Les véhicules dits intelligents actuellement développés par la plupart des constructeurs automobiles, ainsi que les véhicules autonomes nécessitent des informations sur le contexte dans lequel ils évoluent. Certaines de ces informations (par exemple la courbure de la route, la forme des intersections, les limitations de vitesses) sont fournies en temps réel par le système de navigation qui exploite les données de cartes routières numériques. Des défauts résultant de l’évolution du réseau routier ou d’imprécisions lors de la collecte de données peuvent être contenus dans ces cartes numériques et entraîner le dysfonctionnement des systèmes d’aide à la conduite. Les recherches menées dans cette thèse visent à rendre le véhicule capable d’évaluer, de manière autonome et en temps réel, l’intégrité des informations fournies par son système de navigation. Les véhicules de série sont désormais équipés d’un grand nombre de capteurs qui transmettent leurs mesures sur le réseau central interne du véhicule. Ces données sont donc facilement accessibles mais de faible précision. Le défi de cette thèse réside donc dans l’évaluation de l’intégrité des informations cartographiques malgré un faible degré de redondance et l’absence de données fiables. On s’adresse à deux types de défauts cartographiques : les défauts structurels et les défauts géométriques. Les défauts structurels concernent les connections entre les routes (intersections). Un cas particulier de défaut structurel est traité : la détection de ronds-points qui n’apparaissent pas dans la carte numérique. Ce défaut est essentiel car il est fréquent (surtout en Europe) et perturbe le fonctionnement des aides à la conduite. Les ronds-points sont détectés à partir de la forme typique de la trajectoire du véhicule lorsqu’il les traverse, puis sont mémorisés pour avertir les aides à la conduite aux prochains passages du véhicule sur la zone. Les imprécisions de représentation du tracé des routes dans la carte numérique sont quant à elles désignées comme défauts géométriques. Un formalisme mathématique est développé pour détecter ces défauts en comparant l’estimation de la position du véhicule d’après la carte à une autre estimation indépendante de la carte. Cette seconde estimation pouvant elle aussi être affectée par un défaut, les anciens trajetsdu véhicule sur la même zone sont utilisés. Un test statistique est finalement utilisé pour améliorer la méthode de détection de défauts géométriques dans des conditions de mesures bruitées. Toutes les méthodes développées dans le cadre de cette thèse sont évaluées à l’aide de données réelles. / Several Intelligent Vehicles capabilities from Advanced Driving Assistance Systems (ADAS) to Autonomous Driving functions depend on a priori information provided by navigation maps. Whilst these were intended for driver guidance as they store road network information, today they are even used in applications that control vehicle motion. In general, the vehicle position is projected onto the map to relate with links in the stored road network. However, maps might contain faults, leading to navigation and situation understanding errors. Therefore, the integrity of the map-matched estimates must be monitored to avoid failures that can lead to hazardous situations. The main focus of this research is the real-time autonomous evaluation of faults in navigation maps used in intelligent vehicles. Current passenger vehicles are equipped with proprioceptive sensors that allow estimating accurately the vehicle state over short periods of time rather than long trajectories. They include receiver for Global Navigation Satellite System (GNSS) and are also increasingly equipped with exteroceptive sensors like radar or smart camera systems. The challenge resides on evaluating the integrity of the navigation maps using vehicle on board sensors. Two types of map faults are considered: Structural Faults, addressing connectivity (e.g., intersections). Geometric Faults, addressing geographic location and road geometry (i.e. shape). Initially, a particular structural navigation map fault is addressed: the detection of roundabouts absent in the navigation map. This structural fault is problematic for ADAS and Autonomous Driving. The roundabouts are detected by classifying the shape of the vehicle trajectory. This is stored for use in ADAS and Autonomous Driving functions on future vehicle trips on the same area. Next, the geometry of the map is addressed. The main difficulties to do the autonomous integrity monitoring are the lack of reliable information and the low level of redundancy. This thesis introduces a mathematical framework based on the use of repeated vehicle trips to assess the integrity of map information. A sequential test is then developed to make it robust to noisy sensor data. The mathematical framework is demonstrated theoretically including the derivation of definitions and associated properties. Experiments using data acquired in real traffic conditions illustrate the performance of the proposed approaches.

