Jacques Sau

Modelling of irrigation channel dynamics for controller design

Irrigation canals are complex hydraulic systems difficult to control. Design methods have been developed using linear control theory. To use this tool, a linear model of the dynamic is needed. This paper presents a simple model for canal reach dynamics. A reach transfer matrix is obtained by linearisation of Saint Venant equations near a steady flow regime. The accuracy of this transfer matrix is evaluated, in frequency and time domain.

Particle filter-based real-time estimation and prediction of traffic conditions

Real-time estimation and short-term prediction of traffic conditions is one of major concern of traffic managers and ITS-oriented systems. Model-based methods appear now as very promising ways in order to reach this purpose. Such methods are already used in process control (Kalman filtering, Luenberger observers). In the application presented in this paper, due to the high non linearity of the traffic models, particle filter (PF) approach is applied in combination with the well-known first order macroscopic traffic model. Not only shall we show that travel time prediction is successfully realized, but also that we are able to estimate, in real time, the motorway traffic conditions, even on points with no measurement facilities, having, in a way, designed a virtual sensor.

The root locus method: application to linear stability analysis and design of cooperative car-following models

A global framework for linear stability analyses of traffic models, based on the dispersion relation root locus method, is presented and is applied taking the example of a broad class of car-following (CF) models. This approach is able to analyse all aspects of the dynamics: long waves and short wave behaviours, phase velocities and stability features. The methodology is applied to investigate the potential benefits of connected vehicles, i.e. V2V communication enabling a vehicle to send and receive information to and from surrounding vehicles. We choose to focus on the design of the coefficients of cooperation which weights the information from downstream vehicles. The coefficients tuning is performed and different ways of implementing an efficient cooperative strategy are discussed. Hence, this paper brings design methods in order to obtain robust stability of traffic models, with application on cooperative CF models.

Original article: Data assimilation for real-time estimation of hydraulic states and unmeasured perturbations in a 1D hydrodynamic model

Abstract: Water management, in a variety of contexts and objectives, is a very important issue gaining increasing attention worldwide. In some places and during some periods, this is due to the scarcity of the water resource, and increasing competition for its use. In some others, it can be risk reduction due to flood events, or optimization of hydropower production along rivers. Hydraulic modeling, system analysis and automatic control are now parts of most water management projects. In order to operate hydraulic devices on irrigation canals or rivers, detailed information on the hydraulic state of the system must be available. This is particularly true when the control algorithms are based on Linear Quadratic Gaussian or Predictive Control approaches, using full state space models. Usually, the only known quantities are water levels, measured at limited locations. Sometimes, the discharge is known at specific locations (cross devices with gates, weirs, or hydropower turbines). The design of an observer is a very useful tool for reconstructing unmeasured data, such as discharges or water levels at other locations, unknown perturbations, such as inflows or outflows, and model parameters such as Manning-Strickler or hydraulic device discharge coefficients. Several approaches are able to provide such observers. The paper illustrates and compares the use of sequential Kalman Filter and sequential Particle Filter State Observer on these water management problems. Four scenarios have been selected to test the filters, based on twin experiences or using real field data. Both approaches proved to be efficient and robust. The Kalman Filter is very fast in terms of calculation time and convergence. The Particle Filter can handle the non-linear features of the model.

Integrating the Impact of Rain into Traffic Management: Online Traffic State Estimation Using Sequential Monte Carlo Techniques

A new approach to the integration of the effects of inclement weather into traffic management strategies is presented. Adverse weather conditions are a critical factor affecting traffic operations and safety. Previously, a methodology for the analysis of the impact of rain has been addressed, and this impact on key traffic indicators (e.g., free-flow speed, capacity) has been quantified. As a result of these quantification studies, a first parameterization of the fundamental diagram according to rain intensity has been proposed. Since the fundamental diagram represents the basis of many simulation tools, the goal is to develop weather-responsive traffic state estimation tools that can be useful for control applications and traffic management. More precisely, the online traffic state estimation takes place within a Bayesian framework with particle-filtering techniques (i.e., sequential Monte Carlo simulations) in combination with a parameterized first-order macroscopic model. This approach has already been validated for sensor diagnosis and accident detection. In this paper the goal is to show how the integration of the weather effects can improve this efficient tool. The approach is validated with real-world data from the ring road section in Lyon, France (eight sensors from a homogeneous section). The results from different scenarios show the benefits of integrating the impact of rain into traffic state estimation. Strategies to detect a rain event in time and space are also suggested.

Cooperative highway traffic

As cooperative systems (connected vehicles) enable communication and the exchange of information between vehicles and infrastructure, the communication capabilities are expected to lead to better active traffic management on urban motorways. Technological constraints must be the basis for any management strategy. If communication has been analytically proved to help stabilize traffic flow at a microscopic level, then realistic communication strategies should be evaluated by taking into consideration multiple perturbations such as sensor faults and driver cooperation. In this study, a three-layer multiagent framework was used to model and control the homogenization of traffic flow. The physical layer coordinated vehicle dynamics on the basis of a cooperative car-following model. This layer included cooperation derived from the communication and trust layers that, respectively, managed information and its reliability. Simulation results highlight the positive impacts of communication and control on the stability of traffic flow.

