Hierarchical control scheme for multi-terminal high voltage direct current power networks.

Soutenance de thèse de doctorat le 7 Avril 2015, 10h30 à CentraleSupelec (Gif-sur-Yvette) Amphi F3-06

This thesis focuses on the hierarchical control for a multi-terminal high voltage direct current (MT-HVDC) grid suitable for the integration of large scale renewable energy sources. The proposed control scheme is composed of 4 layers, from the low local control at the power converters in the time scale of units of ms; through distributed droop control (primary control) applied in several terminals in the scale of unit of seconds; and then to communication based Model Predictive Control (MPC) that assures the load flow and the steady state voltage/power plan for the whole system, manage large scale storage and include weather forecast (secondary control); finally reaching the higher level controller that is mostly based on optimization techniques, where economic aspects are considered in the same time as longer timespan weather forecast (tertiary control).

Concerning the converters' level, special emphasis is placed on DC/DC bidirectional converters. The main task of these devices is to link several DC grids with different voltages, in analogous form as the use of transformers for AC grids. In this thesis, three different topologies are studied in depth: two phases dual active bridge (DAB), the three phases DAB, and the use of the Modular Multilevel Converter (MMC) technology as DC/DC converter. For each topology a specific non-linear control is presented and discussed. In addition, the DC/DC converter can provide other important services as its use as a direct current circuit breaker (DC-CB), which is a capital device for the future development of MT-HVDC networks. This is possible thanks to the fact that the DC-DC converters studied here include an AC stage, and therefore there exist instants in which the current passes through zero, and consequently we can open the switches when a fault occurs in the network in a safer way. Several operation strategies are studied for these topologies used as DC-CB.

With respect to primary control, which is the responsible to maintain the DC voltage control of the grid, we have studied several control philosophies: master/slave, voltage margin control and droop control. Finally we have chosen to use droop control, among other reasons, because the communication between nodes is not required. Two different approaches have been studied for the droop control. Firstly, we have considered that dynamics of converters (AC/DC) are negligible (too fast compared to the network), and in a second step, based on these _rst results, we have studied the dynamics of droop control coupled to the AC/DC converters. Voltage source converters (VSC) are used as AC/DC converters in this approach.

Relative to the secondary control, its main goal is to schedule power transfer between the network nodes providing voltage and power references to local and primary controllers, providing steady state response to disturbances and managing power reserves. In this part we have proposed a new approach to solve the power flow problem (non-linear equations) based on the contraction mapping theorem, which gives the possibility to use more than one bus for the power balance (slack bus) instead of the classic approach based on the Newton-Raphson (NR) method. In addition with the method prosed in this thesis the unique existence of solution is guarantee when some feasible constraints are fulfilled. Secondary control plays a very important role in practical applications, in particular when including time varying power sources, as renewable ones. In such cases, it is interesting to consider storage devices in order to improve the stability and the efficiency of the whole system. Due to the sample time of secondary control is on the order of minutes, it is also possible to consider different kinds of forecast (weather, load,..) and to achieve additional control objectives, based on managing storage reserves. All these characteristics encourage the use of a model predictive control (MPC) approach to design this task. In this context, several possibilities of optimization objective were considered, like to minimize transmission losses or to avoid power network congestions.

The main task of tertiary control is to manage the load flow of the whole HVDC grid in order to achieve economical optimization, especially relevant with the presence of storage devices. This control level provides power references to the secondary controller. In this thesis we were able to maximize the economic profit of the system by acting on the spot market, and by optimizing the use of storage devices. In this level it is again used the MPC approach, but acting in a higher time scale, and in a complementary way of the secondary objectives. With the aim of implementing the hierarchical control philosophy ex-plained in this thesis, we have built an experimental test bench. This platform has 4 terminals interconnected via a DC grid, and connected to the main AC grid through VSC power converters. This DC grid can work at a maximum of 400 V, and with a maximum allowed current of 15 A. The local VSC converters are controlled by the dSPACE software package. Also, in this network a supervisor PC (secondary controller) is included, which communicates with each dSPACE software of each VSC through a National Instruments CompactRIO programmable automation controller, which combines embedded real-time and FPGA technology, thought a local area network (internet).

Composition du jury:

Dr. Françoise Lamnabhi-Lagarrigue    L2S-CNRS, CentraleSupélec (Directrice)
Dr. Gilney Damm  L2S (Co-encadrant)
Dr. Abdelkrim Benchaib  ALSTOM GRID  (Co-encadrant)
Prof. Séddik Bacha  Université de Grenoble Alpes  (Rapporteur)
Prof. Aleksandar Stankovic  TUFTS University  (Rapporteur)
Dr. Jean-Luc Thomas  CNAM  (Examinateur)
Dr. Adrià Junyent Ferré  Imperial College London  (Examinateur)
Dr. Fernando Dorado Navas  Universidad de Sevilla  (Examinateur)
M. Stéphan Lelaidier  ALSTOM GRID  (Invité)

S³ Working memory in random neural networks

Séminaire le 3 Avril 2015, 10h30 à CentraleSupelec (Gif-sur-Yvette) Salle des séminaires du L2S
Gilles Wainrib, ENS, Computer Science Department

Numerous experimental studies investigate how neural representations of a signal depend on its past context. Although synaptic plasticity and adaptation may play a crucial role to shape this dependence, we study here the hypothesis that this dependence upon past context may be also explained by dynamical network effects, in particular due to the recurrent nature of neural networks connectivity.