Nas últimas décadas, fatores como maior facilidade de crédito e aumento da renda média dos brasileiros motivaram o crescimento das vendas de veículos novos. Com isso, a quantidade de veículos em circulação também aumentou significativamente. Como consequência desse crescimento observa-se o aumento de congestionamento nas grandes cidades, maior número de acidentes de trânsito com vítimas fatais, dentre outros problemas. Neste contexto surge o Sistema de Transporte Inteligente (ITS), que oferece serviços e produtos que utilizam a comunicação de dados no intuito de melhorar o sistema de transporte. Neste sistema, os veículos, os equipamentos móveis e a infraestrutura nas proximidades das vias de tráfego podem transferir dados entre si, formando assim uma rede de comunicação de dados denominada VANET. Esta comunicação permite a implantação de uma série de serviços e soluções de segurança, informação e entretenimento no trânsito. Para gerenciar o roteamento de dados nestas redes de topologia altamente dinâmica, uma série de algoritmos baseados em ACO (Ant Colony Optmization) vem sendo criados. Estes algoritmos são baseados no comportamento das formigas ao saírem de seus ninhos em busca de alimento, as quais tendem a escolher o caminho mais curto entre ninho e alimento. No entanto, observa-se que na literatura não há um padrão para comparação de desempenho destes algoritmos heurísticos, sendo comumente comparados entre si ou com algoritmos de redes MANETs. Neste trabalho foi criado um benchmark com várias instâncias de roteamento multiobjetivo em redes VANETs que podem ser utilizadas para teste de outros algoritmos. Os simuladores de mobilidade e de rede foram configurados para que os cenários de simulação se aproximassem de redes VANETs reais. A área de simulação escolhida para cada cenário foi uma região localizada próxima ao centro da cidade de Curitiba/PR, na qual várias densidades de veículos foram distribuídas de duas formas distintas, uma aleatória e outra tendenciosa a obter maior fluxo nas grandes avenidas. Foram também aplicados o modelo de propagação three log distance sozinho e combinado com o modelo de desvanecimento de Nakagami. Em cada cenário, os veículos origem e destino foram mantidos fixos em lados opostos da área de simulação. Para cada instante da simulação foi aplicado o algoritmo Dijkstra para obter o menor caminho entre origem e destino para a transmissão de pacotes de dados. Um algoritmo de roteamento multiobjetivo baseado em ACO foi proposto e seus resultados foram comparados com o benchmark. Os caminhos encontrados pelo ACO apresentaram maiores números de saltos e, consequentemente, custos superiores aos encontrados pelo algoritmo de Dijkstra. Um benchmark com vários cenários foi criado. As simulações destes cenários mostraram a influência de diversos fatores na conectividade de uma rede VANET, como a densidade de veículos, suas localizações geográficas e modelo de propagação usado. Os resultados obtidos são promissores e apontam a importância na escolha dos modelos de simulação. Tais resultados incentivam o uso de algoritmos heurísticos para roteamento de dados em redes VANET. / In the last decades, we have witnessed an increasing sale of new cars, driven by extensive credit availability and the growth of average income. Hence, the number of vehicles on the roads has increased. Due to this high density of vehicles, the traffic jams as well as fatal accidents are increasing. In order to reduce such factors, the Intelligent Transportation Systems (ITS) aroused, offering connected services and products related to entertainment and road safety. In this system, vehicles, mobile equipments and the infrastructure in the neighborhood of the traffic ways can transfer data to each other, thus creating a network called VANET (Vehicular Ad-hoc Network). To optimize the packets routing in these dynamic networks, several Ant Colony Optmization (ACO) - based algorithms have been proposed. Such algorithms are inspired by the foraging behavior of ants, which are capable of finding the shortest paths from food sources to the nest. However, there are no performance evaluation standards in the recent literature. The algorithms are often compared to each other or with MANET’s algorithms. In this dissertation, a bench-mark of several routing instances for VANETs was created. These benchmarks can be used for testing routing algorithms. The mobility and network simulators were configured in order to create real-world VANET-like scenarios. The geographical area chosen for the scenarios was near to Curitiba downtown. Different vehicle densities were distributed in two way: purely random and biased in such a way that avenues receive higher vehicle flows. The three log-distance path loss model was applied to each scenario, sometimes combined with the Nakagami fading model. In each scenario the source and destination vehicles are fixed on opposite sides of the simulated area. For each simulation time step, the Dijkstra algorithm was run to find the shortest path data transmission between source and destination. A multiobjective ACO-based algorithm was proposed and compared with the Dijkstra algorithm. The paths found by ACO include higher number of hops than those found by the Dijkstra algorithm. A benchmark with several scenarios was created. The scenario’s simulations show the importance of several factors in the VANET connectivity, such as vehicle density, geographical location and propagation models. The results are promising and indicate the importance of choosing appropriated simulation models.

