Inverse Static Analysis of Massive Parallel Arrays of Three-State Actuators via Artificial Intelligence

Massive parallel robots (MPRs) driven by discrete actuators are force regulated robots that undergo continuous motions despite being commanded through a finite number of states only. Designing a real-time control of such systems requires fast and efficient methods for solving their inverse static analysis (ISA), which is a challenging problem and the subject of this thesis. In particular, five Artificial intelligence methods are proposed to investigate the on-line computation and the generalization error of ISA problem of a class of MPRs featuring three-state force actuators and one degree of revolute motion.

Role of αvβ3 – Integrin and TLR2 in the innate response to Herpes Simplex Virus infection and Delivery of retargeted oncolytic Herpes Simplex via carrier cells

In the first part of my thesis I studied the mechanism of initiation of the innate response to HSV-1. Innate immune response is the first line of defense set up by the cell to counteract pathogens infection and it is elicited by the activation of a number of membrane or intracellular receptors and sensors, collectively indicated as PRRs, Patter Recognition Receptors. We reported that the HSV pathogen-associated molecular patterns (PAMP) that activate Toll-like receptor 2 (TLR2) and lead to the initiation of innate response are the virion glycoproteins gH/gL and gB, which constitute the conserved fusion core apparatus across the Herpesvirus. Specifically gH/gL is sufficient to initiate a signaling cascade which leads to NF-κB activation. Then, by gain and loss-of-function approaches, we found that αvβ3-integrin is a sensor of and plays a crucial role in the innate defense against HSV-1. We showed that αvβ3-integrin signals through a pathway that concurs with TLR2, affects activation/induction of interferons type 1, NF-κB, and a polarized set of cytokines and receptors. Thus, we demonstrated that gH/gL is sufficient to induce IFN1 and NF-κB via this pathway. From these data, we proposed that αvβ3-integrin is considered a class of non-TLR pattern recognition receptors. In the second part of my thesis I studied the capacity of human mesenchymal stromal cells isolated by fetal membranes (FM-hMSCs) to be used as carrier cells for the delivery of retargeted R-LM249 virus. The use of systemically administrated carrier cells to deliver oncolytic viruses to tumoral targets is a promising strategy in oncolytic virotherapy. We observed that FM-hMSCs can be infected by R-LM249 and we optimized the infection condition; then we demonstrate that stromal cells sustain the replication of retargeted R-LM249 and spread it to target tumoral cells. From these preliminary data FM-hMSCs resulted suitable to be used as carrier cells

Porous silicon technology for integrated microsystems

With the development of micro systems, there is an increasing demand for integrable porous materials. In addition to those conventional applications, such as filtration, wicking, and insulating, many new micro devices, including micro reactors, sensors, actuators, and optical components, can benefit from porous materials. Conventional porous materials, such as ceramics and polymers, however, cannot meet the challenges posed by micro systems, due to their incompatibility with standard micro-fabrication processes. In an effort to produce porous materials that can be used in micro systems, porous silicon (PS) generated by anodization of single crystalline silicon has been investigated. In this work, the PS formation process has been extensively studied and characterized as a function of substrate type, crystal orientation, doping concentration, current density and surfactant concentration and type. Anodization conditions have been optimized for producing very thick porous silicon layers with uniform pore size, and for obtaining ideal pore morphologies. Three different types of porous silicon materials: meso porous silicon, macro porous silicon with straight pores, and macro porous silicon with tortuous pores, have been successfully produced. Regular pore arrays with controllable pore size in the range of 2µm to 6µm have been demonstrated as well. Localized PS formation has been achieved by using oxide/nitride/polysilicon stack as masking materials, which can withstand anodization in hydrofluoric acid up to twenty hours. A special etching cell with electrolytic liquid backside contact along with two process flows has been developed to enable the fabrication of thick macro porous silicon membranes with though wafer pores. For device assembly, Si-Au and In-Au bonding technologies have been developed. Very low bonding temperature (~200 degrees C) and thick/soft bonding layers (~6µm) have been achieved by In-Au bondi ng technology, which is able to compensate the potentially rough surface on the porous silicon sample without introducing significant thermal stress. The application of the porous silicon material in micro systems has been demonstrated in a micro gas chromatograph system by two indispensable components: an integrated vapor source and an inlet filter, wherein porous silicon performs the basic functions of porous media: wicking and filtration. By utilizing a macro porous silicon wick, the calibration vapor source was able to produce a uniform and repeatable vapor generation for n-decane with less than a 0.1% variation in 9 hours, and less than a 0.5% variation in rate over 7 days. With engineered porous silicon membranes the inlet filter was able to show a depth filtration with nearly 100% collection efficiency for particles larger than 0.3µm in diameter, a low pressure-drop of 523Pa at 20sccm flow rate, and a filter capacity of 500µg/cm2.

ARMin - robot for rehabilitation of the upper extremities

Task-oriented repetitive movements can improve motor recovery in patients with neurological or orthopaedic lesions. The application of robotics can serve to assist, enhance, evaluate, and document neurological and orthopaedic rehabilitation. ARMin is a new robot for arm therapy applicable to the training of activities of daily living in clinics. ARMin has a semiexoskeletal structure with six degrees of freedom, and is equipped with position and force sensors. The mechanical structure, the actuators and the sensors of the robot are optimized for patient-cooperative control strategies based on impedance and admittance architectures. This paper describes the mechanical structure, the control system, the sensors and actuators, safety aspects and results of a first pilot study with hemiplegic and spinal cord injured subjects.

