Gabriel Wainer

Gabriel Wainer
Carleton University


21 Octobre 2019, 14.00
salle des commissions
Campus St. Jérôme

Discrete-Event Modeling and Simulation Methodologies: Past, Present and Future

Modeling and Simulation methods have been used to better analyze the behavior of complex physical systems and it is now common to use simulation as a part of the scientific and technological discovery process. M&S advanced thanks to the improvements in computer technology, which, in many cases, resulted in the development of simulation software using ad-hoc techniques.

Formal M&S appeared in order to try to improve the development task of very complex simulation systems. Some of these techniques proved to be successful in providing a sound base for the development of discrete-event simulation models, improving the ease of model definition and enhancing the application development tasks; reducing costs and favoring reuse.

The DEVS formalism is one of these techniques, which proved to be successful in providing means for modeling while reducing development complexity and costs. DEVS model development is based on a sound theoretical framework. The independence of M&S tasks made possible to run DEVS models on different environments (personal computers, parallel computers, real-time equipment, and distributed simulators) and middleware.

We will present a historical perspective of discrete-event M&S methodologies, showing different modeling techniques. We will introduce DEVS origins and general ideas, and compare it with some of these techniques. We will then show the current status of DEVS M&S, and we will discuss a technological perspective to solve current M&S problems (including real-time simulation, interoperability and model-centered development techniques). We will show some examples of the current use of DEVS, including applications in different fields.

We will finally show current open topics in the area, which include advanced methods for centralized, parallel or distributed simulation, the need of real-time modeling techniques, and our view in these fields.

Gabriel Abba

Gabriel Abba
Université de Lorraine


23 Mai 2019, 14.00
salle des commissions
Campus St. Jérôme

De la commande des robots bipèdes à celle des humanoïdes

Le but de ce séminaire est de présenter la conception et la commande de robots bipèdes. De l’analyse des modèles de ces robots et de celle de la marche, on peut déduire une commande hybride non linéaire stabilisant la marche. A partir du cas le plus simple d’un robot plan au cas des bipèdes 3D, on arrive progressivement à déterminer les conditions nécessaires pour élaborer la commande d’un robot humanoïde ainsi que de résumer les principales caractéristiques de ces robots.

Angelo Alessandri (DIPTEM, Università degli Studi di Genova)

Angelo Alessandri
Università degli Studi di Genova


25 Avril 2019, 14.00
salle X133
Campus Toulon

Moving Horizon Methods for Constrained and Unconstrained Estimation of Dynamic Systems


Moving-horizon estimation (MHE) for dynamic systems relies on the simple idea of using a limited amount of most recent information to estimate the state variables at the current time instant, thus ensuring intrinsic robustness. The literature on MHE is vast since it has followed a favorable trend induced by the success of the model predictive control. MHE has been applied to estimation of many kind of systems in different application fields. The talk will concern the various approaches adopted to perform MHE to systems with linear, nonlinear, and switching dynamics and different optimization tools.

Hassan Hammouri (LAGEPP, UMR CNRS 5007)

Hassan Hammouri
Université Claude Bernard, Lyon


28 Mars 2019, 14.00
salle des commissions, bât Polytech
Campus de St. Jérôme

Existence de contrôles analytiques pour une équation algébro-différentielle analytique


Le problème est initialement posé, et résolu par J-P. Gauthier et I. Kupka [G-K] pour résoudre le problème d’existence d’entrées analytiques qui rendent observable un système analytique. Dans cet exposé, il s’agit d’une part d’étendre le résultat existant, d’autre part de réduire la preuve. Pour résoudre ce problème [G-K] ont fait appel à la stratification d’un certain cône tangent en tant qu’ensemble sous-analytique. Dans ce travail, je propose une preuve basée toujours sur la stratification des ensembles sous-analytique, mais qui ne nécessite pas le passage par la stratification du cône tangent.

Frédéric Mazenc (INRIA et L2S)

Frédéric Mazenc
Centralesupélec, Gif-sur-Yvette


21 Mars 2019, 14.00
salle des commissions, bât Polytech
Campus de St. Jérôme

Model reduction and predictor control


Recent results pertaining to the stabilizations of systems with long delays in the input will be presented. In particular, the so-called prediction model approach and prediction approaches will be described and applied to time-varying systems. Fundamental new tools used to established the results will be explained. In particular, a technique of stability analysis called « trajectory based approach » and its recent extensions will be presented.

