Made available in DSpace on 2014-08-20T14:37:00Z (GMT). No. of bitstreams: 1 Tese_Fioravante_Jaekel_Santos.pdf: 15601631 bytes, checksum: d7a242ff4e1d6221c5e6d1c8129d40a5 (MD5) Previous issue date: 2011-03-31 / Agricultural practices, among others, contribute to reduce the environment quality, through the unbalance between the CO2 fixation and emission to the atmosphere. The importance of soil carbon stocking surpasses the matter of carbon sequestration in itself, since the kinds of soil management that rise their stocks in soils also have a positive effect on their quality as a whole. The soil carbon stocks evolution in a specific site as a function of its past and future uses can be studied through mathematical models as Century. The linking of this model with Remote Sensing and Geographical Information Systems makes it possible to analyze and model the carbon stocks spatial distribution and variability in a certain geographic area. In this study a catchment basin (Piraizinho stream, Rio Grande do Sul State) located in the Brazilian Pampas, traditionally exploited with cattle grazing on native grassland, was selected to analyze and model the soil carbon stocks changes due to other land uses that have arisen in the region in the last twenty years. In order to achieve this, a semi-detailed soil survey was performed, followed by a spatial and temporal analysis of the land uses, based on aerial photographs and satellite images as well as field survey and interviews with landowners and technicians. To establish the parameters and validate the model, a soil sampling was performed based on the soil units and different types of land uses. The data spatial modeling was done on a GIS base using the soil units map and relief in a such a way as to point model the carbon stock through the erosion variable, where to each pixel it was assigned a soil loss using the USLE. Regosols (RReh and RLd units), Acrisols (PVd unit), Luvisols (TXo1, TXo2, TXp, TCo units), Chernozems (MEo unit), Vertisols (VEo unit), and Gleysols (GMve unit) were mapped in an average gently rolling landscape. In 1967 98,75% of the area was used solely by extensive cattle grazing on native grassland, by 2009 this was reduced to 82,67%, due to rice and soybeans crops as well as eucalyptus and acacia foresting. The topsoil carbon stocks varied as a function of the soil units and kinds of use. The largest carbon stocks under traditional grazing simulated to the year 1990 were found on soil units MEo (89,04 Mg ha-1 C) and TCo (79,87 Mg ha-1 C), while soil units RLd (27,48 Mg ha-1 C) and PVd (41,73 Mg ha-1 C) were responsible for the lowest stocks. The simulations also indicate that the no till system with soybeans in summer and rye in winter is not able to maintain the carbon stocks in any of the soil units modeled. The continuous use of this system for 50 years can lead to the consumption of 35% of the carbon stocked in the topsoils (without considering VEo and GMve soil units that were not simulated). The Voisin rational grazing system, that has been adopted by 45 years in a property in the area, showed a higher potential than the traditional system without fire to accumulate carbon in the soils. / As práticas agrícolas, entre outros, contribuem para a redução da qualidade do ambiente, através do desequilíbrio entre fixação e emissão de CO2 para a atmosfera. A importância da estocagem de C no solo extrapola a questão do sequestro em si, uma vez que manejos que intensificam o aumento dos seus estoques no solo também possibilitam a melhoria da sua qualidade como um todo. A evolução pontual dos estoques de carbono do solo em função dos usos passados e futuros pode ser estudada através de modelos matemáticos como o Programa Century. A associação desse com o Sensoriamento Remoto (SR) e Sistemas de Informações Geográficas (SIG) torna possível a análise e a modelagem da sua distribuição e variabilidade espacial em áreas geográficas definidas. Nesse estudo foi selecionada uma bacia hidrográfica (do Arroio Piraizinho, Bagé, RS), situada no Pampa Gaúcho, tradicionalmente utilizada com pecuária extensiva em campo nativo, para analisar e modelar a evolução dos estoques de carbono do solo em função de alterações no uso do solo através de uma nova matriz produtiva que está começando a se implantar na região. Para tal, inicialmente foi realizado um levantamento