Nas últimas décadas, fatores como maior facilidade de crédito e aumento da renda média dos brasileiros motivaram o crescimento das vendas de veículos novos. Com isso, a quantidade de veículos em circulação também aumentou significativamente. Como consequência desse crescimento observa-se o aumento de congestionamento nas grandes cidades, maior número de acidentes de trânsito com vítimas fatais, dentre outros problemas. Neste contexto surge o Sistema de Transporte Inteligente (ITS), que oferece serviços e produtos que utilizam a comunicação de dados no intuito de melhorar o sistema de transporte. Neste sistema, os veículos, os equipamentos móveis e a infraestrutura nas proximidades das vias de tráfego podem transferir dados entre si, formando assim uma rede de comunicação de dados denominada VANET. Esta comunicação permite a implantação de uma série de serviços e soluções de segurança, informação e entretenimento no trânsito. Para gerenciar o roteamento de dados nestas redes de topologia altamente dinâmica, uma série de algoritmos baseados em ACO (Ant Colony Optmization) vem sendo criados. Estes algoritmos são baseados no comportamento das formigas ao saírem de seus ninhos em busca de alimento, as quais tendem a escolher o caminho mais curto entre ninho e alimento. No entanto, observa-se que na literatura não há um padrão para comparação de desempenho destes algoritmos heurísticos, sendo comumente comparados entre si ou com algoritmos de redes MANETs. Neste trabalho foi criado um benchmark com várias instâncias de roteamento multiobjetivo em redes VANETs que podem ser utilizadas para teste de outros algoritmos. Os simuladores de mobilidade e de rede foram configurados para que os cenários de simulação se aproximassem de redes VANETs reais. A área de simulação escolhida para cada cenário foi uma região localizada próxima ao centro da cidade de Curitiba/PR, na qual várias densidades de veículos foram distribuídas de duas formas distintas, uma aleatória e outra tendenciosa a obter maior fluxo nas grandes avenidas. Foram também aplicados o modelo de propagação three log distance sozinho e combinado com o modelo de desvanecimento de Nakagami. Em cada cenário, os veículos origem e destino foram mantidos fixos em lados opostos da área de simulação. Para cada instante da simulação foi aplicado o algoritmo Dijkstra para obter o menor caminho entre origem e destino para a transmissão de pacotes de dados. Um algoritmo de roteamento multiobjetivo baseado em ACO foi proposto e seus resultados foram comparados com o benchmark. Os caminhos encontrados pelo ACO apresentaram maiores números de saltos e, consequentemente, custos superiores aos encontrados pelo algoritmo de Dijkstra. Um benchmark com vários cenários foi criado. As simulações destes cenários mostraram a influência de diversos fatores na conectividade de uma rede VANET, como a densidade de veículos, suas localizações geográficas e modelo de propagação usado. Os resultados obtidos são promissores e apontam a importância na escolha dos modelos de simulação. Tais resultados incentivam o uso de algoritmos heurísticos para roteamento de dados em redes VANET. / In the last decades, we have witnessed an increasing sale of new cars, driven by extensive credit availability and the growth of average income. Hence, the number of vehicles on the roads has increased. Due to this high density of vehicles, the traffic jams as well as fatal accidents are increasing. In order to reduce such factors, the Intelligent Transportation Systems (ITS) aroused, offering connected services and products related to entertainment and road safety. In this system, vehicles, mobile equipments and the infrastructure in the neighborhood of the traffic ways can transfer data to each other, thus creating a network called VANET (Vehicular Ad-hoc Network). To optimize the packets routing in these dynamic networks, several Ant Colony Optmization (ACO) - based algorithms have been proposed. Such algorithms are inspired by the foraging behavior of ants, which are capable of finding the shortest paths from food sources to the nest. However, there are no performance evaluation standards in the recent literature. The algorithms are often compared to each other or with MANET’s algorithms. In this dissertation, a bench-mark of several routing instances for VANETs was created. These benchmarks can be used for testing routing algorithms. The mobility and network simulators were configured in order to create real-world VANET-like scenarios. The geographical area chosen for the scenarios was near to Curitiba downtown. Different vehicle densities were distributed in two way: purely random and biased in such a way that avenues receive higher vehicle flows. The three log-distance path loss model was applied to each scenario, sometimes combined with the Nakagami fading model. In each scenario the source and destination vehicles are fixed on opposite sides of the simulated area. For each simulation time step, the Dijkstra algorithm was run to find the shortest path data transmission between source and destination. A multiobjective ACO-based algorithm was proposed and compared with the Dijkstra algorithm. The paths found by ACO include higher number of hops than those found by the Dijkstra algorithm. A benchmark with several scenarios was created. The scenario’s simulations show the importance of several factors in the VANET connectivity, such as vehicle density, geographical location and propagation models. The results are promising and indicate the importance of choosing appropriated simulation models.