Probability hypothesis density filtering for real-time traffic state estimation and prediction

The probability hypothesis density (PHD) methodology is widely used by the research community for the purposes of multiple object tracking. This problem consists in the recursive state estimation of several targets by using the information coming from an observation process. The purpose of this paper is to investigate the potential of the PHD filters for real-time traffic state estimation. This investigation is based on a Cell Transmission Model (CTM) coupled with the PHD filter. It brings a novel tool to the state estimation problem and allows to estimate the densities in traffic networks in the presence of measurement origin uncertainty, detection uncertainty and noises. In this work, we compare the PHD filter performance with a particle filter (PF), both taking into account the measurement origin uncertainty and show that they can provide high accuracy in a traffic setting and real-time computational costs. The PHD filtering framework opens new research avenues and has the abilities to solve challenging problems of vehicular networks.

A probabilistic hypothesis density filter for traffic flow estimation in the presence of clutter

Prediction of traffic flow variables such as traffic volume, travel speed or travel time for a short time horizon is of paramount importance in traffic control. Hence, the data assimilation process in traffic modeling for estimation and prediction plays a key role. However, the increasing complexity, non-linearity and presence of various uncertainties (both in the measured data and models) are important factors affecting the traffic state prediction. To overcome this problem, new methodologies have been proposed. With this aim, in this paper we propose the use of the Probability Hypothesis Density (PHD) filter for traffic estimation. This methology is intensively studied, developed and improved for the purposes of multiple object tracking and consists in the recursive state estimation of several targets by using the information coming from an observation process. However, some issues need to be studied, especially the impact of the clutter (false alarm) intensity. The goal of this paper is to expose the potential of the PHD filters for real-time traffic state estimation and the choice of an appropriate clutter intensity. This investigation is based on a Cell Transmission Model (CTM) coupled with the PHD filter. It brings a novel tool to the state estimation problem and allows one to estimate the densities in traffic networks. In this work, we compare this PHD filter with the particle filter (PF) which has been successfully applied in traffic control and conclude that the PHD filter can be seen as a relevant alternative that opens new research avenues.

State-space linear stability analysis of platoons of cooperative vehicles

ABSTRACT Platoons of vehicles relying on inter-vehicle communication and control mechanisms have a great potential, e.g. to reduce fuel consumption and increase capacity. Linear stability analysis can be performed to investigate the propagation of perturbations in such cooperative platoons. Existing work, however, does not consider the stability analysis of the linearised global cooperative platoon system. Our car-following model framework is based on the cooperative relation introduced by Wilson, R. E. [2008. “Mechanisms for Spatio-Temporal Pattern Formation in Highway Traffic Models.” Philosophical Transactions of the Royal Society A (2008) 366, 2017–2032]. We present the state-space representation of the linearised dynamical system. We prove analytically, and illustrate in simulations, that a platoon of vehicles is always linearly-stable provided that the rational relations [Wilson, R. E., and J. A. Ward. 2011. “Car-Following Models: Fifty Years of Linear Stability Analysis: A Mathematical Perspective.” Transportation Planning and Technology 13, 2167–2176] hold and the coefficients of cooperation are non-negative. We present a brief analysis of controllability, considering heterogeneity in the control inputs. Finally, we show that for a closed platoon, the conditions for linear stability are the same as the conditions for stability of a one-dimensional infinite traffic flow.

Analyse pour le calage de modèles hydrauliques à surface libre : une approche par les théories des systèmes linéaires et de l’automatique

Tout projet mettant en œuvre un modele necessite une phase de calage. Pour un modele hydrodynamique base sur les equations de Saint-Venant, cela signifie ajuster des coefficients de frottement, des coefficients de debits aux ouvrages frontaux et lateraux, et des termes d’infiltration. Plusieurs methodes basees sur la minimisation d’un critere existent. Quelque soit la methode retenue, il y a une premiere question importante mais souvent negligee a se poser : « est-il possible d’identifier les parametres desires a partir des donnees disponibles ? » Cette premiere question peut etre reliee a la notion de « sensibilite ». Elle depend du lien reliant les parametres aux sorties, mais aussi des precisions souhaitees sur ces parametres et de la precision disponible sur les mesures. Le concept d’equifinalite souligne le fait que, dans certains cas, plusieurs jeux de parametres peuvent fournir les memes sorties d’un modele, modulo des incertitudes donnees. Ce concept est sujet a caution, car il est utilise pour critiquer un modele, intrinsequement, alors qu’il est tout autant lie au jeu de donnees utilise pour le calage. Une question importante est donc « quel serait le jeu de donnees optimal pour le calage ? » Le concept de « pire cas » peut apporter des elements de reponse a cette question, sous certaines hypotheses. L’article presente des methodologies permettant de repondre partiellement a ces questions, et les illustre sur un exemple de la litterature. Cela montre qu’un algorithme de minimisation peut etre utile, mais que savoir repondre aux deux questions posees ci-dessus l’est tout autant.