Les travaux présentés dans ce mémoire de thèse s’inscrivent dans le cadre des Systèmes de TransportIntelligents (STI). Bien que les premières études sur ces systèmes ont commencé dans les années 60, leurdéveloppement reposant sur les techniques de l’information et de la communication, a atteint sa maturitédans le début des années 80. Les STI, sont composés de différents systèmes et intègrent différents concepts(systèmes embarqués, capteurs intelligents, autoroutes intelligentes, . . .) afin d’optimiser le rendementdes infrastructures routières et répondre aux problèmes quotidiens des congestions. Ce mémoire présentequatre contributions dans le cadre du trafic routier et aborde les problèmes de l’estimation et de lacommande afin d’éliminer les problèmes de congestions « récurrentes ». Le premier point traite unproblème crucial dans le domaine des STI qui est celui de l’estimation. En effet, la mise en oeuvre delois de commande pour réguler le trafic impose de disposer de l’ensemble des informations concernantl’évolution de l’état du trafic. Dans ce contexte, deux algorithmes d’estimation sont proposés. Le premierrepose sur l’emploi du modèle METANET et les techniques de modes de glissement d’ordre supérieur. Lesecond est basé sur les CTM (Cell Transmission Models). Plusieurs études comparatives avec les filtresde Kalman sont proposées. La seconde contribution concerne la régulation du trafic. L’accent est mis surle contrôle d’accès isolé en utilisant les algorithmes issus du mode de glissement d’ordre supérieur. Cettecommande est enrichie en introduisant une commande intégrée combinant le contrôle d’accès et le routagedynamique. L’ensemble des résultats, validé par simulation, est ensuite comparé aux stratégies classiquesnotamment le contrôle d’accès avec l’algorithme ALINEA. La troisième contribution traite des problèmesde coordination. En effet, l’objectif est d’appliquer le principe de la commande prédictive pour contrôlerplusieurs rampes d’accès simultanément. L’ensemble des contributions ont été validées en utilisant desdonnées réelles issues en grande partie de mesures effectuées sur des autoroutes françaises. Les résultatsobtenus ont montré un gain substantiel en termes de performances tels que la diminution du trajet, dutemps d’attente, de la consommation énergétique, ainsi que l’augmentation de la vitesse moyenne. Cesrésultats permettent d’envisager plusieurs perspectives nouvelles de développement des recherches dansce domaine susceptibles d’apporter des solutions intéressantes. / The works presented in this PhD dissertation fit into the framework of Intelligent TransportationSystems. Although the beginnings of these systems have started since the 60s, their development, basedon information and communication technologies, has reached maturity during the early 80s. The ITS usesthe intelligence of different systems (embedded systems, intelligents sensors, intelligents highways, etc.)in order to optimize road infrastructures performances and respond to the daily problems of congestions.The dissertation presents four contributions into the framework of road traffic flow and tackles theestimation and control problems in order to eliminate or at least reduce the “recurrent" congestionsphenomena. The first point treats the problem of traffic state estimation which is of most importance inthe field of ITS. Indeed, the implementation and performance of any control strategy is closely relatedto the ability to have all needed information about the traffic state describing the dynamic behavior ofthe studied system. Two estimation algorithms are then proposed. The first one uses the “metanet"model and high order sliding mode techniques. The second is based on the so-called Cell TransmissionModels. Several comparative studies with the Kalman filters, which are the most used in road traffic flowengineering, are established in order to demonstrate the effectiveness of the proposed approaches. Thethree other contributions concern the problem of traffic flow control. At first, the focus is on the isolatedramp metering using an algorithm based on the high order sliding mode control. The second contributiondeals with the dynamic traffic routing problem based on the high order sliding mode control. Such controlstrategy is enriched by introducing the concept of integration, in the third contribution. Indeed, integratedcontrol consists of a combination of several traffic control algorithms. In this thesis the proposed approachcombines an algorithm of on-ramp control with a dynamic traffic routing control. The obtained results arevalidated via numerical simulations. The validated results of the proposed isolated ramp metering controlare compared with the most used ramp metering strategy : ALINEA. Finally, the last contributiontreats the coordination problems. The objective is to coordinate several ramps which cooperate andchange information in order to optimize the highway traffic flow and reduce the total travel time in theapplied area. All these contributions were validated using real data mostly from French freeways. Theobtained results show substantial gains in term of performances such as travel time, energetic consumptiondecreasing, as well as the increasing in the mean speed. These results allow to consider several furtherworks in order to provide more interesting and efficient solutions in the ITS field.

