Les diélectriques solides sont les éléments constitutifs de base des isolants utilisés dans les composants ou les systèmes du génie électrique et de l'électronique de puissance. Principalement à cause des tendances à la miniaturisation, ces isolants sont soumis à des contraintes sans cesse croissantes (électriques, mécaniques et thermiques) qui peuvent dégrader ou induire un vieillissement prématuré des diélectriques. Ceci peut conduire à la défaillance ou à un claquage de la structure, phénomènes qui l'on doit prendre en compte et étudier. De manière plus précise, la charge globale stockée à l'intérieur du matériau diélectrique, généralement dénommée charge d'espace, est directement liée à ces processus de dégradation. Ainsi, il devient alors nécessaire d'analyser le comportement de la charge d'espace lorsque le matériau est contraint dans des conditions proches d'un environnement réel d'utilisation. Parmi les techniques existantes, la (F)LIMM est une méthode thermique dédiée à l'analyse de la charge d'espace dans des fines couches diélectriques (avec une épaisseur comprise entre 5µm et 50µm), avec la possibilité de réaliser des cartographies en 3-D lorsque le faisceau laser est focalisé à la surface de l'échantillon étudié. Les premières caractérisations par cette technique étaient liées à la détermination de la distribution de la charge d'espace dans des films minces qui avaient été préalablement soumis à un champ électrique continu, puis analysés hors tension (volt-off). Cette procédure dite " off-line ", et mise en œuvre pendant la dépolarisation de l'échantillon, est restée longtemps la seule façon d'analyser la charge d'espace. De nouveaux développements sont devenus nécessaires plus récemment pour apporter des réponses au domaine industriel pour lequel une caractérisation du comportement dans des conditions d'utilisation réelles faisait défaut. Dans ce but, une version " sous tension " ou " on-line " du banc expérimental (F)LIMM conventionnel a été développée. Ainsi, la mesure du courant (F)LIMM se réalise maintenant de manière simultanée à l'application d'un champ électrique continu externe sur l'échantillon. Outre la mesure in-situ et l'analyse sous tension, ces nouveaux développements présentent un avantage supplémentaire consistant en la possibilité d'évaluer et de calibrer les profils de température. Dans ce travail, les modifications expérimentales réalisées sur le banc test de mesure sont tout d'abord détaillées et validées. Puis, la procédure de calibration du model thermique proposé est expliquée et testée. Ainsi, la simulation les courants (F)LIMM devient possible et l'on constate une bonne adéquation avec les courants expérimentaux enregistrés. Enfin, quelques applications à des films polymères minces en PEN et en PP sont décrites. Les résultats obtenus prouvent alors que ces évolutions expérimentales et théoriques sont efficaces pour l'étude du comportement de la charge d'espace " sous tension ". / Solid dielectrics are basic elements of most insulations used in devices or systems in electrical engineering or in power electronics. Mainly due to current trends in downsizing, these insulations being subjected to increasing stresses (electrical, mechanical or thermal) may degrade and age dielectrically. This may lead to failure and breakdown, which have to be addressed. More specifically, the global charge store inside the dielectric and generally called space charge is directly linked to these degradation processes. It is therefore necessary to analyze this space charge behavior when the material is at the same time stressed under conditions close to those of the real environment of use. Among many existing techniques, the (F)LIMM is a thermal wave method dedicated to the space charge analysis of thin dielectric films (with a thickness from 5µm to 50µm), with a possibility of 3-D cartographies when the beam is focused onto the sample under study. Its first characterizations were related to the determination of the space charge distribution in thin films that were first submitted to an external DC field and next analyzed in a volt-off way (no voltage applied). This off-line procedure remained the only way of space charge investigations for quite a while. New developments became necessary in order to provide answers to the industry for which a characterization close to the actual conditions of use was lacking. For this purpose, an on-line version of the conventional (F)LIMM set-up was developed. It allows a measurement of the (F)LIMM current when simultaneously a DC bias field is applied to the sample. This development presents, apart from allowing an in-situ and on-line analysis, another important advantage, namely the possibility for assessing or calibrating the temperature profile. In this work, the experimental modifications made to the measuring test rig are first detailed and validated. Then, the calibration procedure of the thermal model developed is explained and tested. As a consequence, the possibility of simulating the (F)LIMM currents is possible and shows a very good agreement with the experimental registered ones. Finally, some applications to PEN and PP polymer thin films are described and results got prove the efficiency of these new developments for space charge on-line investigation.