Context-Aware Adaptive Opportunistic Routing in Mobile Ad-hoc and Sensor Networks

Mobile ad-hoc networks (MANETs) and wireless sensor networks (WSNs) have been attracting increasing attention for decades due to their broad civilian and military applications. Basically, a MANET or WSN is a network of nodes connected by wireless communication links. Due to the limited transmission range of the radio, many pairs of nodes in MANETs or WSNs may not be able to communicate directly, hence they need other intermediate nodes to forward packets for them. Routing in such types of networks is an important issue and it poses great challenges due to the dynamic nature of MANETs or WSNs. On the one hand, the open-air nature of wireless environments brings many difficulties when an efficient routing solution is required. The wireless channel is unreliable due to fading and interferences, which makes it impossible to maintain a quality path from a source node to a destination node. Additionally, node mobility aggravates network dynamics, which causes frequent topology changes and brings significant overheads for maintaining and recalculating paths. Furthermore, mobile devices and sensors are usually constrained by battery capacity, computing and communication resources, which impose limitations on the functionalities of routing protocols. On the other hand, the wireless medium possesses inherent unique characteristics, which can be exploited to enhance transmission reliability and routing performance. Opportunistic routing (OR) is one promising technique that takes advantage of the spatial diversity and broadcast nature of the wireless medium to improve packet forwarding reliability in multihop wireless communication. OR combats the unreliable wireless links by involving multiple neighboring nodes (forwarding candidates) to choose packet forwarders. In opportunistic routing, a source node does not require an end-to-end path to transmit packets. The packet forwarding decision is made hop-by-hop in a fully distributed fashion. Motivated by the deficiencies of existing opportunistic routing protocols in dynamic environments such as mobile ad-hoc networks or wireless sensor networks, this thesis proposes a novel context-aware adaptive opportunistic routing scheme. Our proposal selects packet forwarders by simultaneously exploiting multiple types of cross-layer context information of nodes and environments. Our approach significantly outperforms other routing protocols that rely solely on a single metric. The adaptivity feature of our proposal enables network nodes to adjust their behaviors at run-time according to network conditions. To accommodate the strict energy constraints in WSNs, this thesis integrates adaptive duty-cycling mechanism to opportunistic routing for wireless sensor nodes. Our approach dynamically adjusts the sleeping intervals of sensor nodes according to the monitored traffic load and the estimated energy consumption rate. Through the integration of duty cycling of sensor nodes and opportunistic routing, our protocol is able to provide a satisfactory balance between good routing performance and energy efficiency for WSNs.

Particulate and phytoplankton absorption during POLARSTERN cruises ANT-XXVI/3 and ANT-XXVIII/3

The euphotic depth (Zeu) is a key parameter in modelling primary production (PP) using satellite ocean colour. However, evaluations of satellite Zeu products are scarce. The objective of this paper is to investigate existing approaches and sensors to estimate Zeu from satellite and to evaluate how different Zeu products might affect the estimation of PP in the Southern Ocean (SO). Euphotic depth was derived from MODIS and SeaWiFS products of (i) surface chlorophyll-a (Zeu-Chla) and (ii) inherent optical properties (Zeu-IOP). They were compared with in situ measurements of Zeu from different regions of the SO. Both approaches and sensors are robust to retrieve Zeu, although the best results were obtained using the IOP approach and SeaWiFS data, with an average percentage of error (E) of 25.43% and mean absolute error (MAE) of 0.10 m (log scale). Nevertheless, differences in the spatial distribution of Zeu-Chla and Zeu-IOP for both sensors were found as large as 30% over specific regions. These differences were also observed in PP. On average, PP based on Zeu-Chla was 8% higher than PP based on Zeu-IOP, but it was up to 30% higher south of 60°S. Satellite phytoplankton absorption coefficients (aph) derived by the Quasi-Analytical Algorithm at different wavelengths were also validated and the results showed that MODIS aph are generally more robust than SeaWiFS. Thus, MODIS aph should be preferred in PP models based on aph in the SO. Further, we reinforce the importance of investigating the spatial differences between satellite products, which might not be detected by the validation with in situ measurements due to the insufficient amount and uneven distribution of the data.

Autonomous vehicle control systems for safe crossroads

This article presents a cooperative manoeuvre among three dual mode cars – vehicles equipped with sensors and actuators, and that can be driven either manually or autonomously. One vehicle is driven autonomously and the other two are driven manually. The main objective is to test two decision algorithms for priority conflict resolution at intersections so that a vehicle autonomously driven can take their own decision about crossing an intersection mingling with manually driven cars without the need for infrastructure modifications. To do this, the system needs the position, speeds, and turning intentions of the rest of the cars involved in the manoeuvre. This information is acquired via communications, but other methods are also viable, such as artificial vision. The idea of the experiments was to adjust the speed of the manually driven vehicles to force a situation where all three vehicles arrive at an intersection at the same time.

Evaluation and implementation of a hybrid cross-platform application combining mobile web-technologies and native device functions

