)


The talk will discuss cooperative control of multi-agents on a sphere and in the Euclidean space. We will first consider the control law design of agents on a sphere, and analyze the stability, scaling, and geometry properties, and discuss future directions. Then, for agents evolving in the Euclidean space, we will consider coupled agents with second-order dynamics. The state of a single agent includes both position and velocity, thus generally, the agents have both velocity coupling and position couplings (VCPC); and if we consider different VCPC, then interesting yet difficult problems arise. We then discuss two aspects of analysis on consensus convergence , and future directions.

Bio. Wei Li received the Ph.D. degree in Automatic Control from Shanghai Jiao Tong University, Shanghai, China, in 2008.
From 2009 to 2010, he was a Post-Doctoral Research Associate with the Department of Electrical Engineering, The University of Texas at Dallas, Dallas, TX, USA. Since 2010, he has been an Associate Professor with the Department of Control and Systems Engineering, Nanjing University, Nanjing, China. His current research interests include robotics, autonomous mobile robots, decentralized control, cooperative control of mobile robotic agents, and wireless sensor networks. Dr. Li is an Associate Editor of Asian Journal of Control. He is a Senior Member of IEEE.

Séminaire d’Automatique du plateau de Saclay : Observer synthesis under time-varying sampling for Lipschitz nonlinear systems

Séminaire le 13 Février 2017, 10h00 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Lucien Etienne (L2S, CentraleSupelec)


The problem of observation of continuous-time nonlinear Lipschitz systems under time-varying discrete measurements is studied. This class of systems naturally occurs when continuous processes are observed through digital sensors and information is sent via a network to a computer for state estimation. Since network introduces uncertainties in the sampling time, the observer must be designed so to take these uncertainties into account. Here two classes of observation scheme are studied. First an impulsive observers, which make instantaneous correction when information is received, is considered. Then a Luenberger-like observer with a piece wise constant correction term is studied. For both classes of observer, generic conditions are provided. Then a restriction of the generic conditions is used to establish tractable conditions that allows the synthesis of an observer gain.
Bio. Lucien Etienne received a M.Sc. Degree in Applied Mathematics at the INSA Rouen in 2012 and a joint Ph.D. in Automatic Control from the university of L'Aquila and the university of Cergy-Pontoise in 2016. After a Post-doc at INRIA Lille on observer synthesis for sampled data system, he is currently Post-doc at L2S CentralSupéléc working on switched systems for embedded control under mixed stochastic/deterministic timing uncertainty.

His research interests include switched and hybrid systems, Observer synthesis and sampled data systems.

S³ seminar : Stochastic Quasi-Newton Langevin Monte Carlo

Séminaire le 10 Février 2017, 10h30 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Umut Şimşekli, (LTCI, Télécom ParisTech)


Recently, Stochastic Gradient Markov Chain Monte Carlo (SG-MCMC) methods have been proposed for scaling up Monte Carlo computations to large data problems. Whilst these approaches have proven useful in many applications, vanilla SG-MCMC might suffer from poor mixing rates when random variables exhibit strong couplings under the target densities or big scale differences. In this talk, I will present a novel SG-MCMC method that takes the local geometry into account by using ideas from Quasi-Newton optimization methods. These second order methods directly approximate the inverse Hessian by using a limited history of samples and their gradients. Our method uses dense approximations of the inverse Hessian while keeping the time and memory complexities linear with the dimension of the problem. I will provide formal theoretical analysis where it is shown that the proposed method is asymptotically unbiased and consistent with the posterior expectations. I will finally illustrate the effectiveness of the approach on both synthetic and real datasets. This is a joint work with Roland Badeau, Taylan Cemgil and Gaël Richard. arXiv: https://arxiv.org/abs/1602.03442

S³-PASADENA seminar : Detecting confounding in multivariate linear models via spectral analysis

Séminaire le 31 Janvier 2017, 10h30 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Dominik Janzing, Max Planck Institute for Intelligent Systems, Tuebingen, Germany