Neste trabalho estudamos uma dinâmica estocástica de partículas de duas espécies baseada em células. Basicamente, incorporamos algumas inovações em um modelo unidimensional proposto e resolvido por R. da Silva et al. (Physica A, 2015), que considera que em um célula, na ausência de partículas da espécie contrária, a partícula vai pra frente com uma probabilidade p, que representaria um campo na direção longitudinal de um corredor e fica na própria célula com q=1-p. Contudo, essa probabilidade p é reduzida de acordo com a concentração de partículas contrárias. Nosso trabalho não apenas estendeu o problema pra duas dimensões como também incluiu aspectos relativos a colisão e o espalhamento para células vizinhas. Nossos resultados são divididos em duas situações: a) Espécie contrária permanece imóvel funcionando como obstáculos b) Espécie contrária em movimento. Na primeira situação podemos ver uma interessante transição na distribuição dos tempos de travessia em função das concentrações dos obstáculos, por monitorar a curtose da distribuição. Quando a espécie contrária se movimenta, vemos que o tempo de destilação entre as partículas (tempo para que as espécies estejam geograficamente separadas no corredor) depende do parâmetro ligado ao espalhamento transversal das partículas, parâmetro este, que não influencia no caso das partículas paradas. Finalmente nós colocamos as partículas em um sistema com condições periódicas de contorno. Neste caso, podemos observar o aparecimento de padrões de bandas longitudinais ao campo, exatamente como ocorrem em problemas de coloides carregados sob a ação de campos longitudinais e em modelos de pedestres em corredores. Mostramos como o sistema relaxa para tal tipo de estado estacionário utilizando um adequado parâmetro de ordem ligado a segregação das partículas. Nosso modelo, diferentemente dos modelos para pedestres, não se baseia em equações tipo Langevin. Nossa abordagem é totalmente estocástica e por esse ponto de vista ainda mais fundamental e geral, podendo ser estendida para mais modelos de partículas em fluxos contrários. Nossa solução vem tanto através de simulações Monte Carlo bem como soluções das equações diferenciais parciais que descrevem o sistema e que são oriundas das recorrências estabelecidas para os caminhantes aleatórios. As simulações Monte Carlo e soluções via EDP mostram boa concordância em todos os aspectos analisados, tanto qualitativa quanto quantitativamente. / In this work we study a stochastic dynamic of particles of two types based on cells. Basically we incorporate some innovations on a one-dimensional model proposed and solved by R. da Silva et al. (Physica A, 2015) which considers that in the absence of particles of the opposite species in the cell a particle goes toward the next cell with probability p and returns to the previous cell with probability q = 1 p. However this motion probability linearly decreases with the relative density of the contrary species. Our work not only expands the problem for two dimensions but also includes collision aspects by adding scattering to the neighbouring cells. Our results are divided into two di erent categories: a) One of the species remain xed in their places which means that such particles will work as obstacles; b) Both species can move in the environment. In the rst situation we can observe, by monitoring the kurtosis, that an interesting transition of the crossing time distribution arises as the concentration of the obstacles increases. When both species can move we can observe that the distillation time (spent time for the complete geographical separation of the species in the corridor) depends on the parameter related to the perpendicular scattering of the particles. This same parameter has shown no in uence over the time distributions in the rst situation. Finally we implement periodic boundary conditions in the eld's direction. In this case we are able to observe the arising of band patterns parallel to the eld's direction exactly as it does with oppositely charged colloids under the in uence of a uniform electric eld or pedestrian dynamics in corridors. We also show how the system relax to such stationary state by using a suitable order parameter related to the particles segregation. Di erently from other pedestrian dynamics models, our model is not based on a Langevin-type equation. Our approach is totally stochastic and from this point of view, more fundamental and general to be extended to more types of models considering particles under counter ow. Our solution is obtained by both Monte Carlo simulations and numerical integration of partial di erential equations (PDE) from recurrence relation of the directed random walkers. The Monte Carlo simulations and the solutions of the PDE show a good agreement in all aspects analysed both qualitatively and quantitatively.

Esta pesquisa tem por objetivo geral analisar o uso da tecnologia digital para facilitar a integração creche-família. Foi desenvolvida de acordo com os pressupostos da abordagem qualitativa de pesquisa (FLICK, 2004; SEVERINO, 2007; LÜDKE e ANDRÉ, 1986) e, mais especificamente, como estudo de caso (YIN, 2005) a partir dos dados empíricos obtidos na creche municipal de São Bernardo do Campo/SP, onde a pesquisadora exerce o cargo de diretora escolar. As pesquisas teóricas sobre a Educação de crianças de 0 a 3 anos e sobre os usos da tecnologia digital nos contextos escolares buscaram situar os conceitos centrais que envolviam a proposta. A pesquisa empírica foi realizada por meio de questionário enviado a todas as famílias das crianças atendidas na creche e por entrevistas semi-estruturadas realizadas com mães/pais e professoras. Pela análise dos dados levantados, constatou-se que a maioria das famílias das crianças matriculadas na creche pesquisada tem acesso à internet, principalmente via telefone celular, e a utiliza diária ou semanalmente. Outra constatação foi de que são poucas as famílias que acessam o Blog da instituição para acompanhar o trabalho desenvolvido com seus filhos. Por outro lado, a maior parte das famílias, que respondeu ao questionário, afirmou que utilizaria ferramentas de comunicação via internet para se comunicar com a creche e sugeriu o aplicativo WhatsApp para potencializar a integração creche-família. Em termos de resultados, a pesquisa identifica que as famílias e professoras da creche estão imersas na cultura digital, fazem uso de dispositivos móveis e desejam usar esses dispositivos para interações frequentes com a creche a fim de, no caso das famílias, acompanhar virtualmente o desenvolvimento de seus filhos. Neste sentido, considera-se que a tecnologia digital pode facilitar e potencializar a comunicação e a integração entre creche e família e que as iniciativas de propostas para essa maior integração devem partir da equipe da creche. / This researchs general goal is to analyze the use of digital technology to facilitate day care/family integration. It has been developed in consonance with the theoretical assumptions of the Qualitative Research approach (FLICK, 2004; SEVERINO, 2007; LÜDKE and ANDRÉ, 1986) and, more specifically, with the assumptions of the Case Study approach (YIN, 2005) applied on the empirical data obtained from the municipal day-care center of São Bernardo do Campo/SP, where this researcher holds the position of headteacher. Our theoretical researches about the Education of children under 3 years old and the uses of digital technology in school context aim to shed light on the main concepts that involve our general goal. The empirical research is based on a survey form sent to the families of all the children that attend the mentioned day-care center and on semistructured interviews with their parents and teachers. Through the analysis of the collected data, it has been possible to realize that the majority of those families have access to the internet, mainly through smartphones, and use it on a daily or weekly basis. Another realization the empirical analysis made possible was that few are the families that access the institutional Blog to keep up with the work developed with their children. On the other hand, the majority of the families who answered the survey said they would use internetbased communication tools to be in touch with the day-care center, and suggested that WhatsApp could improve the day care/family integration. In terms of results, this research shows that the families and daycare teachers are immersed in digital culture, make regular use of mobile devices and that the family wish to employ them on everyday interaction with the day-care center in order to keep up with their childrens development. Therefore, we consider that digital technology can facilitate and improve communication and integration between day-care center and families, and such initiatives should come from the daycare staff.

