Séminaire d'Automatique du plateau de Saclay

Séminaire le 4 Décembre 2019, 10h00 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Paolo Mason & Hendra Nurdin

10:00-11:00 Paolo Mason (Chargé de recherche, CNRS, L2S, CentraleSupélec)

Title: Controllability of the Schrodinger equation via adiabatic methods

Abstract: In this presentation I will consider the approximate controllability problem for the bilinear Schrodinger equation. In particular I will focus on the application of adiabatic techniques in presence of conical eigenvalues intersections of the Hamiltonian operator. These methods allow to design, in a constructive way, control laws capable of (approximately) steering the system from an eigenstate of the Hamiltonian to an arbitrary target state (or, more precisely, to an arbitrary density distribution). The relationship between our results and other controllability results for the bilinear Schrodinger equation will be discussed, as well as the connection between adiabatic and singular perturbation techniques. Finally, in order to justify the applicability of the presented results, I will provide physically meaningful classes of Hamiltonian operators
for which eigenvalues intersections are generically conical.

Biography: Paolo Mason was born in Dolo, Italy, in 1978. He received the Laurea degree in mathematics from the University of Padova, Italy, in 2002, and the Ph.D. degree from SISSA, Trieste, Italy, in 2006. Since 2009 he works as a “chargé de recherche” (researcher) for CNRS at the Laboratoire des Signaux et Systèmes, Gif-sur- Yvette, France. His research interests include geometric control theory, quantum control and hybrid systems. 

11:00-12:00 Hendra Nurdin (Senior Lecturer, School of Electrical Engineering and Telecommunications, UNSW, Australia)

Title: Learning nonlinear input-output maps with dissipative quantum systems

Abstract: In this seminar, I will describe a theoretical framework for learning of nonlinear input-output maps with fading memory by dissipative quantum systems, as a quantum counterpart of the theory of approximating such maps using classical dynamical systems. Such a theory can provide the foundation for harnessing of dissipative quantum systems for applications such as nonlinear systems modelling and signal processing. In particular, the theory identifies the properties required for a class of dissipative quantum systems to be universal, in the sense that any input-output map with fading memory can be approximated arbitrarily closely by an element of this class. We then introduce an example class of dissipative quantum systems that is provably universal. Some numerical examples will be presented.

The seminar is based on joint work with J. Chen (Quantum Information Processing, 18(7):198 (2019)) 

Biography: Dr Hendra I. Nurdin received the bachelor's degree in electrical engineering from Institut Teknologi Bandung, Bandung, Indonesia, the master's degree in engineering mathematics from the University of Twente, Enschede, the Netherlands, and the Ph.D. degree in engineering and information science from the Australian National University (ANU), Canberra, ACT, Australia, in 2007. From 2007 to 2011, he was a Research Fellow and then an Australian Research Council APD Fellow with the ANU before joining the University of New South Wales, Australia, in 2012. His research interests include quantum systems, quantum feedback control, stochastic systems and stochastic control, and applications of control theory to microgrids and renewable energy systems. He  is a coauthor, with Naoki Yamamoto, of the Springer research monograph “Linear dynamical quantum systems: Analysis, synthesis, and control” (2017).

Modélisation et commande prédictive flexible de la demande en chauffage des bâtiments raccordés à des réseaux de chaleur

Madame Nadine AOUN
Soutenance de thèse de doctorat le 2 Décembre 2019, 14h00 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40

Composition du jury proposé:

M. Guillaume Sandou Professeur - CentraleSupélec Directeur de Thèse
M. Khalil El Khoury Professeur - Université Libanaise Examinateur
Mme Sihem Guernouti Chargée de Recherche - Cerema Examinatrice
M. Marc Petit Professeur - CentraleSupélec Examinateur
M. Vittorio Verda Professeur - Politechnico di Torino Examinateur
M. Etienne Wurtz Directeur de Recherche - CNRS-CEA-INES Examinateur
M. Bruno Lacarrière Professeur - IMT Atlantique Rapporteur
M. Stéphane Ploix Professeur - Université Grenoble-Alpes Rapporteur
M. Roland Bavière Ingénieur de recherche - CEA Co-encadrant
M. Mathieu Vallée Ingénieur de recherche - CEA Co-encadrant
M. David Canal Ingénieur de Recherche - Ademe Invité

Résumé: La gestion de la demande en chauffage des bâtiments raccordés à des réseaux de chaleur s'effectue au niveau de la sous-station au moyen d'un contrôleur qui détermine la température de départ de l'eau alimentant le circuit de chauffage interne (variable de contrôle).

Dans les stratégies de contrôle classiques, cette variable de contrôle est régie par la température extérieure via une courbe de chauffe : lorsque cette dernière chute, la première est relevée. Ce mode de contrôle est appelé « régulation par loi d’eau ».

En dépit de ses incontestables atouts, parmi lesquels une implémentation simple et une réponse fiable à la demande en chauffage, la régulation par loi d’eau ne tient pas compte de l'inertie thermique du bâtiment et ne permet donc pas une modulation de sa demande.

La modulation de la demande en chauffage se définit comme l'action de contrôle consistant à modifier de manière stratégique et raisonnable les conditions de confort thermique interne à des fins économiques. Il s’agit d’une mesure essentielle dans les stratégies de contrôle flexibles qui envisagent le déplacement des charges et l’effacement des pics pour ainsi permettre un engagement durable des ressources énergétiques favorisant la pénétration de l’énergie renouvelable et de récupération à l’échelle urbaine.

Ces travaux de thèse visent à développer et à évaluer numériquement une stratégie de contrôle prédictif et flexible de la demande en chauffage, basée sur un modèle physique simplifié du système de bâtiment et applicable de manière non intrusive dans les réseaux de chaleur.

Tout d'abord, un simulateur thermique dynamique de bâtiment résidentiel, chauffé par un circuit de radiateurs connecté à une sous-station de réseau de chaleur, est développé. Il permet la définition de plusieurs cas d’études de bâtiments représentatifs du parc immobilier résidentiel Français et constitue l’environnement expérimental virtuel de nos travaux de recherche.

Ensuite, une méthodologie permettant d’obtenir un modèle orienté-contrôle et d’ordre réduit de bâtiment avec son système de chauffage est proposée. Elle commence par la définition de la structure du modèle en se basant sur des connaissances physiques, puis consiste en l'identification des paramètres par optimisation méta-heuristique à l'aide des données générées par le simulateur. La structure du modèle est destinée à capter les dynamiques thermiques à court terme. Ainsi, elle comporte des capacités thermiques pour les éléments à forte inertie thermique : la masse interne et le circuit de chauffage. L'approche d'identification paramétrique évalue la possibilité de réaliser cette tâche en ne s’appuyant que sur des données non intrusives et disponibles au niveau de la sous-station, notamment en s’interdisant d’utiliser des mesures de température intérieure.

Enfin, la stratégie de contrôle prédictif est implémentée pour planifier la température de départ de l'eau de chauffage à partir de prévisions des conditions météorologiques et des variations des prix de l’énergie. Le contrôleur flexible est conçu pour résoudre un problème d’optimisation linéaire sous contraintes, selon le principe de l’horizon fuyant. Il incorpore les équations linéarisées du modèle d’ordre réduit du bâtiment dans la formulation du problème et fait un compromis optimal entre les coûts de consommation d'énergie et l'inconfort thermique, le degré de flexibilité pour la modulation de la demande en chauffage étant défini par l’intermédiaire de paramètres de réglage dédiés.

Cette thèse est cofinancée par l'Agence De l'Environnement et de la Maîtrise de l'Energie (ADEME) et le Commissariat à l´Energie Atomique et aux Energies Alternatives (CEA). Elle est préparée au sein du Laboratoire des Systèmes Energétiques et Démonstrateurs territoriaux (LSED) du CEA et en collaboration avec le Laboratoire des Signaux et Systèmes (L2S), unité mixte de recherche du Centre National de la Recherche Scientifique (CNRS), de l’Ecole CentraleSupélec et de l’Université Paris-Sud.