Short Bio: Gilles Wainrib is assistant professor in the Computer Science Department at Ecole Normale Supérieure and his research interests range from theoretical biology to applied mathematics and artificial intelligence.

Talk 1: Level set methods for seismic full waveform inversion. Talk 2: Some inverse problems for cargo container screening.

Séminaire le 30 Mars 2015, 13h30 à CentraleSupelec (Gif-sur-Yvette) Salle des séminaires du L2S
Oliver Dorn, Foreign Guest DIGITEO

Bio: Oliver Dorn is currently Lecturer at the School of Mathematics at The University of Manchester. He has obtained his PhD in Applied Mathematics from The University in Muenster, Germany, in 1997, followed by various postdoctoral research stays in the US and Canada. From 2002 until 2007 he was awarded a Ramon y Cajal fellowship at Universidad Carlos III de Madrid, where he became full professor (Profesor Titular) in 2008. He has visited Supélec and Université Paris Sud frequently for longer periods, and has published more than 50 papers in internationally competitive journals and conference proceedings. He is a professional member of SIAM (Society for Industrial and Applied Mathematics), IEEE (Institute of Electrical and Electronics Engineers) and EAGE (European Association of Geoscientists and Engineers) and is on the advisory board of the Journal 'Inverse Problems'. Right now he is at L2S as Foreign Guest DIGITEO.

S³ - Inverse problems in signal and image processing and S³ - Bayesian inference framework: from basic to advanced Bayesian computation

Séminaire le 27 Mars 2015, 10h30 à CentraleSupelec (Gif-sur-Yvette) Salle des séminaires du L2S
Ali Mohammad Djafari, CNRS, L2S

In signal and image processing community, we can distinguish two categories:
- Those who start from the observed signals and images and do classical processing: filtering for denoising, change detection, contour detection, segmentation, compression, …
- The second category called “model based”, before doing any processing try first to understand from where those signals and images come from and why they are here . So, first defining what quantity has been at the origin of those observations, then modeling their link by “forward modeling” and finally doing inversion. This approach is often called “Inverse problem approach”. Then, noting the “ill-posedness” of the inverse problems, many “Regularization methods” have been proposed and applied successfully. However, deterministic regularization has a few limitations and recently the Bayesian inference approach has become the main approach for proposing unsupervised methods and effective solutions in many real applications. Interestingly, even many classical methods have found better understanding when re-stated as inverse problem. The Bayesian approach with simple prior models such as Gaussian, Generalized Gaussian, Sparsity
enforcing priors or more sophisticated Hierarchical models such as Mixture models, Gaussian Scale Mixture or Gauss-Markov-Potts models have been proposed in different applications of imaging systems with great success. However, Bayesian computation still is too costly and need more practical algorithms than MCMC. Variational Bayesian Approximation (VBA) methods have recently became a standard for computing the posterior means in unsupervized methods.
Interestingly, we show that VBA includes Joint Maximum A Posteriori (JMAP) and Expectation-Maximization (EM) as special cases. VBA is much faster than MCMC methods, but, it gives only access to the posterior means.
This talk gives an overview of these methods with examples in Deconvolution (simple or blind, signal or image) and  in Computed Tomography (CT).

Bio: http://djafari.free.fr/index.htm

Séminaire S³

PDE-­‐based inversion method with no forward solver for inverse medium scattering problems

Séminaire le 20 Mars 2015, 14h00 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Dr. Yu Zhong

Bio: Dr. Yu Zhong received his B.S. and M.S. degrees in electronic engineering from Zhejiang University, Hangzhou, China, in 2003 and 2006, respectively, and the Ph.D. degree from the National University of Singapore, Singapore, in 2010.
He is currently a Scientist in Institution of High Performance Computing (IHPC), A*STAR, Singapore.
His research interests mainly are inverse-­‐scattering problems and electromagnetic modeling on composite materials. He is a regular visitor at the Laboratoire des Signaux et Systèmes (L2S) in Gif-­sur-­‐Yvette, France as an invited senior scientific expert since 2012.
Talk 1: PDE-­‐based inversion method with no forward solver for inverse medium scattering problems
A new partial differential equation (PDE) based inversion method for inverse medium scattering problems is proposed in this talk, which does not need to solve any forward problem. The proposed method is the subspace-­‐based optimization method (SOM) in the differential-­‐equation frame. The finite difference scheme is used to discretized the Helmholtz equation, and the twofold subspace-­‐based regularization scheme, as in the integral equation based SOM, is applied in this PDE-­‐based inversion method to stabilize the solver. By using such a PDE-­‐based inversion method, the Green’s funciton for the domain of interests is no longer needed. Representative numerical tests are presented to verify the efficacy of the proposed method.

Talk 2: New integral equation and new partial differential equation for inverse medium scattering problems with strong scatterers 

In this talk, we propose two new equations, an integral equation (IE) and a partial differential equation (PDE), for solving inverse medium scattering problems (IMSP) with strong scatterers. First, we present a new integral equation, which could effectively reduce the globle wave contribution in estimating the contrast (the difference between permittivities of the scatterers and the known background) compared to the original Lippmann-­‐Schwinger equation. Using such a new IE in the IE-­‐based inversion method one is able to solve the highly nonlinear IMSP with strong scatterers.
Subsequently, the connection between the PDE-­‐based inversion method (in Talk 1), using the Helmholtz equation, and the conventional IE based inversion method, using the Lippmann-­‐Schwinger equation, is discussed. With such a connection and the new IE, we propose a new PDE,
using which the PDE-­‐based inversion method can also solve the highly nonlinear IMSP. At last, we discuss the pros and cons of both PDE-­‐ and IE-­‐based inversion methods.

