With the advent of ubiquitous computing, interaction design has broadened its object of inquiry into how smart computational artifacts inconspicuously act in  people’s everyday lives. Although user-centered design approaches remains useful for exploring how people cope with interactive systems, they cannot explain how this new breed of artifact participates in people’s sociality. User-centered design approach assumes that humans control interactive systems, but neglects any possibility of the agency of smart artifacts.Following Actor-network Theory, this research recognizes that artifacts and humans share the capacity of influencing society and meshing with each other, constituting hybrid social actors. From that standpoint, the research offers a triadic structure of networked social interaction as a methodological basis to investigate how smart devices perceive their social setting and adaptively mediate people’s interactions within activities. These triadic units of analysis account for the interactions within and between humannonhuman collectives in the actor-network. The within interactions are those that hold together humans and smart artifacts inside a collective and put forward the collective’s assembled meaning for other actors in the network.The between interactions are those that occur among collectives and characterize the dominant relational model of the actor-network. This triadic approach was modeled and used to analyze the interactions of participants in three empirical studies of social activities with communal goals, each mediated by a smart artifact that enacted – signified – a balanced distribution of obligations and privileges among subjects. Overall, the studies found that actor-networks exhibit a social viscosity that hinders people’s interactions. This is because when people try to collectively accomplish goals, they offer resistance to one another. The studies also show that the intervention of smart artifacts can facilitate the achievement of cooperative and collaborative interaction between actors when the artifacts enact the dominant moral principles which prompt the preservation of social balance, enhance the network’s information integrity, and are located at the focus of activity. The articulation of Actor-Network Theory principles with interaction design methods opens up the traditional user-artifact dyad towards triadic collective enactment by embracing diverse kinds of participants and practices, thus facilitating the design of enhanced sociality. The theoretical conceptual framework, methods and tools developed in this research are relevant to product and interaction designers, ubiquitous computing and Science and technology studies (STS) researchers; as well as practitioners interested in making the interaction between humans and computers smarter, more intuitive and more enjoyable. The outcome of this research is addressed to designers, engineers and urban planners who are responsible for the design of the new generation of computational artifacts that will cohabitate today’s transportation terminals and vehicles, offices, parks, streets, among other social spaces of the constructed world.

Wind power is one of the key players in the European Union target of 20% share of energy from renewable sources in 2020 to reduce greenhouse gas emissions and dependence on energy supply. In order to accomplish such target, some European countries are placing the wind farms further from the coast to exploit the higher wind speeds at open sea. However the traditional transmission system for offshore wind farms based on High Voltage Alternate Current (HVAC) is not appropriate for long distances. Hence, this research work was carried out to examine and determine the most suitable transmission system for offshore wind farms in terms of technical, environmental and economic costs. The methodological approach employed in the investigations was arranged in two parts. First, a quantitative and qualitative method of literature review involving critical examination of documented information and experiences of the two commercially available transmission systems: HVAC and HVDC (High Voltage Direct Current); second, a quantitative method that investigates and analyzes an adequate transmission system for a propose an offshore wind cluster in terms of economic and environmental costs. From the analysis, HVDC based on Voltage Source Converters (VSC) exhibited technical superiority over HVAC for long distance submarine transmission, control of active and reactive power, grid support and connection of asynchronous grids. Furthermore, in the proposed offshore wind farm cluster of 1 GW of installed capacity located at 150 km from the point of connection, after a distance of 90 km HVDC becomes more economic due to reduced system losses, and after a distance of 125 km HVDC presents lower environmental impacts associated with the raw materials used in the transmission system. At such point, particular to each transmission project, VSC-HVDC surpasses HVAC in technical, environmental and economic terms.