Mots clés: réseaux de chauffage urbains, modélisation de bâtiments, commande prédictive, identification paramétrique, simulation thermique dynamique, programmation linéaire mixte


Abstract: In District Heating Systems (DHSs), buildings space-heating (SH) demand management takes place at the substation level by means of a controller that determines the supply water temperature to the internal SH circuit (control variable).

In conventional control strategies, this control variable is governed by the outdoor temperature via a heating curve: when the latter drops, the former is raised. This is referred to as Weather-Compensation Control (WCC).

Despite its widely recognized assets which include simple implementation and reliable fulfilment of the SH demand, WCC does not account for the building system thermal inertia, and consequently it does not allow modulation of the demand.

SH demand modulation is the control action of strategically and reasonably altering the indoor thermal comfort conditions for economic benefits. It is a key measure in flexible demand control strategies, which seek loads shifting and peaks shaving to allow sustainable commitment of energy resources in favour of renewable power penetration and waste heat recovery at urban scale.

The work presented in this thesis aims at developing, and numerically evaluating, a flexible Model Predictive Control (MPC) strategy for SH demand, non-intrusively applicable in DHSs.

Firstly, a thermal dynamic simulator of a residential building with a radiator SH circuit connected to a DHS substation is developed. It allows defining multiple case study buildings representative of the French residential stock and constitutes the virtual experimental environment for the research.

Then, a methodology to obtain a control-oriented Reduced-Order Model (ROM) for the building and its SH system is proposed. It starts by defining the ROM structure based on physical knowledge, and proceeds to parameters identification by meta-heuristic optimization using data generated by the simulator. The ROM structure is intended to capture short-term thermal dynamics, thus it features thermal capacitances for elements with considerable thermal inertia: the internal mass and the SH circuit. The parametric identification approach assesses the possibility of achieving this task using strictly non-intrusive data available at the substation level, notably not relying on indoor temperature measurements.

Finally, MPC is implemented to schedule the supply water temperature in anticipation of weather conditions and energy price variations. The flexible controller is designed to solve a constrained linear optimization problem according to the receding horizon principle. It embeds the linearized ROM equations within the problem formulation and makes an optimal trade-off between energy consumption costs and thermal discomfort, the degree of flexibility to modulate SH demand being defined through dedicated tuning parameters.

This PhD thesis is jointly funded by ADEME and CEA. It is prepared at CEA - LSED laboratory in collaboration with L2S, a CNRS - CentraleSupélec - Université Paris-Sud joint laboratory.

Keywords: District heating systems, Building modelling, Model predictive control, parametric identification, thermal dynamic simulation, Mixed-integer linear programming

Séminaire d'Automatique du plateau de Saclay

Séminaire le 29 Novembre 2019, 10h00 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Elena Panteley & Ubirajara F. Moreno

10:00-11:00 Elena Panteley (Director of research, CNRS, L2S, CentraleSupélec)

Title: Dynamic consensus in heterogeneous networks

Abstract: In this talk we present an analysis framework for the study of synchronization of nonlinear systems interconnected over networks described by directed graphs.  Systems may have different dynamical models or the same model with different parameters.  We introduce the concept of dynamic consensus and we characterize the synchronization behavior of the network in terms of the stability properties of two interconnected dynamical systems that evolve in orthogonal spaces: one corresponds to the synchronization error dynamics and the second to the so-called emergent dynamics. Such an approach allows not only to formulate conditions for practical asymptotic synchronization of heterogeneous networked systems, but also to characterize their collective behavior.  In the end we present two extensions of the proposed approach to the analysis of networks of Stuart-Landau oscillators.

Biography. Elena Panteley was born in Leningrad, USSR. She received the M.Sc. and Ph.D. degrees in applied mathematics from the State University of St. Petersburg, St. Petersburg, Russia. She is a Director of Research (DR 2) of the French National Centre of Scientific Research (CNRS), Laboratoire de Signaux et Systemes, Yvette, France. From 1986 to 1998, she held a research position with the Institute for Problem of Mechanical Engineering, the Academy of Science of Russia, St. Petersburg. During 1998, she was an Associate Researcher with the Center for Control Engineering and Computation, University of California at Santa Barbara. During 1999, she was with the INRIA Rhone Alpes, Monbonnot, France. She is coauthor of over 90 scientific articles and book chapters. Her research interests include stability of nonlinear time-varying systems, control of electromechanical systems, nonlinear, and robust control.

11:00-12:00 Ubirajara F. Moreno (Professor, Department of Automation and Systems (DAS), Federal University of Santa Catarina (UFSC))

Title: Modelling and Simulation of Online Social Networks

Abstract: From the first Internet-based social networking applications designed to get people in contact and make friends, to social networks made up of over 2 billion users, the combination of communication networks, portable devices and AI has changed the way People interact and make decisions. The extent of this influence could be observed not only in marketing, and social behavior but also in referendums and elections, leading to distortion of democratic manifestations and representations. The aim of this presentation is show that an approach based on Systems & Control could be applied to modelling and analisys of the behavior of social networks, as well as, to assess some regulatory policies. The analysis are based on simulation of this models on small and large scale networks.

Biography: Since 2004, Ubirajara F. Moreno is a professor at the Department of Automation and Systems (DAS) at Federal University of Santa Catarina (UFSC). In the period of 2014 to 2016 he served as Director of the Campus of Blumenau from the Federal University of Santa Catarina. He participated in several national and international cooperation projects (PROCAD, BRAFITEC, CAPES / COFFECUB, ANP-PRH-34, CNPq-CNRS, CAPES / GRICES, CAPES/SIU, among others), having coordinated a project of Cooperation Brazil / Cuba: CAPES / MES, 108/2010: entitled Development of an Embedded System of Low Cost for Industrial Monitoring and Diagnosis and a cooperation project Brazil / Portugal: CAPES / FCT 353/13 entitled PICC - Integrated Controllers and Communications Project for NCS. In 2016 he was the national committee chairman of the 1st IFAC Conference on Cyber-Physical & Human Systems held in Florianópolis. Currently coordinates the BRAFITEC project, Brazil / France, entitled Franco-Brazilian Cooperation Project for the Training of Engineers in the Control and Automation Area, in partnership with CentraleSupelec, EC Nantes, UFRGS and UNICAMP, and participates in a Brazil/Norway project of international cooperation on Modeling and Control Strategies with Application in Offshore Oil and Gas Production . He participated in joint research projects with the industry in the development of monitoring systems (Petrobras, Bosch Rexroth).The research interests of prof. Ubirajara Franco Moreno are: Networked Control Systems, Cooperative Robotics, Systems Monitoring and Human Machine Interaction. From August/2019 to April/2020 he is in a sabbatical stay, at L2S, working with Françoise Lamnabhi-Lagarrigue, on Cyber-Physical and Human Systems (CPHS) interaction in Industry 4.0.