Este estudo tem como objetivo investigar o desenvolvimento psíquico de gêmeos abrigados no decorrer do primeiro ano de vida. Foram separados ao nascer, tendo como histórico inicial de vida, para um, a ausência do contato materno logo após o nascimento, tendo permanecido hospitalizado até os oito meses de vida e seguido para o abrigamento; e para o outro, a descontinuidade do contato materno e familiar aos dez meses de vida, seguido pelo abrigamento. No desenvolvimento deste estudo, os gêmeos estavam com cinco anos de idade, ainda abrigados, na mesma instituição e próximos da adoção. Trata-se de uma pesquisa qualitativa, do tipo estudo de caso, embasada pelo método psicanalítico, à luz dos pensamentos de D. W. Winnicott. Por meio do diagnóstico psicológico do tipo compreensivo e interventivo, com o emprego das técnicas projetivas hora de jogo e procedimento de desenhos-estórias, buscou-se a compreensão da vida intrapsíquica dos gêmeos. Instrumentos auxiliares como o estudo documental do histórico familiar, como também entrevistas realizadas no abrigo, colaboraram para o alcance de uma visão totalizadora dos casos. A partir da análise dos dados, observaram-se marcantes divergências do funcionamento psíquico dos gêmeos: aquele institucionalizado desde o nascimento apresenta maior interação ambiental e facilita o emprego de intervenções, diante de angústias depressivas emergentes, mostra comportamentos regressivos; o outro separado do lar aos dez meses, diante de angústias despertadas, utiliza-se de mecanismos defensivos de negação e evitação, cujas intervenções, por muitas vezes, são evitadas de modo a afastar a quebra do sistema defensivo operante. Quanto ao sistema defensivo, o primeiro apresenta comportamentos regressivos e fragilidade frente ao ambiente, além de dissociações entre movimentos regressivos e evoluídos; o segundo, mostra comportamentos agressivos e oposicionistas frente a limites e regras, como recursos para domínio e controle ambiental e agressões e violência nas relações, desencadeando dificuldades de interação com o ambiente, permanecendo isolado. Além dos aspectos clínicos observados, é possível conjeturar sobre a influência de aspectos psicopatológicos institucionais, especialmente quanto à clivagem dos gêmeos. Esta pesquisa também permitiu vislumbrar que prejuízos ao desenvolvimento também se desencadeiam de lares vitimizadores e não só de contextos institucionais; e que o processo de adoção deve ser mais bem estruturado, a fim de minimizar estados ansiógenos decorrentes, que observamos nos gêmeos, bem como a devolução da criança ao abrigo após sua adoção, como ocorreu com o segundo gêmeo. Dessa forma, concluiu-se que as crianças, mesmo em situação de abrigamento, apresentam aspectos saudáveis propulsores para a saúde e podem aproveitar experiências boas do ambiente, além de apresentarem capacidade para amar e estabelecer relações afetivas; por isso o contexto institucional merece assistência, principalmente psicológica, a fim de favorecer o ambiente e o desenvolvimento das crianças abrigadas. / This essay is aimed at investigating the psychic development of twins institutionalized since before one year of age. Parted at birth, one child is sent to a hospital where it stays for eight months before is sent to an institution. The child undergoes the absence of the early motherly contact. As for the other child, there is the early motherly contact and from the family, however, discontinued. At ten months of age, the child is then sent to the institution. The study is set up when the twins are five years old, and are still inmates at the same institution, and are on the verge of being adopted. This research is qualitative - a case study – grounded on the psychoanalytic method - according to D.W. Winnicott’s lights. By using comprehensive and intervening psychological diagnosis, as well as projective techniques such as Hour Game and the Drawing-stories procedure, it aims at understanding the intra-psychic life of the twins. Additional tools such as the documented study of the family history, along with the interviews carried out at the shelter, contribute to a thorough view of the cases. As data analysis is performed, expressive differences in the functioning psychic of the twins come to light: on the one hand, the child sent to the institution only just after birth shows more interaction with the environment, when confronted with emerging depressive anguish shows regressive behaviour. Also, the child makes it easy for the implementation of interventions. On the other hand, the child parted from its home at ten months of age reacts otherwise. This, when confronted by desperate anguish, utilises defence mechanisms such as negation and avoidance, whose interventions are, in most cases, avoided in order to maintain the ongoing defensive system. When it comes down to the defence system, of the two twins, the first shows regressive behaviour and fragility –as it faces the environment and dissociation in regressive and evolved movements. The second twin shows aggressive behaviour and opposition to limits and regulations as means of controlling the environment followed by aggression and violence in the relationships leading into difficulty in interacting and isolation from the environment. In addition to the clinical aspects, the influence of the psychopathological institutional aspects might also be pondered, such as the split-off the twins. The research, too, permits visualising what damages to the development come from oppressive homes and not only from institutional contexts, and that the adopting process should be thought over, and restructured in order to soothe the state of anxiety, as is observed in the twins during adoption, or the sending back to shelter of the second child, after its adoption. Thus, we conclude that even children under shelter can show healthy aspects, in that the children will benefit from the good experiences of the environment, and be able to love and hold affective relations. For this reason, the institutional context should be given assistance, principally psychological assistance, and then offers the children under shelter an appropriate environment which favours their development.