S³ - Analysis of remote sensing multi-sensor heterogeneous images

Séminaire le 20 Mars 2015, 10h30 à CentraleSupelec (Gif-sur-Yvette) Salle des séminaires du L2S
Jorge Prendes, IRIT Toulouse and SONDRA CentraleSupelec

Remote sensing images are images of the Earth acquired from planes or satellites. In recent years the technology enabling this kind of images has been evolving really fast. Many different sensors have been developed to measure different properties of the earth surface, including optical images, SAR images and hyperspectral images. One of the interest of this images is the detection of changes on datasets of multitemporal images. Change detection has been thoroughly studied on the case where the dataset consist of images acquired by the same sensor. However, having to deal with datasets containing images acquired from different sensors (heterogeneous images) is becoming very common nowadays.
In order to deal with heterogeneous images, we proposed a statistical model which describe the joint distribution of the pixel intensity of the images, more precisely a mixture model. On unchanged areas, we expect the parameter vector of the model to belong to a manifold related to the physical properties of the objects present on the image, while on areas presenting changes this constraint is relaxed. The distance of the model parameter to the manifold can be thus be used as a similarity measure, and the manifold can be learned using ground truth images where no changes are present. The model parameters are estimated through a collapsed Gibbs sampler using a Bayesian non parametric approach combined with a Markov random field.
In this talk I will present the proposed statistical model, its parameter estimation, and the manifold learning approach. The results obtained with this method will be compared with those of other classical similarity measures.

Bio: Jorge Prendes was born in Santa Fe, Argentina in 1987. He received the 5 years Eng. degree in Electronics Engineering with honours from the Buenos Aires Institute of Technology (ITBA), Buenos Aires, Argentina in July 2010. He worked on Signal Processing at ITBA within the Applied Digital Electronics Group (GEDA) from July 2010 to September 2012. Currently he is a Ph.D. student in Signal Processing in SONDRA laboratory at Supélec, within the cooperative laboratory TéSA and the Signal and Communication Group of the Institut de Recherche en Informatique de Toulouse (IRIT). His main research interest include image processing, applied mathematics and pattern recognition.


Séminaire S³

S³: Rare event simulation: a Point Process interpretation with application in probability and quantile estimati

Séminaire le 13 Mars 2015, 10h00 à CentraleSupelec (Gif-sur-Yvette) Salle des séminaires du L2S
Clément Walter, CEA and University Paris Diderot

Clément Walter, 25, graduated from Mines ParisTech in 2013. Beforehand he attended preparatory classes in Lycée Sainte-Geneviève (branch Maths and Physics). For his master degree he specialised in Geostatistics and started working with CEA as an intern on emulation of complex computer codes (especially kriging) for rare event simulation and estimation. He has then pursued his work in a PhD under the direction of Pr. Josselin Garnier, focusing on multilevel splitting methods.

S³ L0 optimization for DOA and sparse channel estimation

Séminaire le 6 Mars 2015, 10h00 à CentraleSupelec (Gif-sur-Yvette) Salle des séminaires du L2S
Adilson Chinatto, Univeristy of Campinas, BR and Ecole Normale Supérieure de Cachan, FR

Bio: Adilson Chinatto received a degree in Electrical Engineering in 1997 and Masters in 2011, both from the University of Campinas (Unicamp), Brazil. He worked as hardware, software and firmware development engineer for optical transmission equipment in the companies AsGa and CPqD in Brazil. He is a co-founder of Espectro Ltd., a Brazilian design house for hardware and software, focused in signal processing. Nowadays he is coordinator of a High Performance GPS Receiver Project at Espectro Ltd. funded by the Brazilian National Counsel of Technological and Scientific Development (CNPq). He has experience in electrical engineering with emphasis on telecommunication systems, digital signal processing and smart antennas, working mainly with development and implementation of programmable logic devices (FPGA). He is currently finishing his Ph.D. at Unicamp, working with sparse and compressive sensing signal processing.

Séminaire S³

S³ Bayesian Tomography

Séminaire le 6 Mars 2015, 10h00 à CentraleSupelec (Gif-sur-Yvette) Salle des séminaires du L2S
John Skilling, Maximum Entropy Data Consultants Ltd, UK

Bio: John Skilling was awarded his PhD in radio astronomy in 1969.  Through the 1970s and 1980s he was a lecturer in applied mathematics at Cambridge University, specialising in data analysis.  He left to concentrate on consultancy work, originally using maximum entropy methods but moving to Bayesian methodology when algorithms became sufficiently powerful.  John has been a prominent contributor to the “MaxEnt” conferences since their beginning in 1981.  He is the discoverer of the nested sampling algorithm which performs integration over spaces of arbitrary dimension, which is the basic operation dictated by the sum rule of Bayesian calculus.