Este Proyecto de Fin de Carrera presenta un prototipo de aplicación móvil híbrida multi-plataforma para Android y iOS. Las aplicaciones móviles híbridas son una combinación de aplicaciones web móviles y aplicaciones móviles nativas. Se desarrollan parcialmente con tecnologías web y pueden acceder a la capa nativa y sensores del teléfono. Para el usuario se presentan como aplicaciones nativas, ya que se pueden descargar de las tiendas de aplicaciones y son instaladas en el dispositivo. El prototipo consiste en la migración del módulo de noticias financieras de las aplicaciones actuales para móviles de una compañía bancaria reimplementándolo como aplicación híbrida utilizando uno de los entornos de desarrollo disponibles en el mercado para este propósito. El desarrollo de aplicaciones híbridas puede ahorrar tiempo y dinero cuando se pretende alcanzar más de una plataforma móvil. El objetivo es la evaluación de las ventajas e inconvenientes que ofrece el desarrollo de aplicaciones híbridas en términos de reducción de costes, tiempo de desarrollo y resultado final de la aplicación. El proyecto consta de varias fases. Durante la primera fase se realiza un estudio sobre las aplicaciones híbridas que podemos encontrar hoy en día en el mercado utilizando los ejemplos de linkedIn, Facebook y Financial times. Se hace hincapié en las tecnologías utilizadas, uso de la red móvil y problemas encontrados. Posteriormente se realiza una comparación de distintos entornos de desarrollo multi-plataforma para aplicaciones híbridas en términos de la estrategia utilizada, plataformas soportadas, lenguajes de programación, acceso a capacidades nativas de los dispositivos y licencias de uso. Esta primera fase da como resultado la elección del entorno de desarrollo más adecuado a las exigencias del proyecto, que es PhoneGap, y continua con un análisis más detallado de dicho entorno en cuanto a su arquitectura, características y componentes. La siguiente fase comienza con un estudio de las aplicaciones actuales de la compañía para extraer el código fuente necesario y adaptarlo a la arquitectura que tendrá la aplicación. Para la realización del prototipo se hace uso de la característica que ofrece PhoneGap para acceder a la capa nativa del dispositivo, esto es, el uso de plugins. Se diseña y desarrolla un plugin que permite acceder a la capa nativa para cada plataforma. Una vez desarrollado el prototipo para la plataforma Android, se migra y adapta para la plataforma iOS. Por último se hace una evaluación de los prototipos en cuanto a su facilidad y tiempo de desarrollo, rendimiento, funcionalidad y apariencia de la interfaz de usuario. ABSTRACT. This bachelor's thesis presents a prototype of a hybrid cross-platform mobile application for Android and iOS. Hybrid mobile applications are a combination of mobile web and mobile native applications. They are built partially with web technologies and they can also access native features and sensors of the device. For a user, they look like native applications as they are downloaded from the application stores and installed on the device. This prototype consists of the migration of the financial news module of current mobile applications from a financial bank reimplementing them as a hybrid application using one of the frameworks available in the market for that purpose. Development of applications on a hybrid way can help reducing costs and effort when targeting more than one platform. The target of the project is the evaluation of the advantages and disadvantages that hybrid development can offer in terms of reducing costs and efforts and the final result of the application. The project starts with an analysis of successfully released hybrid applications using the examples of linkedIn, Facebook and Financial Times, emphasizing the different used technologies, the transmitted network data and the encountered problems during the development. This analysis is followed by a comparison of most popular hybrid crossplatform development frameworks in terms of the different approaches, supported platforms, programming languages, access to native features and license. This first stage has the outcome of finding the development framework that best fits to the requirements of the project, that is PhoneGap, and continues with a deeper analysis of its architecture, features and components. Next stage analyzes current company's applications to extract the needed source code and adapt it to the architecture of the prototype. For the realization of the application, the feature that PhoneGap offers to access the native layer of the device is used. This feature is called plugin. A custom plugin is designed and developed to access the native layer of each targeted platform. Once the prototype is finished for Android, it is migrated and adapted to the iOS platform. As a final conclusion the prototypes are evaluated in terms of ease and time of development, performance, functionality and look and feel.