Quelques problèmes dans l’analyse et la commande des réseaux électriques avec des charges à puissance constante

Monsieur Juan E. Machado
Soutenance de thèse de doctorat le 22 Novembre 2019, 15h00 à CentraleSupelec (Gif-sur-Yvette) Salle des séminaires du L2S

Composition du jury proposé:

Mr. Romeo Ortega L2S (CNRS, UMR 8506) Directeur de thèse
Mr. Luca Greco Université Paris-Sud (UMR8506) Examinateur
Mr. Robert Griñó Universitat Politècnica de Catalunya Examinateur
Mme Françoise Lamnabhi-Lagarrigue L2S (CNRS, UMR 8506) Examinateur
Mr. John W. Simpson-Porco University of Waterloo Rapporteur
Mr. Aleksandar Stankovic Tufts University Rapporteur

 

Resumé: La croissante demande d’énergie électrique a conduit à la conception de systèmes électriques de grande complexité où les combustibles fossiles constituent la principale source d’énergie. Néanmoins, les préoccupations environnementales poussent à un changement majeur dans les pratiques de production d’électricité, avec un passage marqué des énergies fossiles aux énergies renouvelables et des architectures centralisées à distribuées. Les problèmes de stabilité dus à la présence de ce qu’on appelle les Charges à Puissance Constante (CPLs) constituent l’un des principaux défis auxquels sont confrontés les systèmes électriques distribués. On sait que ces charges, que l’on trouve couramment dans les installations de technologie de l’information et de la communication, réduisent l’amortissement effectif des circuits qui les alimentent, ce qui peut provoquer des oscillations de tension, voire une chute. Dans cette thèse, les principales contributions sont centrées sur la compréhension et la résolution de divers problèmes rencontrés dans l’analyse et le contrôle de systèmes électriques contenant des CPLs. Les contributions sont énumérées comme suit. (i) Des conditions simplement vérifiables sont proposées pour certifier la non existence d’états en régime permanent pour des réseaux multi-ports, à courant alternatif avec une distribution de CPLs. Ces conditions, qui reposent sur les inégalités matricielles linéaires, permettent d’écarter les valeurs des puissances des charges qui produiraient certainement un effondrement de la tension sur l’ensemble du réseau. (ii) Pour des modèles généraux de certains systèmes électriques modernes, y compris les réseaux de transmission à courant continu haute tension et les microréseaux, il est montré que, si des équilibres existent, il existe un équilibre caractéristique à haute tension qui domine tous les autres. En outre, dans le cas des systèmes d’alimentation en courant alternatif sous l’hypothèse de découplage standard, cet équilibre caractéristique s’avère stable à long terme. (iii) Une classe de systèmes port-Hamiltoniens, dans laquelle les variables de contrôle agissent directement sur l’équation de puissance, est explorée. Il est démontré que ces systèmes sont décalés de manière passive lorsque leurs trajectoires sont contraintes à des ensembles facilement définissables. Ces dernières propriétés sont exploitées pour analyser la stabilité de leurs équilibres intrinsèquement non nuls. Il a également été montré que la stabilité des réseaux électriques à courant continu multiports et des générateurs synchrones, tous deux connectés à des CPLs, peuvent naturellement être étudiée dans le cadre proposé. (iv) Le problème de la régulation de la tension de sortie du convertisseur buck-boost alimentant une CPL non connu est résolu. L’un des principaux obstacles à la conception de commandes linéaires classiques provient du fait que le modèle du système est de phase non minimale par rapport à chacune de ses variables d’état. Cette thèse rapporte un contrôleur adaptatif non linéaire capable de rendre un équilibre souhaité asymptotiquement stable; de plus, une estimation de la région d’attraction peut être calculée. (v) La dernière contribution concerne l’amortissement actif d’un système d’alimentation de petite taille à courant continu avec une CPL. Au lieu de connecter des éléments passifs peu pratiques et énergétiquement inefficaces au réseau existant, l’ajout d’un convertisseur de puissance contrôlé est exploré. La contribution principale rapportée ici est la conception d’une loi de contrôle non linéaire basée sur l’observateur pour le convertisseur. La nouveauté de la proposition réside dans le fait qu’il n’est pas nécessaire de mesurer le courant électrique du réseau ni la valeur de la CPL, soulignant ainsi son applicabilité pratique. L’efficacité du schéma de contrôle est ensuite validée par des expériences sur un réseau à courant continu réel.

Abstract: The continuously increasing demand of electrical energy has led to the conception of power systems of great complexity that may extend even through entire countries. In the vast majority of large-scale power systems the main primary source of energy are fossil fuels. Nonetheless, environmental concerns are pushing a major change in electric energy production practices, with a marked shift from fossil fuels to renewables and from centralized architectures to more distributed ones. One of the main challenges that distributed power systems face are the stability problems arising from the presence of the so-called Constant Power Loads (CPLs). These loads, which are commonly found in information and communication technology facilities, are known to reduce the effective damping of the circuits that energize them, which can cause voltage oscillations or even voltage collapse. In this thesis, the main contributions are focused in understanding and solving diverse problems found in the analysis and control of electrical power systems containing CPLs. The contributions are listed as follows. (i) Simply verifiable conditions are proposed to certify the non existence of steady states in general, multi-port, alternating current (AC) networks with a distributed array of CPLs. These conditions, which are based on Linear Matrix Inequalities, allow to discard the values of the loads’ powers that would certainly produce a voltage collapse in the whole network. (ii) For general models of some modern power systems, including High-Voltage Direct Current transmission networks and microgrids, it is shown that if equilibria exist, then there is a characteristic high-voltage equilibrium that dominates, entry-wise, all the other ones. Furthermore, for the case of AC power systems under the standard decoupling assumption, this characteristic equilibrium is shown to be long-term stable. (iii) A class of port-Hamiltonian systems, in which the control variables act directly on the power balance equation, is explored. These systems are shown to be shifted passive when their trajectories are constrained to easily definable sets. The latter properties are exploited to analyze the stability of their intrinsically non zero equilibria. It is also shown that the stability of multi-port DC electrical networks and synchronous generators, both with CPLs, can be naturally studied with the proposed framework. (iv) The problem of regulating the output voltage of the versatile DC buck-boost converter feeding an unknown CPL is addressed. One of the main obstacles for conventional linear control design stems from the fact that the system’s model is nonminimum phase with respect to each of its state variables. As a possible solution to this problem, this thesis reports a nonlinear, adaptive controller that is able to render a desired equilibrium asymptotically stable; furthermore an estimate of the region of attraction can be computed. (v) The last contribution concerns the active damping of a DC small-scale power system with a CPL. Instead of connecting impractical, energetically inefficient passive elements to the existing network, the addition of a controlled DC-DC power converter is explored. The main contribution reported here is the design of a nonlinear, observer-based control law for the converter. The novelty of the proposal lies in the non necessity of measuring the network’s electrical current nor the value of the CPL, highlighting its practical applicability. The effectiveness of the control scheme is further validated through experiments on a real DC network.

Mots clés: Réseaux électriques, Charges à Puissance Constante, Analyse, Commande

Keywords: Power systems, Constant Power Loads (CPLs), Analysis, Control

"Séminaire d'Automatique du plateau de Saclay" of iCODE

Séminaire le 25 Octobre 2019, 10h00 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Antoine Girard & Sajad Naderi

10:00-11:00 Antoine Girard (L2S, Gif-Sur-Yvette, France)

Title: A Symbolic Control Approach to the Programming of Cyber-Physical Systems

Abstract: Autonomous vehicles, intelligent buildings or robots promise to transform the everyday life of our society in all its dimensions (transport, housing, industry, health, assistance to the elderly ...). These systems are examples of cyber-physical systems (CPS) resulting from the integration of computer components and physical processes. The development of these systems is often complex (due to cyber-physical interactions) and with critical safety requirements.
In this talk, I will present the first steps towards developing a framework for CPS programming that will enable fast and safe development of their functionality through a high-level programming language. The originality of the approach is to consider that programs are not intended to be executed on the digital platform made up of computer components, but on the cyber-physical platform, which additionally includes the physical part of the system. Thus, high-level programs do not specify the behavior of the computer components but directly that of the cyber-physical system. Then, an automatic synthesis tool uses a model of the physical process to generate low-level control algorithms that enforce the specified behavior.
I will introduce a high-level language for CPS directly inspired by the formalism of hybrid automata. Following the paradigm of 'correct by construction synthesis', low-level control algorithms are synthesized by symbolic control techniques. The key concept of symbolic control is that of the symbolic model, which is a dynamic finite state system, obtained by abstracting physical trajectories on a finite set of symbols. When symbolic and physical dynamics are formally linked by a behavioral relation (e.g., simulation or bisimulation), controllers synthesized for the symbolic model using discrete synthesis techniques can be refined to controllers certified for the physical system. I will provide illustrating examples from the domain of autonomous vehicles.