The evolutionary pressure to prevent re-replication by inactivating licensing during S phase leaves higher-eukaryotes with large genomes, such as human cells, vulnerable to replication stresses. Origins licensed in G1 must be sufficient to complete replication as new origins cannot be licensed in response to irreversible replication fork stalling. Interdisciplinary approaches between cellular biology and biophysics predict that replication of the genome is routinely incomplete in G2, even in the absence of external stressors. The frequency of converging replication forks that never terminate due to irreversible stalling (double fork stall), which result in a segment of unreplicated DNA, was modelled using high quality origin-mapping data in HeLa and IMR-90 cells. From this, hypotheses were generated that related an increase in unreplicated segments of DNA to reduced functional origin number. Presented in this thesis is the confirmation of this relation by quantifying chromosome mis-segregation and DNA damage responses when origin number was reduced using RNAi against licensing factors. The number of ultrafine anaphase bridges and 53BP1 nuclear bodies are in remarkable concordance with the theoretical predictions for the number of double fork stalls, indicating that cells are able to tolerate under-replication through such post-replicative cellular responses. 53BP1 preferentially binds to chromatin associated with large replicons, and functions synergistically with dormant origins to protect the stability of the genome. Additional candidates, inspired by common fragile site research, have also been characterised as responders to spontaneous under-replication, and include FANCD2 and MiDAS, which function in early mitosis to facilitate completion of replication before cells enter anaphase. In conclusion, a series of mechanisms that sequentially function throughout the cell cycle protects the stability of the human genome against inevitable spontaneous under-replication brought about by its large size.

As informações de mortalidade são úteis para avaliar a situação de saúde de uma população. Dados de mortalidade confiáveis produzidos por um sistema de informação de saúde nacional constituem uma ferramenta importante para o planejamento de saúde. Em muitos países, sobretudo em desenvolvimento, o sistema de informação de mortalidade continua precário. Apesar dos esforços feitos em Moçambique para melhoria das estatísticas de mortalidade, os desafios ainda prevalecem em termos de tecnologias de informação, capacidade técnica de recursos humanos e em termos de produção estatística. O SIS-ROH é um sistema eletrônico de registro de óbitos hospitalares de nível nacional, implementado em 2008 e tem uma cobertura de apenas 4% de todos os óbitos anuais do país. Apesar de ser um sistema de nível nacional, ele presentemente funciona em algumas Unidades Sanitárias (US), incluindo o Hospital Central da Beira (HCB). Dada a importância deste sistema para monitorar o padrão de mortalidade do HCB e, no geral, da cidade da Beira, este estudo avalia a qualidade do SIS-ROH do HCB. É um estudo descritivo sobre a completitude, cobertura, concordância e consistência dos dados do SIS-ROH. Foram analisados 3.009 óbitos de menores de 5 anos ocorridos entre 2010 e 2013 e regsitrados no SIS-ROH e uma amostra de 822 Certificados de Óbitos (COs) fetais e de menores de 5 anos do HCB. O SIS-ROH apresentou uma cobertura inferior a 50% calculados com os dados de mortalidade estimados pelo Inquérito Nacional de Causas de Morte (INCAM). Verificamos a utilização de dois modelos diferentes de CO (modelo antigo e atual) para o registro de óbitos referentes ao ano de 2013. Observou-se completitude excelente para a maioria das variáveis do SISROH. Das 25 variáveis analisadas dos COs observou-se a seguinte situação: 9 apresentaram completitude muito ruim, sendo elas relativas à identificação do falecido (tipo de óbito e idade), relativas ao bloco V em que dados da mãe devem ser obrigatoriamente preenchidos em caso de óbitos fetais e de menores de 1 ano (escolaridade, ocupação habitual, número de filhos tidos vivos e mortos, duração da gestação) e relativas às condições e às causas de óbito (autópsia e causa intermédiacódigo); 3 variáveis apresentaram completitude ruim relativas à identificação do falecido (NID) e relativas às condições e causas de morte (causa intermédia - descrição e causa básica - código); 9 apresentaram completitude regular relativas à identificação do falecido (data de nascimento e idade), relativas ao bloco V (idade da mãe, tipo de gravidez, tipo de parto, peso do feto/bebé ao nascer, morte do feto/bebé em relação ao parto) e relativas às condições e causa de óbito (causa direta- código, causa básica descrição); 2 apresentaram completitude bom relativas à identificação do falecido (sexo e raça/cor) e, por último, 2 apresentaram completitude excelente relativas ao local de ocorrência de óbito (data de internamento e data de óbito ou desaparecimento do cadáver). Algumas variáveis do SIS-ROH e dos COS apresentaram inconsistências. Observou-se falta de concordância para causa direta entre o SIS-ROH e os COs. Conclusão: Moçambique tem feito esforços para aprimorar as estatísticas de mortalidade, porém há lacunas na qualidade; a análise rotineria dos dados pode identificar essas lacunas e subsidiar seu aprimoramento. / The mortality information is useful to assess the health status of a population. Reliable mortality data produced by a national health information system is an important tool for health planning. In many countries, especially developing countries, the mortality information system is still precarious. Despite efforts in Mozambique to improve mortality statistics, challenges still prevail in terms of information technology, technical capacity and human resources and statistical production. The SIS-ROH is an electronic system of national-level hospital deaths registration, implemented in 2008 and has a coverage of only 4% of all annual deaths in the country. Despite being a national system, it currently works in some health units (US), including Beira Central Hospital (HCB). Given the importance of this system to monitor the mortality pattern of HCB and, in general, the city of Beira, this study evaluates the quality of SIS-ROH HCB. It is a descriptive study on the completeness, coverage, compliance and consistency of the SIS-ROH data and examined a sample of 822 HCB deaths Certificates (COs) of fetal and children under 5 years of age. We find the use of two different models of CO (former and current model) for the registration of deaths related to the year 2013. We observed excellent completeness for most SIS-ROH variables. Of the 25 variables of COs there was the following situation: 9 had very bad completeness, which were relating to the identification of the deceased (type of death and age) on the V block in the mother\'s data, where must be filled in case of stillbirths and children under 1 year of age (education, usual occupation, number of living children taken and killed, gestational age) and on the conditions and causes of death (autopsy and intermediate-code causes); 3 variables had bad completeness concerning the identification of the deceased (NID) and on the conditions and causes of death (intermediate cause - description and basic cause - code); 9 showed regular completeness concerning the identification of the deceased (date of birth and age) on the V block (mother\'s age, type of pregnancy, mode of delivery, weight of the fetus / baby birth, death of the fetus / baby compared to delivery) and on the conditions and causes of death (direct cause code, basic cause description); 2 showed good completeness concerning the identification of the deceased (sex and race / color) and, finally, 2 showed excellent completeness concerning the place of occurrence of death (date of admission and date of death or the disappearance corpse). The SIS-ROH had coverage below 50% calculated on mortality data estimated by the National Survey of Causes of Death (INCAM). Some SIS-ROH variables and COS showed inconsistencies. There was a lack of agreement to direct cause between SIS-ROH and COs.