Bayesian scene analysis for multi-camera 3D tracking and camera positioning

Los sistemas de seguimiento mono-cámara han demostrado su notable capacidad para el análisis de trajectorias de objectos móviles y para monitorización de escenas de interés; sin embargo, tanto su robustez como sus posibilidades en cuanto a comprensión semántica de la escena están fuertemente limitadas por su naturaleza local y monocular, lo que los hace insuficientes para aplicaciones realistas de videovigilancia. El objetivo de esta tesis es la extensión de las posibilidades de los sistemas de seguimiento de objetos móviles para lograr un mayor grado de robustez y comprensión de la escena. La extensión propuesta se divide en dos direcciones separadas. La primera puede considerarse local, ya que está orientada a la mejora y enriquecimiento de las posiciones estimadas para los objetos móviles observados directamente por las cámaras del sistema; dicha extensión se logra mediante el desarrollo de un sistema multi-cámara de seguimiento 3D, capaz de proporcionar consistentemente las posiciones 3D de múltiples objetos a partir de las observaciones capturadas por un conjunto de sensores calibrados y con campos de visión solapados. La segunda extensión puede considerarse global, dado que su objetivo consiste en proporcionar un contexto global para relacionar las observaciones locales realizadas por una cámara con una escena de mucho mayor tamaño; para ello se propone un sistema automático de localización de cámaras basado en las trayectorias observadas de varios objetos móviles y en un mapa esquemático de la escena global monitorizada. Ambas líneas de investigación se tratan utilizando, como marco común, técnicas de estimación bayesiana: esta elección está justificada por la versatilidad y flexibilidad proporcionada por dicho marco estadístico, que permite la combinación natural de múltiples fuentes de información sobre los parámetros a estimar, así como un tratamiento riguroso de la incertidumbre asociada a las mismas mediante la inclusión de modelos de observación específicamente diseñados. Además, el marco seleccionado abre grandes posibilidades operacionales, puesto que permite la creación de diferentes métodos numéricos adaptados a las necesidades y características específicas de distintos problemas tratados. El sistema de seguimiento 3D con múltiples cámaras propuesto está específicamente diseñado para permitir descripciones esquemáticas de las medidas realizadas individualmente por cada una de las cámaras del sistema: esta elección de diseño, por tanto, no asume ningún algoritmo específico de detección o seguimiento 2D en ninguno de los sensores de la red, y hace que el sistema propuesto sea aplicable a redes reales de vigilancia con capacidades limitadas tanto en términos de procesamiento como de transmision. La combinación robusta de las observaciones capturadas individualmente por las cámaras, ruidosas, incompletas y probablemente contaminadas por falsas detecciones, se basa en un metodo de asociación bayesiana basado en geometría y color: los resultados de dicha asociación permiten el seguimiento 3D de los objetos de la escena mediante el uso de un filtro de partículas. El sistema de fusión de observaciones propuesto tiene, como principales características, una gran precisión en términos de localización 3D de objetos, y una destacable capacidad de recuperación tras eventuales errores debidos a un número insuficiente de datos de entrada. El sistema automático de localización de cámaras se basa en la observación de múltiples objetos móviles y un mapa esquemático de las áreas transitables del entorno monitorizado para inferir la posición absoluta de dicho sensor. Para este propósito, se propone un novedoso marco bayesiano que combina modelos dinámicos inducidos por el mapa en los objetos móviles presentes en la escena con las trayectorias observadas por la cámara, lo que representa un enfoque nunca utilizado en la literatura existente. El sistema de localización se divide en dos sub-tareas diferenciadas, debido a que cada una de estas tareas requiere del diseño de algoritmos específicos de muestreo para explotar en profundidad las características del marco desarrollado: por un lado, análisis de la ambigüedad del caso específicamente tratado y estimación aproximada de la localización de la cámara, y por otro, refinado de la localización de la cámara. El sistema completo, diseñado y probado para el caso específico de localización de cámaras en entornos de tráfico urbano, podría tener aplicación también en otros entornos y sensores de diferentes modalidades tras ciertas adaptaciones. ABSTRACT Mono-camera tracking systems have proved their capabilities for moving object trajectory analysis and scene monitoring, but their robustness and semantic possibilities are strongly limited by their local and monocular nature and are often insufficient for realistic surveillance applications. This thesis is aimed at extending the possibilities of moving object tracking systems to a higher level of scene understanding. The proposed extension comprises two separate directions. The first one is local, since is aimed at enriching the inferred positions of the moving objects within the area of the monitored scene directly covered by the cameras of the system; this task is achieved through the development of a multi-camera system for robust 3D tracking, able to provide 3D tracking information of multiple simultaneous moving objects from the observations reported by a set of calibrated cameras with semi-overlapping fields of view. The second extension is global, as is aimed at providing local observations performed within the field of view of one camera with a global context relating them to a much larger scene; to this end, an automatic camera positioning system relying only on observed object trajectories and a scene map is designed. The two lines of research in this thesis are addressed using Bayesian estimation as a general unifying framework. Its suitability for these two applications is justified by the flexibility and versatility of that stochastic framework, which allows the combination of multiple sources of information about the parameters to estimate in a natural and elegant way, addressing at the same time the uncertainty associated to those sources through the inclusion of models designed to this end. In addition, it opens multiple possibilities for the creation of different numerical methods for achieving satisfactory and efficient practical solutions to each addressed application. The proposed multi-camera 3D tracking method is specifically designed to work on schematic descriptions of the observations performed by each camera of the system: this choice allows the use of unspecific off-the-shelf 2D detection and/or tracking subsystems running independently at each sensor, and makes the proposal suitable for real surveillance networks with moderate computational and transmission capabilities. The robust combination of such noisy, incomplete and possibly unreliable schematic descriptors relies on a Bayesian association method, based on geometry and color, whose results allow the tracking of the targets in the scene with a particle filter. The main features exhibited by the proposal are, first, a remarkable accuracy in terms of target 3D positioning, and second, a great recovery ability after tracking losses due to insufficient input data. The proposed system for visual-based camera self-positioning uses the observations of moving objects and a schematic map of the passable areas of the environment to infer the absolute sensor position. To this end, a new Bayesian framework combining trajectory observations and map-induced dynamic models for moving objects is designed, which represents an approach to camera positioning never addressed before in the literature. This task is divided into two different sub-tasks, setting ambiguity analysis and approximate position estimation, on the one hand, and position refining, on the other, since they require the design of specific sampling algorithms to correctly exploit the discriminative features of the developed framework. This system, designed for camera positioning and demonstrated in urban traffic environments, can also be applied to different environments and sensors of other modalities after certain required adaptations.

High-fidelity piezoelectric loudspeaker

This project reports on a literature review about piezoelectric loudspeakers and on an experimental research about how to improve some features of a particular horned piezoelectric tweeter. The work involves an investigation of the performance and principle of operation of piezoelectric loudspeakers to understand how the sound is generated and what its main parameters are. Also, previous research papers about how to improve this type of speakers are reported. The knowledge gained was used to reconsider and re-purpose a particular piezoelectric transducer. After characterising the original state of the device with acoustical and electroacoustical measurements, some improvements were implemented. Moreover, interesting conclusions were reached based on the results of the tests that were carried out. A structural study with a scanning laser was then completed. These sections demonstrated the need for providing the speaker with a rear suspension that guides the vibration of the membrane. Finally, an inverse filter was designed in order to get a flat output response. After simulating the results with Matlab, validating experiments were run in the anechoic chamber with great success. RESUMEN. Este proyecto consta de un estudio detallado sobre piezoelectricidad y altavoces piezoeléctricos, así como de una parte experimental consistente en mejorar algunas características de un altavoz piezoeléctrico particular: un tweeter de bocina. El estudio profundiza en cuáles son los principios de funcionamiento y los principales parámetros de este tipo de altavoces. Con el conocimiento adquirido a partir de trabajos de previos sobre el tema e investigación bibliográfica se ha llevado a cabo la parte experimental. Esta parte ha requerido de una serie de medidas acústicas y electroacústicas para, primero, caracterizar el altavoz en su estado original y para posteriormente buscar y validar posibles mejoras, principalmente en la respuesta en frecuencia. Además, se ha realizado un estudio estructural del diafragma a partir de medidas tomadas con un vibrómetro laser Doppler. De estos tres procesos se concluyó que el altavoz bajo estudio tiene un problema en el soporte del cristal piezoeléctrico y se demostró la necesidad de equipar el dispositivo de una suspensión trasera que controle el movimiento del diafragma. Finalmente, se ha diseñado e implementado en Matlab un filtro inverso, con el objetivo de conseguir una respuesta plana a la salida del altavoz. Su funcionamiento fue validado en la cámara anecoica satisfactoriamente.