semidetalhado dos solos da bacia hidrográfica; após, uma análise espacial e temporal dos usos do solo através de fotos aéreas e imagens de satélite, apoiada por levantamentos de campo e entrevistas com proprietários de terras e técnicos da região. Para modelagem dos cenários de carbono, com base nas unidades de solos e nos usos existentes, foi realizada uma coleta de amostras de solo para parametrizar e validar o modelo Century 4.0. A espacialização dos dados foi feita montando uma base SIG com o mapa de solos e do relevo de forma a modelar ponto a ponto o estoque de carbono através da variável erosão, onde em cada pixel foi atribuída uma perda de solo calculada através da USLE. Foram encontrados na área unidades de mapeamento de Neossolos (RReh, RLd), Argissolos (PVd), Luvissolos (TXo1, TXo2, TXp, TCo), Chernossolos (MEo), Vertissolos (VEo) e Gleissolos (GMve), com relevo em média suave ondulado. Até o ano de 1967 98,75% da área da BHP era utilizada somente com pecuária extensiva em campo nativo, que foi reduzida a 82,67% em 2009, com a entrada das culturas de arroz, soja, eucalipto e acácia. Os estoques de C na camada arável dos solos da BHP são afetados de modo variável em função do uso do solo e da unidade de mapeamento. Os maiores estoques de C sob campo nativo com pastoreio tradicional simulados em 1990 ocorreram nas unidades MEo (89,04 Mg ha-1 C) e TCo (79,87 Mg ha-1 C), enquanto que os solos das unidades RLd (27,48 Mg ha-1 C) e PVd (41,73 Mg ha-1 C) produziram as menores capacidades de armazenamento de C. A simulação indica também que o sistema com plantio direto de soja no verão e aveia preta no inverno não mantém os níveis de C dos solos estudados, mostrando um consumo dos estoques em todas as unidades de mapeamento. A manutenção deste tipo de uso por 50 anos pode significar o consumo de 35% dos estoques de C dos solos nas unidades de mapeamento simuladas (todas à exceção da VEo e GMve). O sistema de pastoreio racional Voisin apresentou uma capacidade superior à do pastoreio tradicional sem queima em promover o acúmulo de C nos solos estudados.

Le cancer du sein est le cancer le plus fréquent chez la femme et son incidence a doublé ces 30 dernières années. Les facteurs environnementaux à effet perturbateur endocrinien, tels que les dioxines émises par les activités de combustion industrielle, sont suspectés d'augmenter le risque de cancer du sein. L'alimentation et l'inhalation sont les deux voies majeures d'exposition aux dioxines chez l'Homme. Les données épidémiologiques sur le sujet sont non concluantes, et il existe des limites méthodologiques. Ce travail doctoral avait pour objectif d'étudier l'impact de l'exposition aux dioxines sur le risque de cancer du sein dans la cohorte E3N, en répondant aux limites des études existantes.Nous avons évalué l'exposition alimentaire aux dioxines puis estimé le risque de cancer du sein associé parmi les femmes de la cohorte E3N. Nous avons ensuite développé un score d'exposition basé sur un système d'information géographique, associant la distance à la source, la durée d'exposition et la fréquence de vent dominant, afin d'évaluer l'exposition environnementale aux dioxines à chaque adresse des femmes entre 1990 et 2008. Le risque de cancer du sein associé au score d'exposition cumulé a été estimé dans une étude cas-témoins nichée dans la cohorte E3N, parmi les femmes ayant résidé en Rhône-Alpes. Aucune association n'a été observée, à l'exception d'une diminution du risque de cancer du sein hormono-indépendant, retrouvée de façon significative dans l'étude alimentaire. Cette dernière observation est cohérente avec des données expérimentales. Dû à un manque de puissance statistique pour l'exposition aérienne, nos résultats demandent confirmation au niveau national / Breast cancer (BC) is the first malignancy among women. Its incidence has doubled over the past 30 years. Environmental factors with endocrine disruptive properties, such as dioxins emitted from industrial combustion processes, are suspected to affect BC risk. Ingestion of contaminated food and inhalation are the major exposure routes in humans. Epidemiological evidence on the association between dioxin exposure and BC risk remains inconclusive due to methodological limitations. The aim of the thesis was to investigate the association between dioxin exposure and BC risk in the E3N prospective