Assunto bastante abordado quando se trata de Sistemas Inteligentes de Transportes (ITS), a identificação veicular - utilizada em grande parte das aplicações de ITS deve ser entendida como um conjunto de recursos de hardware, software e telecomunicações, que interagem para atingir, do ponto de vista funcional, o objetivo de, conseguir extrair e transmitir, digitalmente, a identidade de um veículo. É feita tanto por sistemas que transmitem e recebem uma identidade digital quanto por sistemas que, instalados na infraestrutura da via, são capazes de reconhecer a placa dos veículos circulantes. Quando se trata da identificação automática por meio do reconhecimento da placa veicular, os estudos têm se concentrado sobremaneira nas tecnologias de processamento de imagens, não abordando - em sua maioria - uma visão sistêmica, necessária para compreender de maneira mais abrangente todas as variáveis que podem interferir na eficácia da identificação. Com o objetivo de contribuir para melhor entender e utilizar os sistemas de reconhecimento automático de placas veiculares, este trabalho propõe um modelo sistêmico, em camadas, para representar seus componentes. Associada a esse modelo, propõe uma classificação para os diversos tipos de falhas que podem prejudicar seu desempenho. Uma análise desenvolvida com resultados obtidos em testes realizados em campo com sistemas de identificação de placas voltados à fiscalização de veículos aponta resultados relevantes e limitações para obter correlações entre variáveis, em função dos diversos fatores que podem influenciar os resultados. Algumas entrevistas realizadas apontam os tipos de falhas que ocorrem com mais frequência durante a operação desses sistemas. Finalmente, este trabalho propõe futuros estudos e apresenta um glossário de termos, que poderá ser útil a novos pesquisadores. / The automatic vehicle identification is an important feature of Intelligent Transportation Systems (ITS) and is used in most ITS applications. The identification process is comprised of a group of interacting resources that involves hardware, software and telecommunication to, digitally, extract and transmit the identity of vehicles. At least two technologies may be used in the vehicle identification process: on-board devices transmitting a digital identity or systems installed on the road infrastructure, which identify and read the vehicle license plate. As far as vehicle license plate recognition is concerned, studies have been greatly focused on image processing technologies and have not addressed the problem in a systemic approach, which is very important for understanding all variables that can interfere with the effectiveness of identification. Having this approach in mind and intending to contribute for a better performance, this paper proposes a layer model representation of those systems as well as a failure type classification associated with it. An analysis, based on a significant set of results obtained from field tests of systems with plate recognition capabilities for law enforcement, shows important results as well as limitations to obtain mathematical correlation of variables. Interviews conducted with supply actors of such systems in Brazil point out the most significant sources of failures that occur during operation. Finally, the text presents potential topics for research and organizes a glossary of terms that may be useful to future researchers.