Bayesian scene analysis for multi-camera 3D tracking and camera positioning

Los sistemas de seguimiento mono-cámara han demostrado su notable capacidad para el análisis de trajectorias de objectos móviles y para monitorización de escenas de interés; sin embargo, tanto su robustez como sus posibilidades en cuanto a comprensión semántica de la escena están fuertemente limitadas por su naturaleza local y monocular, lo que los hace insuficientes para aplicaciones realistas de videovigilancia. El objetivo de esta tesis es la extensión de las posibilidades de los sistemas de seguimiento de objetos móviles para lograr un mayor grado de robustez y comprensión de la escena. La extensión propuesta se divide en dos direcciones separadas. La primera puede considerarse local, ya que está orientada a la mejora y enriquecimiento de las posiciones estimadas para los objetos móviles observados directamente por las cámaras del sistema; dicha extensión se logra mediante el desarrollo de un sistema multi-cámara de seguimiento 3D, capaz de proporcionar consistentemente las posiciones 3D de múltiples objetos a partir de las observaciones capturadas por un conjunto de sensores calibrados y con campos de visión solapados. La segunda extensión puede considerarse global, dado que su objetivo consiste en proporcionar un contexto global para relacionar las observaciones locales realizadas por una cámara con una escena de mucho mayor tamaño; para ello se propone un sistema automático de localización de cámaras basado en las trayectorias observadas de varios objetos móviles y en un mapa esquemático de la escena global monitorizada. Ambas líneas de investigación se tratan utilizando, como marco común, técnicas de estimación bayesiana: esta elección está justificada por la versatilidad y flexibilidad proporcionada por dicho marco estadístico, que permite la combinación natural de múltiples fuentes de información sobre los parámetros a estimar, así como un tratamiento riguroso de la incertidumbre asociada a las mismas mediante la inclusión de modelos de observación específicamente diseñados. Además, el marco seleccionado abre grandes posibilidades operacionales, puesto que permite la creación de diferentes métodos numéricos adaptados a las necesidades y características específicas de distintos problemas tratados. El sistema de seguimiento 3D con múltiples cámaras propuesto está específicamente diseñado para permitir descripciones esquemáticas de las medidas realizadas individualmente por cada una de las cámaras del sistema: esta elección de diseño, por tanto, no asume ningún algoritmo específico de detección o seguimiento 2D en ninguno de los sensores de la red, y hace que el sistema propuesto sea aplicable a redes reales de vigilancia con capacidades limitadas tanto en términos de procesamiento como de transmision. La combinación robusta de las observaciones capturadas individualmente por las cámaras, ruidosas, incompletas y probablemente contaminadas por falsas detecciones, se basa en un metodo de asociación bayesiana basado en geometría y color: los resultados de dicha asociación permiten el seguimiento 3D de los objetos de la escena mediante el uso de un filtro de partículas. El sistema de fusión de observaciones propuesto tiene, como principales características, una gran precisión en términos de localización 3D de objetos, y una destacable capacidad de recuperación tras eventuales errores debidos a un número insuficiente de datos de entrada. El sistema automático de localización de cámaras se basa en la observación de múltiples objetos móviles y un mapa esquemático de las áreas transitables del entorno monitorizado para inferir la posición absoluta de dicho sensor. Para este propósito, se propone un novedoso marco bayesiano que combina modelos dinámicos inducidos por el mapa en los objetos móviles presentes en la escena con las trayectorias observadas por la cámara, lo que representa un enfoque nunca utilizado en la literatura existente. El sistema de localización se divide en dos sub-tareas diferenciadas, debido a que cada una de estas tareas requiere del diseño de algoritmos específicos de muestreo para explotar en profundidad las características del marco desarrollado: por un lado, análisis de la ambigüedad del caso específicamente tratado y estimación aproximada de la localización de la cámara, y por otro, refinado de la localización de la cámara. El sistema completo, diseñado y probado para el caso específico de localización de cámaras en entornos de tráfico urbano, podría tener aplicación también en otros entornos y sensores de diferentes modalidades tras ciertas adaptaciones. ABSTRACT Mono-camera tracking systems have proved their capabilities for moving object trajectory analysis and scene monitoring, but their robustness and semantic possibilities are strongly limited by their local and monocular nature and are often insufficient for realistic surveillance applications. This thesis is aimed at extending the possibilities of moving object tracking systems to a higher level of scene understanding. The proposed extension comprises two separate directions. The first one is local, since is aimed at enriching the inferred positions of the moving objects within the area of the monitored scene directly covered by the cameras of the system; this task is achieved through the development of a multi-camera system for robust 3D tracking, able to provide 3D tracking information of multiple simultaneous moving objects from the observations reported by a set of calibrated cameras with semi-overlapping fields of view. The second extension is global, as is aimed at providing local observations performed within the field of view of one camera with a global context relating them to a much larger scene; to this end, an automatic camera positioning system relying only on observed object trajectories and a scene map is designed. The two lines of research in this thesis are addressed using Bayesian estimation as a general unifying framework. Its suitability for these two applications is justified by the flexibility and versatility of that stochastic framework, which allows the combination of multiple sources of information about the parameters to estimate in a natural and elegant way, addressing at the same time the uncertainty associated to those sources through the inclusion of models designed to this end. In addition, it opens multiple possibilities for the creation of different numerical methods for achieving satisfactory and efficient practical solutions to each addressed application. The proposed multi-camera 3D tracking method is specifically designed to work on schematic descriptions of the observations performed by each camera of the system: this choice allows the use of unspecific off-the-shelf 2D detection and/or tracking subsystems running independently at each sensor, and makes the proposal suitable for real surveillance networks with moderate computational and transmission capabilities. The robust combination of such noisy, incomplete and possibly unreliable schematic descriptors relies on a Bayesian association method, based on geometry and color, whose results allow the tracking of the targets in the scene with a particle filter. The main features exhibited by the proposal are, first, a remarkable accuracy in terms of target 3D positioning, and second, a great recovery ability after tracking losses due to insufficient input data. The proposed system for visual-based camera self-positioning uses the observations of moving objects and a schematic map of the passable areas of the environment to infer the absolute sensor position. To this end, a new Bayesian framework combining trajectory observations and map-induced dynamic models for moving objects is designed, which represents an approach to camera positioning never addressed before in the literature. This task is divided into two different sub-tasks, setting ambiguity analysis and approximate position estimation, on the one hand, and position refining, on the other, since they require the design of specific sampling algorithms to correctly exploit the discriminative features of the developed framework. This system, designed for camera positioning and demonstrated in urban traffic environments, can also be applied to different environments and sensors of other modalities after certain required adaptations.