We study a model where one target variable Y is correlated
with  a vector X:=(X_1,...,X_d) of predictor variables  being potential causes of Y.
We describe  a method that infers to what extent the statistical dependences between X and Y
are due to the influence of X on Y and to what extent due to a hidden common cause
(confounder) of X and Y. The method is based on an independence assumption stating that, in the absence of confounding,
the vector of regression coefficients describing the influence of each X on Y has 'generic orientation'
relative to the eigenspaces  of the covariance matrix of X. For the special case of a scalar confounder we show that confounding typically spoils this generic orientation in a characteristic way that can be used to quantitatively estimate the amount of confounding.
I also show some encouraging experiments with real data, but the method is work in progress and critical comments are highly appreciated.

Postulating 'generic orientation' is inspired by a more general postulate stating that
P(cause) and P(effect|cause) are independent objects of Nature and therefore don't contain information about each other [1,2,3],
an idea that inspired several causal inference methods already, e.g. [4,5].

[1] Janzing, Schoelkopf: Causal inference using the algorithmic Markov condition, IEEE TIT 2010.
[2] Lemeire, Janzing: Replacing causal faithfulness with the algorithmic independence of conditionals, Minds and Machines, 2012.
[3] Schoelkopf et al: On causal and anticausal learning, ICML 2012.
[4] Janzing et al: Telling cause frome effect based on high-dimensional observations, ICML 2010.
[5] Shajarisales et al: Telling cause from effect in deterministic linear dynamical systems, ICML 2015.

Feedback transformations of underactuated mechanical systems for trajectory planning: case studies in non-prehensile manipulation

Séminaire le 26 Janvier 2017, 14h30 à CentraleSupelec (Gif-sur-Yvette) Amphi F3-09
Anton SHIRIAEV (Department of Engineering Cybernetics, NTNU, Norway)


The talk is aimed at discussion of challenges present in developing model based trajectory planning algorithms for underactuated mechanical systems. Nonlinearity of system’s dynamics and presence of one or several passive degrees of freedom are among several structural properties that are difficult to handle in a trajectory search. Other challenges are related to different formats of representation of feasible trajectories, where some choices can be better suited for further stabilization or advantageous in sensitivity analysis with respect to uncertainty in system parameters. The author is interested to emphasize the attention on two relatively new points in the problem that have recently helped in solving a series of longstanding manipulation problems in robotics: change of coordinates and feedback transformation for mechanical systems are useful to formulate as dependent on an individual trajectory even though such a feasible behavior is not found yet. The arguments are illustrated by new analytical results and the case study in non-prehensile manipulation.

Feedback transformations of underactuated mechanical systems for trajectory planning: case studies in non-prehensile manipulation

Séminaire le 26 Janvier 2017, 14h30 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Anton SHIRIAEV (Department of Engineering Cybernetics, NTNU, Norway)


The talk is aimed at discussion of challenges present in developing model based trajectory planning algorithms for underactuated mechanical systems. Nonlinearity of system’s dynamics and presence of one or several passive degrees of freedom are among several structural properties that are difficult to handle in a trajectory search. Other challenges are related to different formats of representation of feasible trajectories, where some choices can be better suited for further stabilization or advantageous in sensitivity analysis with respect to uncertainty in system parameters. The author is interested to emphasize the attention on two relatively new points in the problem that have recently helped in solving a series of longstanding manipulation problems in robotics: change of coordinates and feedback transformation for mechanical systems are useful to formulate as dependent on an individual trajectory even though such a feasible behavior is not found yet. The arguments are illustrated by new analytical results and the case study in non-prehensile manipulation.