O rápido crescimento das regiões urbanas tem impacto significativo nos sistemas de tráfego e transportes. Políticas de gerenciamento e estratégias de planejamento alternativas são claramente necessárias para o tratamento da capacidade limitada, e cada vez mais deficitária, das redes viárias. O conceito de Sistemas Inteligentes de Transportes (ITS) surge neste cenário; mais do que procurar aumentar a capacidade por meio de modificações físicas na infraestrutura, sua premissa baseia-se na utilização de tecnologias avançadas de comunicação e computação para melhor gerir os recursos de tráfego e transportes atuais. Influenciar o padrão do comportamento dos usuários é um desafio que tem estimulado muita pesquisa na área de ITS, onde fatores humanos passam a ter grande importância na modelagem, simulação e avaliação dessa abordagem inovadora. Este trabalho tem como foco a utilização de Sistemas Multiagentes (MAS) na representação dos sistemas de tráfego e transporte, com base nas novasmedidas de desempenho impostas pelas tecnologias ITS. As características de agentes têm grande potencial para representar componentes geográfica e funcionalmente distribuídos, como a maioria dos elementos no domínio da aplicação. Uma arquitetura BDI (beliefs, desires, intentions) é apresentada como alternativa a modelos tradicionais, usados para representar o comportamento do motorista em simulação microscópica, considerando-se a representação explícita dos estados mentais dos usuários. Os conceitos básicos de ITS e MAS são apresentados, assim como exemplos de aplicações relacionados com o tema do trabalho. Esta foi a motivação para a extensão de um simulador microscópico existente, no sentido de incorporar as características dos MAS para melhorar a representação dos motoristas. Assim, a demanda é gerada a partir de uma população de agentes, resultando da decisão sobre a rota e o tempo de partida ao longo de vários dias. O modelo estendido, que passa a suportar a interação de motoristas BDI, foi efetivamente implementado e foram executados diferentes experimentos para testar a abordagem em cenários de tráfego urbano. MAS permite uma abordagem direcionada a processos que facilita a construção de representações modulares, robustas, e extensíveis, características pouco presentes em abordagens voltadas ao resultado. Suas premissas de abstração permitem uma associação direta entre modelo e implementação. Incerteza e variabilidade são assim tratadas de maneira mais intuitiva, uma vez que arquiteturas cognitivas permitem uma fácil representação do comportamento humano na estrutura do motorista. Desta forma, MAS estende a simulação microscópica de tráfego no sentido de melhor representar a complexidade inerente às tecnologias ITS. / The rapid growth of urban areas has a significant impact on traffic and transportation systems. New management policies and planning strategies are clearly necessary to cope with the more than ever limited capacity of existing road networks. The concept of Intelligent Transportation System (ITS) arises in this scenario; rather than attempting to increase road capacity by means of physical modifications to the infrastructure, the premise of ITS relies on the use of advanced communication and computer technologies to handle today’s traffic and transportation facilities. Influencing users’ behaviour patterns is a challenge that has stimulated much research in the ITS field, where human factors start gaining great importance to modelling, simulating, and assessing such an innovative approach. This work is aimed at using Multi-agent Systems (MAS) to represent the traffic and transportation systems in the light of the new performance measures brought about by ITS technologies. Agent features have good potentialities to represent those components of a system that are geographically and functionally distributed, such as most components in traffic and transportation. A BDI (beliefs, desires, and intentions) architecture is presented as an alternative to traditional models used to represent the driver behaviour within microscopic simulation allowing for an explicit representation of users’ mental states. Basic concepts of ITS and MAS are presented, as well as some application examples related to the subject. This has motivated the extension of an existing microscopic simulation framework to incorporate MAS features to enhance the representation of drivers. This way demand is generated from a population of agents as the result of their decisions on route and departure time, on a daily basis. The extended simulation model that now supports the interaction of BDI driver agents was effectively implemented, and different experiments were performed to test this approach in commuter scenarios. MAS provides a process-driven approach that fosters the easy construction of modular, robust, and scalable models, characteristics that lack in former result-driven approaches. Its abstraction premises allow for a closer association between the model and its practical implementation. Uncertainty and variability are addressed in a straightforward manner, as an easier representation of humanlike behaviours within the driver structure is provided by cognitive architectures, such as the BDI approach used in this work. This way MAS extends microscopic simulation of traffic to better address the complexity inherent in ITS technologies.

Les véhicules modernes sont équipés de périphériques permettant d'automatiser des tâches (changement de vitesse de transmission, régulation de vitesse, etc.) ou de fournir des services à l'utilisateur (aide à la conduite, détection d'obstacles, etc.). Les communications entre les véhicules permettent d'élargir ces services grâce à la collaboration de plusieurs véhicules (prévention des accidents, gestion du trafic routier, etc.). La multiplication de ces périphériques, de leurs interfaces et protocoles rend l'échange de données plus complexe. Par ailleurs, la communication inter- véhicules est plus contraignante à cause de la haute mobilité des véhicules. Dans cette thèse, nous proposons la conception d'un canal de communication Connect to All (C2A) qui permet d'assurer l'interopérabilité entre les périphériques embarqués dans un véhicule. En effet, il détecte la connexion à chaud d'un équipement, le reconnaît et lui permet d'échanger des données avec les autres périphériques connectés. La conception du canal commence par la modélisation de ce canal en utilisant deux techniques différentes (l'outil de modélisation et de vérification UPPAAL et le Langage de Description et de Spécification (LDS)). La vérification des modèles proposés a pour but de valider le fonctionnement. Ensuite, nous détaillons une implémentation réelle du canal sur une carte embarquée qui vise à démontrer la faisabilité du concept d'interopérabilité de C2A.Nous avons aussi étudié les effets de la mobilité dans