cohort, filling current methodological gaps.First, we assessed the association between estimated dietary dioxin exposure and BC risk among women from the E3N cohort. Second, we developed a geographic information system (GIS)-based metric to assess airborne dioxin exposure at the individual address level, including proximity to and technical characteristics of industrial sources, exposure duration and prevailing wind frequency. The metric was then applied to each E3N women’s addresses from 1990 to 2008, and airborne dioxin exposure was estimated for cases and matched controls from a cohort sub-population (the Rhône-Alpes region). Third, we estimated BC risk associated with cumulative airborne dioxin exposure. Overall, no statistically significant association was observed, except for a decrease in hormone-independent BC risk. The latter was significant for dietary dioxin exposure. For airborne exposure, we might have lacked statistical power and confirmation at the national level is required. The inverse association with ER-negative BC risk is consistent with experimental evidence

Cette thèse est réalisée dans le cadre d'un contrat CIFRE avec l'agence d'urbanisme del'agglomération stéphanoise, Epures.L’ambition de cette recherche est de proposer une méthode « objective » et« systématique » de lecture de l’héritage matériel d'un territoire, ce qui imposed’abandonner tous critères de sélection discriminant : l’ensemble des traces matériellesqui forme cet héritage peut être déduit de l’analyse d’unités minimales et objectives(parcellaire, bâti, réseau viaire) , dont l’apparition, la disparition et la transformation dans letemps peuvent être retracées par l’exploitation de sources spécifiques en particulier descartes anciennes disponibles. D’autres unités, linéaires, ponctuelles ou surfaciquespeuvent être mises en évidence par une confrontation directe avec le terrain.L'analyse a été réalisée sur la partie ligérienne de l'ancienne vallée industrielle du Gier.La méthode met en oeuvre un Système d'Information Géographique (SIG) historique quiautorise l’intégration, la réinterprétation et le croisement d’informations provenant desources variées : sérielles, textuelle, cartographiques ou iconographiques. Cet outil facilitele partage et la mise à jour des données produites.Les résultats obtenus offrent une lecture de l'évolution de la vallée du Gier à l'échelle desunités parcellaires et bâties depuis deux siècles. Ils proposent une nouvelle approche duterritoire et de son évolution qui permet de questionner les politiques patrimoniales oud'aménagement existantes. Etendue à l'ensemble de l'espace stéphanois ou à d'autresterritoires et régulièrement mis à jour, la méthode peut permettre de constituer unobservatoire de leur évolution. / This thesis is carried out within the context of a contract CIFRE with the agency of townplanning of the agglomeration of Saint-Etienne, Epures.The ambition of this research is to propose a method “objective” and “systematic” to readthe material heritage of a territory, which forces to give up all criteria of selectiondiscriminating: the whole of the material traces which forms this heritage can be deducedfrom the analysis of minimal units and objectify (plots, built elements, road network), ofwhich appearance, disappearance and the transformation in time can be recalled by theexploitation of specific sources in particular of the old maps available. Other units, linear,specific or surface can be highlighted by a direct confrontation with the land.The analysis was carried out on the “ligérienne” part of the old industrial valley of Gier.The method implements a Geographical Information system (GIS) historical whichauthorizes integration, the reinterpretation and the crossing of information coming fromvaried sources: serial, textual, cartographic or iconographic. This tool facilitates thedivision and the update of the produced data.The results obtained have offered a reading of the evolution of the valley of Gier on theplot scale for two centuries. They propose a new approach of the territory and its evolutionwhich makes it possible to question the patrimonial policies or of installation existing.Extended to the whole space of the agglomeration of Saint-Etienne or to other territoriesand regularly updated, the method can make it possible to constitute an observatory oftheir evolution.

