Accueil  >  GERI  >  SMRT  >  séminaires 1993-2013  >  séminaires 2012 

séminaires 2012

Modélisation des réseaux de transport

Vendredi 14 décembre 2012

de 14h à 17h, Marne-la-Vallée, Bâtiment Le Descartes2, 2 rue de la Butte Verte, 93166 Noisy-le-Grand Cedex, Salle Pascal. Plan d’accès  

  • Markos Papageorgiou Exploiting the Fundamental Diagram of Urban Networks for Feedback-Based Gating
    Traffic signal control for urban road networks has been an area of intensive research efforts for several decades, and various algorithms and tools have been developed and implemented to increase the network traffic flow efficiency. Despite the continuous advances in the field of traffic control under saturated conditions, novel and promising developments of simple concepts in this area remains a significant challenge, because some proposed approaches that are based on various meta-heuristic optimization algorithms can hardly be used in a real-time environment. To address this problem, the recently developed notion of network fundamental diagram (NFD) for urban networks is exploited to improve mobility in saturated traffic conditions. Specifically, a simple control design model is first developed, followed by the appropriate design of a simple, but highly efficient feedback regulator for well-targeted gating action. As a case study, the proposed methodology is applied to the urban network of Chania, Greece, using microscopic simulation. The results show that, with the proposed feedback gating scheme, the total delay in the network decreases significantly; the mean speed increases accordingly; while the reliability of the traffic conditions across different simulation replications features also a strong increase; compared to the base (non-gating) case. Additional investigations demonstrate that equivalent quality results are achievable even with very few real-time measurements from the oversaturated network, i.e. at a strongly reduced implementation cost compared to what is needed for the capture of the complete NFD. Related future work is outlined.
  • Nadir Farhi An algebraic Approach for Performance Bound Calculus on Transportation Networks
    We propose an adaptation of the basic techniques of the deterministic network calculus theory to the road traffic flow theory. Network calculus is a theory based on min-plus algebra. It uses algebraic techniques to compute performance bounds in communication networks, such as maximum end-to-end delays and backlogs. The objective of this article is to investigate the application of such techniques for determining performance bounds on road networks, such as maximum bounds on travel times. The main difficulty to apply the network calculus theory on road networks is the modeling of interaction of cars inside one road, or more precisely the congestion phase. We propose a traffic model for a single-lane road without passing, which is compatible with the network calculus theory. The model permits to derive a maximum bound of the travel time of cars through the road. Then, basing on that model, we explain how to extend the approach to model intersections and large-scale networks.

Mardi 23 octobre 2012

De 9h à 11h45 à Marne-la-Vallée (Salle MLV 6 LVMT - plan d'accès)

  • Mohamed Abdel-Aty, Prof. of Engineering, Dept. of Civil, Environmental & Construction Engineering, University of Central Florida An Overview of the Multi-Level Analysis in Traffic Safety
    In the first part of the seminar, an overview of various traffic safety studies by the presenter is provided. These studies vary in focus from the macroscopic to microscopic levels. Traffic safety data are inherently multi-level, and therefore analytical techniques need to address this phenomenon. Traditional crash prediction models are incapable of taking into account the multilevel data structure. Therefore they suffer from a common underlying limitation that each observation (e.g., a crash or a vehicle involvement) in the estimation procedure corresponds to an individual situation in which the residuals exhibit independence. However, this “independence” assumption may often not hold true since multilevel data structures exist extensively because of the traffic data collection and clustering process. Disregarding the possible within-group correlations may lead to the production of models with unreliable parameter estimates and statistical inferences. In the second half of the seminar, a 5×ST-level hierarchy is proposed to represent the general framework of multilevel data structures in traffic safety. The involvement and emphasis for different sub-groups of these levels depend on the different research purposes. To properly accommodate the potential cross-level heterogeneity and spatiotemporal correlation due to the multilevel data structure, a Bayesian hierarchical approach that explicitly specifies multilevel structure and reliably yields parameter estimates is introduced. Using Bayesian hierarchical models, the results from several case studies by the presenter are highlighted.