The focus of this thesis is the development and implementation of a neural network model predictive controller to be used for controlling the integrated recycle heat exchanger (Intrex) in a 300MW circulating fluidized bed (CFB) boiler. Discussion of the development of the controller will include data collection and preprocessing, controller design and controller tuning. The controller will be programmed directly into the plant distributed control system (DCS) and does not require the continuous use of any third party software. The intrexes serve as the loop seal in the CFB as well as intermediate and finishing superheaters. Heat is transferred to the steam in the intrex superheaters from the circulating ash which can vary in consistency, quantity and quality. Fuel composition can have a large impact on the ash quality and in turn, on intrex performance. Variations in MW load and airflow settings will also impact intrex performance due to their impact on the quantity of ash circulating in the CFB. Insufficient intrex heat transfer will result in low main steam temperature while excessive heat transfer will result in high superheat attemperator sprays and/or loss of unit efficiency. This controller will automatically adjust to optimize intrex ash flow to compensate for changes in the other ash properties by controlling intrex air flows. The controller will allow the operator to enter a target intrex steam temperature increase which will cause all of the intrex air flows to adjust simultaneously to achieve the target temperature. The result will be stable main steam temperature and in turn stable and reliable operation of the CFB.

ABSTRACT The repair and strengthening of concrete bridge members with CFRP has become increasingly popular over recent years. However, significant research is still needed in order to develop more robust guidelines and specifications. The research project aims to assist with improving design prosedures for damaged concrete members with the use of CFRP. This document summarizes the analysis and testing of full-scale 40’ foot long prestressed concrete (PSC) bridge girders exposed to simulated impact damage and repaired with carbon fiber reinforced polymers (CFRP) materials. A total of five AASHTO type II bridge girders fabricated in the 1960’s were taken from an existing bridge, and tested at the Florida Department of Transportation FDOT structures lab in Tallahassee, Florida. The test specimens were tested under static loading to failure under 4-point bending. Different CFRP configurations were applied to each of the girders. Each of the test girders performed very well as each of them held a higher capacity than the control girder. The repaired girders 5, 6 and 7 surpassed the control girder’s capacity by 10.88%, 15.9% and 11.39%. These results indicate that repairing laterally damaged prestressed concrete bridge girders with CFRP is an effective way to restore the girders flexural capacity.

L'objectif de cette thèse est de réaliser des missions diverses de navigation autonome en environnement intérieur et encombré avec des robots terrestres. La perception de l'environnement est assurée par un banc stéréo embarqué sur le robot et permet entre autres de calculer la localisation de l'engin grâce à un algorithme d'odométrie visuelle. Mais quand la qualité de la scène perçue par les caméras est faible, la localisation visuelle ne peut pas être calculée de façon précise. Deux solutions sont proposées pour remédier à ce problème. La première solution est l'utilisation d'une méthode de fusion de données multi-capteurs pour obtenir un calcul robuste de la localisation. La deuxième solution est la prédiction de la qualité de scène future afin d'adapter la trajectoire du robot pour s'assurer que la localisation reste précise. Dans les deux cas, la boucle de commande est basée sur l'utilisation de la commande prédictive afin de prendre en compte les différents objectifs de la mission : ralliement de points, exploration, évitement d'obstacles. Une deuxième problématique étudiée est la navigation par points de passage avec évitement d'obstacles mobiles à partir des informations visuelles uniquement. Les obstacles mobiles sont détectés dans les images puis leur position et vitesse sont estimées afin de prédire leur trajectoire future et ainsi de pouvoir anticiper leur déplacement dans la stratégie de commande. De nombreuses expériences ont été réalisées en situation réelle et ont permis de montrer l'efficacité des solutions proposées. / The aim of this thesis is to perform various autonomous navigation missions in indoor and cluttered environments with mobile robots. The environment perception is ensured by an embedded stereo-rig and a visual odometry algorithm which computes the localization of the robot. However, when the quality of the scene perceived by the cameras is poor, the visual localization cannot be computed with a high precision. Two solutions are proposed to tackle this problem. The first one is the data fusion from multiple sensors to perform a robust computation of the localization. The second solution is the prediction of the future scene quality in order to adapt the robot's trajectory to ensure that the localization remains accurate. In the two cases, the control loop is based on model predictive control, which offers the possibility to consider simultaneously the different objectives of the mission : waypoint navigation, exploration, obstacle avoidance. A second issue studied is waypoint navigation with avoidance of mobile obstacles using only the visual information. The mobile obstacles are detected in the images and their position and velocity are estimated in order to predict their future trajectory and consider it in the control strategy. Numerous experiments were carried out and demonstrated the effectiveness of the proposed solutions.