Séminaire S³

FREEDM Concept: Power Distribution System Modernization with Solid State Electronics and Integrated Renewables

Séminaire le 5 Mars 2015, 11h00 à CentraleSupelec (Gif-sur-Yvette) Salle des séminaires du L2S
Prof. Iqbal Husain

This FREEDM concept provides an architecture for a future electric power distribution system that is suitable for plug-and-play of distributed renewable energy and distributed energy storage devices. Motivated by the success of the (information) Internet, the FREEDM concept envisions the “Energy Internet,” a system that enables flexible energy sharing for consumers in a residential distribution system. One of the core technologies for the FREEDM system is the solid-state transformer (SST) which is a power electronic device, but makes use of high frequency transformer for isolated AC-AC power conversion. The SST not only dramatically decreases the volume and weight compared with the conventional power transformer, but also behave like a smart transformer with the advantage of power flow control, reactive power compensation capability, and potential fault current limitation. Besides, the regulated low voltage DC/AC bus of SST could be used as an interface to distributed renewable resources and storage devices such as wind, solar, charging stations, and DC micro-grid. The SST is an "Energy Router" proposed to provide such a compatible and flexible interface for energy routing. In addition to the role of a traditional ac transformer, the SST can also provide a low voltage (LV) DC bus. The bi-directional power flow capability of the SST provides possibilities to feed locally generated power back to the grid. Islanding can also be achieved when the HV side is disconnected from grid. The SST has been designed and built based on 13kV SiC MOSFET and JBS diode and is to be connected to medium voltage (MV) distribution network up to 7.2kV and output both a 380V low voltage (LV) dc and 120/240V ac.   

This talk will present the hardware research activities and results aimed at demonstrating the FREEDM concept followed by the power electronic controls for the SST. The experimental results of the developed DC/AC hybrid microgrid built to demonstrate the SST's ability to integrate different renewable energy sources and to improve grid quality will also be discussed.


Dr. Iqbal Husain is the Director of the Future Renewable Electric Energy Delivery and Management (FREEDM) Engineering Research Center and the ABB Distinguished Professor at North Carolina State University where he joined in Fall 2011. Prior to coming to NC State he was serving as a faculty member at the University of Akron, Ohio for seventeen years. He was a visiting Professor at Oregon State University in 2001, and also lectured at Texas A&M University. Prof. Husain also serves as the Director of the Advanced Transportation Energy Center (ATEC).He received the B.Sc. degree from Bangladesh University of Engineering and Technology, Bangladesh, and the M.S. and Ph.D. degrees from Texas A&M University, College Station, Texas. 

Prof. Husain’s research interests are in the areas of control and modeling of electrical drives, design of electric machines, development of power conditioning circuits, microinverters for distributed power generation, inverter controls for grid synchronization, and modeling and control of electric and hybrid vehicle systems. The primary application of Prof. Husain’s work is in the transportation, automotive, aerospace, power system and renewable energy industries.

Prof. Husain was the General Co-Chair for Energy Conversion Congress & Expo (ECCE) 2012 in Raleigh, NC, the chairman of the IEEE-IAS Transportation Systems Committee, and the chairman of the IEEE-IAS Electric Machines Committee. Prof. Husain received the 2006 SAE Vincent Bendix Automotive Electronics Engineering Award, the 2004 College of Engineering Outstanding Researcher Award, the 2000 IEEE Third Millennium Medal, the 1998 IEEE-IAS Outstanding Young Member award, and several IEEE-IAS prize paper awards. He was elected to be an IEEE Fellow in 2009.



Réduction active des excitations NVH d’une machine électrique par la stratégie de commande

Soutenance de thèse de doctorat le 2 Mars 2015, 14h00 à CentraleSupelec (Gif-sur-Yvette) Amphi F3-05

Les travaux de cette thèse s’inscrivent dans le contexte du développement de groupes motopropulseurs électriques pour véhicules automobiles. Ces développements ont mis en lumière les problématiques liées aux phénomènes acoustiques et vibratoires des machines électriques, spécifiques à l'application automobile. L'objectif de ces travaux est la mise en place de lois de commande dédiées à la réduction d'harmoniques de courant à l'origine d'harmoniques vibratoires.
Pour cela, les différents éléments du système ont été modélisés, puis les paramètres du modèle électrique de la machine ont été recalés au moyen de mesures expérimentales.
Deux stratégies de commande ont ensuite été développées puis mises en place. D'une part, un régulateur dédié à l'harmonique de courant visé a été synthétisé par optimisation Hinfini. D'autre part, la modélisation de l'origine de cet harmonique comme provenant d'un signal perturbateur externe a permis la compensation de celui-ci au moyen d'un observateur. La robustesse des lois de commande a été analysée vis-à-vis des variations des modèles nominaux.
L'efficacité de ces deux stratégies sur la réduction des harmoniques de courant visés et des harmoniques vibratoires correspondants a été validée expérimentalement sur banc d'essai.

Composition du Jury :

M. Michel BASSET                           Université de Haute-Alsace              Rapporteur

M. Dominique BEAUVOIS                  CentraleSupélec                                 Professeur - Co-encadrant

M. Demba DIALLO                              Université Paris-Sud                           Examinateur        

M. Gwennaël FAVENNEC                   Ingénieur Renault                               Ingénieur – Co-encadrant

M. Emmanuel GODOY                        CentraleSupélec                                 Professeur – Directeur de la thèse

M. Mickaël HILAIRET                          Université de Franche Comté           Rapporteur

M. Louis HUMBERT                             Ingénieur BMW                                  Examinateur

M. Vincent LANFRANCHI              Université Technologique   de Compiègne            Examinateur

Robust nonlinear control from continuous time to sampled-data with aerospace applications.

