Dissertations / Theses on the topic '3D widget'

Les surfaces tactiles ayant connu un grand essor ces dernières années, le grand public les utilise quotidiennement pour de multiples tâches, telles que la consultation d'e-mail, la manipulation de photos, etc. En revanche, très peu d'applications 3D existent sur ces dispositifs, alors que de telles applications pourraient avoir un grand potentiel dans des domaines variés, telles que la culture, l'architecture, ou encore l'archéologie. La difficulté majeure pour ce type d'applications est d'interagir avec un espace défini en trois dimensions à partir d'une modalité d'interaction définie en deux dimensions. Les travaux effectués dans cette thèse explorent l'association entre surfaces tactiles et manipulation de contenu 3D pour le grand public. Les premières études ont été réalisées afin de comprendre comment l'utilisateur réagit pour manipuler un objet virtuel 3D avec une surface tactile sans lui imposer de techniques d'interaction particulières. De par les connaissances acquises sur les utilisateurs, les travaux suivants présentent l'élaboration de nouvelles techniques d'interaction ainsi que leur évaluation.

Creating interactive prototypes can be a long and difficult process. It requires expertise in various fields. Prior work in developing interactive prototypes minimize time required to make a prototype, but generally sacrifice fidelity for fluidity. Advances in 3D printing create new opportunities to prototype with greater fidelity and fluidity. We investigate the use of several kinds of sensors, including IR photo interrupters, IR photo reflectors, push button switches, and potentiometers, to create interactive prototypes. We first design a library of 3D printable interaction components, buttons, sliders, and knobs using those sensors then we develop software to transform interaction events into events in computer programs. The combinations of interaction components and sensing devices are evaluated based on their durability and ability to be printed into prototypes and used as human-computer interface devices.

Video super-resolution (VSR) aims to recover a realistic high-resolution (HR) frame from its corresponding center low-resolution (LR) frame and several neighbouring supporting frames. The neighbouring supporting LR frames can provide extra information to help recover the HR frame. However, these frames are not aligned with the center frame due to the motion of objects. Recently, many video super-resolution methods based on deep learning have been proposed with the rapid development of neural networks. Most of these methods utilize motion estimation and compensation models as preprocessing to handle spatio-temporal alignment problem. Therefore, the accuracy of these motion estimation models are critical for predicting the high-resolution frames. Inaccurate results of motion compensation models will lead to artifacts and blurs, which also will damage the recovery of high-resolution frames. We propose an effective wide activated separate 3 dimensional (3D) Convolution Neural Network (CNN) for video super-resolution to overcome the drawback of utilizing motion compensation models. Separate 3D convolution factorizes the 3D convolution into convolutions in the spatial and temporal domain, which have benefit for the optimization of spatial and temporal convolution components. Therefore, our method can capture temporal and spatial information of input frames simultaneously without additional motion evaluation and compensation model. Moreover, the experimental results demonstrated the effectiveness of the proposed wide activated separate 3D CNN.

Yes
What has generated the unprecedented fascination with the Internet? What future lies ahead for computing as the Internet and its associated infrastructure expand? Will the network be able to cope with rising demands for carrying capacity and response speed? Will it change the way scientists, designers, artists, computer professionals, and home users work in the future? These are some of the wideranging questions being asked about the Internet and World Wide Web.

Thesis (S.M.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2007.
Includes bibliographical references (p. 85-89).
This thesis describes a method for real-time vision-based localization in human-made environments. Given a coarse model of the structure (walls, floors, ceilings, doors and windows) and a video sequence, the system computes the camera pose (translation and rotation) in model coordinates with an accuracy of a few centimeters in translation and a few degrees in rotation. The system has several novel aspects: it performs 6-DOF localization; it handles visually cluttered and dynamic environments; it scales well over regions extending through several buildings; and it runs over several hours without losing lock. We demonstrate that the localization problem can be split into two distinct problems: an initialization phase and a maintenance phase. In the initialization phase, the system determines the camera pose with no other information than a search region provided by the user (building, floor, area, room). This step is computationally intensive and is run only once, at startup. We present a probabilistic method to address the initialization problem using a RANSAC framework. In the maintenance phase, the system keeps track of the camera pose from frame to frame without any user interaction.
(cont.) This phase is computationally light-weight to allow a high processing frame rate and is coupled with a feedback loop that helps reacquire "lock" when lock has been lost. We demonstrate a simple, robust geometric tracking algorithm based on correspondences between 3D model lines and 2D image edges. We present navigation results on several real datasets across the MIT campus with cluttered, dynamic environments. The first dataset consists of a five-minute robotic exploration across the Robotics, Vision and Sensor Network Lab. The second dataset consists of a two-minute hand-held, 3D motion in the same lab space. The third dataset consists of a 26-minute exploration across MIT buildings 26 and 36.
by Olivier Koch.
S.M.

