High mass loss rates in evolved stars make them the major contributors to recycling processed material back into the interstellar medium. This mass loss creates large circumstellar shells, rich in molecular material. This dissertation presents millimeter and submillimeter studies of the end stages of low mass and high mass stars in order to probe their molecular content in more detail. In low mass stars, the molecular material is carried on into the planetary nebula (PN) stage. Observations of CS, HCO⁺, and CO in planetary nebulae (PNe) of various post-asymptotic giant branch ages have shown that molecular abundances in these objects do not significantly vary with age, as previously thought. More detailed observations of the slightly oxygen-rich PN NGC 6537 resulted in the detection of CN, HCN, HNC, CCH, CS, SO, H₂CO, HCO⁺ and N₂H⁺, as well as numerous ¹³C isotopologues. Observations of the middle-aged PN M2-48 showed the presence of CN, HCN, HNC, CS, SO, SO₂, SiO, HCO⁺, N₂H⁺, and several ¹³C isotopologues. These observations represent the first detections of CS, SO, SO₂, and SiO in any planetary nebula. The implications of these observations are discussed. A 1 mm spectral survey of the supergiant star NML Cygni has been carried out with the Arizona Radio Observatory Submillimeter Telescope resulting in the observation of 102 emission features arising from 17 different molecules and 4 unidentified features. The line profiles observed in this circumstellar shell are asymmetric and vary between different molecules, akin to what has been seen in another supergiant, VY Canis Majoris. The non-LTE radiative transfer code ESCAPADE has been used to model molecular abundances in the various asymmetric outflows of VY Canis Majoris, showing just how chemically and kinematically complex these supergiant circumstellar envelopes really are.

The goal of Evolutionary Robotics (ER) is the development of automatic processes for the synthesis of robot control systems using evolutionary computation. The idea that it may be possible to synthesise robotic control systems using an automatic design process is appealing. However, ER is considerably more challenging and less automatic than its advocates would suggest. ER applies methods from the field of neuroevolution to evolve robot control systems. Neuroevolution is a machine learning algorithm that applies evolutionary computation to the design of Artificial Neural Networks (ANN). The aim of this thesis is to assay the practical characteristics of neuroevolution by performing bulk experiments on a set of Reinforcement Learning (RL) problems. This thesis was conducted with the view of applying neuroevolution to the design of neurocontrollers for small low-cost Autonomous Underwater Vehicles (AUV). A general approach to neuroevolution for RL problems is presented. The is selected to evolve ANN connection weights on the basis that it has shown competitive performance on continuous optimisation problems, is self-adaptive and can exploit dependencies between connection weights. Practical implementation issues are identified and discussed. A series of experiments are conducted on RL problems. These problems are representative of problems from the AUV domain, but manageable in terms of problem complexity and computational resources required. Results from these experiments are analysed to draw out practical characteristics of neuroevolution. Bulk experiments are conducted using the inverted pendulum problem. This popular control benchmark is inherently unstable, underactuated and non-linear: characteristics common to underwater vehicles. Two practical characteristics of neuroevolution are demonstrated: the importance of using randomly generated evaluation sets and the effect of evaluation noise on search performance. As part of these experiments, deficiencies in the benchmark are identified and modifications suggested. The problem of an underwater vehicle travelling to a goal in an obstacle free environment is studied. The vehicle is modelled as a Dubins car, which is a simplified model of the high-level kinematics of a torpedo class underwater vehicle. Two practical characteristics of neuroevolution are demonstrated: the importance of domain knowledge when formulating ANN inputs and how the fitness function defines the set of evolvable control policies. Paths generated by the evolved neurocontrollers are compared with known optimal solutions. A framework is presented to guide the practical application of neuroevolution to RL problems that covers a range of issues identified during the experiments conducted in this thesis. An assessment of neuroevolution concludes that it is far from automatic yet still has potential as a technique for solving reinforcement problems, although further research is required to better understand the process of evolutionary learning. The major contribution made by this thesis is a rigorous empirical study of the practical characteristics of neuroevolution as applied to RL problems. A critical, yet constructive, viewpoint is taken of neuroevolution. This viewpoint differs from much of the reseach undertaken in this field, which is often unjustifiably optimistic and tends to gloss over difficult practical issues.