Giovanni MATTEI
Soutenance de thèse de doctorat le 13 Février 2015, 10h30 à CentraleSupelec (Gif-sur-Yvette) Amphi Ampère

La thèse porte sur le développement des techniques non linéaires robustes de stabilisation et commande des systèmes avec perturbations de modèle. D’abord, on introduit les concepts de base de stabilité et stabilisabilité robuste dans le contexte des systèmes non linéaires. Ensuite, on présente une méthodologie de stabilisation par retour d’état en présence d’incertitudes qui ne sont pas dans l’image de la commande («unmatched»). L’approche récursive du «backstepping» permet de compenser les perturbations «unmatched» et de construire une fonction de Lyapunov contrôlée robuste, utilisable pour le calcul ultérieur d’un compensateur des incertitudes dans l’image de la commande («matched»). Le contrôleur obtenu est appelé «recursive Lyapunov redesign». Ensuite, on introduit la technique de stabilisation par «Immersion & Invariance» comme outil pour rendre un contrôleur non linéaire, robuste par rapport à des dynamiques non modelées. La première technique de contrôle non linéaire robuste proposée est appliquée au projet d’un autopilote pour un missile air-air et au développement d’une loi de commande d’attitude pour un satellite avec appendices flexibles. L’efficacité du «recursive Lyapunov redesign» est mis en évidence dans les deux cas d’étude considérés. En parallèle, on propose une méthode systématique de calcul des termes incertains basée sur un modèle déterministe d’incertitude. La partie finale du travail de thèse est relative à la stabilisation des systèmes sous échantillonnage. En particulier, on reformule, dans le contexte digital, la technique d’Immersion et Invariance. En premier lieu, on propose des solutions constructives en temps continu dans le cas d’une classe spéciale des systèmes en forme triangulaire «feedback form», au moyen de «backstepping» et d’arguments de domination non linéaire. L’implantation numérique est basée sur une loi multi-échelles, dont l’existence est garantie pour la classe des systèmes considérée. Le contrôleur digital assure la propriété d’attractivité et des trajectoires bornées. La loi de commande, calculée par approximation finie d’un développement asymptotique, est validée en simulation de deux exemples académiques.

Composition du jury    
Prof. Salvatore MONACO, Università La Sapienza - DIAG "Antonio Ruberti" (Directeur)
Dr. Dorothée NORMAND-CYROT, L2S-CNRS, Supélec                               (Directeur)
Prof. Jean-Pierre BARBOT, ENSEA Cergy Pontoise (Rapporteur)
Prof. Giovanni ULIVI, DIA Università degli Studi Roma Tre (Rapporteur)
Prof. Stefano BATTILOTTI, Università La Sapienza - DIAG "Antonio Ruberti" (Examinateur)
Dr. Vincent FROMION, INRA Jouy en Josas (Examinateur)
Dr. Romeo ORTEGA, L2S-CNRS, Supélec (Examinateur)
Dr. Silviu NICULESCU, L2S-CNRS, Supélec (Invité)


Online adaptive optimal control of a hybrid electric vehicle

Séminaire le 5 Février 2015, 14h30 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Catalin Stefan Teodorescu

Par Doctor Catalin Stefan Teodorescu Project engineer, Flanders' Mechatronics Technology Centre (FMTC), www.fmtc.be A division of Flanders Make, www.flandersmake.be In general, the hybrid driveline concept consists of using simultaneously 2 power sources: the internal combustion engine (ICE) on the one hand, and an electric power source, e.g., a supercapacitor (SC), on the other hand. In the above figure, the red arrows represent power flow of the two illustrated power sources: specifically, it can be generative or regenerative. The latter translates the ability of the SC to store some energy, which can potentially be used in order to achieve significant fuel consumption reduction: e.g. by avoiding wasting energy when the driver pushes the break pedal and instead storing it. In this presentation we will investigate the following problem. In order to be able to cope with the actual demand of the driver expressed in terms of demanded power load (see the above figure), we have an infinite number of choices/combinations concerning the power split ICE versus SC. However, only one (or a few) of these trajectories are optimal in terms of fuel consumption and this is what is being investigated. Two control solutions are presented: Dynamic Programming, mainly used for benchmarking purpose and Equivalent Consumption Minimization Strategy. Results show the latter to be a strong candidate towards online, real-time implementation.

Développement de nouvelles méthodes itératives de reconstruction tomographique pour réduction des artefacts métalliques et réduction de la dose en imagerie dentaire

Long Chen
Soutenance de thèse de doctorat le 5 Février 2015, 14h30 à CentraleSupelec (Gif-sur-Yvette) Salle des séminaires du L2S

Cette thèse est constituée de deux principaux axes de recherche portant sur l'imagerie dentaire par la tomographie à rayons X : le développement de nouvelles méthodes itératives de reconstruction tomographique afin de réduire les artefacts métalliques et la réduction de la dose délivrée au patient. Afin de réduire les artefacts métalliques, nous prendrons en compte le durcissement du spectre des faisceaux de rayons X et le rayonnement diffusé. La réduction de la dose est abordée dans cette thèse en diminuant le nombre des projections traitées.