High-throughput sequencing has greatly influenced the amount of data produced and biological questions asked and answered. Sequencing approaches have also enabled rapid development of related technological fields such as single-cell and spatially resolved expression profiling. The introductory parts of this thesis give an overview of the basic molecular and technological apparatus needed to analyse the transcriptome in cells and tissues. This is succeeded by a summary of present investigations that report recent advancements in RNA profiling. RNA integrity needs to be preserved for accurate gene expression analysis. A method providing a low-cost alternative for RNA preservation was reported. Namely, a low concentration of buffered formaldehyde was used for fixation of human cell lines and peripheral blood cells (Paper I). The results from bulk RNA sequencing confirmed gene expression was not negatively impacted with the preservation procedure (r2>0.88) and that long-term storage of such samples was possible (r2=0.95). However, it is important to note that a small population of cells overexpressing a limited amount of genes can skew bulk gene expression analyses making them sufficient only in carefully designed studies. Therefore, gene expression should be investigated at the single cell resolution when possible. A method for high-throughput single cell expression profiling termed microarrayed single-cell sequencing was developed (Paper II). The method incorporated fluorescence-activated cell sorting, sample deposition and profiling of thousands of barcoded single cells in one reaction. After sample attachment to a barcoded array, a high-resolution image was taken which linked the position of each array barcode sequence to each individual deposited cell. The cDNA synthesis efficiency was estimated at 17.3% while detecting 27,427 transcripts per cell on average. Additionally, spatially resolved analysis is important in cell differentiation, organ development and pathological changes. Current methods are limited in terms of throughput, cost and time. For that reason, the spatial transcriptomics method was developed (Paper III). Here, the barcoded microarray was used to obtain spatially resolved expression profiles from tissue sections using the same imaging principle. The mouse olfactory bulb was profiled on a whole-transcriptome scale and the results showed that the expression correlated well (r2=0.94-0.97) as compared to bulk RNA sequencing. The method was 6.9% efficient, reported signal diffusion at ~2 μm and accurately deconvoluted layer-specific transcripts in an unbiased manner. Lastly, the spatial transcriptomics concept was applied to profile human breast tumours in three dimensions (Paper IV). Unbiased clustering revealed previously un-annotated regions and classified them as parts of the immune system, providing a detailed view into complex interactions and crosstalk in the whole tissue volume. Spatial tumour classification divulged that certain parts of the tumour clearly classified as other subtypes as compared to bulk analysis providing useful data for current practice diagnostics. The last part of the thesis discusses a look towards the future, how the presented methods could be used, improved upon or combined in translational research.

QC 20170109

Climate change during the last 2.5 million years is characterised by glacial-interglacial cycles of fluctuating sea level and temperature increasing in magnitude and duration towards the present day. The central North Sea preserves these glacial-interglacial cycles in an expanded sedimentary sequence creating a high resolution palaeo-climatic record. Basin-wide, low-resolution 3D seismic data, covering more than 80,000 km2 of the central North Sea, is combined with high-resolution, broadband 3D seismic, regional 2D seismic and local ultra-high resolution seismic from the Dogger Bank windfarm development zone in order to investigate in full the sedimentary sequence. The evolution of the basin is analysed along with the preserved geomorphological landforms in order to build a framework for the development of the North Sea and its changing palaeo-environments from the inception of the Quaternary (2.58 Ma) until the extensive glacial unconformity formed during the Elsterian (0.48 Ma).At the onset of the Quaternary the structure of the North Sea was that of an elongate marine basin, rapidly infilled from the south by continued progradation of the large clinoformal deposits of the southern North Sea deltaic system. The basin rapidly decreased in extent and depth however it was not until around 1.1 Ma that the broad, shallow shelf of the present day was fully established. A revision of the current seismic stratigraphy is proposed, identifying four new Members within the Aberdeen Ground Formation taking into account the development of the basin through time. Powerful downslope gravity currents dominated the basin during much of the early Quaternary, although a well-established, anti-clockwise tidal gyre acted to gently modify the gravity currents. Iceberg scouring was nearly continual from the onset of the Quaternary until grounded ice sheets began to penetrate into the basin from 1.7 Ma, more than half a million years before any previous estimates. Effects of confluence of the British and Fennoscandian ice sheets are observed from 1.3 Ma. The tunnel valleys of the Dogger Bank represent a continuation of the North Sea tunnel valley network, interacting with both older glaciotectonic thrusting and younger glaciotectonic folded deformation.