  • Jean-Patrick Lebacque (Ifsttar/GRETTIA) Méthodes pour la réconciliation de données pour les modèles de trafic du second ordre de type GSOM

jeudi 10 mai 2012

de 14h à 17h à l’Ifsttar Bron. Plan d'accès et transports

  • Archived ITS Data as a Foundation for Science and Innovation Robert Bertini (Portland State University, USA)14h-15h30
    Abstract Over the past 20 years the transportation field has witnessed a data revolution—some might say that we have transitioned from a data «desert» to a data «ocean». Join Prof. Robert Bertini as he traces aspects of his research career over the past 17 years and describes the ways in which this data «ocean» has enabled and influenced his work. We know that most transportation management, logistics, and operations systems, as well as intelligent transportation systems (ITS) applications are founded upon a data infrastructure, and sometimes these data are archived in a systematic manner. If managed properly, archived ITS data can provide a unique foundation for scientific discovery in the traffic and transportation field. Through careful processing and using innovative visualization techniques, transportation data can reveal fundamental traffic principles, behavior and phenomena, such as freeway merging at on-ramps and lane-drops and oscillations. Transport data may also be used to enable the optimization of the deployment of transportation resources, such as freeway sensors or bus stops.  Further, transport data, if analysed carefully can provide the basis for conducting evaluations and assessments of transportation system congestion and performance, at different geographic levels, such as facility, corridor or region. Finally, the inception of a «connected» transportation environment with vehicles and infrastructure devices communicating at high frequency, we are moving to yet another level of transportation data availability. But numerous questions remain about who will own and manage these data, how will privacy be maintained and more. What does the future hold in the field of transportation enabled by an avalanche of high resolution data?
  • Macroscopic Modelling of Intersections - Recent Advances Jean-Patrick Lebacque Ifsttar/GRETTIA 15h30-17h

jeudi 15 mars 2012

de 14h à 17h à l’École des Ponts ParisTech  (amphithéâtre Navier)
6-8 rue Blaise Pascal cité Descartes Champs sur Marne. Plan d'accès

  • Pedestrian dynamics: Experiment, Modeling and Application Armin Seyfried, Centre de recherche de Julich (Allemagne) 14h-15h30
    Abstract Multifunctional buildings in combination with a wide range of large-scale public events present new challenges for the quality of security concepts. To prevent critical situations optimal crowd management needs accurate and up-to-date information about the current status. To improve the safety of mass events we developed in the project Hermes an evacuation assistant performing faster than real-time simulations in order to identify in advance overcrowded areas. Such an application needs quantitatively validated models for pedestrian dynamics. But even the empirical data base to validate the models is contradictory and rudimentary.
    The first part of the talk gives an overview about the project. In the second part an introduction to the empirical data base of pedestrian dynamics is provided. After this critical review I introduce basic quantities of pedestrian streams (flow, density and velocity) along the measurement methods. Trajectories from well controlled experiments with up to 400 participants are used to quantify basic relations, like the velocity-density relation, and to examine how different measurement methods influence these relations. The third part focuses on modeling approaches continuous in space. Ideas of a force-based model are outlined. The model represents the velocity dependent shape of pedestrians by ellipses changing the size of their semiaxis with speed and provides a reliable and quantitative description of pedestrian dynamics.
  • Téléphonie mobile et qualité de service en TC Vincent Aguiléra LVMT / ENPC 15h30-17h
    Résumé Le fonctionnement particulier du réseau de téléphonie mobile en zone souterraine permet d'estimer en temps réel des indicateurs relatifs à la  qualité du service fourni, comme le respect de la grille horaire en  station ou la densité de voyageurs dans les trains.
    Les premiers résultats  de traitements de ces données, effectués dans le cadre d'un accord de  partenariat entre Orange R&D et le LVMT, seront présentés.