Jean-Michel Coron (LJLL, UPMC)

Jean-Michel Coron
Université Pierre et Marie Curie, Paris


11 Octobre 2018, 14.00
salle de conférences Gérard JAUMES, bât Polytech
Campus de St. Jérôme

Stabilisation rapide de systèmes de contrôle

On présente des résultats et des méthodes sur la stabilisation rapide, voire en temps petit, de systèmes de contrôle modélisés par des équations différentielles ordinaires (avec une application à un glisseur) ou par des équations aux dérivées partielles en dimension 1 d’espace (systèmes hyperboliques, équations de la chaleur, équation de
Burgers visqueuse).

ZhiWu Li (Xidian University, China)

ZhiWu Li
Xidian University


21 Juin 2018, 15.30
salle des commissions, bât Polytech
Campus de St. Jérôme


Deadlock Analysis and Control of Resource Allocation Systems: Structural and Reachability Graph Approaches

This talk exposes the recent advances of deadlock problems in resource
allocation systems using Petri nets. The pertinent methodologies are
categorized by structural analysis and reachability graph analysis
techniques. The former, without enumerating the reachable states of a
system, utilize structural objects to derive a liveness-enforcing
supervisor, while its structure can be compact. The latter can usually
lead to an optimal supervisor with a minimal control structure subject
to a full state enumeration and solution to integer linear programming
problems. Open issues in this area are outlined.

Stéphane Gaubert (CMAP, École Polytechnique)

Stéphane Gaubert
École Polytechnique

21 Juin 2018, 14.00
salle des commissions, bât Polytech
Campus de St. Jérôme


Tropical analysis of timed Petri nets with priorities and application to performance evaluation of an emergency call center

We analyze a timed Petri net model of an emergency call center which processes calls with different levels of priority. The counter variables of the Petri net represent the cumulated number of events as a function of time. We show that these variables are determined by a piecewise linear dynamical system. We also prove that computing the stationary regimes of the associated uid dynamics reduces to the problem of computing a tropical prevariety, i.e., to solving a polynomial system over a tropical (min-plus) semifield. This leads to explicit formulae expressing the throughput of the uid system as a piecewise linear function of the resources, revealing the existence of different congestion phases. Numerical experiments show that the analysis of the fluid dynamics yields a good approximation of the real throughput. In this way, tropical geometry allows one to identify bottleneck resources. This works originates from a case study, concerning the analysis of the new organization of reception of the 17-18-112 emergency calls in the Paris area, currently deployed by Préfecture de Police. This is a joint work with Xavier Allamigeon and Vianney Boeuf.

Ahmed Chemori (LIRMM – CNRS University of Montpellier, France)

Ahmed Chemori
Université de Montpellier

24 Mai 2018, 14.00
salle des commissions, bât Polytech
Campus de St. Jérôme


Control of Complex Robotic Systems:
Challenges, Design and Experiments

Robotics was initially and for a long time guided by needs in industry. Indeed, the early years of robotics was largely focused on robot manipulators, used mainly for simple and repetitive automation tasks. The first industrial robot manipulator appeared in 1961 in the assembly lines of General Motors. The early control systems for robot manipulators were designed to control independently each axis of the robot as a Single-Input-Single-Output (SISO) linear system. Linear automatic control theory was then extensively used in this basic solution, where the coupling dynamics between the different axes of the robot were often neglected and the robot model significantly simplified. Beyond these issues, the main barriers to progress were the high cost of computation, the lack of good sensors, and the lack of fundamental understanding of robot dynamics. However, the progress of robotics and automation as well as their associated innovative applications has required the consideration of more and more complex tasks needing high performances. These challenging tasks required a deeply understanding of complex nonlinear dynamics of robots. Besides, it has also motivated the development of new theoretical advances in different control fields (robust, adaptive, etc.), which has consequently enabled more sophisticated applications. Nowadays, robotic control systems are highly advanced, including manipulation robotics, underwater robotics, aerial robotics, mobile robotics, medical robotics, parallel robotics, wearable robotics, humanoid robotics and more others. In this lecture the main challenges related to control of robotic systems will be emphasized, and illustrated through different applications in robotics. For each of these fields, the motivations and the need of developing advanced control schemes will be first highlighted. Then some proposed advanced control solutions will be introduced and illustrated through real-time experiments.