Les télécommunications sont devenus indispensables dans notre monde actuel. De plus, le volume des données échangées ne cesse de croître. En effet, nous pouvons transmettre nos photos, nos vidéos au monde entier. Nonobstant, nous ne voulons pas attendre pour les avoir, ce qui exige un débit de données très important et par conséquent des signaux avec des bandes passantes plus larges. Les satellites de télécommunications doivent donc s’adapter, c'est pourquoi nous proposons dans ces travaux la recherche de filtre à large bande avec une recherche de compacité et de faibles pertes. Nous nous sommes intéressés à l'utilisation de matériaux céramiques qui permettent d'obtenir de bonnes performances vis à vis de nos besoins. Notre travail est aussi rendu possible par le développement de procédés de fabrication additifs, comme par exemple la stéréolithographie, qui va nous permettre de nous affranchir fortement de règles de dessin contraignantes que nous pourrions avoir en utilisant des procédés classiques. Nous avons développé des filtres avec de larges bandes passantes autour de 4GHz. Une première étude nous a permis de rechercher des concepts qui permettent d'obtenir de forts couplages, conditions sine qua non pour réaliser ces filtres. Plusieurs concepts sont présentés ainsi que leur fabrication et leur mesures. Nous avons ainsi démontré expérimentalement que les concepts proposés, à base de pièces monoblocs céramiques, sont capables de produire des filtres à bandes passantes supérieures à 60 % (voire même 110 % pour une version améliorée)
Every day, the data exchanges increase thanks to the new technologies. We can keep our files, our pictures, our videos online to have an access anywhere on the planet (for now). In this way, the data output of the telecommunication systems has to be increased in order to satisfy the more and more demanding users. One way to allow this is to increase the bandwidths of the different signals, making possible to transmit more data at the same time. In this work, we will develop wide bandpass filters dedicated to space telecommunications. For that purpose, we need them to be compact, with low insertion loss and a limited number of parts to assemble. Consequently, we are interested to use resonators made with ceramic materials that permits to reach such properties. Moreover, these materials are compatible with stereolithography, an additive manufacturing process. Such technology is here very useful for our purpose since its design freedom allows the creation of almost all kind of geometries. To realize such wide bandpass filters, we need strong couplings between the different resonators and also for the accesses, so we will present our studies focused on reaching these specific objectives. Then, we will present different designs of wide bandpass filter around 4GHz. After different generation of ceramic based components, we are be able to experimentally create a 60% bandwidth (even 100% for our last version) very compact bandpass filter filling the objectives of this PhD thesis

Le travail présenté dans cette thèse s'inscrit à la fois dans le domaine de l'informatique musicale et dans celui de la réalité virtuelle. Nous présentons tout d'abord les avantages des environnements virtuels immersifs et de l'interaction graphique 3D pour les instruments multi-processus et en particulier pour le live-looping hiérarchique, un nouveau modèle d'instrument que nous définissons. Nous proposons ensuite des solutions pour la visualisation et la manipulation des processus sonores dans ces environnements virtuels. Afin de représenter les processus sonores, nous introduisons les widgets réactifs 3D qui s'appuient sur des mappings audiovisuels que nous évaluons. Nous développons un nouveau dispositif d'interaction 3D adapté à l'interaction musicale, appelé Piivert, ainsi que des techniques associées, dont les tunnels afin de manipuler les widgets réactifs 3D. Finalement, nous mettons en œuvre le live-looping hiérarchique avec l'instrument immersif Drile, en utilisant les techniques développées précédemment, et nous étudions plusieurs scènes pour les performances musicales immersives.

The most common application of hypertext today is found on the World Wide Web, with the numbers of sites and potential users increasing continually. Hypertext systems are characterized by hyperlinks that allow users nonsequential access to the documents contained within them. Because users are not constrained to read through these documents in a linear manner, new problems may arise not found in traditional paper versions. These problems are generally characterized by a sense of being "lost" within a hypertext system. The user may not know exactly where they are or how to get where they want to be. They have lost many of the visual cues that indicate position found in ordinary texts like books. Navigational aids or overview maps have been suggested as a means to help counteract this problem. These navigational aids can take a standard table-of-contents and extend it dynamically or provide a completely new paradigm of browsing. To that extent, many new varieties of maps (including three-dimensional ones) have been developed but not thoroughly studied. As well, it may be theorized that users of differing cognitive abilities may be helped or hindered by such devices. An empirical study was performed to investigate the effect of multidimensional maps. Three different navigational aids were examined which varied the way pages are displayed along one, two or three dimensions. Two hypertext systems were also where one was roughly twice the size of the other. The participants were given a search task twice to examine performance on page revisits. Finally, three cognitive tests were given to view the effects of individual differences. These included a spatial ability, verbal ability and visual memory test. The results indicated that no performance differences existed between the different navigational aids. However, a significant interaction was present between the maps and the type of Web site; smaller Web sites benefited from the 3D navigational aid. It is theorized that an observed effect for hypertext system was due to site complexity as opposed to size. The results from the cognitive ability measures were mixed. People with low verbal ability scores took longer to locate answers. People with high spatial ability scores found more answers and had scores that were less sensitive to the type of navigational aid used. No significant differences were discovered between people of high and low visual memory abilities.
Master of Science