Submitted by Marlene Santos (marlene.bc.ufg@gmail.com) on 2017-10-03T17:52:20Z No. of bitstreams: 2 Tese - Fabiana Gonçalves dos Reis - 2017.pdf: 5970849 bytes, checksum: e7504d09f87c5499323f6d7c66e10a0f (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Approved for entry into archive by Luciana Ferreira (lucgeral@gmail.com) on 2017-10-04T12:07:46Z (GMT) No. of bitstreams: 2 Tese - Fabiana Gonçalves dos Reis - 2017.pdf: 5970849 bytes, checksum: e7504d09f87c5499323f6d7c66e10a0f (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Made available in DSpace on 2017-10-04T12:07:46Z (GMT). No. of bitstreams: 2 Tese - Fabiana Gonçalves dos Reis - 2017.pdf: 5970849 bytes, checksum: e7504d09f87c5499323f6d7c66e10a0f (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Previous issue date: 2017-08-09 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES / Neurological disorders are a group of conditions that manifest early in the development of the child, so that delayed neuropsychomotor development (ADNPM) and intellectual disability (ID) can impair cognitive, language, motor and social behavior. The etiology of ID is quite heterogeneous and prenatal, perinatal and postnatal factors are associated with an increased risk of this deficiency. However, 30 to 50% of the cases remain with the unknown etiology. In this context, the main objective of this study was to identify submicroscopic genomic alterations (<5Mb) by means of microarray chromosomal analysis (CMA) in patients with clinical indication of ADNPM and/or ID, sent by attending physicians of the state public health network of Goiás. We analyzed 149 cases of both sexes. Of the total number of patients, 47 had the diagnosis clarified using cytogenetics by G banding. Of 102 patients with an incomplete diagnosis, 72 agreed to participate in the present study and, therefore, performed the CMA. The elucidation of the diagnosis by CMA was possible in 22 patients. Among the results obtained, three rare cases were selected to compose this thesis. The first case is from a patient in whom a de novo microduplication of 0.45 Mb in the 5q35.2q35.3 region containing the NSD1 gene was identified. The effect of the dosing of this gene has been related to Sotos Syndrome and its inverted phenotype. The second case shows the molecular detection of an absent allele on the X chromosome and the presence of 28 CGG repeats in FMR1 gene in the present allele. The CMA showed that the patient had a de novo microdeletion of 4.176 Mb in the Xq27.3-q28 region that affected 34 genes, including five genes (TMEM185A, TMEM257, FMR1, IDS, and FMR2) that were directly correlated with ID phenotypes and neurological disorders. The third case is a de novo microdeletion of 1.59 Mb in the 1p32.3 region involving the DHCR24 gene, which causes a gene dosage effect influencing the activation of enzymes that cause desmosterolosis, which is a desmosterol conversion disorder in cholesterol. Thus, the results of this thesis showed the relevance of the use of the CMA technology to diagnose patients with clinical signs of ADNPM and/or ID that presented karyotype without alterations, evidencing the importance of this technology for public health. / Os distúrbios neurológicos constituem um grupo de condições que se manifestam precocemente durante o desenvolvimento da criança, de forma que o atraso do desenvolvimento neuropsicomotor (ADNPM) e a deficiência intelectual (DI) podem acarretar prejuízo às funções cognitivas, de linguagem, motricidade e comportamento social. A etiologia da DI é bastante heterogênea e fatores pré-natais, perinatais e pós-natais estão associados ao aumento do risco dessa deficiência. No entanto, 30 a 50% dos casos permanecem com a etiologia desconhecida. Neste contexto, o objetivo principal deste estudo foi identificar alterações genômicas submicroscópicas (<5Mb) por meio da análise cromossômica por microarranjos (CMA) em pacientes com indicação clínica de ADNPM e/ou DI, encaminhados por médicos assistentes da rede pública de saúde do Estado de Goiás. Foram