Biography: Antoine Girard is a senior researcher at CNRS and a member of the Laboratory of Signals and Systems. He received the Ph.D. degree in applied mathematics from Grenoble Institute of Technology, in 2004. From 2004 to 2006, he held postdoctoral positions at University of Pennsylvania and Université Grenoble-Alpes. From 2006 to 2015, he was an Assistant/Associate Professor at the Université Grenoble-Alpes. His main research interests deal with analysis and control of hybrid systems with an emphasis on computational approaches, formal methods and applications to cyber-physical systems. Antoine Girard received the George S. Axelby Outstanding Paper Award from the IEEE Control Systems Society in 2009. In 2014, he was awarded CNRS Bronze Medal. In 2015, he was appointed as a junior member of the Institut Universitaire de France (IUF). In 2016, he was awarded an ERC Consolidator Grant. In 2018, he received the first Test of Time Award from the International Conference on Hybrid Systems: Computation and Control and the European Control Award.

11:00-12:00 Sajad Naderi (Eindhoven University of Technology, The Netherlands)

Title: Model order reduction for linear time delay systems based on energy functionals

Abstract: In this talk, I first present a model order reduction approach for asymptotically stable linear time systems with point-wise delays. This approach, which can be regarded as an extension of existing balanced model order reduction techniques for linear delay-free systems, is based on energy functionals that characterize observability and controllability properties of time delay system. This type of approach provides an a priori bound on the reduction error. Moreover, the resulting reduced model is an asymptotically stable time delay system with the same delay-structure as the original model. In the second part of the presentation, I introduce an extended model order reduction technique for time delay systems. This extension is beneficial when the preservation of physical interconnection structures or uncertainties is desired.

Biography: Sajad Naderi received his MSc in control systems from the school of electrical and computer engineering at the University of Tehran, Iran. For his MSc thesis, he worked on the design and implementation of nonlinear adaptive controllers for the speed control of PMSM drives. He is currently pursuing a PhD degree within the dynamics and control group of the mechanical engineering department at Eindhoven University of Technology, The Netherlands. His PhD research focuses on model order reduction of infinite-dimensional systems, with application to managed pressure drilling automation. In the scope of this industrial project, he has spent 1.5 years of his PhD at the Norwegian company Kelda Drilling Controls in Porsgrunn, Norway.

Contribution à l'estimation d'état par méthodes ensemblistes ellipsoidales et zonotopiques

Dory MERHY
Soutenance de thèse de doctorat le 24 Octobre 2019, 14h00 à CentraleSupelec (Gif-sur-Yvette) Amphi F3-06

Membres du jury :

Mme Cristina STOICA MANIU          CentraleSupélec - L2S          Directrice de thèse
M. Teodoro ALAMO                         Université de Séville            Co-encadrant
M. Eduardo F. CAMACHO                  Université de Séville         Co-encadrant
M. Antoine GIRARD                            CNRS - L2S                          Examinateur
M. Christophe COMBASTEL                Université de Bordeaux      Examinateur
M. Dan SELISTEANU                       Université de Craiova             Rapporteur
M. José Manuel CARO BRAVO               Université de Huelva              Rapporteur
M. Nacim RAMDANI                        Université d'Orléans            Rapporteur

 

Résumé : Dans le contexte des systèmes dynamiques, cette thèse développe des techniques d'estimation d'état ensemblistes pour des différentes classes de systèmes. On considère pour cela le cas d'un système standard linéaire invariant dans le temps soumis à des perturbations, des bruits de mesure et des incertitudes inconnus, mais bornés. Dans une première étape, une technique d'estimation d'état ellipsoïdale est étendue, puis appliquée sur un modèle d'octorotor utilisé dans un contexte radar. Une extension de cette approche ellipsoïdale d'estimation d'état est proposée pour des systèmes descripteurs. Dans la deuxième partie, nous proposons une méthode fondée sur la minimization du P-rayon d'un zonotope, appliquée à un modèle d'octorotor. Cette méthode est ensuite étendue pour traiter les systèmes affines par morceaux. Dans la continuité des approches précédentes, un nouveau filtre de Kalman sous contraintes zonotopiques est proposé dans la dernière partie de cette thèse. En utilisant la forme duale d'un problème d'optimisation, l'algorithme projète l'état sur un zonotope qui forme l'envelope de l'ensemble des contraintes auxquelles l'état est soumis. La complexité de l'algorithme est ensuite améliorée en remplaçant le zonotope initial par une forme réduite en limitant son nombre de générateurs.

Title : Contribution to ellipsoidal and zonotopic set-membership state estimation

Abstract : In the context of dynamical systems, this thesis focuses on the development of robust set-membership state estimation procedures for different classes of systems. We consider the case of standard linear time-invariant systems, subject to unknown but bounded perturbations and measurement noises. The first part of this thesis builds upon previous results on ellipsoidal set-membership approaches. An extended ellipsoidal set-membership state estimation technique is applied to a model of an octorotor used for radar applications. Then, an extension of this ellipsoidal state estimation approach is proposed for descriptor systems. In the second part, we propose a state estimation technique based on the minimization of the P-radius of a zonotope, applied to the same model of the octorotor. This approach is further extended to deal with piecewise affine systems. In the continuity of the previous approaches, a new zonotopic constrained Kalman filter is proposed in the last part of this thesis. By solving a dual form of an optimization problem, the algorithm projects the state on a zonotope forming the envelope of the set of constraints that the state is subject to. Then, the computational complexity of the algorithm is improved by replacing the original possibly large-scale zonotope with a reduced form, by limiting its number of generators.

Mots-clés : Estimation d'état ensembliste, inégalité matrice linéaire, zonotopes, ellipsoides

Keywords : Set-membership state estimation, linear matrix inequality, zonotopes, ellipsoids

 

«Trajectoires optimales et commande prédictive d'un quadricoptère pour la réalisation de plans de vol cinématographiques » « Optimal trajectory planning and predictive control for cinematographic flight plans with quadrotors »

Gauthier ROUSSEAU
Soutenance de thèse de doctorat le 18 Octobre 2019, 10h00 à

  CentraleSupélec (Gif-sur-Yvette) - Bâtiment Bouygues - Amphi SC.071

 

Membres du jury:

Mme Cristina STOICA MANIU            L2S/CentraleSupélec Directrice de thèse
M. Mathieu BABEL  Parrot Drones Co-encadrant
Mme Sihem TEBBANI L2S/CentraleSupélec Co-encadrante
M. Nicolas MARTIN Parrot Drones Co-encadrant - invité
M. Sylvain BERTRAND ONERA Examinateur
M. Pedro CASTILLO-GARCIA Université de Technologie de Compiègne Examinateur
M. Nicolas LANGLOIS ESIGELEC/IRSEEM Examinateur
M. Nicolas PETIT  Mines ParisTech Rapporteur
M. Didier THEILLIOL Université de Lorraine Rapporteur

 

Résumé: Cette thèse s'intéresse à la réalisation autonome de plans de vol cinématographiques par un quadrotor équipé d'une caméra. Ces plans de vol consistent en une série de points de passage à rejoindre successivement, en adoptant diverses méthodes de prise de vue et en respectant des références de vitesse ainsi que des couloirs de vols. Une étude approfondie de la dynamique du quadrotor est tout d'abord proposée, et utilisée pour construire un modèle linéarisé du drone autour de l'équilibre de vol stationnaire. L'analyse de ce modèle linéaire permet de mettre en évidence l'impact de l'inertie des rotors du drone dans sa dynamique, notamment l'apparition d'un comportement à non minimum de phase en roulis ou tangage, lorsque les moteurs sont inclinés. Dans un second temps, deux algorithmes de génération de trajectoires lisses, faisables et adaptées à la cinématographie sont proposés. La faisabilité de la trajectoire est garantie par le respect de contraintes sur ses dérivées temporelles, adaptées pour la cinématographie et obtenues grâce à l'étude du modèle non linéaire du drone. Le premier repose sur une optimisation bi-niveaux d'une trajectoire polynomiale par morceaux, dans le but de trouver la plus rapide des trajectoires à minimum de jerk permettant d'accomplir la mission. Le second algorithme consiste en la génération de trajectoires B-spline non-uniformes à durée minimale. Pour les deux solutions, une étude de l’initialisation du problème d'optimisation est présentée, de même qu'une analyse de leurs avantages et limitations. Pour ce faire, elles sont notamment confrontées à des simulations et vols extérieurs. Enfin, une loi de commande prédictive est proposée pour asservir les mouvements de la caméra embarquée de manière douce mais précise.