Séminaire d’Automatique du plateau de Saclay : What drives the quality of local public goods in Africa? Disentangling social capital and ethnic divisions

Séminaire le 24 Janvier 2017, 11h00 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Guillaume Hollard (Département d’Economie, Ecole polytechnique)


Two important lines of research shaped our understanding of the ability of communities to engage in collective action. The first line proposes ethnic division as a key determinant, with more ethnically heterogeneous countries having lower economic performances and levels of public goods. Thus, we expect to find better schools where ethnic fractionalization is low. The second line of research focuses on social capital as a major determinant of the ability to engage in collective action.We expect that trust among community members, a widely-used measure of social capital, is an important and positive determinant of school quality. The present work aims to disentangle the relative effects of ethnic fractionalization and social capital on school quality. We use instrumental variable estimations to address reverse causality and other endogeneity issues. We instrument both social capital and ethnic fractionalization by using historical information on the settlement patterns of ethnic groups in Sub-Saharan Africa. Our empirical strategy is implemented by combining four datasets, including Afrobarometer, covering 16 Sub-Saharan countries. We run our analysis at the district level, with more than 1000 districts covered. We find an important and positive effect of trust on the practical aspects of schooling, such as maintaining buildings or providing textbooks. A one percent increase in the level of trust increases the quality of local public goods by 0.18 to 1.05 percent, depending on the measure of school quality under consideration. In sharp contrast, ethnic fractionalization is found to have a very limited effect, if any. We propose a simple model of public good provision that explores a channel by which social capital and ethnic division may (or may not) have an impact on the provision of local public goods such as schools. Our results suggest that policies designed to enhance social capital are likely to have a positive effect on schools and local public goods in general.

Bio. Directeur de recherche au CNRS et professeur associé à l'école polytechnique. Spécialisé dans l'analyse de la décision et du comportement. Responsable du laboratoire d'économie expérimentale.

Séminaire d’Automatique du plateau de Saclay : Moral hazard with mean field type interactions

Séminaire le 24 Janvier 2017, 10h00 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Thibault Mastrolia (CMAP, Ecole Polytechnique)


We investigate a moral hazard problem in finite time with lump-sum and continuous payments, involving infinitely many Agents, with mean field type interactions, hired by one Principal. By reinterpreting the mean-field game faced by each Agent in terms of a mean field FBSDE, we are able to rewrite the Principal’s problem as a control problem for McKean-Vlasov SDEs. We solve completely and explicitly the problem in special cases, going beyond the usual linear-quadratic framework.

Bio. Après avoir effectué un magistère de mathématiques à l'université de Strasbourg puis le M2 MASEF de l'université Paris-Dauphine, j'ai poursuivi mes études par un doctorat au sein de cette université sous la direction d'Anthony Réveillac et de Dylan Possamaï autour du calcul de Malliavin, des EDSR et de leurs applications en finance. J'ai ensuite été recruté comme Maître de conférences en probabilités et mathématiques financières à l'Ecole Polytechnique. Actuellement, je travaille autour de problèmes de théorie des contrats et de leurs applications.

S³ Seminar: Adapting to unknown noise level in super-resolution

Séminaire le 20 Janvier 2017, 10h30 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Claire Boyer (LSTA, UPMC)


We study sparse spikes deconvolution over the space of complex-valued measures when the input measure is a finite sum of Dirac masses. We introduce a new procedure to handle the spike deconvolution when the noise level is unknown. Prediction and localization results will be presented for this approach. An insight on the probabilistic tools used in the proofs could be briefly given as well.

S³ seminar : Inverse problems for speech production

Séminaire le 20 Janvier 2017, 10h30 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Benjamin Elie (LORIA, IADI)


Studies on speech production are based on the extraction and the analysis of the acoustic features of human speech, and also on their relationships with the articulatory and phonatory configurations realized by the speaker. An interesting tool, which will be the topic of the talk, to make such researches is the articulatory synthesis, which consists in the numerical simulation of the mechanical and acoustical phenomena that are involved in speech production. The aim is to numerically reproduce a speech signal that contains the observed acoustic features with regards to the actual articulatory and phonatory gestures of the speaker. Using the articulatory approach may lead to a few problems that will be tackled in this talk, and to which possible solutions will be discussed. Firstly, the different articulatory gestures realized in natural speech should be precisely observed. For that purpose, the first part of the talk focuses on methods to acquire articulatory films of the vocal tract by MRI techniques with a fast acquisition rate via sparse techniques (Compressed Sensing). The aim is, in fine, to build an articulatory and a coarticulation model. The investigation of the acoustical phenomena involved in natural speech require to separate the contributions of the different acoustic sources in the speech signal. The periodic/aperiodic decomposition of the speech signal is the subject of the second part of the talk. The challenge is to be able to study the acoustic properties of the frication noise that is generated during the production of fricatives, and also to quantify the amount of voicing produced during fricatives. Finally, in order to directly use the analysis by synthesis methods, it is interesting to estimate the articulatory configurations of the speaker from the acoustic signal. This is the aim of the acoustic-articulatory inversion for copy synthesis, which is the third part of the talk. Direct applications of these problems for the study of speech production and phonetics will be presented.