analisados 149 casos, de ambos os sexos. Do total de pacientes, 47 tiveram o diagnóstico esclarecido utilizando-se a citogenética por bandeamento G. Dos 102 com diagnóstico não concluído, 72 concordaram em participar da presente pesquisa e, portanto, realizaram o CMA. A elucidação do diagnóstico pelo CMA foi possível em 22 pacientes. Dentre os resultados obtidos, foram selecionados três casos raros para compor esta tese. O primeiro caso é de um paciente em que foi identificada uma microduplicação de novo de 0,45 Mb na região 5q35.2q35.3, contendo o gene NSD1. O efeito da dosagem deste gene tem sido relacionado à síndrome de Sotos e ao seu fenótipo invertido. O segundo caso apresenta a detecção molecular de um alelo ausente no cromossomo X e a presença de 28 repetições CGG no gene FMR1 no alelo presente. O CMA mostrou que a paciente tem uma microdeleção de novo de 4,176 Mb na região Xq27.3-q28 que afetou 34 genes, dentre estes, cinco genes (TMEM185A, TMEM257, FMR1, IDS, and FMR2) que foram correlacionados diretamente com os fenótipos de DI e distúrbios neurológicos. O terceiro caso é de uma microdeleção de novo de 1,59 Mb na região 1p32.3 envolvendo o gene DHCR24, que acarreta um efeito de dosagem gênica influenciando na ativação de enzimas que causam a desmosterolose, que é um transtorno de conversão do desmosterol em colesterol. Assim, os resultados desta tese mostraram a relevância da utilização da tecnologia do CMA para diagnosticar pacientes com indicação clínica de ADNPM e/ou DI que apresentaram cariótipo sem alterações, evidenciando a importância desta tecnologia para a saúde pública.

SmartWatches are watches with the property that they can connect to internet.The interest and market for smartwatches has increased remarkably over the pastfew years. However, most of today's solutions cover only the frequency bandsBluetooth (BT) and Global Positioning System(GPS). A desired property of thenext generation of smartwatches is to include additional frequency bands suchas Global System for Mobile Communication(GSM) and Third Generation (3G)cellular bands, so that the Smart Watch can become independent of a smartphone.This sets new challenges for antenna design in small regions. The limited sizetogether with the increased bandwidth demand causes the major problem.The purpose of this thesis is to investigate the antenna concept for wrist wear-able devices that has earlier been developed by the RF team at Sunway Commu-nication AB in Stockholm. The main objective of this project is to study howthis antenna concept can be changed to achieve a given set of requirements onfrequency, bandwidth and eciency. Characteristic Mode Analysis and antennaquality factor calculated from the input impedance, are tools used to help under-stand and improve the characteristics of the antenna.A decomposition method and a parametric study were initially applied to theplanar version of the antenna concept. We investigated then the inuence of theposition of the feeds and the slot that separates the two antenna elements. Anoptimised design of the planar antenna was achieved based on the results fromthose studies. Four prototypes of this antenna were built and measured in the RFlaboratory in Kista.The possibility of miniaturization was studied through simulations by analysingthe eect of using a higher dielectric constant material in the antenna. Two dif-ferent arrangements were investigated. The eect of bending the antenna to t ona human wrist was also investigated, where a frequency shift was observed in theresults. The change in bandwidth of the dierent modes is related to the changein the stored energy of the antenna caused by the bending.We made a prototype of the planar antenna that shows a good agreement withsimulation results and fulls the basic requirements on frequency, bandwidth andeciency. We also found that the addition of a shorting pin showed a bandwidthimprovement of the high-band port and we showed through simulation that it ispossible to reduce the size of the antenna by adding a layer of PCABS material.

