Avec l'augmentation du nombre de capteurs et d'actuateurs dans les avions et le développement de liaisons de données fiables entre les avions et le sol, il est devenu possible d'améliorer la sécurité et la fiabilité des systèmes à bord en appliquant des techniques d'analyse en temps réel. Cependant, étant donné la disponibilité limité des ressources de calcul embarquées et le coût élevé des liaisons de données, les solutions architecturelles actuelles ne peuvent pas exploiter pleinement toutes les ressources disponibles, limitant leur précision.Notre but est de proposer un algorithme distribué de prédiction de panne qui pourrait être exécuté à la fois à bord de l'avion et dans une station au sol tout en respectant un budget de communication. Dans cette approche, la station au sol disposerait de ressources de calcul rapides et de données historiques et l'avion disposerait de ressources de calcul limitées et des données de vol actuelles.Dans cette thèse, nous étudierons les spécificités des données aéronautiques et les méthodes déjà existantes pour produire des prédictions de pannes à partir de ces dernières et nous proposerons une solution au problème posé. Notre contribution sera détaillé en trois parties.Premièrement, nous étudierons le problème de prédiction d'événements rares créé par la haute fiabilité des systèmes aéronautiques. Beaucoup de méthodes d'apprentissage en classification reposent sur des jeux de données équilibrés. Plusieurs approches existent pour corriger le déséquilibre d'un jeu de donnée et nous étudierons leur efficacité sur des jeux de données extrêmement déséquilibrés.Deuxièmement, nous étudierons le problème d'analyse textuelle de journaux car de nombreux systèmes aéronautiques ne produisent pas d'étiquettes ou de valeurs numériques faciles à interpréter mais des messages de journaux textuels. Nous étudierons les méthodes existantes basées sur une approche statistique et sur l'apprentissage profond pour convertir des messages de journaux textuels en une forme utilisable en entrée d'algorithmes d'apprentissage pour classification. Nous proposerons notre propre méthode basée sur le traitement du langage naturel et montrerons comment ses performances dépassent celles des autres méthodes sur un jeu de donnée public standard.Enfin, nous offrirons une solution au problème posé en proposant un nouvel algorithme d'apprentissage distribué s'appuyant sur deux paradigmes d'apprentissage existant, l'apprentissage actif et l'apprentissage fédéré. Nous détaillerons notre algorithme, son implémentation et fournirons une comparaison de ses performances avec les méthodes existantes / With the rise of the number of sensors and actuators in an aircraft and the development of reliable data links from the aircraft to the ground, it becomes possible to improve aircraft security and maintainability by applying real-time analysis techniques. However, given the limited availability of on-board computing and the high cost of the data links, current architectural solutions cannot fully leverage all the available resources limiting their accuracy.Our goal is to provide a distributed algorithm for failure prediction that could be executed both on-board of the aircraft and on a ground station and that would produce on-board failure predictions in near real-time under a communication budget. In this approach, the ground station would hold fast computation resources and historical data and the aircraft would hold limited computational resources and current flight's data.In this thesis, we will study the specificities of aeronautical data and what methods already exist to produce failure prediction from them and propose a solution to the problem stated. Our contribution will be detailed in three main parts.First, we will study the problem of rare event prediction created by the high reliability of aeronautical systems. Many learning methods for classifiers rely on balanced datasets. Several approaches exist to correct a dataset imbalance and we will study their efficiency on extremely imbalanced datasets.Second, we study the problem of log parsing as many aeronautical systems do not produce easy to classify labels or numerical values but log messages in full text. We will study existing methods based on a statistical approach and on Deep Learning to convert full text log messages into a form usable as an input by learning algorithms for classifiers. We will then propose our own method based on Natural Language Processing and show how it outperforms the other approaches on a public benchmark.Last, we offer a solution to the stated problem by proposing a new distributed learning algorithm that relies on two existing learning paradigms Active Learning and Federated Learning. We detail our algorithm, its implementation and provide a comparison of its performance with existing methods