Abstract: This thesis deals with the autonomous performance of cinematographic flight plans with camera equipped quadrotors. These flight plans consists in a series of waypoints to join while adopting various camera behaviors, along with speed references and flight corridors. First, an in depth study of the nonlinear dynamics of the drone is proposed, which is then used to derive a linear model of the system near the hovering equilibrium. An analysis of this linear model allows us to emphasize the impact of the inertia of the propellers when the latter are tilted, such as the apparition of a non minimum phase behavior of the pitch or dynamics. Then, two algorithms are proposed to generate smooth and feasible trajectories suited for cinematography. The feasibility of the trajectory is ensured by constraints on its time derivatives, suited for cinematography and obtained with the use of the nonlinear model of the drone. The first algorithm proposed in this work is based on a bi-level optimization of a piecewise polynomial trajectory and try to find the fastest feasible minimum jerk trajectory to perform the flight plan. The second algorithm consists in the generation of feasible, minimum time, non uniform B-spline. For both solutions, a study of the initialization of the optimization problem is proposed, as well as a discussion about their advantages and limitations. To this aim, they are notably confronted to simulations and outdoor flight experiments. Finally, a predictive control law is propose to smoothly but accurately control the on-board camera.

 

 

 

 

Séminaire d'Automatique du plateau de Saclay

Séminaire le 17 Octobre 2019, 10h00 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Islam Boussaada & Ludovic Sacchelli

10:00-11:00 Islam Boussaada (Inria Saclay, Equipe DISCO & L2S, Gif-Sur-Yvette)

Title: Coalescence and Splitting Mechanisms of Spectral Values and their Effect on Stability: Towards a New Framework for Reduced Complexity Pole-placement Design for Time-Delay Systems

Abstract: For linear delay-differential equations, a question of ongoing interest is to determine conditions on the equation parameters that guarantee exponential stability and stabilization of solutions. This talk starts by a review of an old design method for time-delay systems called finite pole-placement. Its advantages and limitations shall be stressed. Next, some recent results showing a link between the stable manifold and the manifold corresponding to a given multiplicity of a spectral value shall be presented, hence enabling a spectral abscissa assignment. After a motivation of the tracking of multiple spectral values for analysis/control perspectives, some existing links between Birkhoff’s interpolation problem and a result due to Pólya and Szegö on the number of quasipolynomial's roots in a horizontal strip shall be revisited. Later, hints of an analytic proof of the dominancy of the quasipolynomial's root will be presented, setting up a reduced-complexity delayed stabilizing design. Sensitivity of the control design with respect to the parameters' variation will be discussed. To overcome the sensitivity of multiple roots, an extension of the approach to real distinct pole assignment shall be presented. Finally, various reduced order examples will illustrate the applicative perspectives of the proposed control approach.

Biography: Islam Boussaada received his Master in Mathematics from University Tunis II, and an M.Sc. degree in Pure Mathematics from University Paris 7 in 2004. In December 2008, he defended his Ph.D. degree in Mathematics from University of Rouen Normandy. In June 2016, he received his HDR degree (French Habilitation) in Physics from University Paris Saclay-University Paris Sud. In 2010, IB was appointed for two years as a post-doctoral fellow in the control of time-delay systems at L2S, Supelec-CNRS-University Paris Sud. Since 2012, he has been an associate professor at IPSA and an associate researcher at MODESTY Team of L2S. Since September 2017, IB is appointed permanent researcher at DISCO Team and full professor at IPSA where he headed the Aeronautical and Aerospace Systems department from September 2017 till May 2019.
Since September 2018 untill August 2020, IB is a researcher in temporary secondment at Inria Saclay-DISCO Team. His research interests belong to the qualitative theory of dynamical systems and its application in control problems. It includes stability analysis and stabilization of linear/nonlinear dynamical systems, analysis of parametric systems, analysis of delay induced dynamics, nonhyperbolic dynamics, analysis of algebraic dierential systems, control of active vibrations, dynamics of biochemical networks. IB is co-author of a monograph and co-editor of a contributed book, both published in Springer series, as well as co-author of more than 60 peer-reviewed publications. He co-organized the 4th GDRI DelSys's Workshop on Observing and Controlling Complex Dynamical Systems (November 2015), as well as the 1st GDRI Spa-Disco's workshop on Delays and Constraints in Distributed Parameters Systems (November 2017), both funded by CNRS and held at CentraleSupelec (Gif sur Yvette). At the occasion of the 20th World Congress of the International Federation of Automatic Control (IFAC) (Toulouse, July 2017), IB co-organized an invited session "Frequency domain Techniques for Time-delay Systems". At the occasion of the 13th-15th IFAC Workshop on Time-delay Systems (Istanbul 2016, Budapest 2018, Sinaia 2019), IB co-organized thematic sessions on Spectral Methods for Rightmost Roots Characterization in LTI Time-delay Systems. Since September 2018, IB is co-leading the national research group GT OSYDI of the CNRS/GDR MACS and is a deputy director of the IRS iCODE Institute of the University Paris Saclay.

11:00-12:00 Ludovic Sacchelli (Lehigh University, Pennsylvania, USA)

Title: Stabilization of non-uniformly observable system


Abstract: A common strategy in dynamic output feedback stabilization is to apply a state feedback to an observer in order to stabilize the coupled state-observer system. It is well known that global stabilizability, paired with uniform observability, implies semi-global stabilisability by dynamic output feedback. However in many generic cases, the system is not uniformly observable, and usual strategies for semi-global stabilization break down. New approaches need to be explored to resolve this issue. We will present case studies to give an outlook for the challenges raised by this problem and highlight a promising answer based on the idea of unitary embeddings of control systems.

Biography: Ludovic Sacchelli is a visiting assistant professor in the Mathematics Department of Lehigh University, in Bethlehem, Pennsylvania. He graduated from Ecole Normale Supérieure de Cachan and received his master's degree in analysis of PDEs from Paris-Sud University in 2015. He obtained a Ph.D. in applied mathematics from Ecole Polytechnique in 2018 on the topic of sub-Riemannian geometry. Ludovic spent the following year as a postdoc in the Electrical Engineering Department of University of Toulon (LIS Lab) before departing for his current position. His research interests lie in sub-Riemannian geometry, control theory and observability.