L'évaluation des données de sécurité du médicament reste encore limitée dans les pays en voie de développement. La création du Centre National de Pharmacovigilance du Vietnam en 2009 a permis l'enregistrement des notifications spontanées d'effets indésirables (EIs) permettant une analyse quantitative et qualitative et de générer des éventuels signaux de pharmacovigilance. Les objectifs principaux de la thèse étaient : (i) décrire le système de pharmacovigilance vietnamien, et comparer les données à travers quelques exemples, au système de pharmacovigilance français, (ii) appliquer un ensemble d'approches pharmaco-épidémiologiques pour identifier les risques médicamenteux au Vietnam et (iii) améliorer la sous-notification des EIs. Nous introduisons pour la première fois, un bilan complet du système de pharmacovigilance vietnamien avec une série de suggestions pour les pays partageant le même contexte de ressources limitées. Quelques résultats et défis pour le développement durable du système ont également été discutés. Secondairement, nous avons utilisé les bases de données de pharmacovigilance vietnamienne et française pour comparer le profil de notification pour 2 types d'EIs : anaphylaxie et syndrome de Steven-Johnson et nécrolyse épidermique toxique (SSJ/NET) d'origine médicamenteuse. Nous avons généré les premiers signaux de pharmacovigilance vietnamienne : l'allopurinol, la carbamazépine, les médicaments traditionnels, la colchicine, l'acide valproïque et le méloxicam ont généré des signaux pour le SSJ/NET - déjà connus dans littérature. Par ailleurs, nous avons retrouvé des signaux significatifs pour le cefixime et le paracétamol. Pour l'anaphylaxie, nous avons identifié 4873 (13.2%) cas dans la base vietnamienne pendant la période 2010-2016 avec une tendance à l'augmentation au cours du temps. Les antibiotiques (notamment céphalosporines de 3ème génération) sont les causes principales de l'anaphylaxie médicamenteuse au Vietnam. De plus, des signaux pour certains médicaments comme l'alpha-chymotrypsine, l'amoxicilline/sulbactame et les solutions de glucose ou électrolytes peuvent être typiques pour les pays en voie de développement. Enfin, l'identification des EIs à partir des données de laboratoire de biologie de l'hôpital pourrait améliorer le taux de notification d'EI au Vietnam. [...] / The understanding and quantitative analysis in drug safety domain among developing countries remain still limited. The creation of the National Drug Information and Adverse Drug Reaction Monitoring Centre (NDIADRMC) of Vietnam in 2009 was a landmark for the pharmacovigilance activities in this country and allowed the registration of spontaneous adverse drug reaction (ADR) reports. The accumulation of National Pharmacovigilance Database of Vietnam (NPDV) requires a quantitative and qualitative analysis and generation of pharmacovigilance signals to better protect Vietnamese people's health. The main objectives of the thesis are: (i) describe the Vietnamese pharmacovigilance system, with the comparison to another developed pharmacovigilance system as a reference (France), (ii) apply a set of pharmacoepidemiological approaches to identify the specific drug-related risks, and (iii) improve the under-reporting issue in Vietnam. Firstly, we introduce a full-detailed overview of Vietnamese pharmacovigilance system with a series of lessons learned for the other countries sharing the similar limited-resource context. Some achievements and challenges for the sustainable development of the system were also equitably discussed. Secondly, we used the Vietnamese and French pharmacovigilance databases to compare the differences in characteristics of two types of ADR: anaphylaxis and Steven-Johnson's syndrome and toxic epidermal necrolysis (SJS/TEN) induced by medications. Next, we generated the first Vietnamese pharmacovigilance signals. The signals of drug-induced SJS/TEN were generated with allopurinol, carbamazepine, traditional or herbal drugs, colchicine, valproic acid and meloxicam which were similar to the other previously studies in literature. Furthermore, we also found the significant signals of cefixime and paracetamol. For drug-induced anaphylaxis, we identified 4873 (13.2%) cases in the Vietnamese database during the period 2010-2016 with an increasing trend over time. The antibiotics (especially the third-generation cephalosporins) were the main causes of drug-induced anaphylaxis in Vietnam. In addition, the signals were generated with several drugs such as alpha-chymotrypsin, amoxicillin/sulbactam and glucose or electrolyte solutions which were typical for the resource-restricted countries. On the other hand, the identification of ADR through the screening the laboratory test results could help to increase the ADR reporting rate in Vietnam. [...]