The main challenge of this Thesis is to reliably preserve quality of service (QoS) contracts in component-based software systems under changing conditions of system execution. In response to this challenge, the presented contribution is twofold. The first is a model for component-based software applications, QoS contracts and reconfiguration rules as typed attributed graphs, and the definition of QoS-contracts semantics as state machines in which transitions are performed as software reconfigurations. Thus, we effectively use (formal) models at runtime to reliably reconfigure software applications for preserving its QoS contracts. More specifically, we show the feasibility of exploiting design patterns at runtime in reconfiguration loops to fulfill expected QoS levels associated to specific context conditions. We realize this formal model through a component-based architecture and implementation that can be used as an additional layer of SCA middleware stacks to preserve the QoS contracts of executed applications.  The second contribution is the characterization of adaptation properties to evaluate self-adaptive software systems in a standardized and comparable way. By its own nature, the adaptation mechanisms of self-adaptive software systems are essentially feedback loops as defined in control theory. Thus, it results reasonable to evaluate them using the standard properties used to evaluate feedback loops, re-interpreting these properties for the software domain. We define the relibility of our formal model realization in terms of a subset of the characterized adaptation properties, and we show that these properties are guaranteed in this realization.

This research is part of the project entitled ""CANOPOR"". Overall, this project aims to carry out studies of forest canopy dynamics through Airborne LiDAR scanning. From this perspective, this study focus on developing a methodology for segmentation automatic or semi-automatic tree crowns in tropical rain forest. The study area corresponds to Paracou experimental in French Guiana. On this scenario, a combination of combination of high resolution optical images and LiDAR scanners is used. The present document provides full details about the pre-treatment process, images treatment, results and comparisons with different segmentation approaches. The validation method is also discussed. Results revealed that the use of LiDAR and variables such as elevation and intensity are a valuable source of information for understanding the structure of complex ecosystems in tropical forests. 

Abstract: The success of machine learning algorithms depends on the representation of the data used. Specific domain knowledge can be used to design good representations. However, these representations are limited to a specific problem or task, and to the amount of available labeled data. Another approach is to automatically learn generic priors that can be used in different tasks and context. In the field of natural language processing, recent work has been done in obtaining such priors by learning useful vector representation of words from unlabeled data. The representations can then be used to improve existing natural language processing systems. These word vectors are obtained using special neural network architectures trained on billions of tokens. However, most of these models are learned and evaluated on English language corpora. In this work, Word2vec, a recent neural network based toolkit for learning word representations is used on German language data. The goal is to evaluate the learned representations of words in different language processing and information retrieval tasks. In particular, a semantic-syntactic evaluation set is constructed for the German language. In addition to that, the learned word vector representations are used as features for a classifier of German language business documents. The learned features outperformed existing handcrafted features and performed similar to other state-of-the-art approaches. Resumen: El éxito de los algoritmos de aprendizaje automático depende de la representación de los datos utilizados. conocimiento de dominio específico se puede utilizar para diseñar buenas representaciones. SIN EMBARGO, representaciones de síntesis se limitan a un problema o tarea específica, y que la cantidad de datos disponibles etiquetados. Otro enfoque es aprender automáticamente priores genéricos hizo kann utilizado en diferentes tareas y el contexto. En el campo del procesamiento del lenguaje natural, los trabajos recientes se ha hecho en la obtención de distribuciones previas examinados por el aprendizaje de la representación vector útil de las palabras de datos no etiquetados. Las representaciones a continuación, se pueden utilizar para mejorar el sistema de procesamiento de lenguaje natural existente. Estos vectores de palabras se obtienen utilizando arquitecturas de redes neuronales especiales entrenados en miles de millones de fichas. Sin embargo la mayoría de los modelos de síntesis se aprenden y se evalúa el corpus idioma Inglés. En este trabajo, Word2vec, un conjunto de herramientas basadas red neuronal reciente para el aprendizaje de representación de palabra se utiliza en los datos de idioma alemán. El objetivo es evaluar las representaciones aprendido de palabras en diferentes tareas de procesamiento del lenguaje y la recuperación de información. En particular, una evaluación semántico-sintáctica conjunto se construye para el idioma alemán. Además de eso, las representaciones vectoriales de palabras aprendidas se utilizan como características para un clasificador de documentos comerciales idioma alemán. Las características aprendidas superaron características artesanales existentes y funcionando de modo similar a otros enfoques del estado de la técnica.