la communication inter-véhiculaires grâce à une approche hybride mixant le routage et un service de localisation. Cette approche offre un mécanisme qui permet de réduire les coûts de la localisation des véhicules tout en augmentant les performances de routage. En plus, nous comparons deux applications de cette approche : Hybrid Routing and Grid Location Service (HRGLS) et Hybrid Routing and Hierarchical Location Service (HRHLS) avec des approches originelles pour démontrer la valeur ajoutée. Cette approche est enrichie avec un algorithme de prédiction de mobilité. Ce dernier permet de mieux cerner le déplacement des véhicules en les estimant. De même, l'approche hybride avec prédiction de mobilité Predictive Hybrid Routing and Hierarchical Location Service (PHRHLS) est comparée à HRHLS et l'approche originelle afin de révéler les bénéfices de la prédiction de mobilité. / Modern vehicles are equipped with various devices that aim to automate tasks (shift transmission, cruise control, etc.) or to provide services to the user (driver assistance, obstacle detection, etc.). Communications between vehicles help to expand these services through the collaboration of several vehicles (accident prevention, traffic management, etc.). The proliferation of these devices, their interfaces and protocols makes the data exchange more complex. In addition, inter-vehicle communication is more restrictive because of the vehicles' high mobility.In this work, we propose the design of a communication channel Connect to All (C2A) that ensures the interoperability between embedded devices in a vehicle. In fact, it detects the equipment connection, recognizes it and allows it to exchange data with other devices. The channel design starts by the modelling step using two different techniques (the model checker tool UPPAAL and the Specification and Description Language (SDL). Then, we validate the designed models. We also detail a concrete implementation of the channel on an embedded chip that aims to show the C2A interoperability concept feasibility.We also studied the mobility effects in the inter-vehicular communication through a hybrid approach mixing routing and location-based service. This approach provides a mechanism to reduce vehicle-tracking costs while increasing routing performances. Moreover, we compare two applications of this approach: Hybrid Routing and Grid Location Service (HRGLS) and Hybrid Routing and Hierarchical Location Service (HRHLS) with classical approaches to prove the added value. Then, this approach is improved with a mobility prediction algorithm. The latter allows a better understanding of the vehicle movements by estimating them. Similarly, the hybrid approach with mobility prediction Predictive Hybrid Routing and Hierarchical Location Service (PHRHLS) is compared with the basic approach and HRHLS in order to show the mobility prediction advantages.

Les avancées technologiques, dues à l’avènement des nouvelles technologies d’information etde communication, ont donné naissance au concept des Systèmes de Transport Intelligents (STI).Les objectifs de telles applications consistent à apporter des solutions efficaces pour faire face auxproblèmes quotidiens des phénomènes de congestion. L’importance ainsi que les enjeux socioéconomiquesposés par les congestions imposent d’introduire des solutions innovantes utilisantles avancées récentes dans le domaine de la commande. Les travaux présentés dans cette thèse sesituent dans le cadre des STI et traitent des problèmes de la commande du trafic autoroutier et surles Voies Rapides Urbaines (VRU). Parmi les techniques de commande utilisée, nos travaux se focalisentprincipalement sur le contrôle d’accès isolé. L’objectif d’une telle action de régulation estd’agir sur le débit des rampes d’entrée, via des feux de signalisation, afin de maintenir la densitésur la voie principale aux alentours d’un seuil critique permettant ainsi, une utilisation optimale del’infrastructure autoroutière ou des VRU. L’algorithme proposé repose sur l’utilisation conjointede la platitude différentielle et le concept de la commande par mode glissant d’ordre supérieur. Laprincipale caractéristique de la platitude réside dans sa capacité à assurer une génération de trajectoiressans intégration d’aucune équation différentielle du modèle étudié. L’intérêt de la commandepar mode glissant d’ordre supérieur est de permettre le suivi de trajectoires d’une manière robustemême en présence d’incertitudes et de perturbations typiques aux systèmes de trafic. La pertinencede l’approche proposée est validée via un ensemble de simulations avec des données réelles d’uneportion de l’autoroute A6 du périphérique de Paris. De plus, la validation a été enrichie par l’évaluationde performances basée sur des critères couramment utilisés par les exploitants. L’ensembledes résultats ouvre la voie à plusieurs perspectives d’amélioration et de généralisation de cettecommande à des réseaux routiers plus complexes. / The technological advances, due to the advent of the new information and communication technologieshave given rise to the Intelligent Transportation Systems (ITS) concept. The objectivesof such applications are to provide effective solutions to deal with the daily problems of congestion.The importance as well as the socio-economic challenges raised by congestion requires theintroduction of innovative solutions based on the latest advances in the automatic control field. Theworks presented in this thesis lie in the frame of ITS and treat the problems of the freeway andUrban Express Routes (UER) control. Among the used control techniques, our works focus mainlyon the isolated ramp metering. The objective of this control measurement is to act on the on-rampflow, through traffic lights, in order to keep the traffic density on the mainstream section around acritical threshold allowing then an optimal use of the freeway or UER infrastructures. The proposedalgorithm rests on the jointly use of differential flatness and high order sliding mode control(HOSMC) concept. The main characteristic of the differential flatness lies in its ability to providea trajectory generation, without integration of any differential equation of the studied model. Onthe other hand, the advantage of HOSMC is to allow a robust trajectory tracking even in the case ofthe presence of uncertainties and disturbances which are typical to traffic systems. The relevanceof the proposed approach is validated through a set of numerical simulations using real-data froma part of the A6 freeway from Paris ring. In addition, the validation step has been enriched by theperformance evaluation based on a set of criteria commonly used by the freeway practitioners. Theobtained results paves the way to several perspectives in order to improve the proposed controlapproach and its generalization for more complex freeway networks.