Techniques champ proche: Du système antennaire à l'application sur le terrain

Mohammed SERHIR
Habilitation à Diriger des Recherches (HDR) le 10 Octobre 2019, 14h00 à

Cette soutenance aura lieu à CentraleSupélec – Campus de Paris-Saclay

Lieu: Amphi IV  dans le Bât. Eiffel

Composition du jury:

M. Jean-Yves DAUVIGNAC Professeur à l’Université de Nice-Sophia Antipolis, LEAT Rapporteur
M. Raphael GILLARD Professeur à l’INSA de Rennes, IETR Rpporteur
M. Sébastien LAMBOT Professeur à l’Université Catholique de Louvain, Louvain Georadar Research Centre Rpporteur
M. Dominique LESSELIER Directeur de Recherche CNRS, L2S Examinateur
M. Lionel PICHON Directeur de Recherche CNRS, GeePs Examinateur
M. Jean-Michel GEFFRIN Ingénieur de Recherche CNRS, Institut Fresnel Examinateur
M. Régis GUINVARC'H Professeur à CentraleSupélec, SONDRA Invité

Résumé:

Dans ce travail de synthèse, je donne un aperçu de mes activités de recherche en électromagnétisme appliqué. Axées sur la modélisation comportementale, mes premières contributions visaient à proposer une passerelle entre la simulation numérique et la caractérisation expérimentale des antennes. Dans un deuxième temps, j’évoque mes apports dans le thème de la caractérisation d’antennes en champ proche à travers la méthode matricielle pour le calcul du champ lointain. Cette activité expérimentale est à l’origine de plusieurs collaborations nationales et internationales. Enfin, je développe mes travaux sur l’imagerie micro-onde appliquée au radar de sol. J’y présente les conclusions de la modélisation du problème direct par la méthode DGTD et la résolution du problème inverse par Linear Sampling Method. Les applications étudiées sont la détection des systèmes racinaires des arbres et la localisation de câbles souterrains en situation réelle. Les travaux de recherche sur l’imagerie du proche sous-sol sont pluridisciplinaires et constituent le point de convergence de mes activités précédentes y compris les développements antennaires dédiés au radar de sol.

 

Ps:Vous êtes cordialement invités au pot qui suivra dans la salle du conseil dans l'aile B4 du Bât. Breguet

Control of automated vehicles and their influence on traffic

Séminaire le 3 Octobre 2019, 14h00 à CentraleSupelec (Gif-sur-Yvette) Salle des séminaires du L2S
Karl H. Johansson

Abstract: Automated and connected road vehicles enable large-scale control and optimisation of the transport system with the potential to radically improve fuel efficiency, decrease the environmental footprint, and enhance safety. In this talk we will focus on automated heavy-duty vehicle platooning, which is currently being implemented and evaluated by several truck manufacturers world-wide. We will discuss how to deploy feedback control of individual platoons utilising the cellular communication infrastructure and how such controlled platoons can be used improve overall traffic conditions. It will be argued that the average total variation of traffic density can be reduced and thereby creating incentives for platooning beyond fuel savings and driver support. Extensive experiments done on European highways will illustrate system performance and safety requirements. The presentation will be based on joint work with collaborators at KTH and at the truck manufacturers Scania and Volvo.

Biography: Karl H. Johansson is Professor at the School of Electrical Engineering and Computer Science, KTH Royal Institute of Technology. He received MSc and PhD degrees from Lund University. He has held visiting positions at UC Berkeley, Caltech, NTU, HKUST Institute of Advanced Studies, and NTNU. His research interests are in networked control systems, cyber-physical systems, and applications in transportation, energy, and automation networks. He has received several best paper awards and other distinctions from IEEE, IFAC and ACM. He has been awarded Distinguished Professor with the Swedish Research Council and Wallenberg Scholar with the Knut and Alice Wallenberg Foundation. He has received the Future Research Leader Award from the Swedish Foundation for Strategic Research and the triennial Young Author Prize from IFAC. He is Fellow of the IEEE and the Royal Swedish Academy of Engineering Sciences, and he is IEEE Distinguished Lecturer.

Control of automated vehicles and their influence on traffic

Séminaire le 3 Octobre 2019, 14h00 à CentraleSupelec (Gif-sur-Yvette) Salle des séminaires du L2S
Karl H. Johansson

Abstract: Automated and connected road vehicles enable large-scale control and optimisation of the transport system with the potential to radically improve fuel efficiency, decrease the environmental footprint, and enhance safety. In this talk we will focus on automated heavy-duty vehicle platooning, which is currently being implemented and evaluated by several truck manufacturers world-wide. We will discuss how to deploy feedback control of individual platoons utilising the cellular communication infrastructure and how such controlled platoons can be used improve overall traffic conditions. It will be argued that the average total variation of traffic density can be reduced and thereby creating incentives for platooning beyond fuel savings and driver support. Extensive experiments done on European highways will illustrate system performance and safety requirements. The presentation will be based on joint work with collaborators at KTH and at the truck manufacturers Scania and Volvo.

Biography: Karl H. Johansson is Professor at the School of Electrical Engineering and Computer Science, KTH Royal Institute of Technology. He received MSc and PhD degrees from Lund University. He has held visiting positions at UC Berkeley, Caltech, NTU, HKUST Institute of Advanced Studies, and NTNU. His research interests are in networked control systems, cyber-physical systems, and applications in transportation, energy, and automation networks. He has received several best paper awards and other distinctions from IEEE, IFAC and ACM. He has been awarded Distinguished Professor with the Swedish Research Council and Wallenberg Scholar with the Knut and Alice Wallenberg Foundation. He has received the Future Research Leader Award from the Swedish Foundation for Strategic Research and the triennial Young Author Prize from IFAC. He is Fellow of the IEEE and the Royal Swedish Academy of Engineering Sciences, and he is IEEE Distinguished Lecturer.

Commande de systèmes plats avec contraintes et Applications de la Commande sans Modèle aux quadrotors et au Cloud Computing

Madame Maria BEKCHEVA
Soutenance de thèse de doctorat le 11 Juillet 2019, 14h30 à

   CentraleSupélec, 3 rue Joliot Curie, 91192, Gif-sur-Yvette 
Salle : Amphi II, Bât. Eiffel

 

Composition du jury proposé :

M. Hugues MOUNIER Université Paris-Sud Directeur de thèse
M. Luca GRECO Université Paris-Sud Co-directeur de thèse
M. Emmanuel DELALEAU ENIB Rapporteur
M. Didier THEILLIOL CRAN Rapporteur
Mme Mireille BAYART CRISTAL Examinateur
M. Michel FLIESS Laboratoire LIX - Ecole Polytechnique Examinateur
M. Cédric JOIN CRAN Examinateur
M. Silviu Iulian NICULESCU CNRS- L2S-CentraleSupelec Examinateur

Résumé : 

La première partie de la thèse est consacrée à la commande avec contraintes de systèmes différentiellement plats. Deux types de systèmes sont étudiés : les systèmes non linéaires de dimension finie et les systèmes linéaires à retards. Nous présentons une approche unifiée pour intégrer les contraintes d'entrée/état/sortie dans la planification des trajectoires. Pour cela, nous spécialisons les sorties plates (ou les trajectoires de référence) sous forme de courbes de Bézier. En utilisant la propriété de platitude, les entrées/états du système peuvent être exprimés sous la forme d'une combinaison de sorties plates (courbes de Bézier) et de leurs dérivées. Par conséquent, nous obtenons explicitement les expressions des points de contrôle des courbes de Bézier d'entrées/états comme une combinaison des points de contrôle des sorties plates. En appliquant les contraintes souhaitées à ces derniers points de contrôle, nous trouvons les régions faisables pour les points de contrôle de Bézier de sortie, c'est-à-dire un ensemble de trajectoires de référence faisables. Ce cadre permet d’éviter le recours, en général fort coûteux d’un point de vue informatique, aux schémas d’optimisation.     Pour résoudre les incertitudes liées à l'imprécision de l'identification et modélisation des modèles et les perturbations, nous utilisons la commande sans modèle (Model Free Control-MFC) et dans la deuxième partie de la thèse, nous présentons deux applications démontrant l'efficacité de notre approche : Nous proposons une conception de contrôleur qui évite les procédures d'identification du système du quadrotor tout en restant robuste par rapport aux perturbations endogènes (la performance de contrôle est indépendante de tout changement de masse, inertie, effets gyroscopiques ou aérodynamiques) et aux perturbations exogènes (vent, bruit de mesure). Pour atteindre notre objectif en se basant sur la structure en cascade d'un quadrotor, nous divisons le système en deux sous-systèmes de position et d'attitude contrôlés chacun indépendamment par la commande sans modèle de deuxième ordre dynamique. Nous validons notre approche de contrôle avec trois scénarios réalistes : en présence d'un bruit inconnu, en présence d’un vent variant dans le temps et en présence des variations inconnues de masse, tout en suivant des manœuvres agressives. Nous utilisons la commande sans modèle et les correcteurs « intelligents » associés, pour contrôler (maintenir) l'élasticité horizontale d'un système de Cloud Computing. Comparée aux algorithmes commerciaux d’Auto-Scaling, notre approche facilement implémentable se comporte mieux, même avec de fluctuations aigües de charge. Ceci est confirmé par des expériences sur le cloud public Amazon Web Services (AWS).