The excavation for the 53-story reinforced concrete One Museum Park West building (OMPW) in Chicago was constructed using combined support systems. The building was supported by caissons and a perimeter wall formed by secant piles which transitioned to tangent piles below the bottom of the cut. After installing the perimeter wall and deep foundations for the building, the excavation for the reinforced concrete central core was supported by a circular cofferdam made of sheet piles and steel ring beams. This excavation was made using conventional bottom-up excavation techniques. The excavation for the rest of the building was made using a top-down method and was laterally braced with reinforced concrete floor slabs structurally connected to the secant pile walls and the reinforced concrete core. Field performance data during construction of the OMPW basements were collected with settlement points, inclinometers and strain gages installed in the concrete basement slabs. Significant settlements were observed adjacent to the excavation during wall and deep foundation installation, central core construction and top-down excavation. The influence of concrete material time-dependence of the floor slabs on the performance during top-down excavation is evaluated and a method for the separation of temperature-related strains from the excavation-induced strains in the concrete is described. The construction of the basements is simulated with a 3D numerical structural model to study the influence of the concrete material time-dependence of the floor slabs in the excavation-induced movements resulting from top-down construction. Computed strains in the floor slabs are compared with the observed performance data collected from the strain gages. For use in numerical simulation of the entire excavation process, hypoplasticity constitutive parameters for clays (HC model) at the element test level were calibrated with tests conducted on high quality block samples obtained from the excavation of Block 37 project in Chicago. They included oedometer tests, bender element measurements, and triaxial tests following different stress probes, and were further refined with seismic cone penetration results.  The entire OMPW excavation is simulated with the calibrated parameters with a fully coupled-flow deformation, three-dimensional finite element model. Influence of the construction of secant pile walls, concrete material time-dependence, and nonlinear concrete behavior of the perimeter pile walls on the performance of the OMPW excavation is determined. Conclusions about top-down construction are drawn in light of the measured and simulated performance of the OMPW excavation.

The Interaction of electromagnetic waves with dielectric bodies and metals has been extensively studied because of its importance to problems including propagation through rain or snow, scattering by and detection of air borne particles, coupling to missiles with plasma plumes or dielectric-filled apertures, performance of communication antennas in the presence of dielectric and magnetic inhomogeneities, and medical diagnostics and power absorption in biological bodies. Computational electromagnetics methods (CEM) offer and indispensable tool for calculating the electromagnetic scattering from an internal field distribution of arbitrarily shaped, inhomogeneous, dielectric bodies. The aim of this thesis is the study and simulation of a RF coils system design by developing a novel parallel fast Method of Moments (MoM) modeling approach suitable for the simulation of dielectric bodies and metals. The parallel fast MoM implementation uses volume and surface basis functions with special properties appropriate for the representation of flux current densities for perfect electric conductors (PEC) and dielectrics. The results obtained with our modeling method were confirmed by comparisons with analytical solutions and other commercial software results, yielding very good agreement. The RF coil is employed in high field Magnetic Resonance Imaging (MRI) to obtain high quality brain images. Among all the clinical imaging techniques, MRI stands as a noninvasive technique that provides accurate, detailed anatomic images, which has had a major impact in the diagnosis of human diseases. MRI is a widely use soft-tissue imaging modality that has involved over the past several years into a powerful and versatile medical diagnostic tool capable of providing in-vivo diagnostic images of human anatomy. Current research areas in MRI system design are driven by the need to obtain detailed high resolution images with improved image signal-to-noise ratio (SNR) at a given magnetic field strength. One of the most critical factor that influences the quality and resolution of the MRI is the homogeneity of the RF field. To this end, this requirement demands the development of high performance MRI radio frequency (RF) coils and a standard procedure for enhancing the uniformity of the field directly at the modeling stage of the RF Coil.