Calibración de sensores de fuerza bajo estímulos sinusoidales: contribuciones a su caracterización

El presente trabajo tiene como objetivo el desarrollo de un patrón primario para la calibración de sensores de fuerza bajo excitaciones sinusoidales. Con consecuencia de dicho desarrollo se establecerá un método de calibración de sensores de fuerza en condiciones dinámicas que permitirá la caracterización de estos sensores en dichas condiciones y determinar la incertidumbre asociada. Este patrón se basa en la definición directa de fuerza como masa por aceleración. Para ello se carga el sensor con distintas cargas calibradas y se somete a distintas aceleraciones mediante un excitador de vibraciones. Dichas aceleraciones se generan para frecuencias desde 5 Hz a 2400 Hz. La aceleración se mide mediante un vibrómetro láser con trazabilidad a la unidad de longitud (longitud de onda del láser). Al ser una medición completamente dinámica se necesita un sistema de adquisición de datos multicanal para la toma de datos en tiempo real. Este sistema adquiere las señales eléctricas provenientes del vibrómetro láser, del sensor a caracterizar y del acelerómetro para mediciones auxiliares. Se ha dispuesto de cuatro sensores de fuerza para realizar ensayos, un sensor piezoeléctrico y tres sensores resistivos. En este trabajo se han estudiado los factores de influencia y se ha implementado un método de calibración para minimizar los mismos, así como también se han establecido las correcciones a realizar. Para la caracterización dinámica del sensor se ha partido de un modelo de oscilador armónico amortiguado forzado, se ha establecido la metodología para la determinación de sus parámetros de caracterización y se ha estudiado su validez. También se ha realizado una comparación entre los resultados obtenidos para condiciones estáticas y dinámicas. ABSTRACT The aim in the current work is the development of a primary standard for force sensors calibration under sinusoidal excitations. As consequence of this development a method for force sensors calibration under dynamic conditions will be established that will allow these sensors characterization for such conditions and the determination of their associated uncertainty. This standard is based on the direct definition of force as mass multiplied by acceleration. To do so, the sensor is loaded with different calibrated loads and is maintained under different accelerations by means of a vibration shaker. These accelerations are generated with frequencies from 5 Hz up to 2400 Hz. The acceleration is measured by means of a laser vibrometer with traceability to the unit of length (laser wavelength). As the measurement is totally dynamic a multiple channel data acquisition system is required for data acquisition in real time. This system acquires the electrical signals outputs coming from the laser vibrometer, the sensor to be characterised and two accelerometers for additional measurements. Four force sensors, one piezoelectric sensor and three resistive sensors, have been available to perform the tests. During this work the influence factors have been studied and a calibration method to minimise these factors have been implemented as well as the corrections to be performed have been established. As the starting point for the sensor dynamic characterization, a model for a forced damped harmonic oscillator has been used, a method for the characterizing parameters determination has been established and its validity has been studied. A comparison between results for static and dynamic conditions has been performed as well.

Métodos diferenciales e interferométricos para la evaluación de deformaciones de la corteza terrestre mediante técnicas GNSS e InSAR = Differential and interferometric methods for evaluation of crustal deformation using GNSS and InSAR techniques