Mots-clés :

Platitude différentielle, Commande sans modèle, Commande des systèmes avec contraintes, Quadrotors, Cloud Computing.

Imaging with Electromagnetic Waves and Fields, from Eddy Current to Microwave

Séminaire le 4 Juillet 2019, 10h00 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Yu Zhong

Abstract: Imaging problems with electromagnetic waves and fields are of great interest due to non-intrusive inspection enabled by such imaging methods. In this talk, two major imaging methods in two different frequency bands will mainly be discussed, eddy current imaging at low frequency and microwave imaging at resonant frequency regime. As these two types of problems are nonlinear and unstable, from mathematical perspectives, one will show, in each, how these difficulties are specifically handled.

In the first part, the physical mechanism of eddy current inspection will be discussed, followed by a full description of an inspection system. An imaging method that could work with the measured eddy current signals will then be proposed. It includes a forward model for eddy current interactions with defects, an experimental signal calibration model, a defect model for inversion, and an optimization scheme. It will be shown how these bricks work together to provide imaging results from phaseless eddy current signals.
In the second part, the highly nonlinear inverse scattering problems (ISPs) will be shown how to be efficiently tackled by the recently proposed contraction integral equation for inversion (CIE-I), in both three-dimensional (3-D) problems and 2-D problems with phaseless data. With the CIE-I, the non-linearity of ISPs is largely remedied by suppressing multiple scattering effects within the inversions, without compromising the physical model accuracy. This is very important when handling the computationally costly 3-D ISPs, since  each iteration of inversion might cost many computational resources. Compared to conventional imaging methods with the well-known Lippmann-Schwinger integral equation (LSIE), this new imaging method with CIE-I shows much better performance when tackling both 3-D ISPs and 2-D ones with phaseless data, w.r.t. resolvability against non-linearity and convergence speed.

 

Biography: Yu Zhong received the B.E. and M.E. degrees in electronic engineering from Zhejiang University, Hangzhou, China, in 2003 and 2006, respectively, and the Ph.D. degree in electrical and computer engineering from the National University of Singapore, Singapore, in 2010. He was a Research Engineer and a Fellow with the National University of Singapore, from 2009 to 2013, then involved in a French-Singaporean MERLION Cooperative Program. Since 2014, he has been a Scientist with the Institute of High Performance Computing (IHPC), Agency for Science, Technology and Research, Singapore. He has been regularly invited to the Laboratoire des Signaux et Systèmes (L2S), Gif-sur-Yvette, France, as Senior Scientific Expert once per year since 2012. He was invited as a Visiting Professor to University of Trento, Italy, in June 2018. His current research interests include numerical methods for inverse problems associated with waves and fields, electromagnetic and acoustic modeling with complex materials, and non-destructive testing.

Probabilité et Mécanique Quantique: Loi de Bayes, Estimation de paramètres

Séminaire le 9 Mai 2019, 11h00 à CentraleSupelec (Gif-sur-Yvette) Salle des séminaires du L2S
Clément Pellegrini

Abstract. Dans cet exposé nous reviendrons sur le modèle mathématique décrivant l'expérience de Serge Haroche: "quantum non-demolition experiment" pour lequel il a reçu le prix Nobel de Physique. A travers ce modèle nous verrons comment la loi de Bayes apparait naturellement dans le contexte de la mécanique quantique: notamment dans le contexte des mesures indirectes. Nous verrons ensuite comment nous pouvons faire de l'estimation de paramètres sur ces modèles et comment on peut parler de stabilité du filtre sous-jacent. Cet exposé ne demande pas de prérequis de mécanique quantique, nous introduirons les concepts de base nécessaires.

Bio. Clément Pellegrini, Maitre de conférences à l'université Paul Sabatier Toulouse III depuis 2009
Post-doctorat sous la direction de Francesco Petrucionne à Durban 2008-2009
Doctorat sous la direction de Stéphane Attal à l'université Claude Bernard Lyon: thèse soutenue en 2008

COMMANDE PREDICTIVE POUR LE VEHICULE AUTONOME

Iris BALLESTEROS-TOLOSANA
Soutenance de thèse de doctorat le 26 Janvier 2018, 14h00 à CentraleSupelec (Gif-sur-Yvette) Amphi F3-05

The thesis work contained in this manuscript is dedicated to the Advanced Driving Assistance Systems, which has become nowadays a strategic research line in many car companies. This kind of systems can be seen as a first generation of assisted or semi-autonomous driving, that will set the way to fully automated vehicles. The first part of this memory focuses on the analysis and control of lateral dynamics control applications - Autosteer by target tracking and the Lane Centering Assistance System (LCA). In this framework, safety plays a key role, bringing into focus the application of different constrained control techniques for linear parameter-varying (LPV) models. Model Predictive Control (MPC) and Interpolation Based Control (IBC) have been the ones privileged in the present work. In addition, it is a critical feature to design robust control systems that ensure a correct behavior under system’s variation of parameters or in the presence of uncertainty. Robust Positive Invariance (RPI) theory tools are considered to design robust LPV control strategies with respect to large vehicle speed variations and curvature of the road changes. The second axis of this thesis is the optimization-based trajectory planning for overtaking and lane change in highways with anti-collision enhancements. To achieve this goal, an exhaustive description of the possible scenarios that may arise is presented, allowing to formulate an optimization problem which maximizes passenger comfort and ensures system constraints’ satisfaction.

Mots-clés :

COMMANDE,PREDICTIVE,VEHICULE,TEMPO,TRAJECTOIRE

Composition du jury proposé

M. Antonios TZES   NYU Abu Dhabi   Rapporteur

M. Arben CELA   Université Paris-Est, ESIEE Paris   Rapporteur

M. Saïd MAMMAR   Université d'Evry Val-d'Essonne   Examinateur

M. Eduardo FERNANDEZ CAMACHO   Universidad de Sevilla  Examinateur

M. François FAUVEL   Renault SAS   Examinateur

M. Sorin OLARU   CentraleSupelec   Directeur de these

M. Pedro RODRIGUEZ AYERBE   CentraleSupelec   CoDirecteur de these

M. Renaud DEBORNE   Renault SAS   Invité

M. Guillermo PITA GIL   Renault SAS   Invité

Evaluation de performance d’une ligne ferroviaire suburbaine partiellement équipée d’un automatisme CBTC

Juliette POCHET
Soutenance de thèse de doctorat le 12 Janvier 2018, 10h30 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40