This thesis presents a predictive energy management controller for long-haul hybrid trucks. In a receding horizon control framework, the vehicle speed reference, battery energy reference, and engine on/off decision are optimized over a prediction horizon. A mixed-integer quadratic program (MIQP) is formulated by performing modelling approximations and by including the binary engine on/off decision in the optimal control problem. The branch-and-bound algorithm is applied to solve this problem. Simulation results show fuel consumption reductions between 10-15%, depending on driving cycle, compared to a conventional truck. The hybrid truck without the predictive control saves significantly less. Fuel consumption is reduced by 3-8% in this case. A sensitivity analysis studies the effects on branch-and-bound iterations and fuel consumption when varying parameters related to the binary engine on/off decision. In addition, it is shown that the control strategy can maintain a safe time gap to a leading vehicle. Also, the introduction of the battery temperature state makes it possible to approximately model the dynamic battery power limitations over the prediction horizon. The main contributions of the thesis are the MIQP control problem formulation, the strategy to solve this with the branch-and-bound method, and the sensitivity analysis.

This thesis describes the theory of model predictive control and application of the theory to synchronous drives. It shows explicit and on-line solutions and compares the results with classical vector control structure.

Confidentiel / Confidential

Un groupe de robots collaboratifs peut gérer des tâches qui sont difficiles, voire impossibles, à accomplir par un seul. On appelle un ensemble de robots coopérant un système multi-agents (SMA). L'interaction entre agents est un facteur clé dans la commande coopérative qui pose d'importants défis théoriques et pratiques. L'une des tâches du contrôle coopératif est le consensus dont l'objectif est de concevoir des protocoles de commande afin de parvenir à un accord entre leurs états respectifs. Cette thèse améliore la navigation pour les SMA, tout en tenant compte de certaines contraintes pratiques (modèle du robot et contraintes temporelles) dans la conception de contrôleurs coopératifs pour chaque agent, de manière décentralisée. Dans cette thèse, deux directions sont étudiées. D'une part, le taux de convergence est une spécification de performance importante pour la conception du contrôleur pour un système dynamique. La convergence rapide est toujours recherchée pour améliorer les performances et la robustesse. La plupart des algorithmes de consensus existants se concentrent sur la convergence asymptotique, où le temps d'établissement est infini. Cependant, de nombreuses applications nécessitent une convergence rapide généralement caractérisée par une stratégie de commande à temps fini. De plus, la commande à temps fini autorise certaines propriétés intéressantes, mais le temps de stabilisation dépend des conditions initiales des agents. L'objectif ici est de concevoir un protocole de consensus leader-follower à temps fixe pour les SMA décrits en temps continu. Ce problème est étudié en utilisant la théorie de la stabilisation à temps fixe, qui garantit que le temps de stabilisation est borné quelles que soient les conditions initiales. Les contrôleurs et les observateurs à modes glissants sont conçus pour que chaque agent résolve le problème du consensus à temps fixe lorsque le leader est dynamique. D'autre part, par rapport aux systèmes à temps continu, le problème du consensus dans un cadre à temps discret convient mieux aux applications pratiques en raison de la limitation des ressources de calcul pour chaque agent. Le modèle de commande prédictive (MPC) permet de gérer les contraintes de commande et d'état des systèmes. Dans cette thèse, cette méthode est appliquée pour traiter le problème du consensus en temps discret en laissant chaque agent résoudre, à chaque étape, un problème de commande optimale contraint impliquant uniquement l'état des agents voisins. Les performances de suivi sont également améliorées dans cette thèse en ajoutant de nouveaux termes à partir du MPC classique. Les contrôleurs proposés sont simulés et implémentés sur un groupe composé de plusieurs robots réels en utilisant ROS (Robotic Operating System). Dans cette thèse, quelques solutions correspondant au problème de la connexion entre plusieurs robots mobiles de manière décentralisée, du réglage des périodes d'échantillonnage et des paramètres de contrôle sont également abordées. / Nowadays, robots have become increasingly important to investigate hazardous and dangerous environments. A group of collaborating robots can often deal with tasks that are difficult, or even impossible, to be accomplished by a single robot. Multiple robots working in a cooperative manner is called as a Multi-Agent System (MAS). The interaction between agents to achieve a global task is a key in cooperative control. Cooperative control of MASs poses significant theoretical and practical challenges. One of the fundamental topics in cooperative control is the consensus where the objective is to design control protocols between agents to achieve a state agreement. This thesis improves the navigation scheme for MASs, while taking into account some practical constraints (robot model and temporal constraints) in the design of cooperative controllers for each agent, in a fully decentralized way. In this thesis, two directions are investigated. On one hand, the convergence rate is an important performance specification to design the controller for a dynamical system. As an important performance measure for the coordination control of MASs, fast convergence is always pursued to achieve better performance and robustness. Most of the existing consensus algorithms focus on asymptotic convergence, where the settling time is infinite. However, many applications require a high speed convergence generally characterized by a finite-time control strategy. Moreover, finite-time control allows some advantageous properties but the settling time depend on the initial states of agents. The objective here is to design a fixed-time leader-follower consensus protocol for MASs described in continuous-time. This problem is studied using the powerful theory of fixed-time stabilization, which guarantee that the settling time is upper bounded regardless to the initial conditions. Sliding mode controllers and sliding mode observers are designed for each agent to solve the fixed-time consensus tracking problem when the leader is dynamic. On the other hand, compared with continuous-time systems, consensus problem in a discrete-time framework is more suitable for practical applications due to the limitation of computational resources for each agent. Model Predictive Control (MPC) has the ability to handle control and state constraints for discrete-time systems. In this thesis, this method is applied to deal with the consensus problem in discrete-time by letting each agent to solve, at each step, a constrained optimal control problem involving only the state of neighboring agents. The tracking performances are also improved in this thesis by adding new terms in the classical MPC technique. The proposed controllers will be simulated and implemented on a team of multiple Mini-Lab Enova Robots using ROS (Robotic Operating System) which is an operating system for mobile robots. ROS provides not only standard operating system services but also high-level functionalities. In this thesis, some solutions corresponding to problem of connection between multiple mobile robots in a decentralized way for a wireless robotic network, of tuning of the sampling periods and control parameters are also discussed.