La tomographie par rayons X a pour objectif de reconstruire la cartographie des coefficients d'atténuations d'un objet inconnu de façon non destructive. Les bases mathématiques de la tomographie repose sur la transformée de Radon et son inversion. Néanmoins des artefacts métalliques apparaissent dans les images reconstruites en inversant la transformée de Radon (la méthode de rétro-projection filtrée), un certain nombre d'hypothèse faites dans cette approche ne sont pas vérifiées. En effet, la présence de métaux exacerbe les phénomènes de durcissement de spectre et l'absence de prise en compte du rayonnement diffusé. Nous nous intéressons dans cette thèse aux méthodes itératives issues d'une méthodologie Bayésienne. Afin d'obtenir des résultats de traitement compatible avec une application clinique de nos nouvelles approches, nous avons choisi un modèle direct relativement simple et classique (linéaire) associé à des approches de corrections de données. De plus, nous avons pris en compte l'incertitude liée à la correction des données en utilisant la minimisation d'un critère de moindres carrés pondérés. Nous proposons donc une nouvelle méthode de correction du durcissement du métal sans connaissances du spectre de la source et des coefficients d'atténuation des matériaux. Nous proposons également une nouvelle méthode de correction du diffusé associée sur les mesures sous certaines conditions notamment de faible dose.

En imagerie médicale par tomographie à rayons X, la surexposition ou exposition non nécessaire irradiante augmente le risque de cancer radio-induit lors d'un examen du patient. Il y a donc une demande continue de réduction de la dose de rayons X transmise au patient. Notre deuxième axe de recherche porte donc sur la réduction de la dose en diminuant le nombre de projections. Nous avons donc introduit un nouveau mode d'acquisition possédant un échantillonnage angulaire adaptatif. On utilise pour définir cette acquisition notre connaissance a priori de l'objet. Ce mode d'acquisition associé à un algorithme de reconstruction dédié, nous permet de réduire le nombre de projections tout en obtenant une qualité de reconstruction comparable au mode d'acquisition classique. Enfin, dans certains modes d’acquisition des scanners dentaires, nous avons un détecteur qui n'arrive pas à couvrir l'ensemble de l'objet. Pour s'affranchir aux problèmes liés à la tomographie locale qui se pose alors, nous utilisons des acquisitions multiples suivant des trajectoires circulaires. Nous avons adaptés les résultats développés par l’approche « super short scan » [Noo et al 2003] à cette trajectoire très particulière et au fait que le détecteur mesure uniquement des projections tronquées.

Nous avons évalué nos méthodes de réduction des artefacts métalliques et de réduction de la dose en diminuant le nombre des projections sur les données réelles. Grace à nos méthodes de réduction des artefacts métalliques, l'amélioration de qualité des images est indéniable et il n'y a pas d'introduction de nouveaux artefacts en comparant avec la méthode de l'état de l'art NMAR [Meyer et al 2010]. Par ailleurs, nous avons réussi à réduire le nombre des projections avec notre nouveau mode d'acquisition basé sur un « super short scan » appliqué à  des trajectoires multiples. La qualité obtenue est comparable aux reconstructions obtenues avec les modes d'acquisition classique ou short-scan mais avec une réduction de 20% de la dose radioactive.


Membres du jury

Directeur de thèse Mr RODET Thomas Professeur, ENS Cachan, SATIE
Co-encadrant Mr. GAC Nicolas Maître de conférences, Université Paris-Sud, L2S
Rapporteurs Mr. DESBAT Laurent Professeur des universités, Université Joseph Fourier
  Mr. BLEUET Pierre Ingénieur de recherche CEA, HDR
Examinateurs Mme NGUYEN-VERGER Maï Professeur des universités, Université de Cergy-Pontoise
  Mme MARCOS Sylvie Directeur de recherche, CNRS
Invitée Mme MAURY Colombe Ingénieur de recherche, Trophy, Carestream Dental


S³ Correlation mining in high dimension with limited samples

Séminaire le 30 Janvier 2015, 10h30 à CentraleSupelec (Gif-sur-Yvette) Salle des séminaires du L2S
Alfred Hero

Alfred Hero:
Title: Correlation mining in high dimension with limited samples 

Abstract: Correlation mining arises in many areas of engineering, social sciences, and natural sciences. Correlation mining discovers columns of a random matrix that are highly correlated with other columns of the matrix and can be used to construct a dependency network over columns. However, when the number n of samples is finite and the number p of columns increases such exploration becomes futile due to a phase transition phenomenon: spurious discoveries will eventually dominate. In this presentation I will present theory for predicting these phase transitions and present Poisson limit theorems that can be used to determine finite sample behavior of correlation structure. The theory has application to areas including gene expression analysis, network security, remote sensing, and portfolio selection. 

BioAlfred O. Hero III received the B.S. (summa cum laude) from Boston University (1980) and the Ph.D from Princeton University (1984), both in Electrical Engineering. Since 1984 he has been with the University of MichiganAnn Arbor, where he is the R. Jamison and Betty Williams Professor of Engineering. His primary appointment is in the Department of Electrical Engineering and Computer Science and he also has appointments, by courtesy, in the Department of Biomedical Engineering and the Department of Statistics. From 2008-2013 he was held the Digiteo Chaire d'Excellence at the Ecole Superieure d'Electricite, Gif-sur-Yvette, France. He is a Fellow of the Institute of Electrical and Electronics Engineers (IEEE) and several of his research articles have recieved best paper awards. Alfred Hero was awarded the University of Michigan Distinguished Faculty Achievement Award (2011). He received the IEEE Signal Processing Society Meritorious Service Award (1998), the IEEE Third Millenium Medal (2000), and the IEEE Signal Processing Society Technical Achievement Award (2014). Alfred Hero was President of the IEEE Signal Processing Society (2006-2008) and was on the Board of Directors of the IEEE (2009-2011) where he served as Director of Division IX (Signals and Applications). Alfred Hero's recent research interests are in statistical signal processing, machine learning and the analysis of high dimensional spatio-temporal data. Of particular interest are applications to networks, including social networks, multi-modal sensing and tracking, database indexing and retrieval, imaging, and genomic signal processing.