Séminaire d’Automatique du plateau de Saclay : Fokker-Planck optimal control for stochastic processes

Séminaire le 13 Décembre 2016, 11h00 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Mario Annunziato (Università degli Studi di Salerno)


An innovative framework for the control of stochastic process by means of an optimization problem on the Fokker-Planck equation is presented. The time dependent probability density function (PDF)  as representative of the dynamical state of a stochastic system is used, hence the governing Kolmogorov-Fokker-Planck-type (KFP) equation is employed as a constraint for the minimization of a cost function. The problem to find a controller that minimizes the cost function can be solved by solving an optimality system of time dependent forward and backward partial differential equations. A short review of control objectives, KFP equations and numerical techniques to tackle the optimization problem, is shown by using models from biology, physics, and finance.

Bio. Mario Annunziato is a researcher in Mathematics, in the field of Numerical analysis at "Università degli Studi di Salerno" since 2004. He is also a member of "Gruppo Nazionale di Calcolo Scientifico, Instituto Nazionale di Alta Mathemaica". He has received his Ph.D. degree in Physics at " Università degli Studi di Pisa" in 2000. He obtained a degree of laurea in Physics at "Rome University - La Sapienza" in 1995.

His research interests focus on numerical solutions of time dependent Partial Differential Equations (PDE) and Integral Equations, related to stochastic processes and stochastic optimal control.

Performances et méthodes pour l'échantillonnage comprimé: Robustesse à la méconnaissance du dictionnaire et optimisation du noyau d'échantillonnage

Stéphanie BERNHARDT
Soutenance de thèse de doctorat le 5 Décembre 2016, 14h00 à CentraleSupelec (Gif-sur-Yvette) Amphi F3-05