Sweden has an enormous treasure in its vast number of large-scale historical maps from a period of 400 years made for different purposes, that we call map series. The maps are also very time and regional dependent with respect to their concepts. A large scanning project by Lantmäteriverket will make most of these maps available as raster images. In many disciplines in the humanities and social sciences, like history, human geography and archaeology, historical maps are of great importance as a source of information. They are used frequently in different studies for a variety of problems. A full and systematic analyse of this material from a database perspective has so far not been conducted. During the last decade or two, it has been more and more common to use data from historical maps in GIS-analysis. In this thesis a novel approach to model these maps is tested. The method is based on the modelling of each map series as its own ontology, thus focusing on the unique concepts of each map series. The scope of this work is a map series covering the province of Gotland produced during the period 1693-1705. These maps have extensive text descriptions concerned with different aspects of the mapped features. Via a code marking system they are attached to the maps. In this thesis a semantic analysis and an ontology over all the concepts found in the maps and text descriptions are presented. In our project we model the maps as close to the original structure as possible with a very data oriented view. Furthermore; we demonstrate how this ontology can be used as a conceptual schema for a logical E/R database schema. The Ontology is described in terms of the Protégé meta-model and the E/R schema in UML. The mapping between the two is a set of elementary rules, which are easy for a human to comprehend, but hard to automate. The E/R schema is implemented in a demonstration system. Examples of some different applications which are feasibly to perform by the system are presented. These examples go beyond the traditional use of historical maps in GIS today

This study had the main aim of identifying and analyzing the bad treatments against elderly into the urban limits of the city of Santa Maria / RS, Brazil. Strictly, it meant: placing the relation among existence, frequency, location and types of domestic violence against elderly; identifying the levels of vulnerability to bad treatment cases reported in the police in a department called Policy of the Elderly; and charting the levels of the amount of elderly in risk through geo information. In order to contribute with the process of the analyses of the bad treatments in a more dynamic way, it was used the geomatics because this area of knowledge gives priority to the multidisciplinary, motivating to perform works by using and developing the information technology. In order to reach the proposed aims, a survey in the termination services like Accident Report (AR) was done in the Police of the Elderly in terms of the years 2007/2008. After getting the information, the following categories were established: age, genre, marital status, neighborhood, victim s literacy and profession besides the type of reported aggression, relation aggressed and aggressor. Possessing that information, it was done a cadastral map placing the aggressors in the city what gives origin to the listed categories. For this, it was used the GIS and the computer program ARCVIEW 9, 3. The database is the table of Attributes for ARCVIEW 9, 3 organized by the Excel table. Each column has a field with spatial and non spatial data that were interrelated allowing visualization. In a whole, 350 chips were analyzed but 138 of them were excluded because they did not fit with the aim of the study. It was observed in the study that from those 212 chips that were investigated for bad treatments against elderly, 92 (44%) lived with the partner. According to WHO (2007), almost half of all women murdered are killed by partner or boyfriend. The majority of the elderly have basic education (63.21%), following by High School (9.43%) scoring 72.64%. Considering some years ago education was not as accessible as it is nowadays, that datum is completely justified. The female genre between 70 to 79 years old was what presented higher levels of bad treatments, scoring 63 women. Married women and widows presented the highest levels. Analyzing the type of aggression, this study verified that the physical aggression happens with more frequency, following by psychological (threats) ones and bad treats. From the 42 neighborhoods into the urban limits of the city of Santa Maria RS, it was verified that the largest concentration of bad treatments against elderly is placed in the neighborhoods: center with 22 complaints; Camobi with 17 complaints; Passo da Areia with 15 complaints. Among the aggressors, 83 were sons, 24 were daughters and 19 were victim s neighbors. Relating the indexes presented by the Police of the Elderly, it is important to clarify that they do not represent the whole amount of the bad treatments there are in our society. There are elderly who do not denounce the violence, the aggressions and the contempt situation in which they live. / Este estudo apresentou como objetivo principal identificar e analisar a prevalência de maus tratos contra idosos no perímetro urbano do município de Santa Maria/RS. Especificamente, buscou-se: espacializar a relação entre existência, frequência localização e tipos de maus tratos domésticos contra idosos; identificar os níveis de vulnerabilidade para casos de maus-tratos notificados através da Delegacia do Idoso; e mapear os níveis de concentração de idosos em situação de risco, analisando com a aplicação de geotecnologias da informação. Para contribuir com o processo de análise dos maus tratos tornando-o mais dinâmico, utilizou-se a Geomática, por ser uma área de