The primary objective of this study is to develop new computational methods for robust design optimization (RDO) and reliability-based design optimization (RBDO) of high-dimensional, complex engineering systems. Four major research directions, all anchored in polynomial dimensional decomposition (PDD), have been defined to meet the objective. They involve: (1) development of new sensitivity analysis methods for RDO and RBDO; (2) development of novel optimization methods for solving RDO problems; (3) development of novel optimization methods for solving RBDO problems; and (4) development of a novel scheme and formulation to solve stochastic design optimization problems with both distributional and structural design parameters. The major achievements are as follows. Firstly, three new computational methods were developed for calculating design sensitivities of statistical moments and reliability of high-dimensional complex systems subject to random inputs. The first method represents a novel integration of PDD of a multivariate stochastic response function and score functions, leading to analytical expressions of design sensitivities of the first two moments. The second and third methods, relevant to probability distribution or reliability analysis, exploit two distinct combinations built on PDD: the PDD-SPA method, entailing the saddlepoint approximation (SPA) and score functions; and the PDD-MCS method, utilizing the embedded Monte Carlo simulation (MCS) of the PDD approximation and score functions. For all three methods developed, both the statistical moments or failure probabilities and their design sensitivities are both determined concurrently from a single stochastic analysis or simulation. Secondly, four new methods were developed for RDO of complex engineering systems. The methods involve PDD of a high-dimensional stochastic response for statistical moment analysis, a novel integration of PDD and score functions for calculating the second-moment sensitivities with respect to the design variables, and standard gradient-based optimization algorithms. The methods, depending on how statistical moment and sensitivity analyses are dovetailed with an optimization algorithm, encompass direct, single-step, sequential, and multi-point single-step design processes. Thirdly, two new methods were developed for RBDO of complex engineering systems. The methods involve an adaptive-sparse polynomial dimensional decomposition (AS-PDD) of a high-dimensional stochastic response for reliability analysis, a novel integration of AS-PDD and score functions for calculating the sensitivities of the failure probability with respect to design variables, and standard gradient-based optimization algorithms, resulting in a multi-point, single-step design process. The two methods, depending on how the failure probability and its design sensitivities are evaluated, exploit two distinct combinations built on AS-PDD: the AS-PDD-SPA method, entailing SPA and score functions; and the AS-PDD-MCS method, utilizing the embedded MCS of the AS-PDD approximation and score functions. In addition, a new method, named as the augmented PDD method, was developed for RDO and RBDO subject to mixed design variables, comprising both distributional and structural design variables. The method comprises a new augmented PDD of a high-dimensional stochastic response for statistical moment and reliability analyses; an integration of the augmented PDD, score functions, and finite-difference approximation for calculating the sensitivities of the first two moments and the failure probability with respect to distributional and structural design variables; and standard gradient-based optimization algorithms, leading to a multi-point, single-step design process. The innovative formulations of statistical moment and reliability analysis, design sensitivity analysis, and optimization algorithms have achieved not only highly accurate but also computationally efficient design solutions. Therefore, these new methods are capable of performing industrial-scale design optimization with numerous design variables.

Vid en brand på byggarbetsplatsen finns det risk för personskador, skador på egendom och att byggproduktionen hamnar i ett driftstopp som kan bli kostsamt. Det finns brister vid planeringen av brandskyddet under byggtid (BUB) och vid olika arbetsmoment på byggarbetsplatsen. Planeringen av BUB behöver förbättras utefter hur en byggarbetsplats, och således brandrisker, förändras under byggproduktionen. Ett förslag på hur detta skulle kunna förbättras är att BUB är med i produktionsplaneringen och redovisas på arbetsplatsdispositionsplaner (APD-planer). Arbetsmoment som kan vara en brandrisk borde också planeras tillsammans med BUB för en brandsäkrare byggarbetsplats. Brandskyddsprojektering är idag till stor del inte inkluderat i det digitala arbetssättet som Building Information Modeling (BIM) innebär. Att brandkonsulten inte är med och arbetar i BIM leder till att de fördelar som finns med samverkan mellan olika discipliner inte kan utnyttjas. Den forskning som finns rörande brandskydd i BIM idag fokuserar på projektering av den färdiga byggnaden och hur brandskydd ska kunna inkluderas i BIM. Det finns idag ingen forskning som berör både brandskydd på byggarbetsplatsen och BIM-projektering av byggarbetsplatsen. Detta leder till ett kunskapsglapp mellan BIM som arbetssätt på byggarbetsplatsen och planering av BUB. På grund av de påträffade problemområdena syftade studien till att undersöka förbättringsmöjligheter för brandskyddet under byggtid. Detta genom att kartlägga förutsättningar och tillämpbara områden där BUB kan inkluderas i arbetet med APD-planer och tidsberoende BIM-modeller. Målet med studien var att framta ett underlag till framtida arbetsmetoder där BUB, APD-planer och BIM på byggarbetsplatser används tillsammans. För att uppfylla syftet utfördes en litteraturstudie och en intervjustudie. Litteraturstudien gjordes för att granska publicerat material rörande BUB, byggarbetsplatsplanering med APD-planer samt BIM på byggarbetsplatsen och BIM med hänsyn till brandskydd. Intervjustudien gjordes för att få en inblick i hur personer använder sig av BUB, APD-planer och BIM på byggarbetsplatser (i praktiken). Den information som framkom från teoretisk kunskap och praktiska erfarenheter analyserades för att kartlägga förutsättningar och tillämpbara områden där BUB kan inkluderas i arbetet med APD-planer och tidsberoende BIM-modeller. Detta för att kunna överbrygga gapet som finns mellan BUB, APD-planer och användandet av BIM på byggarbetsplatser. De identifierade områden som framtagits är planering, kommunikation samt kontroll och uppföljning. För att kunna arbeta med