Dans cette thèse, nous nous intéressons à deux méthodes promettant de reconstruire un signal parcimonieux largement sous-échantillonné : l’échantillonnage de signaux à taux d’innovation fini et l’acquisition comprimée. Il a été montré récemment qu’en utilisant un noyau de pré-filtrage adapté, les signaux impulsionnels peuvent être parfaitement reconstruits bien qu’ils soient à bande non-limitée. En présence de bruit, la reconstruction est réalisée par une procédure d’estimation de tous les paramètres du signal d’intérêt. Dans cette thèse, nous considérons premièrement l’estimation des amplitudes et retards paramétrisant une somme finie d'impulsions de Dirac filtrée par un noyau quelconque et deuxièmement l’estimation d’une somme d’impulsions de forme quelconque filtrée par un noyau en somme de sinus cardinaux (SoS). Le noyau SoS est intéressant car il est paramétrable par un jeu de paramètres à valeurs complexes et vérifie les conditions nécessaires à la reconstruction. En se basant sur l’information de Fisher Bayésienne relative aux paramètres d’amplitudes et de retards et sur des outils d’optimisation convexe, nous proposons un nouveau noyau d’échantillonnage. L’acquisition comprimée permet d’échantillonner un signal en-dessous de la fréquence d’échantillonnage de Shannon, si le vecteur à échantillonner peut être approximé comme une combinaison linéaire d’un nombre réduit de vecteurs extraits d’un dictionnaire sur-complet. Malheureusement, dans des conditions réalistes, le dictionnaire (ou base) n’est souvent pas parfaitement connu, et est donc entaché d’une erreur (DB). L’estimation par dictionnaire, se basant sur les mêmes principes, permet d’estimer des paramètres à valeurs continues en les associant selon une grille partitionnant l’espace des paramètres. Généralement, les paramètres ne se trouvent pas sur la grille, ce qui induit un erreur d’estimation même à haut rapport signal sur bruit (RSB). C’est le problème de l’erreur de grille (EG). Dans cette thèse nous étudions les conséquences des modèles d’erreur DB et EG en terme de performances bayésiennes et montrons qu’un biais est introduit même avec une estimation parfaite du support et à haut RSB. La BCRB est dérivée pour les modèles DB et EG non structurés, qui bien qu’ils soient très proches, ne sont pas équivalents en terme de performances. Nous donnons également la borne de Cramér-Rao moyennée (BCRM) dans le cas d’une petite erreur de grille et étudions l’expression analytique de l’erreur quadratique moyenne bayésienne (BEQM) sur l’estimation de l’erreur de grille à haut RSB. Cette dernière est confirmée en pratique dans le contexte de l’estimation de fréquence pour différents algorithmes de reconstruction parcimonieuse. Nous proposons deux nouveaux estimateurs : le Bias-Correction Estimator (BiCE) et l’Off-Grid Error Correction (OGEC) permettant de corriger l'erreur de modèle induite par les erreurs DB et EG, respectivement. Ces deux estimateurs principalement basés sur une projection oblique des mesures sont conçus comme des post-traitements, destinés à réduire le biais d’estimation suite à une pré-estimation effectuée par n’importe quel algorithme de reconstruction parcimonieuse. Les biais et variances théoriques du BiCE et du OGEC sont dérivés afin de caractériser leurs efficacités statistiques. Nous montrons, dans le contexte difficile de l’échantillonnage des signaux impulsionnels à bande non-limitée que ces deux estimateurs permettent de réduire considérablement l’effet de l'erreur de modèle sur les performances d’estimation. Les estimateurs BiCE et OGEC sont tout deux des schémas (i) génériques, car ils peuvent être associés à tout estimateur parcimonieux de la littérature, (ii) rapides, car leur coût de calcul reste faible comparativement au coût des estimateurs parcimonieux, et (iii) ont de bonnes propriétés statistiques.

 

Mots-clés :

échantillonnage, parcimonie, erreur de modèle, bornes bayésiennes, noyaux, signaux impulsionnels

 

Composition du jury

M. Rémy BOYER Université Paris-Sud Directeur de thèse

Mme Sylvie MARCOS CNRS Co-Directeur de thèse

M. Pascal LARZABAL Université Paris-Sud Co-Encadrant de thèse

M. David BRIE Université de Lorraine Rapporteur

M. André FERRARI Université de Côte d'Azur Rapporteur

M. Eric CHAUMETTE ISAE-Supaéro Examinateur

M. Ali MOHAMMAD-DJAFARI CNRS Examinateur

M. Nicolas DOBIGEON Université de Toulouse Examinateur

Broadcasting with delayed CSIT: finite SNR analysis and heterogeneous feedback

Chao HE
Soutenance de thèse de doctorat le 2 Décembre 2016, 14h00 à CentraleSupelec (Gif-sur-Yvette) Amphi F3-05

One of the key techniques for future wireless networks, namely state-feedback enabled interference mitigation, is explored with information theoretical tools under some realistic settings. In order to verify the usefulness of delayed CSIT in harsh situations, we investigate it with finite SNR and/or feedback heterogeneity in a broadcast communication setup. In the first part of this dissertation, we are mainly interested in the finite SNR performance of delayed CSIT and the uniformity across the results in Gaussian channel and in erasure channel. Several relatively simple schemes are proposed in multi-user broadcast channels when states are assumed to be fully known at the destinations but only strictly causally at the transmitter. Enhancement in terms of higher symmetric rates and more input alphabet options are then characterized when analyzing the corresponding regions in Gaussian/erasure cases. In the second part of this dissertation, algorithms adapted to distinct feedback heterogeneities are targeted as only part of communication nodes are involved in the feedback process. In particular, we concentrate on 1) broadcast channel with feedback from partial receivers, 2) broadcast relay channel with feedback at the relay. Improvements on achievable rates are justified in either cases via analysis and some examples.