conhecimento que prioriza a multidisciplinariedade, incentivando a realização de trabalhos através do uso e desenvolvimento da tecnologia da informação. Para atender aos objetivos propostos foi realizado um levantamento nos serviços de denúncia do tipo Boletim de Ocorrência (BO), junto a Delegacia do Idoso referente aos anos de 2007/2008. Após a obtenção dos dados foram estabelecidas as seguintes categorias: idade, sexo, estado civil, bairro, grau de instrução e profissão do agredido, além do tipo de agressão declarada, relação agredido/agressor. Com os dados referidos elaborou-se um mapa cadastral, que identificasse no ambiente urbano do município a espacialização dos agredidos, gerando assim as relações entre as categorias elencadas. Utilizou-se o SIG e programa computacional ARCVIEW 9.3. O Banco de Dados é a tabela de Atributos do aplicativo ARCVIEW 9.3, organizada com a tabela Excel. Cada coluna contém um campo com dados espaciais e não espaciais que foram relacionados entre si permitindo sua visualização. Ao todo, foram analisadas 350 fichas. Destas foram excluídos 138 fichas por não se encaixarem aos objetivos do estudo. Observou-se no estudo que das 212 fichas investigadas de notificação de maus-tratos contra o idoso, 92 (44%) viviam com o companheiro. Segundo a OMS (2007), quase metade das mulheres assassinadas são mortas pelo companheiro ou namorado. A maioria dos idosos possui Ensino Fundamental (63,21%), seguido por Ensino Médio (9,43%) totalizando 72,64%. Este dado é plenamente justificável considerando que a educação há alguns anos não era tão acessível como atualmente. O sexo feminino foi o que apresentou maior incidência de maus tratos na faixa etária de 70 a 79 anos com 63 mulheres. As mulheres casadas (52) e viúvas (67) foram as que apresentaram a maior prevalência. Analisando o tipo de agressão, verificou-se neste estudo, que a agressão física é a mais frequente, seguida de agressões psicológicas/ameaças, e maus tratos. Dos 42 bairros constantes na Divisão administrativa do perímetro urbano do município de Santa Maria RS verificou-se que a maior concentração de maus tratos contra idosos está localizada nos bairros: Centro com 22 denúncias, Camobi com 17; Passo D Areia 15 notificações. Conforme a relação com agressor quem mais agrediu foram 83 filhos, 24 filhas e 19 vizinhos. Relacionando aos números apresentados pela Delegacia do Idoso, é importante esclarecer que não significam a totalidade dos maus tratos existentes em nossa sociedade. Há ainda muitos idosos que não denunciam a situação de violência, agressões e menosprezo em que vivem.

For a little over twenty years, researchers have worked on integrating multi-criteria aggregation procedures (MCAP) to GIS. Several notable contributions have brought this field to what it is today. After studying the course of MCDA-GIS integration through several works, we question the future of such an attempt. Indeed most works that aim for an integration do not survive long after their direct purpose has been fulfilled. We end up understanding through a critical review of the existing systems that technical integration means nothing if it is not visible to the user on an operational level.<p><p>We therefore propose several contributions to improve the usability of MCDA methods in a geographic context. One of our works consists in adapting the PROMETHEE-GAIA methodology to be used on maps for spatially referenced problems. To do so, we define symbols/glyphs that display select parts of the results obtained through the PROMETHEE and GAIA methods. This allows for the comparison of alternatives' profiles and characteristics based on their geographic location which wasn't possible before. This adaptation helps us combine multicriteria and geographic aspects in an entirely new way.<p>We also propose some extensions of the GAIA method to improve the quality of the results and reduce the risk of wrong interpretations to be made due to losses of data. / Doctorat en Sciences de l'ingénieur / info:eu-repo/semantics/nonPublished

Cette étude est une analyse spatiale des vestiges archéologiques datant des origines à la fin de l'Ancien Empire des complexes funéraires de Netjerikhet et de Sekhemkhet à Saqqâra. L'objectif de ce travail est de comprendre l'occupation spatiale, l'enjeu étant de pouvoir traiter la quantité très importante d'informations issues des publications. Pour cerner l'ensemble de la documentation à examiner, une base de données et un système d'information géographique (SIG) ont été créés. Grâce à ces outils informatiques, tous les artefacts ont ainsi pu être enregistrés et une analyse des relations qu'entretiennent les différentes entités recensées a pu être menée pour tenter de comprendre comment s'organise l'espace. Des cartes imprimées permettent de visualiser les résultats de requêtes, d'ailleurs imaginées au préalable pour construire les outils. Cet examen des données, par le biais d'une géodatabase, permet également d'évaluer tout l'intérêt des outils informatiques pour la recherche. / This study is a spatial analysis of the archaeological remains dating from the origins to the end of the Old Kingdom of the funerary complexes of Netjerikhet and Sekhemkhet at Saqqara. The aim of this work is to understand the spatial occupation, the challenge is to process the very large amount of information from publications. To include all of the documentation to be processed, a database and a geographic information system (GIS) have been created. With these software tools, all artefacts were thus recorded and an analysis of the relationships between the different entities listed has been conducted to try to understand how the space is organized. Printed maps allow visualizing query results, queries moreover imagined beforehand to build the tools. This examination of the data through a geodatabase also assesses the interest of these tools for research.