This thesis describes the application of adaptive equalization based on the constant modulus algorithm (CMA) and the decision-feedback minimum mean squared error (DF-MMSE) concept to the two compatible offset QPSK waveforms (FQPSK and SOQPSK-TG) that constitute the ARTM Tier-1 waveforms. An adaptive version of the DF-MMSE equalizer is developed and applied to this application. In the presence of frequency selective multipath interference typically encountered in aeronautical telemetry, both equalization techniques are shown to provide reliable performance for FQPSK and SOQPSK-TG. The performance of both waveforms with the DF-MMSE equalizer is slightly better than that using the CMA equalizer. Implementation trade-offs between the two types of equalizers are discussed.

<p> This paper analyzes the 1971-1976 outmigration pattern of the Canadian female elderly from the 23 Census Metropolitan Area's (CMA's). The migration is conceptualized as a three-level choice process and is represented by the logit model. The major findings are as follows: (1) Elderly females are substantially less migratory than young females but are slightly more mobile than elderly males. In addition,, the elderly females in the Western region tend to be more mobile than those in other regions. (2) Elderly females show less preference for the metro politan destinati ons than the young population; however, among the elderly, females have a stronger preference for metropolitan areas than males. (3) In general, the metropolita~d outmiS!'ants' destination choice pattern is less dispersed for the elderly than for the young. Among the older persons, female migrants have a larger dispersion than male migrants in most CMA's. (4) With respect to metropolitanward elderly migrants from the CMA's, the probability of choosing a particular destination is positively related to population size, brightness , and housing growth, and negatively related to the logarithm of di stance, cultural dissimilarity, coldness, and gross rent. (5)Environmental variables are more important than the housing variables in determining the destination choice pattern of the elderly mi grants. </p> / Thesis / Bachelor of Arts (BA)

Grâce aux progrès réalisés ces dernières années dans la conception de réseaux intelligents, tels que les réseaux centrés sur la personne (WBAN) ou les réseaux sans fils de proximité (WPAN), de nouveaux types d'applications émergent et utilisent des capteurs d'informations capables de relever les paramètres physiologiques, environnementaux et plus particulièrement le positionnement des personnes. Dans ce cadre, nous nous sommes attachés dans ce travail de recherche à la conception et la réalisation d'antennes en polarisation circulaire pouvant être intégrées dans des vêtements, pour les standards de géolocalisation européen Galiléo et Américain GPS. Nous avons utilisé pour ces antennes des structures métamatériaux de type conducteurs magnétiques artificiels, afin d'augmenter les performances en rayonnement et pour diminuer le couplage avec et le corps humain. Une autre voie explorée est l'utilisation d'antennes patchs qui sont de nature faible encombrement. La bande de fréquence du standard de communications par satellite Iridium étant très proche du standard de géolocalisation GPS, nous avons trouvé intéressant de développer une solution de type patch couvrant les deux bandes GPS (1,575 GHz) et Iridium (1,621 GHz). L'antenne devant être intégrée sur une boite crânienne, les niveaux de débit d'absorption spécifique et les modifications du rayonnement sous conformation de l'antenne ont également été étudiés. L'évolution de ce travail a été ensuite d'étudier le comportement de cette antenne posée sur le dessus d'un casque militaire français. Les performances en rayonnement ont été satisfaisantes et ont montré la possibilité d'une telle application.