 

Devant le jury composé de :

 

Michel KIEFFER                         Professeur (Université Paris-Sud)                      Président du jury

Inbar FIJALKOW                        Professeur (ENSEA)                                           Rapporteur

Albert Guillén i FABREGAS       Professeur (Universitat Pompeu Fabra)             Rapporteur

David GESBERT                        Professeur (EURECOM)                                     Examinateur

Michèle Angela WIGGER           Professeur associé (Télécom ParisTech)            Examinateur

Samir M. PERLAZA                   Chargé de recherche (INRIA Lyon)                     Examinateur

Sheng YANG                              Enseignant-Chercheur (CentraleSupélec)           Directeur de thèse

Pablo PIANTANIDA                   Enseignant-Chercheur (CentraleSupélec)           Co-encadrant

Efficacité énergétique et spectrale dans les réseaux 5G

Iyad LAHSEN-CHERIF
Soutenance de thèse de doctorat le 2 Décembre 2016, 10h00 à CentraleSupelec (Gif-sur-Yvette) Salle des séminaires du L2S

La pénurie d'énergie et le manque d'infrastructures dans les régions rurales représentent une barrière pour le déploiement et l'extension des réseaux cellulaires. Les approches et techniques pour relier les stations de base (BSs) entre elles à faible coût et d'une manière fiable et efficace énergiquement sont l'une des priorités des opérateurs. Ces réseaux peu denses actuellement, peuvent évoluer rapidement et affronter une croissance exponentielle due principalement à l'utilisation des téléphones mobiles, tablettes et applications gourmandes en bande passante. La densification des réseaux est l'une des solutions efficaces pour répondre à ce besoin en débit élevé. Certes, l'introduction de petites BSs apporte de nombreux avantages tels que l'amélioration du débit et de la qualité du signal, mais entraîne des contraintes opérationnelles telles que le choix de l'emplacement des noeuds dans ces réseaux de plus en plus denses ainsi que leur alimentation. Les problèmes où la contrainte spatiale est prépondérante sont bien appropriés à la modélisation par la géométrie stochastique qui permet une modélisation réaliste de distribution des BSs. Ainsi, l'enjeu est de trouver de nouvelles approches de gestions d'interférence et de réductions de consommation énergétique dans les réseaux sans fil. Le premier axe de cette thèse s'intéresse aux méthodes de gestion d'interférence dans les réseaux cellulaires se basant sur la coordination entre les BSs, plus précisément, la technique Coordinated MultiPoint Joint Transmission (CoMP-JT). En CoMP-JT, les utilisateurs en bordure de cellules qui subissent un niveau très élevé d'interférences reçoivent plusieurs copies du signal utile de la part des BSs qui forment l'ensemble de coordination. Ainsi, nous utilisons le modèle $r$-$l$ Square Point Process (PP) à fin de modéliser la distribution des BSs dans le plan. Le processus $r$-$l$ Square PP est le plus adapté pour modéliser le déploiement réel des BSs d'un réseaux sans fil, en assurant une distance minimale, ($r$ - $l$), entre les points du processus. Nous discutons l'impact de la taille de l'ensemble de coordination sur les performances évaluées. Ce travail est étendu pour les réseaux denses WiFi IEEE 802.11, où les contraintes de portées de transmission et de détection de porteuse ont été prises en compte. Dans le deuxième axe du travail, nous nous intéressons à l'efficacité énergétique des réseaux mesh. Nous proposons l'utilisation des antennes directionnelles (DAs) pour réduire la consommation énergétique et améliorer le débit de ces réseaux mesh. Les DAs ont la capacité de focaliser la transmission dans la direction du récepteur, assurant une portée plus importante et moins d'énergie dissipée dans toutes les directions. Pour différentes topologies, nous dérivons le nombre de liens et montrons que ce nombre dépend du nombre de secteurs de l'antenne. Ainsi, en utilisant les simulations, nous montrons que le gain, en énergie et en débit, apporté par les DAs peut atteindre 70% dans certains cas. De plus, on propose un modèle d'optimisation conjointe d'énergie et du débit adapté aux réseaux WMNs équipés de DAs. La résolution numérique de ce modèle confortent les résultats de simulation obtenus dans la première partie de cette étude sur l'impact des DAs sur les performances du réseaux en termes de débit et d'énergie consommée. Ces travaux de thèse s'inscrivent dans le cadre du projet collaboratif (FUI16 LCI4D), qui consiste à concevoir et à valider une architecture radio ouverte pour renforcer l'accès aux services broadband dans des lieux ne disposant que d'une couverture minimale assurée par un réseau macro-cellulaire traditionnel.