La actividad volcánica interviene en multitud de facetas de la propia actividad humana, no siempre negativas. Sin embargo, son más los motivos de peligrosidad y riesgo que incitan al estudio de la actividad volcánica. Existen razones de seguridad que inciden en el mantenimiento del seguimiento y monitorización de la actividad volcánica para garantizar la vida y la seguridad de los asentamientos antrópicos en las proximidades de los edificios volcánicos. En esta tesis se define e implementa un sistema de monitorización de movimientos de la corteza en las islas de Tenerife y La Palma, donde el impacto social que representa un aumento o variación de la actividad volcánica en las islas es muy severo. Aparte de la alta densidad demográfica del Archipiélago, esta población aumenta significativamente, en diferentes periodos a lo largo del año, debido a la actividad turística que representa la mayor fuente de ingresos de las islas. La población y los centros turísticos se diseminan predominantemente a lo largo de las costas y también a lo largo de los flancos de los edificios volcánicos. Quizá el mantenimiento de estas estructuras sociales y socio-económicas son los motivos más importantes que justifican una monitorización de la actividad volcánica en las Islas Canarias. Recientemente se ha venido trabajando cada vez más en el intento de predecir la actividad volcánica utilizando los nuevos sistemas de monitorización geodésica, puesto que la actividad volcánica se manifiesta anteriormente por deformación de la corteza terrestre y cambios en la fuerza de la gravedad en la zona donde más tarde se registran eventos volcánicos. Los nuevos dispositivos y sensores que se han desarrollado en los últimos años en materias como la geodesia, la observación de la Tierra desde el espacio y el posicionamiento por satélite, han permitido observar y medir tanto la deformación producida en el terreno como los cambios de la fuerza de la gravedad antes, durante y posteriormente a los eventos volcánicos que se producen. Estos nuevos dispositivos y sensores han cambiado las técnicas o metodologías geodésicas que se venían utilizando hasta la aparición de los mismos, renovando métodos clásicos y desarrollando otros nuevos que ya se están afianzando como metodologías probadas y reconocidas para ser usadas en la monitorización volcánica. Desde finales de la década de los noventa del siglo pasado se han venido desarrollando en las Islas Canarias varios proyectos que han tenido como objetivos principales el desarrollo de nuevas técnicas de observación y monitorización por un lado y el diseño de una metodología de monitorización volcánica adecuada, por otro. Se presenta aquí el estudio y desarrollo de técnicas GNSS para la monitorización de deformaciones corticales y su campo de velocidades para las islas de Tenerife y La Palma. En su implementación, se ha tenido en cuenta el uso de la infraestructura geodésica y de monitorización existente en el archipiélago a fin de optimizar costes, además de complementarla con nuevas estaciones para dar una cobertura total a las dos islas. Los resultados obtenidos en los proyectos, que se describen en esta memoria, han dado nuevas perspectivas en la monitorización geodésica de la actividad volcánica y nuevas zonas de interés que anteriormente no se conocían en el entorno de las Islas Canarias. Se ha tenido especial cuidado en el tratamiento y propagación de los errores durante todo el proceso de observación, medida y proceso de los datos registrados, todo ello en aras de cuantificar el grado de fiabilidad de los resultados obtenidos. También en este sentido, los resultados obtenidos han sido verificados con otros procedentes de sistemas de observación radar de satélite, incorporando además a este estudio las implicaciones que el uso conjunto de tecnologías radar y GNSS tendrán en un futuro en la monitorización de deformaciones de la corteza terrestre. ABSTRACT Volcanic activity occurs in many aspects of human activity, and not always in a negative manner. Nonetheless, research into volcanic activity is more likely to be motivated by its danger and risk. There are security reasons that influence the monitoring of volcanic activity in order to guarantee the life and safety of human settlements near volcanic edifices. This thesis defines and implements a monitoring system of movements in the Earth’s crust in the islands of Tenerife and La Palma, where the social impact of an increase (or variation) of volcanic activity is very severe. Aside from the high demographic density of the archipelago, the population increases significantly in different periods throughout the year due to tourism, which represents a major source of revenue for the islands. The population and the tourist centres are mainly spread along the coasts and also along the flanks of the volcanic edifices. Perhaps the preservation of these social and socio-economic structures is the most important reason that justifies monitoring volcanic activity in the Canary Islands. Recently more and more work has been done with the intention of predicting volcanic activity, using new geodesic monitoring systems, since volcanic activity is evident prior to eruption because of a deformation of the Earth’s crust and changes in the force of gravity in the zone where volcanic events will later be recorded. The new devices and sensors that have been developed in recent years in areas such as geodesy, the observation of the Earth from space, and satellite positioning have allowed us to observe and measure the deformation produced in the Earth as well as the changes in the force of gravity before, during, and after the volcanic events occur. The new devices and sensors have changed the geodetic techniques and methodologies that were used previously. The classic methods have been renovated and other newer ones developed that are now vouched for as proven recognised methodologies to be used for volcanic monitoring. Since the end of the 1990s, in the Canary Islands various projects have been developed whose principal aim has been the development of new observation and monitoring techniques on the one hand, and the design of an appropriate volcanic monitoring methodology on the other. The study and development of GNSS techniques for the monitoring of crustal deformations and their velocity field is presented here. To carry out the study, the use of geodetic infrastructure and existing monitoring in the archipelago have been taken into account in order to optimise costs, besides complementing it with new stations for total coverage on both islands. The results obtained in the projects, which are described below, have produced new perspectives in the geodetic monitoring of volcanic activity and new zones of interest which previously were unknown in the environment of the Canary Islands. Special care has been taken with the treatment and propagation of errors during the entire process of observing, measuring, and processing the recorded data. All of this was done in order to quantify the degree of trustworthiness of the results obtained. Also in this sense, the results obtained have been verified with others from satellite radar observation systems, incorporating as well in this study the implications that the joint use of radar technologies and GNSS will have for the future of monitoring deformations in the Earth’s crust.