La langue de travail s’avère être un des principaux indicateurs de la langue d’usage publique et de l’intégration linguistique des immigrants. Au Québec, et plus particulièrement dans la RMR de Montréal, les pressions démographiques ainsi que celles posées par le statut de la langue anglaise, la mondialisation et l’utilisation de la technologie entraînent une ambiguïté quant à la nécessité de l’utilisation du français au travail. Or, cette étude porte sur la langue de travail de la population immigrante qui travaille dans la RMR de Montréal. Les analyses effectuées permettent de dresser un portrait de la langue de travail des immigrants en 2006, avec, par la suite, une mise en évidence des variables associées aux choix linguistiques des immigrants sur le marché du travail et ce, à l’aide de données tirées du recensement canadien de 2006. Les analyses descriptives réalisées démontrent que la scolarité, le pays de provenance, le contexte social québécois au moment de l’immigration et les variables linguistiques des immigrants sont les variables les plus fortement liées à l’utilisation du français, de l’anglais ou du bilinguisme au travail en 2006. Ces mêmes variables, dans les analyses explicatives, se révèlent également être les facteurs les plus fortement associés aux choix linguistiques des immigrants sur le marché du travail de la RMR de Montréal. Enfin, une étude comme celle-ci permet de constater la situation de la langue de travail chez les immigrants et, si nécessaire, de suggérer des changements aux politiques et règlements encadrant l’immigration et l’intégration linguistiques des immigrants. / Language of work is proven to be an important variable in order to explain how immigrants integrate to a society and in which language they communicate in public. In the province of Québec and especially in the Montreal CMA, the status of French as a widespread language of work remains unclear. This situation is driven by factors such as globalization, increasing use of technology and demographic trends and figures that help to strengthen the use of English in the public sphere. Therefore, this study will focus on examining Montreal CMA immigrants’ language of work. Some analysis will first provide an overview of immigrants’ language of work in 2006. Then, an emphasis will be put on the key variables explaining immigrants’ linguistic choices on the labour market. Those analyses will be performed using the 2006 Canadian census data. Among the main findings of the explanatory analysis, it should be noted that immigrants’ education and linguistic profile, their country of origin and period of immigration are the variables most strongly associated with the use of French, English or bilingualism at work in 2006. The descriptive analysis showed similar conclusions, reinforcing the major influence of the previously mentioned variables. Finally, this study is useful in assessing the situation of immigrants’ language of work and in suggesting changes to policies regulating immigration trends and immigrants’ linguistic integration.

Many algorithms today require a number of parameters to be set in order to perform well in a given application. The tuning of these parameters is often difficult and tedious to do manually, especially when the number of parameters is large. It is also unlikely that a human can find the best possible solution for difficult problems. To be able to automatically find good sets of parameters could both provide better results and save a lot of time. The prominent methods Bayesian optimization and Covariance Matrix Adaptation Evolution Strategy (CMA-ES) are evaluated for automatic parameter tuning in localization algorithms in this work. Both methods are evaluated using a localization algorithm on different datasets and compared in terms of computational time and the precision and recall of the final solutions. This study shows that it is feasible to automatically tune the parameters of localization algorithms using the evaluated methods. In all experiments performed in this work, Bayesian optimization was shown to make the biggest improvements early in the optimization but CMA-ES always passed it and proceeded to reach the best final solutions after some time. This study also shows that automatic parameter tuning is feasible even when using noisy real-world data collected from 3D cameras.