György Terdik:
Title: A new covariance function for spatio-temporal data analysis with application to atmospheric pollution and sensor networking

Abstract: See the attached file.

Bio: György Terdik, received the Ph.D. degree from theKossuth Lajos University, Hungary. He is a Professor  of the Department of Information Technology at the University ofDebrecen, Hungary. His research areas include nonlinear, non-Gaussian time series analysis, Lévyprocesses, and high-speed network modeling, spatio-temporal time series, Dr. Terdik is a member of the editorial board of the

quarterly Publicationes Mathematicae Debrecen.

Márton Ispány:
Title Poisson INAR processes with serial and seasonal correlation 

Abstract: Recently, there has been considerable interest in integer-valued time series models.  Motivation to include discrete data models comes from the need to account for the discrete nature of certain data sets, often counts of events, objects or individuals. Among the most successful integer-valued time series models proposed in the literature we mention the INteger-valued AutoRegressive model of order p (INAR(p)). However, seasonal count processes have not been investigated yet, except one of our new papers. In the talk, we study INAR processes which possess serial and seasonal structure as well. The main properties of the models will be derived such as the stationarity and the autocorrelation function. The conditional least squares and conditional maximum likelihood estimators of the model parameters will be studied and their asymptotical properties will be established. In addition, we would like to discuss the case in which the marginal distributions are Poisson in detail. Monte 
Carlo experiments will be conducted to evaluate and compare the performance of various estimators for finite sample sizes. Real data set on the area of insurance will be applied to evaluate the model performance. 

Bio: Márton Ispány (https://it.inf.unideb.hu/honlap/ispanymarton/english) received the M.Sc.(1989) and PhD (summa cum laude) in Statistics (1997) from University of Debrecen. Since 2007 he has been with the  Department of Information Technology, Faculty of Informatics, University of Debrecen. Since 2012 he has been the head of the department. Márton Ispány 's recent research interests are in branching processes (functional limit theorems, asymptotics for conditional least squares estimation, integer valued autoregression), statistical modelling(generalized SVD, contaminated statistical models, EM algorithm), data mining (decision trees, stochastic algorithms, MCMC, web mining), and applied statistics: econometrics and insurance, cross-country modelling, statistical genetics.


Interval Analysis - Fundamentals and Electromagnetic Engineering Applications

Séminaire le 29 Janvier 2015, 14h00 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Paolo Rocca

Paolo Rocca received the MS degree in Telecommunications Engineering from the University of Trento in 2005 (summa cum laude) and the PhD Degree in Information and Communication Technologies from the same University in 2008. He is currently Assistant Professor at the Department of Information Engineering and Computer Science (University of Trento) and a member of the ELEDIA Research Center. Dr. Rocca is the author/co-author of over 230 peer reviewed papers on international journals and conferences. He has been a visiting Ph.D. student at the Pennsylvania State University (U.S.A.), at the University Mediterranea of Reggio Calabria (Italy), and a visiting researcher at the Laboratoire des Signaux et Systèmes (L2S@ Supélec, France) in 2012 and 2013. Moreover, he has been an Invited Associate Professor at the University of Paris Sud (France) in 2015. Dr. Rocca has been awarded from the IEEE Geoscience and Remote Sensing Society and the Italy Section with the best PhD thesis award IEEE-GRS Central Italy Chapter. His main interests are in the framework of antenna array synthesis and design, electromagnetic inverse scattering, and optimization techniques for electromagnetics. He serves as an Associate Editor of the IEEE Antennas and Wireless Propagation Letters.

Five generations of sliding mode controllers: stages of evolution

Séminaire le 28 Janvier 2015, 10h00 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Prof. Leonid Fridman from Universidad Nacional Autónoma de México (UNAM)

The history and evolution of sliding control will be discussed. The reason for the crisis of the first order sliding modes will be explained. The second order sliding mode control algorithms and their specific features will be presented. The control chattering of the continuous second order super-twisting control algorithm will be discussed. The precision of the arbitrary order sliding mode controllers will be shown. The continuous arbitrary order sliding mode controllers will be presented and discussed. Videos with the experimental illustration of the properties of the main sliding mode algorithms will be presented. Short biography: Leonid M. Fridman received an M.S. degree in mathematics from Kuibyshev (Samara) State University, Samara, Russia, in 1976, a Ph.D. degree in applied mathematics from the Institute of Control Science, Moscow, Russia, in 1988, and a Dr. Sc. degree in control science from Moscow State University of Mathematics and Electronics, Moscow, Russia, in 1998. From 1976 to 1999, he was with the Department of Mathematics, Samara State Architecture and Civil Engineering University. From 2000 to 2002, he was with the Department of Postgraduate Study and Investigations at the Chihuahua Institute of Technology, Chihuahua, Mexico. In 2002, he joined the Department of Control Engineering and Robotics, Division of Electrical Engineering of Engineering Faculty at National Autonomous University of Mexico (UNAM), Mexico. His research interests are variable structure systems. He is currently a Chair of TC on Variable Structure Systems and Sliding mode control of IEEE Control Systems Society. Prof. Fridman is an Associated Editor of the Journal of Franklin Institute, Nonlinear Analysis: Hybrid Systems, and the Conference Editorial Board of IEEE Control Systems Society. He is an author and editor of seven books and 12 special issues and author of more than 350 technical papers on sliding mode control. He is a winner of Scopus prize for the best cited Mexican Scientists in Mathematics and Engineering 2010. He was working as an invited professor in 19 universities and research centers of Argentina, Australia, France, China, Germany, Italy, Israel, and Spain. Web page: verona.fi-p.unam.mx/~lfridman/