Over the past decades, attention has been brought to the importance of indoor air quality and the serious threat of bio-aerosol contamination towards human health. A novel idea to transport hazardous particles away from sensitive areas is to automatically control bio-aerosol concentrations, by utilising airflows from ventilation systems. Regarding this, computational fluid dynamics (CFD) may be employed to investigate the dynamical behaviour of airborne particles, and data-driven methods may be used to estimate and control the complex flow simulations. This thesis presents a methodology for machine-learning based control of particle concentrations in turbulent gas-solid flow. The aim is to reduce concentration levels at a 90 degree corner, through systematic manipulation of underlying two-phase flow dynamics, where an energy constrained inlet airflow rate is used as control variable. A CFD experiment of turbulent gas-solid flow in a two-dimensional corner geometry is simulated using the SST k-omega turbulence model for the gas phase, and drag force based discrete random walk for the solid phase. Validation of the two-phase methodology is performed against a backwards facing step experiment, with a 12.2% error correspondence in maximum negative particle velocity downstream the step. Based on simulation data from the CFD experiment, a linear auto-regressive with exogenous inputs (ARX) model and a non-linear ARX based neural network (NN) is used to identify the temporal relationship between inlet flow rate and corner particle concentration. The results suggest that NN is the preferred approach for output predictions of the two-phase system, with roughly four times higher simulation accuracy compared to ARX. The identified NN model is used in a model predictive control (MPC) framework with linearisation in each time step. It is found that the output concentration can be minimised together with the input energy consumption, by means of tracking specified target trajectories. Control signals from NN-MPC also show good performance in controlling the full CFD model, with improved particle removal capabilities, compared to randomly generated signals. In terms of maximal reduction of particle concentration, the NN-MPC scheme is however outperformed by a manually constructed sine signal. In conclusion, CFD based NN-MPC is a feasible methodology for efficient reduction of particle concentrations in a corner area; particularly, a novel application for removal of indoor bio-aerosols is presented. More generally, the results show that NN-MPC may be a promising approach to turbulent multi-phase flow control.

In underwater tracking and surveillance, the active towed array sonar presents a way of discovering and tracking adversarial submerged targets that try to stay hidden. The configuration consist of listening and emitting hydrophones towed behind a ship. Moreover, it has inherent limitations, and the characteristics of sound in the ocean are complex. By varying the pulse form emitted and the trajectory of the ship the measurement accuracy may be improved. This type of optimization constitutes a sensor management problem. In this thesis, a model of the tracking scenario has been constructed derived from Cramér-Rao bound analyses. A model predictive control approach together with information measures have been used to optimize a filter's estimated state of the target. For the simulations, the MATLAB environment has been used. Different combinations of decision horizons, information measures and variations of the Kalman filter have been studied. It has been found that the accuracy of the Extended Kalman filter is too low to give consistent results given the studied information measures. However, the Unscented Kalman filter is sufficient for this purpose.