Photoconductivity and Electro-Optical Properties of Wide Band Gap Metal Oxides

The PhD activity described in this Thesis was focused on the study of metal-oxide wide-bandgap materials, aiming at fabricating new optoelectronic devices such as solar-blind UV photodetectors, high power electronics, and gas sensors. Photocurrent spectroscopy and DC photocurrent time evolution were used to investigate the performance of prototypes under different atmospheres, temperatures and excitation wavelengths (or dark conditions). Cathodoluminescence, absorption spectroscopy, XRD and SEM were used to assess structural, morphologic, electrical and optical properties of materials. This thesis is divided into two main sections, each describing the work done on a different metal-oxide semiconductor. 1) MOVPE-grown Ga2O3 thin films for UV solar-blind photodetectors and high power devices The semiconducting oxides, among them Ga2O3, have been employed for several decades as transparent conducting oxide (TCO) electrodes for fabrication of solar cells, displays, electronic, and opto-electronic devices. The interest was mainly confined to such applications, as these materials tend to grow intrinsically n-type, and attempts to get an effective p-type doping has consistently failed. The key requirements of TCO electrodes are indeed high electrical conductivity and good transparency, while crystallographic perfection is a minor issue. Furthermore, for a long period no high-quality substrates and epi-layers were available, which in turn impeded the development of a truly full-oxide electronics. Recently, Ga2O3 has attracted renewed interest, as large single crystals and high-quality homo- and hetero-epitaxial layers became available, which paved the way to novel application areas. Our research group spent the last two years in developing a low temperature (500-700°C) MOVPE growth procedure to obtain thin films of Ga2O3 on different substrates (Dept. of Physics and IMEM-CNR at UNIPR). We obtained a significant result growing on oriented sapphire epitaxial films of high crystalline, undoped, pure phase -Ga2O3 (hexagonal). The crystallographic properties of this phase were investigated by XRD, in order to clarify the lattice parameters of the hexagonal cell. First design and development of solar blind UV photodetectors based on -phase was carried out and the optoelectronic performance is evaluated by means of photocurrent spectroscopy. The UV-response is adequately fast and reliable to render this unusual phase a subject of great interest for future applications. The availability of a hexagonal phase of Ga2O3 stable up to 700°C, belonging to the same space group of gallium nitride, with high crystallinity and tunable electrical properties, is intriguing in view of the development of nitride-based devices, by taking advantage of the more favorable symmetry and epitaxial relationships with respect to the monoclinic β-phase. In addition, annealing at temperatures higher than 700°C demonstrate that the hexagonal phase converts totally in the monoclinic one. 2) ZnO nano-tetrapods: charge transport mechanisms and time-response in optoelectronic devices and sensors Size and morphology of ZnO at the nanometer scale play a key role in tailoring its physical and chemical properties. Thanks to the possibility of growing zinc oxide in a variety of different nanostructures, there is a great variety of applications, among which gas sensors, light emitting diodes, transparent conducting oxides, solar cells. Even if the operation of ZnO nanostructure-based devices has been recently demonstrated, the mechanisms of charge transport in these assembly is still under debate. The candidate performed an accurate investigation by photocurrent spectroscopy and DC-photocurrent time evolution of electrical response of both single-tetrapod and tetrapod-assembly devices. During the research done for this thesis, a thermal activation energy enables the performance of samples at high temperatures (above about 300°C). The energy barrier is related to the leg-to-leg interconnection in the assembly of nanotetrapods. Percolation mechanisms are responsible for both the very slow photo-response (minutes to hours or days) and the significant persistent photocurrent. Below the bandgap energy, electronic states were investigated but their contribution to the photocurrent are two-three order of magnitude lower than the band edge. Such devices are suitable for employ in photodetectors as well as in gas sensors, provided that the mechanism by which the photo-current is generated and gas adsorption on the surface modify the conductivity of the material are known.

Estudo da intensificação da coalescência de emulsões de água em óleo com a aplicação de onda estacionária de ultrassom.

Considerando que o petróleo quando extraído dos poços em águas profundas chega a ter teor de água superior a 50% e que antes de ser enviado à refinaria deve ter uma quantidade de água inferior a 1%, torna-se necessário o uso de técnicas de redução da quantidade de água. Durante a extração do petróleo formam-se emulsões de água em óleo que são muito estáveis devido a um filme interfacial contendo asfaltenos e/ou resinas ao redor das gotas de água. Nesse trabalho é apresentada a utilização de ondas estacionárias de ultrassom para realizar a quebra dessas emulsões. Quando gotículas de água com dimensões da ordem de 10m, muito menores que o comprimento de onda, são submetidas a um campo acústico estacionário em óleo, a força de radiação acústica empurra as gotículas para os nós de pressão da onda. Uma célula de coalescência com frequência central ao redor de 1 MHz, constituída por quatro camadas sendo uma piezelétrica, uma de acoplamento sólido, uma com o líquido e outra refletora, foi modelada empregando o método da matriz de transferência, que permite calcular a impedância elétrica em função da frequência. Para minimizar o efeito do gradiente de temperatura entre a entrada e a saída da cavidade da célula, quando está em operação, foram utilizados dois transdutores piezelétricos posicionados transversalmente ao fluxo que são excitados e controlados independentemente. Foi implementado um controlador digital para ajustar a frequência e a potência de cada transdutor. O controlador tem como entrada o módulo e a fase da corrente elétrica no transdutor e como saída a amplitude da tensão elétrica e a frequência. Para as células desenvolvidas, o algoritmo de controle segue um determinado pico de ressonância no interior da cavidade da célula no intervalo de frequência de 1,09 a 1,15 MHz. A separação acústica de emulsões de água em óleo foi realizada em uma planta de laboratório de processamento de petróleo no CENPES/PETROBRAS. Foram testados a variação da quantidade de desemulsificante, o teor inicial de água na emulsão e a influência da vazão do sistema, com uma potência de 80 W. O teor final de água na emulsão mostrou que a aplicação de ultrassom aumentou a coalescência de água da emulsão, em todas as condições testadas, quando comparada a um teste sem aplicação de ultrassom. Identificou-se o tempo de residência no interior da célula de separação como um fator importante no processo de coalescência de emulsões de água e óleo. O uso de desemulsificante químico é necessário para realizar a separação, porém, em quantidades elevadas implicaria no uso de processos adicionais antes do repasse final do petróleo à refinaria. Os teores iniciais de água na emulsão de 30 e 50% indicam que o uso da onda estacionária na coalescência de emulsões não tem limitação quanto a esse parâmetro. De acordo com os resultados obtidos em laboratório, essa técnica seria indicada como uma alternativa para integrar um sistema de processamento primário em conjunto com um separador eletrostático.