BUB och skedesstyrda APD-planer i BIM inom de identifierade områdena finns det krav på vissa grundläggande förutsättningar. För att kunna ta fram APD-planer och göra dessa skedesstyrda i BIM-modellen krävs det utbildning för att få nödvändig kunskap och kompetens hos både projekteringen och produktionen. Det kommer också krävas ett tidigt samarbete mellan produktionen och projekteringen för att täcka alla kompetensområden som behövs vid planering av produktionen av en byggnad med BIM. För att kunna projektera APD-planer i BIM är det även en förutsättning att alla discipliner arbetar i BIM-modellen. Saknas det information, exempelvis landskap, går det inte att göra fullständiga APD-planer över byggområdet i BIM. Att ta fram skedesstyrda APD-planer där BUB inkluderas i en BIM-modell har i denna studie visat sig vara fördelaktigt inom flera områden. Att planera BUB tillsammans med APD-planer ivgör att brandskyddet på byggarbetsplatsen kan planeras med hänsyn till produktionen och att kraven går att genomföra utan att produktionen blir lidande. Med skedesstyrda APD-planer i en BIM-modell blir det möjligt att visualisera förändringarna på byggarbetsplatsen och på så vis även förutspå vilka brandrisker som uppkommer under olika skeden i byggnationen. BUB inkluderat i en BIM-modell kan vara ett verktyg för att förbättra kommunikationen på byggarbetsplatsen. Detta genom att informationen som läggs in blir mer projektspecifik vilket leder till att BUB blir lättare att förmedla, tydligare att följa och mer förståelig. Med bättre information och ökad kunskap om BUB kan kontroll och uppföljning utföras effektivare och därmed förbättra brandskyddet genom att kontrollerna utförs med en högre kvalité. Detta ger förhoppningen om att brandskydd under byggtid blir en större del av planering och projektering av byggarbetsplatsen samt leder till en säkrare byggarbetsplats. / In case of a fire at a construction site there is a risk of personal injury, property damage and stop in the building production that may be costly. There are shortcomings in the planning of fire safety at the construction site and at different construction methods. The planning of fire safety at construction sites are also in need of improvements according to how a construction site, and thus risks of fire, change during construction production. This by planning the fire safety along with the production of the building and that the fire safety is presented on workplace disposition plans. Fire protection design is mostly not included in the digital work method that Building Information Modeling (BIM) implies. The fact that the fire consultant is not involved with and works in BIM means that the benefits of collaboration between different disciplines cannot be utilized. The research on fire protection in BIM today focuses on the design of the completed building and how fire protection can be included in BIM. There is currently no research regarding both fire safety at the construction site and BIM modelling of the construction site. This leads to a gap between BIM as a working method at the construction site and the planning of fire safety. Because of mentioned problem areas, the study’s purpose was to investigate improvements for fire safety on construction sites. This by mapping out conditions and applicable areas where fire safety can be included in the planning of workplace disposition plans in time-dependent BIM models. The aim of the study was to provide a basis for future work methods in which fire safety, workplace disposition plans and BIM at construction sites are used together. To fulfill the purpose, a literature study and an interview study were conducted. The literature study was conducted to review published material regarding fire safety, construction site planning with workplace disposition plans, BIM at the construction site and BIM with fire safety design. Interviews were conducted to gain an insight into how people use fire safety documents, workplace disposition plans and BIM on construction sites (in practice). The information derived from theoretical knowledge and practical experience was analyzed to map the conditions and applicable areas where fire safety can be included in the planning of workplace disposition plans and time-dependent BIM models. This to bridge the gap between fire safety, workplace disposition plans and the use of BIM on construction sites. The identified areas that have been established are planning, communication and control and follow-up. To be able to work with fire safety and workplace disposition plans in time-dependent BIM model within the identified areas, there are some basic conditions that needs to be fulfilled. To be able to develop workplace disposition plans and make these time-dependent in the BIM model, education is needed to get the necessary knowledge and skills in both the design and production. There also needs to be an early collaboration between production and design to cover all areas of expertise needed in planning the production of a building with BIM. To be able to design workplace disposition plans in BIM, it is also a prerequisite that all disciplines work in the BIM model. If there is no information, such as information of the landscape, it is not possible to complete workplace disposition plans over the whole construction site in BIM. To develop time-dependent workplace disposition plans where fire safety are included in a BIM model, this study has proven to be beneficial in several areas. Planning the fire safety together with workplace disposition plans allows the fire protection at the construction site to be planned viregarding production and that the fire safety requirements can be implemented without the production being compromised. Time-dependent workplace disposition plans in a BIM model make it possible to visualize the changes at the construction site, thus predicting fire risks during different stages of the construction. Fire safety included in a BIM model can be a tool for improving communication at the construction site. This is because the information about the fire safety on construction sites inserted becomes more project-specific, which means that the information about the fire safety requirements can be easier to convey, clearer to follow and more understandable. With better information and increased knowledge about fire safety on construction sites, control and follow-up can be performed more effectively and therefore improving the fire safety by performing the controls with a higher quality. This gives the hope that fire protection during construction will be a major part of planning and designing the construction site and that this will lead to a safer construction site.