La thèse porte sur le contrôle des systèmes de systèmes (SdS) et, sur la manière de construire des SdS adaptables et fiables. Ce travail fait partie du laboratoire d’excellence Labex MS2T sur le développement des SdS technologiques. Les SdS sont des systèmes complexes constitués de plusieurs systèmes indépendants qui fonctionnent ensemble pour atteindre un objectif commun. L’ingénierie des SdS est une approche qui se concentre sur la manière de construire et de concevoir des SdS fiables capables de s’adapter à l’environnement dynamique dans lequel ils évoluent. Compte tenu de l’importance du contrôle des systèmes constituants (SC) pour atteindre les objectifs du SdS , la première partie de cette thèse a consisté en une étude bibliographique sur le sujet du contrôle des SdS. Certaines méthodes de contrôle existent pour les systèmes à grande échelle et les systèmes multi-agents, à savoir, le contrôle hiérarchique, distribué et décentralisé peuvent être utiles et sont utilisés pour contrôler les SdS. Ces méthodes ne conviennent pas pour contrôler un SdS dans sa globalité et son évolution, en raison de l’indépendance de leur SC ; alors que les “frameworks” multi-vues conviennent mieux à cet objectif. Une approche de ”framework” générale est proposée pour modéliser et gérer les interactions entre les SC dans un SdS. La deuxième partie de notre travail a consisté à contribuer aux systèmes de transport intelligent. À cette fin, nous avons proposé le gestionnaire de manœuvres coopératives pour les véhicules autonomes (CMMAV), un “framework” qui guide le développement des applications coopératives dans les véhicules autonomes. Pour valider le CMMAV, nous avons développé le gestionnaire de manœuvres latérales coopératives (CLMM), une application sur les véhicules autonomes qui permet d’échanger des demandes afin de coopérer lors de manœuvres de dépassement sur autoroute. Cette application a été validée par des scénarios formels, des simulations informatiques, et testée sur les véhicules autonomes du projet Robotex au laboratoire Heudiasyc. / This thesis focuses on System of Systems (SoS) control, and how to build adaptable and reliable SoS. This work is part of the Labex MS2T laboratory of excellence on technological SoS development. SoS are complex systems that consist of multiple independent systems that work together to achieve a common goal. SoS Engineering is an approach that focuses on how to build and design reliable SoS that can adapt to the dynamic environment in which they operate. Given the importance of controlling constituent systems (CS) in order to achieve SoS objectives, the first part of this thesis involved a literature study about the subject of SoS control. Some control methods exist for large-scale systems and multi-agent systems, namely, hierarchical, distributed, and decentralized control might be useful and are used to control SoS. These methods are not suitable for controlling SoS in its whole, because of the independence of their CS; whereas, multi-views frameworks are more suitable for this objective. A general framework approach is proposed to model and manage the interactions between CS in a SoS. The second part of our work consisted of contributing to Intelligent Transportation Systems. For this purpose, we have proposed the Cooperative Maneuvers Manager for Autonomous Vehicles (CMMAV), a framework that guides the development of cooperative applications in autonomous vehicles. To validate the CMMAV, we have developed the Cooperative Lateral Maneuvers Manager (CLMM), an application on the autonomous vehicles that enables equipped vehicles to exchange requests in order to cooperate during overtaking maneuvers on highways. It was validated by formal scenarios, computer simulations, and tested on the autonomous vehicles of the Equipex Robotex in Heudiasyc laboratory.