Fallopiajaponica (Houtt.) Ronse Decraene (Japanese knotweed), an introduced, invasive, rhizomatous perennial plant, has become an increasing problem for nature conservation and land management in both rural and urban areas in the British Isles. In the native range of the plant, Japan, Taiwan and northern China, a number of varieties are recorded. Three congeners of F. japonica are present in the British Isles, F. sachalinensis, F. japonica var. conipacta and F. baldschuanica in addition to a hybrid F. x bohemica. An investigation by postal survey of the distribution of the hybrid F. x bohemica has identified 131 records for the British Isles. Both male and female plants of F. x bohemica have been recorded. Current understanding suggests that only female plants of F. japonica are present in the British Isles, inferring that the only means of reproduction is through vegetative regeneration. High rates of regeneration were recorded in this study for stem and rhizome material for both F. japonica and F. x bohemica in an aquatic and terrestrial environment. Implications of vegetative regeneration are discussed in terms of current management practices and future methods of control. A combination of digging with a mechanical excavator followed by spraying with the herbicide glyphosate decreased the time required to achieve an effective level of control of F. japonica compared to spraying alone. Fragmentation of the rhizome system through digging resulted in an increase in stem density allowing a more effective delivery of herbicide. Implications in terms of costs for F. japonica treatment on sites awaiting re-development are discussed. Analysis of data collated from surveys of F. japonica in Swansea using a Geographical Information System suggest that the primary habitats infested are waste ground and stream and river banks. Results suggest that disturbance, both by natural means and by human intervention has been the primary cause of spread of F. japonica in the British Isles. Management strategies are proposed which take account of these results and measures are put forward to help prevent future infestations.

[pt] Esta dissertação apresenta um estudo sobre a Logística de Distribuição de Derivados de Petróleo, com o desenvolvimento de um modelo matemático para a determinação do custo logístico de distribuição de estoques estratégicos de combustíveis no Brasil. Este modelo não se propõe a determinar o custo ótimo de distribuição e sim estabelecer uma distribuição racional dos estoques estratégicos de combustíveis, com base na malha real de distribuição. Para o desenvolvimento do modelo, primeiramente foi realizado um diagnóstico da distribuição nacional de derivados energéticos de petróleo, mais especificamente gasolina, diesel, querosene de aviação, óleo combustível e gás liquefeito de petróleo. Para cada um desses produtos determinou-se a malha de distribuição, com os fluxos de produtos entre refinarias e terminais até as bases de combustíveis secundárias, com seus respectivos modais. Em uma segunda etapa foi construída uma base de dados georreferenciada no software TransCAD, com os dados levantados e diagnosticados anteriormente. Este sistema de informação geográfica serviu como suporte para análises logísticas e como facilitador na visualização e entendimento dos resultados. Os resultados dos custos logísticos de distribuição são mostrados por produto, para evidenciar o caráter diferenciado da distribuição de cada derivado. / [en] This master thesis presents a study about the petroleum products distribution system, with the development of a mathematical model for the determination of the logistic costs of strategic oil stocks in Brazil. This model doesn't intend to determine the minimum distribution cost , but establishes a rational distribution of the strategic oil stocks, based in the real distribution network. For the development of the model, firstly a diagnosis of the national distribution was accomplished, more specifically gasoline, diesel, aviation kerosene, combustible oil and liquefied gas of petroleum. For each one of those products, the distribution network was determined, with the flows of products among refineries and terminals until the secondary bases of fuels, with its respective transportation modals. In a second stage, a geographical database was built in the software TransCAD, with all available data . This geographical information system was a support tool for logistic analyses, visualization and better understanding of the results. The results of the total distribution costs are presented by product, to evidence the differentiated character of the distribution of each product.