Les alliages métalliques complexes (CMAs) sont des composés intermétalliques dont la structure cristallographique diffère de celle des alliages conventionnels par le nombre conséquent d'atomes dans la maille (jusqu'à plusieurs milliers d'atomes), généralement arrangés sous forme d'agrégats atomiques de haute symétrie. Ils sont prometteurs pour un certain nombre d'applications technologiques, en particulier les revêtements fonctionnels, en raison de leurs propriétés de surface uniques. Cette thèse a pour objectif, à la fois la détermination de la structure et des propriétés électroniques d’une surface d’un CMA de la famille des clathrates intermétalliques, et des propriétés de mouillage intrinsèques de plusieurs CMAs à base d’aluminium. Dans une première partie, nous nous sommes intéressés aux surfaces de bas indice (100) et (110) du clathrate Ba8Au5.25Ge40.75. Leurs structures atomiques et électroniques ont été déterminées en combinant des expériences de sciences des surfaces et des calculs basés sur la théorie de la fonctionnelle de la densité. La structure tridimensionnelle de Ba8Au5.25Ge40.75, formée d'un réseau de deux types de cages (structure hôte) à base de germanium et d’or, qui emprisonnent les atomes de Ba, induit une nanostructuration de la surface contrôlée par son orientation, puisque le type de cages préservées à la surface diffère pour les surfaces (100) et (110). Dans les deux cas, les atomes de Ba qui protrudent à la surface, ont un rôle primordial pour la stabilité de surface : ils assurent un transfert de charge qui sature les liaisons pendantes des atomes de germanium en surface. Dans une seconde partie, les propriétés intrinsèques de mouillage de plusieurs CMAs à base d’aluminium, ont été déterminées par une approche couplant des mesures de microscopie et des calculs ab initio. Expérimentalement, les angles de contact de gouttes de plomb (métal sonde) sur plusieurs surfaces de CMAs ont été systématiquement mesurés. Les angles précédents étant fonction, entre autres, de l’énergie interfaciale, des calculs d'énergie interfaciale ont été menés, d’une part avec un substrat d’un métal simple, Al(111), et d’autre part sur un substrat de CMA, Al13Co4(100). Les résultats obtenus mettent en évidence une forte influence de la structure de l’interface sur l’énergie interfaciale / Complex metallic alloys (CMAs) are intermetallic compounds possessing a large unit cell containing several tens to hundreds of atoms. Their structure can be described alternatively by the packing of highly symmetric atomic clusters. Clathrate (or cage) compounds are a new class of CMAs having a crystal structure described by a complex arrangement of covalently-bonded cages. The Ba8Au5.25Ge40.75 type-I clathrate is one such cage compound, whose bulk properties have been (and still are) extensively explored for thermoelectric applications. In fact, it is possible to tune the compound electronic structure by a fine control of its bulk composition. Regarding the properties of the Ba8Au5.25Ge40.75 surface, information remains scarce if not inexistent. However, it is known that the surfaces of CMAs often exhibit interesting surface properties. To this end, we have studied two low-index surfaces: BaAuGe(100) and BaAuGe(110) by a combination of experimental (XPS; LEED; STM) and computational (DFT) methods. Experimental results show no evidence for surface segregation and LEED patterns are consistent with (1x1) bulk terminations with no surface reconstruction. The interplay between the 3D nano-caged structure and 2D surfaces is investigated. We demonstrate that the surface structures of the two surfaces considered preserve the bulk structure cages in addition to an ordered arrangement of surface Ba atoms. The two surfaces are formed by a breakage of highly directional covalent bonds present within the framework, hence leading to destabilizing dangling bonds. Ab initio calculations show that the surface structure is stabilized through electron charge transfer from protruding Ba to surface Ge and Au atoms, saturating the dangling bonds. This charge-balance mechanism lifts the possible surface reconstruction envisaged. We reveal how the surface nanostructuration is surface orientation dependent. The results indicate that the surface electronic structure of BaAuGe(110) is impacted by the Au surface concentration. The surface models for BaAuGe(100) and BaAuGe(110) present a metallic character and low work function values, useful for further applications. Such structurally complex surfaces may also be used as templates for novel nanoscale architectures. Further in this work, we also applied the state-of-the-art surface science techniques to investigate the wetting properties of Al-based CMAs. In these experiments, chemically inert Pb element was used as a metal probe. Systematic analysis is done to find the correlation between the wetting properties and the electronic structure properties of these CMAs. Interfacial energy calculations have been performed to model the Pb/CMA interface based on few approaches reported in literature. We have tested these approaches on a moiré patterned Pb(111)/Al(111) interface. This interface is found to be controlled by geometric factors. Hence, an acquired understanding was applied to Pb deposited on Al13Co4(100) (Al-rich side) interface