The purpose of this report is to investigate the performance, measured in terms of Key Performance Indicators (KPIs) of a UMTS radio network (UTRAN) under different traffic and mobility scenarios and develop general guidelines for dimensioning the UMTS network optimally. In this work, simulations are performed using Wines tool, which is a dynamic simulator that models all the radio-resource management functionality of the UTRAN. Two main series of experiments have been performed. In the first series, the distribution of the traffic is homogeneous and the main purpose is to find capacity figures when increasing the traffic density of one of the four possible defined services (WWW, FTP, voice and video-call) and also with a given traffic mix involving all services together. Two mobility profiles are used (pedestrian and vehicular). In the second series of experiments, a more realistic scenario, based on detailed geographical characteristics relevant for signal propagation, and traffic densities based on traffic maps (non homogeneous scenario), has been simulated. The purpose was to analyze the possible differences in capacity between the “ideal” model and the more “realistic” model. For the single-service scenario analysis, circuit-switched services (speech and video-call), it can be observed that the capacity is mainly uplink-limited. The corresponding results for the packet-switched services (FTP, WEB) have shown that capacity is mostly downlink-limited, as it was expected due to the more asymmetrical data rates in the downlink compared to the uplink for both services. For the service mix analysis in both scenarios (homogeneous and non homogeneous) it was found that the capacity is mostly downlink-limited (i.e. the downlink power target level is exceeded before the uplink load target level).

Abstract: The manufacturing brick generates a direct effect in the air quality affecting the health of the surrounding communities of this type of industry. In the Colombian municipality of Nemocón, in the Patio Bonito rural locality, the majority of small and micro-sized brick fields uses craft techniques for manufacturing different types of bricks. The manufacturing plant is basically represented by a kiln and a plot of land as a farmyard; the area where these elements are concentrated is commonly named “Chircal”. Craft brickers usually use kilns of direct fire and upward firing for cooking. The main fuel used in the manufacture of bricks is coal, followed by natural gas and diesel. The present work aims to characterize air quality, taking into account the massive growth of the brick manufacturing activity in the rural locality of Patio Bonito, Nemocón, Colombia, specifically assessing the particulate matter PM10, which is one of the most critical air pollutant due to its impact on health of its inhabitants and on the local environment. The year 2013 was selected for the analysis, because the highest PM10 concentrations of the last five years that happened in that year, according to historical data from the environmental authority of Cundinamarca region. For assessing the air quality in Patio Bonito, a Gaussian air quality model named URBAIR, developed by the University of Aveiro, was applied. For an adequate operation of this model, it was necessary to collect input data as emission sources and meteorological data. The modeling periods were August 8th, November 20th and 24th 2013, when daily averaged with concentrations above 100 µg/m3, limit regulated by Colombian legislation were registered. Moreover, two improvement scenarios were proposed. The first considered the operation of “Colmena” kilns with an emission control system. The second implied Hoffman and Tunnel kilns, because those showed the lowest PM10 emissions. Results allowed to conclude that those improvement measures can reduce the PM10 contribution to air pollution levels. Resumen: La fabricación del ladrillo genera un efecto directo en la calidad del aire afectando la salud de las comunidades circundantes de este tipo de industria. En el municipio colombiano de Nemocón, en la localidad rural de Patio Bonito, la mayoría de las pequeñas ladrilleras utilizan técnicas artesanales para la fabricación de diferentes tipos de ladrillos. La planta de fabricación está básicamente representada por un horno y una parcela de tierra como un patio; el área donde se concentran estos elementos se denomina comúnmente ""Chircal"". Los trabajadores artesanales generalmente usan hornos de fuego directo y fuego ascendente para cocinar. El principal combustible utilizado en la fabricación de ladrillos es el carbón, seguido del gas natural y el diesel. El presente trabajo tiene como objetivo caracterizar la calidad del aire, teniendo en cuenta el crecimiento masivo de la actividad de fabricación de ladrillos en la localidad rural de Patio Bonito, Nemocón, Colombia, evaluando específicamente el material particulado PM10, que es uno de los contaminantes atmosféricos más críticos debido a su impacto en la salud de sus habitantes y en el entorno local. El año 2013 fue seleccionado para el análisis, debido a las mayores concentraciones de PM10 de los últimos cinco años que ocurrieron en ese año, según datos históricos de la autoridad ambiental de la región de Cundinamarca. Para evaluar la calidad del aire en Patio Bonito, se aplicó un modelo Gaussiano de calidad del aire denominado URBAIR, desarrollado por la Universidad de Aveiro. Para un funcionamiento adecuado de este modelo, fue necesario recopilar datos de entrada como fuentes de emisión y datos meteorológicos. Los periodos de modelación fueron el 8 de agosto, el 20 y el 24 de noviembre de 2013, cuando se registró un promedio diario con concentraciones superiores a 100 ?g / m3, límite regulado por la legislación colombiana. Además, se propusieron dos escenarios de mejora. El primero consideró la operación de hornos ""Colmena"" con un sistema de control de emisiones. El segundo implicaba los hornos Hoffman y Tunnel, porque aquellos mostraban las emisiones más bajas de PM10. Los resultados permitieron concluir que estas medidas de mejora pueden reducir la contribución de PM10 a los niveles de contaminación del aire.