In today’s research and market, IT applications for health-care are gaining huge interest of both IT and medical researchers. Cardiovascular diseases (CVDs) are considered the largest cause of death for both men and women regardless of ethnic backgrounds. More efficient treatments and most importantly efficient methods of cardiac diagnosis that examine heart diseases are desired. Electrocardiography (ECG) is an essential method used to diagnose heart diseases. However, diagnosing any cardiovascular disease based on the 12-lead ECG printout from an ECG machine using human eye might seriously impair analysis accuracy. To meet this challenge of today’s ECG analysis methodology, a more reliable solution that can analyze huge amount of patient’s data in real-time is desired. The software solution presented in this article is aimed to reduce the risk while diagnosing cardiovascular diseases (CVDs) by human eye, computation of large-scale patient’s data in real-time at the patient’s location and sending the required results or summary to the doctor/nurse. Keeping in mind the importance of real-time analysis of patient’s data, the software system has built upon small individual algorithms/modules designed for multi-core architecture, where each module is supposed to be processed by an individual core/processor in parallel. All the input and output processes to the analysis system are made automated, which reduces operator’s interaction to the system and thus reducing the cost. The outputs/results of the processing are summarized to smaller files in both ASCII and binary formats to meet the requirement of exchanging the data over Voice and Data Networks.

Raman spectroscopy, together with multivariate statistical analyses, has proven to be a near real-time analytical technique capable of phenotyping cells, tissues and organs. This dissertation will show exclusively the application of the Raman spectroscopy phenotypic profiling method to; (i) microbial toxicity, (ii) ex-vivo organ perfusion, and (iii) subcellular location targeting. Real-time analytical methods for monitoring living biological systems will enable study of the physiological changes associated with growth, genetic manipulations, and adverse environmental conditions. Most existing analytical methods (NMR exempt), though highly accurate, must be performed off-line and most require destruction of the studied sample. These attributes make these methodologies less desirable to the study of physiological changes of cells, tissues, and organs. In this work, Raman spectroscopy has been identified and shown to be a good candidate for real-time analysis mainly because it can be performed: (i) in near real-time, (ii) non-destructively and with minimal sample preparation, (iii) through a glass barrier (i.e., can be performed in situ), and (iv) with minimal spectral interference from water. Here, Raman spectroscopy was used in combination with multivariate statistics to analyze the differing toxic effects of 4-C chain alcohols on E. coli. Good correlations were established between Raman spectra and off-line analytical techniques used to measure: (i) saturated, unsaturated, and cyclopropane fatty acids; (ii) amino acid composition of total protein; and (iii) cell membrane fluidity. Also, Raman 'fingerprint' analysis was used to discriminate among different phenotypic responses of cells. In addition, this methodology was applied to analyze perfusates of organs maintained by the VasoWave® organ perfusion system. Raman fingerprints can be used to assess organ health, and it is believed this data can be used to inform decisions such as whether or not to transplant an organ. Finally, molecular biology techniques were used to design and produce specific protein targets harboring a silver binding domain fusion, which upon release migrate to specific subcellular locations. By employing the related technique of surface-enhanced Raman scattering (SERS), which produces a highly amplified Raman signal in the presence of metallic nanoparticle substrates (e.g., silver nanoparticles), different regions of the E. coli cell structure were studied. The target regions studied by the technique included: (i) outer cell membrane, (ii) periplasm, and the (iii) cytoplasm. / Ph. D.