abstract: Unmanned aerial systems (UASs) have recently enabled novel applications such as passenger transport and package delivery, but are increasingly vulnerable to cyberattack and therefore difficult to certify. Legacy systems such as GPS provide these capabilities extremely well, but are sensitive to spoofing and hijacking. An alternative intelligent transport system (ITS) was developed that provides highly secure communications, positioning, and timing synchronization services to networks of cooperative RF users, termed Communications and High-Precision Positioning (CHP2) system. This technology was implemented on consumer-off-the-shelf (COTS) hardware and it offers rapid (<100 ms) and precise (<5 cm) positioning capabilities in over-the-air experiments using flexible ground stations and UAS platforms using limited bandwidth (10 MHz). In this study, CHP2 is considered in the context of safety-critical and resource limited transport applications and urban air mobility. The two-way ranging (TWR) protocol over a joint positioning-communications waveform enables distributed coherence and time-of-flight(ToF) estimation. In a multi-antenna setup, the cross-platform ranging on participating nodes in the network translate to precise target location and orientation. In the current form, CHP2 necessitates a cooperative timing exchange at regular intervals. Dynamic resource management supports higher user densities by constantly renegotiating spectral access depending on need and opportunity. With these novel contributions to the field of integrated positioning and communications, CHP2 is a suitable candidate to provide both communications, navigation, and surveillance (CNS) and alternative positioning, navigation, and timing (APNT) services for high density safety-critical transport applications on a variety of vehicular platforms. / Dissertation/Thesis / Doctoral Dissertation Electrical Engineering 2020

With the increasing highlighted security concerns in Intelligent Transportation System (ITS), Vision-based Automated Vehicle Recognition (VAVR) has attracted considerable attention recently. A comprehensive VAVR system contains three components: Vehicle Detection (VD), Vehicle Make and Model Recognition (VMMR), and Vehicle Re-identification (VReID). These components perform coarse-to-fine recognition tasks in three steps. The VAVR system can be widely used in suspicious vehicle recognition, urban traffic monitoring, and automated driving system. Vehicle recognition is complicated due to the subtle visual differences between different vehicle models. Therefore, how to build a VAVR system that can fast and accurately recognize vehicle information has gained tremendous attention. In this work, by taking advantage of the emerging deep learning methods, which have powerful feature extraction and pattern learning abilities, we propose several models used for vehicle recognition. First, we propose a novel Recurrent Attention Unit (RAU) to expand the standard Convolutional Neural Network (CNN) architecture for VMMR. RAU learns to recognize the discriminative part of a vehicle on multiple scales and builds up a connection with the prominent information in a recurrent way. The proposed ResNet101-RAU achieves excellent recognition accuracy of 93.81% on the Stanford Cars dataset and 97.84% on the CompCars dataset. Second, to construct efficient vehicle recognition models, we simplify the structure of RAU and propose a Lightweight Recurrent Attention Unit (LRAU). The proposed LRAU extracts the discriminative part features by generating attention masks to locate the keypoints of a vehicle (e.g., logo, headlight). The attention mask is generated based on the feature maps received by the LRAU and the preceding attention state generated by the preceding LRAU. Then, by adding LRAUs to the standard CNN architectures, we construct three efficient VMMR models. Our models achieve the state-of-the-art results with 93.94% accuracy on the Stanford Cars dataset, 98.31% accuracy on the CompCars dataset, and 99.41% on the NTOU-MMR dataset. In addition, we construct a one-stage Vehicle Detection and Fine-grained Recognition (VDFG) model by combining our LRAU with the general object detection model. Results show the proposed VDFG model can achieve excellent performance with real-time processing speed. Third, to address the VReID task, we design the Compact Attention Unit (CAU). CAU has a compact structure, and it relies on a single attention map to extract the discriminative local features of a vehicle. We add two CAUs to the truncated ResNet to construct a small but efficient VReID model, ResNetT-CAU. Compared with the original ResNet, the model size of ResNetT-CAU is reduced by 60%. Extensive experiments on the VeRi and VehicleID dataset indicate the proposed ResNetT-CAU achieve the best re-identification results on both datasets. In summary, the experimental results on the challenging benchmark VMMR and VReID datasets indicate our models achieve the best VMMR and VReID performance, and our models have a small model size and fast image processing speed.