Mots-clés :  Coordinated MultiPoint, Géométrie stochastique , Réseaux mesh, Optimisation

Composition du jury

M. André-Luc BEYLOT        Professeur à l'ENSEEIHT, Toulouse     Rapporteur
M. Anthony BUSSON          Professeur à l'Université de Lyon 1     Examinateur
M. Bernard COUSIN            Professeur à l'Université de Rennes1     Examinateur
M. Steven MARTIN              Professeur à l'Université Paris 11     Examinateur
Mme Lynda MOKDAD         Professeur à l'Université Paris 12     Rapporteur
Mme Véronique VEQUE      Professeur à l'Université Paris 11     Directeur de thèse
Mme Lynda ZITOUNE          Maître de Conférence à l'ESIEE, Paris     Co-encadrant de thèse

S³ seminar: High dimensional sampling with the Unadjusted Langevin Algorithm

Séminaire le 23 Novembre 2016, 10h30 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Alain Durmus (LTCI, Telecom ParisTech)


Recently, the problem of designing MCMC sampler adapted to high-dimensional distributions and with sensible theoretical guarantees has received a lot of interest. The applications are numerous, including large-scale inference in machine learning,  Bayesian nonparametrics, Bayesian inverse problem, aggregation of experts among others. When the density is L-smooth (the log-density is continuously differentiable and its derivative is Lipshitz), we will advocate the use of a “rejection-free” algorithm, based on the discretization of the  Euler diffusion with either constant or decreasing stepsizes. We will present several new results allowing convergence to stationarity under different conditions for the log-density (from the  weakest, bounded oscillations on a compact set and super-exponential in the tails to the log concave).
When the density is strongly log-concave, the convergence of an appropriately weighted empirical measure is also investigated and bounds for the mean square error and exponential deviation inequality for Lipschitz functions will be reported.
Finally, based on optimzation techniques we will propose new methods to sample from high dimensional distributions. In particular, we will be interested  in densities which are not continuously differentiable. Some Monte Carlo experiments will be presented to support our findings.

Séminaire d'Automatique du Plateau de Saclay : Extending spacecraft operational life: Challenges and opportunities for control

Séminaire le 15 Novembre 2016, 11h00 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Prof. Ilya Kolmanovsky (University of Michigan)


To extend spacecraft operational life, control techniques that can accommodate actuator failures, reduce the use of fuel and avoid collisions with debris are of interest. Such techniques may need to exploit “higher order” physical effects, such as forces and torques normally considered as disturbances, approaches that can take advantage of nonlinearities in spacecraft kinematics and dynamics, and handle stationary and moving obstacle avoidance requirements. Hence spacecraft operational life extension problems create many potential opportunities for the application of nonlinear, optimal and constrained/predictive control.

After general remarks on control challenges and opportunities in spacecraft operational life extension problems, the presentation will focus on related recent case studies. 

In particular, it will be shown that for a spacecraft with only two functioning reaction wheels linear controllability of attitude dynamics can be regained, under appropriate assumptions, if solar radiation pressure torques are included in the analysis. This conclusion can be exploited for handling reaction wheel failures based on conventional linear controllers.  Alternative approaches that do not rely on the solar radiation pressure torques but exploit nonlinearities in the spacecraft kinematics and dynamics will also be discussed. Furthermore, an intriguing capability of model predictive controllers to achieve discontinuous stabilization in underactuated spacecraft attitude control problems will be highlighted. We will also touch upon coupled translational and rotational relative motion dynamics of a rigid spacecraft in central gravity field and potential opportunities to move translationally in space by employing attitude control only.  In the final part of the presentation, drift counteraction optimal control problems for systems with drift and/or large disturbances will be introduced, in which the objective is to maximize the time for a system to violate prescribed constraints. Potential applications of drift counteraction optimal control, including for geostationary satellite orbit maintenance and drag induced orbit decay compensation, will be discussed.