Automobile système embarqué est une architecture distribuée de l'ordinateur des applications basées sur. La prolifération des systèmes embarqués dans une automobile a apporté de nombreux avantages,tels que le remplacement du système mécanique ancienne avec capteur électronique en réseau et desactionneurs, par exemple, dans des applications telles suspensions adaptatives. Le remplacement des systèmes mécaniques avec ceux électroniques et l'intégration de nouvelles fonctionnalités dans l'électronique soulève une grave préoccupation, c'est de fournir des garanties que ces systèmes embarqués seront en mesure d'effectuer, même dans des environnements difficiles, en particulier dans un système critique pour la sécurité comme un automobile. De plus, ceux-ci l'actualité informatique applications à la demande, imposée par un processus physique.Par exemple, pour éviter un événement catastrophique comme un accident de la demande de freinage doit répondre aux contraintes de minutage. Ce qui implique que la durée de temps entre l'instance de l'application du frein (à la pédale de frein) et l'instance de l'actionnement au niveau des roues d'un véhicule automobile doit être inférieure à la limite. En outre, l'application de freinage est généralement répartie sur le nombre de nœuds, qui sont embarqués communicants les uns avec les autres en utilisant une ressource de communication partagée. Par conséquent, il est important que nous fournissons des garanties que la demande, individuellement et collectivement, est atteinte de ses contrainte temporelle; qui est dans la composition de plusieurs nœuds embarqués. En outre, la prolifération des applications informatiques est également livré avec une hétérogénéité croissante et la complexité de l'architecture intégrée, ce qui conduira à l'augmentation de la complexité de l'analyse pour les systèmes automobiles.Par conséquent, il ya un besoin croissant d'assurer que ces systèmes automobiles embarqués répondre à des contraintes temporelles et de fournir des garanties de sécurité au cours de leur fonctionnement normal ou lors de situations critiques. Cette thèse vise à développer les analyses d'ordonnançabilité pour systèmes automobiles et les réseaux intégrés, avec le but de faciliter,d'une manière rentable et fiable, la conception et l'analyse des systèmes embarqués automobiles. Les analyses sont élaborées et appliquées dans le contexte de l'automobile; de ​​façon à réduire le risque d'échec en raison de délai: les limites du matériel; frais généraux de mise en œuvre, et les interférences dues à la circulation probaliste / Automotive embedded system is a distributed architecture of computer-based applications. The proliferation of embedded systems in an automobile has brought numerous benefits; such as replacement of old mechanical system with networked electronic sensor and actuators, for example, in applications like adaptive suspensions. The replacement of mechanical systems with electronic onesand the integration of new functionality in electronics raises a serious concern; that is to provide guarantees that these embedded systems will be able to perform, even in harsh environments, particularly in a safety-critical system like an automobile.Moreover, these computer-based applications demand timeliness, imposed by a physical process. For example, to avoid a catastrophic event like a crash the braking application has to meet thetiming-constraints. This implies that the time duration between the instance of application of the brake (at brake pedal) and the instance of actuation at the wheels of an automobile should be less than the deadline. Moreover, the braking application is usually spread over number of embedded nodes, which are communicating with each other using a shared communication resource. Therefore, it is important that we provide some guarantees that an application, individually and collectively, is meeting its timing constraint; that is in the composition of multiple embedded nodes. Moreover, theproliferation of computer-based applications also comes with an increasing heterogeneity and complexity of the embedded architecture; which lead to the increase in the complexity of the analysis for the automotive systems Therefore, there is an increasing need to ensure that these automotive embedded systems meet temporal constraints and provide safety guarantees during their normal operation or during critical situations. This thesis aims at developing the schedulability analyses for automotive systems and embedded networks; with the aim to facilitate, in a cost-effective and reliable manner, the design and analysis of automotive embedded systems. The analyses are developed and applied in the automotive context; so as to reduce the risk of deadline failure due to: hardware limitations ; implementation overheads; and nterference due to probabilistic traffic

Military helicopter operations require precise maneuverability characteristics for performance to be determined for the entire helicopter flight envelope. Historically, these maneuverability analyses are combinatorial in nature and involve human-interaction, which hinders their integration into conceptual design. A model formulation that includes the necessary quantitative measures and captures the impact of changing requirements real-time is presented. The formulation is shown to offer a more conservative estimate of maneuverability than traditional energy-based formulations through quantitative analysis of a typical pop-up maneuver. Although the control system design is not directly integrated, two control constraint measures are deemed essential in this work: control deflection rate and trajectory divergence rate. Both of these measures are general enough to be applied to any control architecture, while at the same time enable quantitative trades that relate overall vehicle maneuverability to control system requirements. The dimensionality issues stemming from the immense maneuver space are mitigated through systematic development of a maneuver taxonomy that enables the operational envelope to be decomposed into a minimal set of fundamental maneuvers. The taxonomy approach is applied to a helicopter canonical example that requires maneuverability and design to be assessed simultaneously. The end result is a methodology that enables the impact of design choices on maneuverability to be assessed for the entire helicopter operational envelope, while enabling constraints from control system design to be assessed real-time.

Esta pesquisa ocupa-se em investigar, à luz da Teoria da Variação laboviana, o emprego de tepe em lugar de vibrante em onset silábico na fala da comunidade bilíngue (português-italiano) de Flores da Cunha (RS), alternância que não se verifica na fala de monolíngues-português. Utilizaram-se dados de dois bancos, VARSUL e BDSer, levantados de entrevistas realizadas em 1990 e 2009, o que possibilitou fazer um estudo em tempo real. A análise da fala de 40 informantes, 20 de cada banco, revelou um decréscimo na frequência total de aplicação da regra: 41% nos dados do VARSUL, 31% nos do BDSer, indício de regressão da regra. Na análise em tempo real, estudo de tendência, observou-se semelhança no papel favorecedor ou desfavorecedor das variáveis controladas, exceto no que diz respeito à variável Idade. O estudo de painel mostrou que os informantes diminuíram a aplicação da regra com o passar dos anos, o que contraria a expectativa de estabilidade do padrão linguístico da fala adulta. / This is a research about the use of tap where a trill is expected in Brazilian Portuguese, which does not occur in the speech of monolinguals, but can be verified in the speech of Portuguese-Italian bilinguals who live in the city of Flores da Cunha (Rio Grande do Sul, Brazil). The analysis follows the Theory of Language Variation, by William Labov. The data come from two different sources, VARSUL and BDSer data banks, which have sociolinguistic interviews recorded in 1990 and 2009. This fact enabled us to carry out a real time study. The analysis of the speech of 40 informants, 20 from each bank, showed a decrease in the total frequency of rule application: 41% in the speech of VARSUL informants, 31% in the speech of BDSer informants, a sign of rule regression. The real time analysis, tendency study, showed that the role of the variables under control has not changed along 20 years, except the role of the variable Age. The pannel study showed a decrease in rule application by the three informants studied. This goes against the expectation of a stable linguistic behavior in adult speech.