Small Primitive Roots and Malleability of RSA Moduli

Séminaire le 13 Janvier 2015, 14h00 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Prof. Jorge Jimenez Urroz from Polytechnic University of Catalonia, Barcelona, Catalunya,

Prof. Jorge Jimenez Urroz from Polytechnic University of Catalonia, Barcelona, Catalunya, Spain will give a seminar at L2S, on Tuesday 13th January 2015, room F.3.09, Supélec, campus of Gif-sur-Yvette.

Title: Small Primitive Roots and Malleability of RSA Moduli

Abstract: We prove that factorization is a malleable problem, in the sense that given an RSA modulus n, partial information on another integer n' independent helps to factorize n.

List of publications : http://www-ma4.upc.edu/~jjimenez/papers.htm

Structure d'information, stratégies de communication et application aux réseaux distribués

Soutenance de thèse de doctorat le 12 Novembre 2014, 14h00 à Telecom ParisTech (Paris)

Cette thèse étudie des problèmes d’optimisation distribuée avec différentes structures d’ob- servations et leurs applications aux réseaux sans fil et aux problèmes de Smart Grids. Spécifique- ment, une structure d’observation asymétrique entre deux agents est considérée, où un premier agent a connaissance complète à propos de la réalisation d’un état aléatoire, et l’autre agent ne connaît rien à propos de cet état. Dans ce contexte, la question est de savoir comment transmettre de l’information depuis le premier agent vers le second agent dans le but d’utiliser de manière optimale les ressources de communication. Plusieurs modèles sont étudiés dans cette thèse. Pour tous, un élément commun est le fait que la source d’information doit être encodée de manière appropriée pour optimiser l’utilisation de la configuration du système. Un premier modèle est étudié où aucun canal de communication n’est disponible entre les agents et ils ont une fonc- tion d’utilité commune. Cependant, le seul moyen de communiquer est via les actions choisies par les agents. Comme les actions ont une influence sur le paiement, l’agent informé encode sa connaissance à propos de l’état dans ses actions, qui seront observées de manière imparfaite par le second agent. Ce dernier décodera l’information et choisira ses actions dans le but de maxi- miser la fonction objectif commune. Nous utilisons des outils de théorie de l’information pour caractériser ce compromis optimal par une contrainte d’information, et appliquons ce scénario à un problème de contrôle de puissance pour un canal à interférence. Notre nouvelle stratégie (le contrôle de puissance codé) donne des gains très prometteurs comparés aux approches classiques.

Dans une seconde partie, nous considérons qu’il existe un canal dédié de communication, c’est- à-dire que les actions de l’agent informé n’ont pas d’influence sur le paiement et sont seulement utiles pour la transmission d’information. De plus, les agents sont supposés avoir des intérêts divergents, si bien que l’agent informé n’a pas nécessairement d’incitation à envoyer tout son savoir à l’agent non informé. La théorie des jeux et les jeux de « Cheap talk » en particulier sont le bon cadre pour analyser ce genre de problème. Nous caractérisons le schéma de signal sur lequel les agents se seront mis d’accord. Ce schéma amènera à un équilibre de Nash, est donc optimisera la façon dont la communication est faite. Ce modèle est d’un intérêt particulier pour les réseaux de véhicules électriques où un véhicule électrique doit envoyer son besoin en terme de puissance de charge à un aggrégateur qui choisira un niveau de charge effectif pour le véhicule électrique. Ce dernier ne se souciera que de son besoin, alors que l’aggrégateur se soucie également de l’état du réseau. Ce modèle aide à optimiser la façon dont le réseau est utilisé.

Enfin, nous considérons un modèle avec plus de deux agents, où le but principal est pour tous les agents de retrouver l’observation parfaite des actions passées de tous les agents. Ceci est d’un intérêt très particulier d’un point de vue de la théorie des jeux pour caractériser les utilités espérées de long terme des agents. Dans ce modèle, nous ajoutons un encodeur qui observe parfaitement toutes les actions passées et aidera les agents à obtenir l’observation parfaite. En fait, ceci sera possible si la bonne contrainte d’information est satisfaite. Nous caractérisons donc cette dernière, en utilisant un schéma de codage hybride combinant des outils classiques de théorie de l’information ainsi que des outils de la théorie des graphes.

Membres du jury
Eitan Altman INRIA Sophia-Antipolis (Examinateur)
Jean-Claure Belfiore Telecom ParisTech (Examinateur)
Matthieu Bloch Georgia Institute of Technology (Examinateur)
Olivier Gossner Paris School of Economics (Rapporteur)
Rida Laraki Université Paris-Dauphine (Rapporteur)
Samson Lasaulce L2S-CNRS (Directeur de thèse)
Michèle Wigger Telecom ParisTech (Examinateur)


Assemblée générale du L2S

Thu, 12/06/2012 -
10:30 to 12:30

St Rémy de Provence
Assemblée générale du laboratoire