Distributed Information Gathering and Estimation in Wireless Sensor Networks

Wenjie LI
Soutenance de thèse de doctorat le 15 Novembre 2016, 10h00 à CentraleSupelec (Gif-sur-Yvette) Amphi F3-06

Wireless sensor networks (WSNs) have attracted much interests in the last decade. The first part of this thesis considers sparse random linear network coding is for data gathering and compression in WSNs. An information-theoretic approach is applied to demonstrate the necessary and sufficient conditions to realize the asymptotically perfect reconstruction under MAP estimation. The second part of the thesis concerns the distributed self-rating (DSR) problem, for WSNs with nodes that have different ability of performing some task (sensing, detection...). The main assumption is that each node does not know and needs to estimate its ability. Depending on the number of ability levels and the communication conditions, three sub-problems have been addressed: i) distributed faulty node detection (DFD) to identify the nodes equipped with defective sensors in dense WSNs; ii) DFD in delay tolerant networks (DTNs) with sparse and intermittent connectivity; iii) DSR using pairwise comparison. Distributed algorithms have been proposed and analyzed. Theoretical results assess the effectiveness of the proposed solution and give guidelines in the design of the algorithm.

Membres du jury :

M. Cédric RICHARD, Professeur (Université de Nice Sophia Antipolis),  Rapporteur
M. Jean-Marie GORCE, Professeur (INSA Lyon), Rapporteur
Mme. Béatrice LAROCHE,  Directeur de recherche (INRA), Examinatrice
M. Fabio FAGNANI,  Professor (Politecnico di Torino), Examinateur
M. Davide DARDARI, Associate Professor (University of Bologna), Examinateur
M. Michel KIEFFER,  Professeur (Université Paris-Sud), Directeur de thèse
Mme. Francesca BASSI, Enseignant-Chercheur (ESME-Sudria), Co-encadrant de thèse

Contributions à l’étude des systèmes non linéaires incertains : Application à la commande de systèmes biotechnologiques

Sihem TEBBANI
Habilitation à Diriger des Recherches (HDR) le 7 Novembre 2016, 14h00 à

L’étude des systèmes non linéaires, et plus particulièrement leur commande, peut être complexe. De plus, les incertitudes sur leurs modèles engendrent généralement une détérioration des performances des lois de commande développées. Cet exposé présentera nos contributions dans cette thématique en termes de modélisation, estimation et commande de systèmes non linéaires incertains. En considérant les incertitudes paramétriques du modèle, nous avons proposé des stratégies robustes vis-à-vis de ces incertitudes. Les principales approches étudiées pour l’estimation sont les observateurs par intervalles et à horizon glissant, et pour la commande, la commande prédictive non linéaire.

Plus spécifiquement, l’exposé mettra l’accent sur nos contributions à la commande de la culture de microorganismes, principalement pour des applications pour le développement durable. Dans tous les cas de culture de microalgues et de bactéries étudiées, nous avons proposé des stratégies efficaces, permettant de maximiser soit la croissance du microorganisme ou la production d’un composé à forte valeur ajoutée, et validées expérimentalement dans certains cas.

Membres du Jury :

M. Olivier BERNARD, INRIA, rapporteur

M. Didier DUMUR, CentraleSupélec, L2S, examinateur

M. Jérôme HARMAND, INRA, rapporteur

M. Nicolas LANGLOIS, ESIGELEC, IRSEEM, examinateur

M. Hugues MOUNIER, Université Paris-Sud, L2S, examinateur

Mme. Isabelle QUEINNEC, CNRS, rapporteur

M. Dan SELISTEANU, Université de Craiova, examinateur

Pages