Au cours des dernières décennies, les constructeurs automobiles ont constamment introduit des innovations technologiques visant à rendre les véhicules plus sûrs. Le niveau de sophistication de ces systèmes avancés d’aide à la conduite s’est accru parallèlement aux progrès de la technologie des capteurs et de la puissance informatique intégrée. Plus récemment, une grande partie de la recherche effectuée par l'industrie et les institutions s'est concentrée sur l'obtention d'une conduite entièrement automatisée. Les avantages sociétaux potentiels de cette technologie sont nombreux, notamment des routes plus sûres, des flux de trafic améliorés et une mobilité accrue pour les personnes âgées et les handicapés. Toutefois, avant que les véhicules autonomes puissent être commercialisés, ils doivent pouvoir partager la route en toute sécurité avec d’autres véhicules conduits par des conducteurs humains. En d'autres termes, ils doivent pouvoir déduire l'état et les intentions du trafic environnant à partir des données brutes fournies par divers capteurs embarqués, et les utiliser afin de pouvoir prendre les bonnes décisions de conduite sécurisée. Malgré la complexité apparente de cette tâche, les conducteurs humains ont la capacité de prédire correctement l’évolution du trafic environnant dans la plupart des situations. Cette capacité de prédiction est rendu plus simple grâce aux règles imposées par le code de la route qui limitent le nombre d’hypothèses; elle repose aussi sur l’expérience du conducteur en matière d’évaluation et de réduction du risque. L'absence de cette capacité à comprendre naturellement une scène de trafic constitue peut-être, le principal défi qui freine le déploiement à grande échelle de véhicules véritablement autonomes sur les routes.Dans cette thèse, nous abordons les problèmes de modélisation du comportement du conducteur, d'inférence sur le comportement des autres véhicules, et de la prise de décision pour la navigation sûre. En premier lieu, nous modélisons automatiquement le comportement d'un conducteur générique à partir de données de conduite démontrées, évitant ainsi le réglage manuel traditionnel des paramètres du modèle. Ce modèle codant les préférences d’un conducteur par rapport au réseau routier (par exemple, voie ou vitesse préférées) et aux autres usagers de la route (par exemple, distance préférée au véhicule de devant). Deuxièmement, nous décrivons une méthode qui utilise le modèle appris pour prédire la séquence des actions à long terme de tout conducteur dans une scène de trafic. Cette méthode de prédiction suppose que tous les acteurs du trafic se comportent de manière aversive au risque, et donc ne peut pas prévoir les manœuvres dangereux ou les accidents. Pour pouvoir traiter de tels cas, nous proposons un modèle probabiliste plus sophistiqué, qui estime l'état et les intentions du trafic environnant en combinant la prédiction basée sur le modèle avec les preuves dynamiques fournies par les capteurs. Le modèle proposé imite ainsi en quelque sorte le processus de raisonnement des humains. Nous humains, savons ce qu’un véhicule est susceptible de faire compte tenu de la situation (ceci est donné par le modèle), mais nous surveillerons sa dynamique pour en détecter les écarts par rapport au comportement attendu. En pratique, la combinaison de ces deux sources d’informations se traduit par une robustesse accrue des estimations de l’intention par rapport aux approches reposant uniquement sur des preuves dynamiques. En dernière partie, les deux modèles présentés (comportemental et prédictif) sont intégrés dans le cadre d´une approche décisionnel probabiliste. Les méthodes proposées se sont vues évalués avec des données réelles collectées avec un véhicule instrumenté, attestant de leur efficacité dans le cadre de la conduite autonome sur autoroute. Bien que centré sur les autoroutes, ce travail pourrait être facilement adapté pour gérer des scénarios de trafic alternatifs. / During the past few decades automakers have consistently introduced technological innovations aimed to make road vehicles safer. The level of sophistication of these advanced driver assistance systems has increased parallel to developments in sensor technology and embedded computing power. More recently, a lot of the research made both by industry and institutions has concentrated on achieving fully automated driving. The potential societal benefits of this technology are numerous, including safer roads, improved traffic flows, increased mobility for the elderly and the disabled, and optimized human productivity. However, before autonomous vehicles can be commercialized they should be able to safely share the road with human drivers. In other words, they should be capable of inferring the state and intentions of surrounding traffic from the raw data provided by a variety of onboard sensors, and to use this information to make safe navigation decisions. Moreover, in order to truly navigate safely they should also consider potential obstacles not observed by the sensors (such as occluded vehicles or pedestrians). Despite the apparent complexity of the task, humans are extremely good at predicting the development of traffic situations. After all, the actions of any traffic participant are constrained by the road network, by the traffic rules, and by a risk-aversive common sense. The lack of this ability to naturally understand a traffic scene constitutes perhaps the major challenge holding back the large-scale deployment of truly autonomous vehicles in the roads.In this thesis, we address the full pipeline from driver behavior modeling and inference to decision-making for navigation. In the first place, we model the behavior of a generic driver automatically from demonstrated driving data, avoiding thus the traditional hand-tuning of the model parameters. This model encodes the preferences of a driver with respect to the road network (e.g. preferred lane or speed) and also with respect to other road users (e.g. preferred distance to the leading vehicle). Secondly, we describe a method that exploits the learned model to predict the future sequence of actions of any driver in a traffic scene up to the distant future. This model-based prediction method assumes that all traffic participants behave in a risk-aware manner and can therefore fail to predict dangerous maneuvers or accidents. To be able to handle such cases, we propose a more sophisticated probabilistic model that estimates the state and intentions of surrounding traffic by combining the model-based prediction with the dynamic evidence provided by the sensors. In a way, the proposed model mimics the reasoning process of human drivers: we know what a given vehicle is likely to do given the situation (this is given by the model), but we closely monitor its dynamics to detect deviations from the expected behavior. In practice, combining both sources of information results in an increased robustness of the intention estimates in comparison with approaches relying only on dynamic evidence. Finally, the learned driver behavioral model and the prediction model are integrated within a probabilistic decision-making framework. The proposed methods are validated with real-world data collected with an instrumented vehicle. Although focused on highway environments, this work could be easily adapted to handle alternative traffic scenarios.