Esta tese objetiva analisar processos de variação linguística em tempo real (LABOV, 1994) no português brasileiro de contato com variedades dialetais italianas, para verificar se existe coerência dialetal (GUY e HINSKENS, 2016) na comunidade de Flores da Cunha (RS). Para tanto, fizemos análise quantitativa conforme metodologia de análise de regra variável (LABOV, 1972), para cotejar o padrão de variação de -ão::-on em tempo real com o de outro processo na mesma comunidade. A pesquisa foi embasada em dados de fala provenientes de dois acervos distintos, o do projeto VARSUL (1990) e do Projeto BDSer (2008/2009). Na „troca‟ de -ão por –on, houve um decréscimo na frequência total de aplicação da regra variável: 30,7% nos dados do VARSUL e 21,6% nos do BDSer, indício de regressão de regra. Para verificar a existência de coerência dialetal, compararam-se os resultados da análise de ão::on com os do estudo de Azeredo (2012), sobre o uso variável de tepe em lugar de vibrante no português de contato com variedades dialetais italianas e observamos que as variantes atribuídas ao contato são, ambas, favorecidas pelos homens de mais idade e com nível mais baixo de escolaridade. Essa correlação entre as variáveis analisadas aponta para a existência de coerência dialetal na comunidade de fala em estudo. / This thesis aims at analyzing linguistic variation processes in real time (LABOV, 1994) in Brazilian Portuguese in contact with Italian dialectal varieties, to verify if there is lethal coherence (GUY and HINSKENS, 2016) in the community of Flores da Cunha (RS). We performed a quantitative analysis according to the methodology of variable rule analysis (LABOV, 1972), to compare the variation pattern of -ão::-on in real time with that of another process in the same community. Our research was based on speech data from two different collections, the VARSUL project (1990) and the BDSer Project (2008/2009). The analysis showed a decrease in the total frequency of use of -on in contexts where -ão is expected: 30.7% in the speech of VARSUL informants and 21.6% in the speech of BDSer informants, a sign of rule regression. We compared the results of the analysis of -ão::-on with the results of Azeredo (2012) about the use of tap where a trill is expected. We observed that older men with low level of schooling favor the variants under analysis. The correlation between the analyzed variables points to the existence of lethal coherence in the speech community under study.

Esta tese objetiva analisar processos de variação linguística em tempo real (LABOV, 1994) no português brasileiro de contato com variedades dialetais italianas, para verificar se existe coerência dialetal (GUY e HINSKENS, 2016) na comunidade de Flores da Cunha (RS). Para tanto, fizemos análise quantitativa conforme metodologia de análise de regra variável (LABOV, 1972), para cotejar o padrão de variação de -ão::-on em tempo real com o de outro processo na mesma comunidade. A pesquisa foi embasada em dados de fala provenientes de dois acervos distintos, o do projeto VARSUL (1990) e do Projeto BDSer (2008/2009). Na „troca‟ de -ão por –on, houve um decréscimo na frequência total de aplicação da regra variável: 30,7% nos dados do VARSUL e 21,6% nos do BDSer, indício de regressão de regra. Para verificar a existência de coerência dialetal, compararam-se os resultados da análise de ão::on com os do estudo de Azeredo (2012), sobre o uso variável de tepe em lugar de vibrante no português de contato com variedades dialetais italianas e observamos que as variantes atribuídas ao contato são, ambas, favorecidas pelos homens de mais idade e com nível mais baixo de escolaridade. Essa correlação entre as variáveis analisadas aponta para a existência de coerência dialetal na comunidade de fala em estudo. / This thesis aims at analyzing linguistic variation processes in real time (LABOV, 1994) in Brazilian Portuguese in contact with Italian dialectal varieties, to verify if there is lethal coherence (GUY and HINSKENS, 2016) in the community of Flores da Cunha (RS). We performed a quantitative analysis according to the methodology of variable rule analysis (LABOV, 1972), to compare the variation pattern of -ão::-on in real time with that of another process in the same community. Our research was based on speech data from two different collections, the VARSUL project (1990) and the BDSer Project (2008/2009). The analysis showed a decrease in the total frequency of use of -on in contexts where -ão is expected: 30.7% in the speech of VARSUL informants and 21.6% in the speech of BDSer informants, a sign of rule regression. We compared the results of the analysis of -ão::-on with the results of Azeredo (2012) about the use of tap where a trill is expected. We observed that older men with low level of schooling favor the variants under analysis. The correlation between the analyzed variables points to the existence of lethal coherence in the speech community under study.