Tackling climate change effects worldwide brings the interest of many donor countries with expertise to (co)develop solutions into beneficiary countries through international cooperation. The solutions proposed can be either too solution-driven, understood as a donor country selling its knowledge / products, or on the other hand, too locally problem-driven, which is desirable in theory, but facing practical limitations such as time or scale constraints when closing to outside influences. A methodology that tries to join both (apparently opposing) sides, with their advantages and disadvantages of problem and solution driven hasn’t been explored yet.For exploring how this can be done, in this research 2 case studies are analyzed: the Bangladesh Delta Plan 2100 (BDP2100), a nation-wide long term holistic plan on the Bangladeshi Delta, and a group of Coastal Water Management initiatives, a set of continuous efforts to adopt livelihoods to the dynamic situation in the coastal regions of Bangladesh. These cases are analyzed through interviews and desk research to determine to what extent are they problem and solution driven compared to proposed frameworks and pointing how can they be improved. From these analyses, a combined framework is proposed where problem and solution driven approaches are joined taking the best from each perspective. The result is meant to be a first guiding tool for decision-makers in donor and beneficiary countries that face the dilemma between taking local and foreign elements for developing solutions. Resumen Hacer frente a los efectos del cambio climático en todo el mundo atrae el interés de muchos países donantes con experiencia para (co) desarrollar soluciones en los países beneficiarios, a través de la cooperación internacional. Las soluciones propuestas pueden ser demasiado orientadas a la solución, entendidas como un país donante que vende sus conocimientos / productos o, por otro lado, demasiado localmente orientadas a los problemas, lo que es deseable en teoría, pero enfrentan limitaciones prácticas como limitaciones de tiempo o escala cuando se cierran a influencias externas. Aún no se ha explorado una metodología que intente unir a ambas partes (aparentemente opuestas), con las ventajas y desventajas de soluciones basadas en desarrollo netamente de problemas y desarrollo netamente de soluciones. Para explorar cómo se puede combinar esto, en esta investigación se analizan 2 casos de estudio: el Bangladesh Delta Plan 2100 (BDP2100), un plan holístico a largo plazo a nivel nacional en el delta de Bangladesh, y un grupo de iniciativas de gestión costera del agua, un conjunto de esfuerzos continuos para adoptar medios de vida a la situación dinámica en las regiones costeras de Bangladesh. Estos casos se analizan a través de entrevistas e investigaciones teóricas para determinar en qué medida están orientados a los problemas y soluciones en comparación con los marcos propuestos, y señalan cómo pueden mejorarse. A partir de estos análisis, se propone un marco combinado donde se unen los enfoques impulsados por problemas y soluciones, tomando lo mejor de cada perspectiva. El resultado pretende ser una primera herramienta de guía para los tomadores de decisiones en los países donantes y beneficiarios que enfrentan el dilema entre tomar elementos locales y extranjeros para desarrollar soluciones.