La localisation d’un véhicule est une étape cruciale dans le développement des véhicules intelligents. Les recherches sur ce sujet ont connu un grand essor ces dernières années. L’accent est souvent porté sur la précision de la localisation, nous présentons ici une méthode de localisation sur carte existante dont l’objectif est d’estimer la position du robot non seulement de façon précise mais également de manière fiable. Pour ce faire, l’algorithme développé se présente en deux étapes principales : une étape de sélection et de perception des informations les plus pertinentes et une étape de mise à jour de la position estimée et de la fiabilité, cette dernière étape permet également de détecter et réparer ou éliminer les précédentes erreurs. La perception de l’environnement est réalisée à travers différents capteurs associés à des détecteurs spécifiques. L’humain utilise aussi différents capteurs pour se localiser et va intuitivement sélectionner le plus performant, s’il fait jour il utilisera ses yeux, sinon son oreille ou le toucher. Nous avons développé une approche similaire pour le robot qui tient compte des contraintes environnementales et de l’estimation actuelle pour sélectionner à chaque instant l’ensemble capteur, indice de la scène et détecteur le plus pertinent. La phase de perception, étant pilotée par un processus Top-Down, peut bénéficier d’informations déjà connues permettant ainsi de se focaliser sur l’indice recherché et d’améliorer les phases de détection et d’associations de données. Cette approche Top-Down s’appuie sur un réseau bayésien. Ce dernier permet de modéliser les interactions entre les différents événements qui se produisent en gérant l’incertitude. Il permet une prise en compte facile des différents événements. Par ailleurs le réseau bayésien présente une grande flexibilité pour l’ajout d’événements supplémentaires pouvant engendrer des non-détections (tels que dysfonctionnement de capteurs, conditions météorologiques, etc.). Les données de l’environnement sont rendues disponibles grâce à une carte géoréférencée préalablement fournie. Avec le développement de cartes disponibles facilement sur internet, cette façon de faire permet d’exploiter au mieux l’information déjà fournie. L’utilisation d’une carte géoréférencée permet d’avoir un référentiel commun entre tous les véhicules ou éléments de l’infrastructure facilitant ainsi l’échange d’informations et ouvrant du coup la possibilité d’interaction simplifiées dans le cas de flottes par exemple. Les résultats montrent que l’approche développée est pertinente pour une localisation précise et fiable aussi bien statique que dynamique. L’ajout de nouveaux capteurs se fait naturellement et sans nécessiter d’heuristique particulière. La localisation obtenue est suffisamment précise et fiable pour permettre des applications de conduite autonome utilisant, entre autres, cet algorithme. / Vehicle localization is a crucial step in the development of smart vehicles. The research in this domain has been growing in recent years. Generally, the effort is focused on the localization accuracy, we present here a localization method on existing map where the objective is to estimate the robot position not only with accuracy but also with confidence. To achieve this , the algorithm developed has two main steps : one, selection and perception of the most relevant informations and two, position estimation and confidence update. This last step also allows to detect and eliminate the previous errors. Environment perception is well achieved, thanks to different sensors associated with specific detectors. Humans use different senses, shifting automatically in order to localize themselves depending on the situation of the environment, for e.g if there is enough illumination we depend on eyes, else the ear or the touch otherwise. We have developed a similar approach for the robot that takes into account the specific environmental constraints and actual position estimation to select at each instant the most relevant set of sensor, landmark and detector. The perception step, led by a top-down process, can use already known informations allowing a focus on the searched landmark and an improvement of the detection and data associations steps. This top-down approach is well implemented, thanks to a Bayesian network. Bayesian network allows to model the interactions between the different probable events with management of the uncertainty. With this network, it is very easy to take into account those different events. Moreover, a Bayesian network has a great flexibility to take into consideration additional events that can cause false detections (like sensor failure, meteorological conditions and others). The environment data is obtained with a Georeferenced map (from GIS). With the already available maps on the internet, allows to exploit an already existing information. The use of a Georeferenced map facilitates the communication of informations between a vehicle and several other vehicles or with an element of the infrastructure, that can be very useful for multi vehicle coordination, for example. The results shows that the developed approach is very accurate and reliable for localization, whether static or dynamic, and can be applied for autonomous driving. Moreover, new sensors can be added at ease.