Esta pesquisa ocupa-se em investigar, à luz da Teoria da Variação laboviana, o emprego de tepe em lugar de vibrante em onset silábico na fala da comunidade bilíngue (português-italiano) de Flores da Cunha (RS), alternância que não se verifica na fala de monolíngues-português. Utilizaram-se dados de dois bancos, VARSUL e BDSer, levantados de entrevistas realizadas em 1990 e 2009, o que possibilitou fazer um estudo em tempo real. A análise da fala de 40 informantes, 20 de cada banco, revelou um decréscimo na frequência total de aplicação da regra: 41% nos dados do VARSUL, 31% nos do BDSer, indício de regressão da regra. Na análise em tempo real, estudo de tendência, observou-se semelhança no papel favorecedor ou desfavorecedor das variáveis controladas, exceto no que diz respeito à variável Idade. O estudo de painel mostrou que os informantes diminuíram a aplicação da regra com o passar dos anos, o que contraria a expectativa de estabilidade do padrão linguístico da fala adulta. / This is a research about the use of tap where a trill is expected in Brazilian Portuguese, which does not occur in the speech of monolinguals, but can be verified in the speech of Portuguese-Italian bilinguals who live in the city of Flores da Cunha (Rio Grande do Sul, Brazil). The analysis follows the Theory of Language Variation, by William Labov. The data come from two different sources, VARSUL and BDSer data banks, which have sociolinguistic interviews recorded in 1990 and 2009. This fact enabled us to carry out a real time study. The analysis of the speech of 40 informants, 20 from each bank, showed a decrease in the total frequency of rule application: 41% in the speech of VARSUL informants, 31% in the speech of BDSer informants, a sign of rule regression. The real time analysis, tendency study, showed that the role of the variables under control has not changed along 20 years, except the role of the variable Age. The pannel study showed a decrease in rule application by the three informants studied. This goes against the expectation of a stable linguistic behavior in adult speech.

Esta pesquisa ocupa-se em investigar, à luz da Teoria da Variação laboviana, o emprego de tepe em lugar de vibrante em onset silábico na fala da comunidade bilíngue (português-italiano) de Flores da Cunha (RS), alternância que não se verifica na fala de monolíngues-português. Utilizaram-se dados de dois bancos, VARSUL e BDSer, levantados de entrevistas realizadas em 1990 e 2009, o que possibilitou fazer um estudo em tempo real. A análise da fala de 40 informantes, 20 de cada banco, revelou um decréscimo na frequência total de aplicação da regra: 41% nos dados do VARSUL, 31% nos do BDSer, indício de regressão da regra. Na análise em tempo real, estudo de tendência, observou-se semelhança no papel favorecedor ou desfavorecedor das variáveis controladas, exceto no que diz respeito à variável Idade. O estudo de painel mostrou que os informantes diminuíram a aplicação da regra com o passar dos anos, o que contraria a expectativa de estabilidade do padrão linguístico da fala adulta. / This is a research about the use of tap where a trill is expected in Brazilian Portuguese, which does not occur in the speech of monolinguals, but can be verified in the speech of Portuguese-Italian bilinguals who live in the city of Flores da Cunha (Rio Grande do Sul, Brazil). The analysis follows the Theory of Language Variation, by William Labov. The data come from two different sources, VARSUL and BDSer data banks, which have sociolinguistic interviews recorded in 1990 and 2009. This fact enabled us to carry out a real time study. The analysis of the speech of 40 informants, 20 from each bank, showed a decrease in the total frequency of rule application: 41% in the speech of VARSUL informants, 31% in the speech of BDSer informants, a sign of rule regression. The real time analysis, tendency study, showed that the role of the variables under control has not changed along 20 years, except the role of the variable Age. The pannel study showed a decrease in rule application by the three informants studied. This goes against the expectation of a stable linguistic behavior in adult speech.