En zone dense, la croissance actuelle du trafic sur les lignes ferroviaires suburbaines conduit les exploitants à déployer des systèmes de contrôle-commande avancés des trains, tels que les systèmes dits « CBTC » (Communication Based Train Control) jusque-là réservés aux systèmes de métro. Les systèmes CBTC mettent en œuvre un pilotage automatique des trains et permettent une amélioration significative des performances. Par ailleurs, ils peuvent inclure un module de supervision de la ligne en charge de réguler la marche des trains en cas d’aléa, améliorant ainsi la robustesse du trafic. Face au problème de régulation, la recherche opérationnelle a produit un certain nombre de méthodes permettant de répondre efficacement aux perturbations, d’une part dans le secteur métro et d’autre part dans le secteur ferroviaire lourd. En tirant profit de l’état de l’art et des avancées faites dans les deux secteurs, les travaux présentés dans ce manuscrit cherchent à contribuer à l’adaptation des fonctions de régulation des systèmes CBTC pour l’exploitation de lignes ferroviaires suburbaines. L’approche du problème débute par la construction de l’architecture fonctionnelle d’un module de supervision pour un système CBTC standard. Nous proposons ensuite une méthode de régulation basée sur une stratégie de commande prédictive et sur une optimisation multi-objectif des consignes des trains automatiques. Afin d’être en mesure d’évaluer précisément les performances d’une ligne ferroviaire suburbaine équipée d’un automatisme CBTC, il est nécessaire de s’équiper d’un outil de simulation microscopique adapté. Nous présentons dans ce manuscrit l’outil SNCF nommé SIMONE qui permet une simulation réaliste du point de vue fonctionnel et dynamique d’un système ferroviaire incluant un système CBTC. Les objectifs des travaux de thèse nous ont naturellement conduits à prendre part, avec l’équipe SNCF, à la spécification, à la conception et à l’implémentation de cet outil. Finalement, grâce à l’outil SIMONE, nous avons pu tester la méthode de régulation proposée sur des scénarios impliquant des perturbations. Afin d’évaluer la qualité des solutions, la méthode multi-objectif proposée a été comparée à une méthode de régulation individuelle basée sur une heuristique simple. La méthode de régulation multi-objectif propose de bonnes solutions au problème, dans la majorité des cas plus satisfaisantes que celles proposées par la régulation individuelle, et avec un temps de calcul jugé acceptable. Le manuscrit se termine par des perspectives de recherche intéressantes.

Membres du jury :

M. Guillaume SANDOU, Professeur, CentraleSupélec, FRANCE - Directeur de thèse
M. Joaquin RODRIGUEZ, Directeur de Recherche, Ifsttar, FRANCE - Rapporteur
M. François DELMOTTE, Professeur, Université d'Artois, FRANCE - Rapporteur
M. Sylvain BARO, Ingénieur, SNCF Réseau, FRANCE - Examinateur
Mme Sihem TEBBANI, Professeure, CentraleSupélec, FRANCE - Examinateur
Mme Evguenia DMITRIEVA, Ingénieur, RATP, FRANCE - Examinateur

S³ seminar : Non-negative orthogonal greedy algorithms for sparse approximation

Séminaire le 8 Décembre 2017, 10h30 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Thanh NGUYEN (CRAN, L2S)

Sparse approximation under non-negativity constraints naturally arises in several applications. Many sparse solvers can be directly extended to the non-negative setting. It is not the case of Orthogonal Matching Pursuit (OMP), a well-known sparse solver, which gradually updates the sparse solution support by selecting a new dictionary atom at each iteration. When dealing with non-negative constraints, the orthogonal projection computed at each OMP iteration is replaced by a non-negative least-squares (NNLS) subproblem whose solution is not explicit. Therefore, the usual recursive (fast) implementations of OMP do not apply. A Non-negative version of OMP (NNOMP) was proposed in the recent literature together with several variations. In my talk, I will first recall the principle of greedy algorithms, in particular NNOMP, and then, I will introduce our proposed improvements, based on the use of the active-set algorithm to address the NNLS subproblems. The structure of the active-set algorithm is indeed intrisically greedy. Moreover, the active-set algorithm can be called with a warm start, allowing us to fastly solve the NNLS subproblems. (Joint work with Charles Soussen (L2S), Jérôme Idier (LS2N), and El-Hadi Djermoune (CRAN).)

Localisation de sources distribuées cohérentes - Méthodes paramétriques et approches inverses

José PICHERAL
Habilitation à Diriger des Recherches (HDR) le 6 Décembre 2017, 14h00 à

Liste des membres du jury

Yide WANG

Pascal CHEVALIER

Jérôme ANTONI

Sylvie MARCOS

Jean-Hugh THOMAS

Philippe FOSTER

Séminaire d'Automatique du Plateau de Saclay : Necessary and sufficient condition for exponential synchronization of nonlinear systems

Séminaire le 30 Novembre 2017, 11h00 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Vincent Andrieu (CNRS Researcher, LAGEP-CNRS, Université de Lyon 1, France)

Based on recent works on transverse exponential stability, some necessary and sufficient conditions for the existence of a (locally) exponential synchronizer are established. We show that the existence of a structured synchronizer is equivalent to the existence of a stabilizer for the individual linearized systems (on the synchronization manifold) by a linear state feedback. This, in turns, is also equivalent to the existence of a symmetric covariant tensor field which satisfies a kind of Lyapunov inequality. Based on this property, we provide the construction of such synchronizer. We discuss then the possibility to achieve global synchronization.

Bio. Vincent Andrieu graduated in applied mathematics from “INSA de Rouen”, France, in 2001. After working in ONERA (French aerospace research company), he obtained a PhD degree from “Ecole des Mines de Paris” in 2005. In 2006, he had a research appointment at the Control and Power Group, Dept. EEE, Imperial College London. In 2008, he joined the CNRS-LAAS lab in Toulouse, France, as a “CNRS-chargé de recherche”. Since 2010, he has been working in LAGEP-CNRS, Université de Lyon 1, France. In 2014, he joined the functional analysis group from Bergische Universitäte Wuppertal in Germany, for two sabbatical years. His main research interests are in the feedback stabilization of controlled dynamical nonlinear systems and state estimation problems. He is also interested in practical application of these theoretical problems, and especially in the field of aeronautics and chemical engineering.

Séminaire d'Automatique du Plateau de Saclay : Observer design for nonlinear systems

Séminaire le 30 Novembre 2017, 10h00 à CentraleSupelec (Gif-sur-Yvette) Salle du conseil du L2S - B4.40
Pauline Bernard (PhD, PSL Reserch University, Systems and Control Center, MINES ParisTech)

Unlike for linear systems, no systematic method exists for the design of observers for nonlinear systems. However, observer design may be more or less straightforward depending on the coordinates we choose to express the system dynamics. In particular, some specific structures, called normal forms, have been identified for allowing a direct and easier observer construction. It follows that a common way of addressing the problem consists in looking for a reversible change of coordinates transforming the expression of the system dynamics into one of those normal forms, design an observer in those coordinates, and finally deduce an estimate of the system state in the initial coordinates via inversion of the transformation. This talk gives contributions to each of those three steps.
First, we show the interest of a new triangular normal form with continuous (non-Lipschitz) nonlinearities. Indeed, we have noticed that systems which are observable for any input but with an order of differential observability larger than the system dimension, may not be transformable into the standard Lipschitz triangular form, but rather into an "only  continuous" triangular form. In this case, the famous high gain observer no longer is sufficient, and we propose to use  homogeneous observers instead.
Another canonical form of interest is the Hurwitz linear form which admits a trivial observer. The question of transforming a nonlinear system into such a form has only been addressed for autonomous systems with the so-called Lunberger or Kazantzis-Kravaris observers. This design consists in solving a PDE and we show here how it can be extended to time-varying/controlled systems.
As for the inversion of the transformation, this step is far from trivial in practice, in particular when the domain and image spaces have different dimensions. When no explicit expression for a global inverse is available, numerical inversion usually relies on the resolution of a minimization problem with a heavy computational cost. That is why we have developed a method to avoid the explicit inversion of the transformation by bringing the observer dynamics (expressed in the canonical form coordinates) back into the initial system coordinates. This is done by dynamic extension, i.e. by adding some new coordinates to the system and transforming an injective immersion into a surjective diffeomorphism.

Bio. Pauline Bernard graduated from MINES ParisTech in 2014 with a Master degree in Applied Mathematics and Automatic Control. In 2017, she obtained her Ph.D. in Mathematics and Automatic Control at PSL Reserch University, prepared at the Systems and Control Center, MINES ParisTech under the supervision of Laurent Praly and Vincent Andrieu.

Pages