Les nouveaux composants à semi-conducteur de type grand gap ont été développés pour des applications de conversion de puissance en raison de leurs hautes fréquences de commutation (de centaine kHz à quelques MHz) et pertes faibles. Afin de bien profiter ses avantages, la technologie des circuits imprimés (PCB) est intéressante pour une intégration à haute densité de puissance grâce à sa flexibilité et son faible coût. Cependant, à cause de la mauvaise conductivité thermique du matériau FR-4 utilisé pour le substrat PCB et la haute densité de puissance réalisée, il est primordial de trouver des solutions thermiques pour améliorer les performances thermiques de la structure de PCB. Dans cette thèse, trois solutions thermiques pour les structures de PCB ont été proposées, y compris des solutions avec des vias thermiques, de cuivre épais sur le substrat de PCB ainsi que des dispositifs de refroidissement thermoélectrique (TEC). Nos études sont basées sur la modélisation électrothermique et la méthode d’éléments finis en 3D. Tout d’abord, l’optimisation des paramètres des vias (diamètre, épaisseur de placage, surface formée par des vias, la distance entre des vias etc.) a été réalisée pour optimiser l’effet de refroidissement. Ensuite, on constate que les performances thermiques des structures de PCB peuvent être améliorées en utilisant cuivre épais sur le substrat de PCB. Cuivre épais augmente le flux thermique latéral dans la couche de cuivre. Les influences de l’épaisseur de cuivre (35 à 500 µm) ont été étudiées. Cette solution est facile à réaliser et peut être combinée à d’autres solutions de refroidissement. Enfin, le dispositif thermoélectrique comme les modules Peltier est une technologie de refroidissement local. Les influences des paramètres de Peltier (Propriétés du matériau thermoélectrique, nombre d’éléments Peltier, distance entre la source de chaleur et les dispositifs Peltier, etc.) ont été identifiées. Il est démontré que des modules Peltier ont l’application potentielle pour le développement d’intégration de PCB attendu que son active contrôle des températures
The emerging wide bandgap (WBG) semiconductor devices have been developed for power conversion applications instead of silicon devices due to higher switching frequencies (from few 100 kHz to several MHz) and lower on-state losses resulting in a better efficiency. In order to take full advantage of the WBG components, PCB technology is attractive for high power density integration thanks to its flexibility and low cost. However, due to poor thermal conductivity of the commonly used material Flame Retardant-4 (FR4), efficient thermal solutions are becoming a challenging issue in integrated power boards based on PCB substrates. So it is of the first importance to seek technological means in order to improve the thermal performances. In this thesis, three main thermal management solutions for PCB structures have been investigated including thermal vias, thick copper thickness on the PCB substrate as well as thermoelectric cooling (TEC) devices. Our studies are based on the electro-thermal modeling and 3D finite element (FE) methods. Firstly, optimization of the thermal via parameters (via diameter, via plating thickness, via-cluster surface, via pattern, pitch distance between vias etc.) has been realized to improve their cooing performances. We presented and evaluated thermal performances of the PCB structures by analyzing the thermal resistance of the PCB substrate with different thermal vias. Secondly, it is found that thermal performances of the PCB structures can be enhanced by using thick copper thickness on top of the PCB substrate, which increases the lateral heat flux along the copper layer. Influences of the copper thickness (35 µm to 500 µm) has been discussed. This solution is easy to realize and can be combined with other cooling solutions. Thirdly, thermoelectric cooler like Peltier device is a solid-state cooling technology that can meet the local cooling requirements. Influences of Peltier parameters (Thermoelectric material properties, number of Peltier elements, distance between the heating source and the Peltier devices etc.) have been identified. All these analyses demonstrate the potential application of Peltier devices placed beside the heating source for PCB structures, which is a benefit for developing the embedding technology in such structures

Cette thèse a été réalisée dans le secteur de l'industrie automobile, en collaboration avec le Groupe Renault et concerne en particulier le développement de systèmes d'aide à la conduite avancés et de véhicules autonomes. Les progrès réalisés par la communauté scientifique durant les dernières décennies, dans les domaines de l'informatique et de la robotique notamment, ont été si importants qu'ils permettent aujourd'hui la mise en application de systèmes complexes au sein des véhicules. Ces systèmes visent dans un premier temps à réduire les risques inhérents à la conduite en assistant les conducteurs, puis dans un second temps à offrir des moyens de transport entièrement autonomes. Les méthodes de SLAM multi-objets actuellement intégrées au sein de ces véhicules reposent pour majeure partie sur l'utilisation de capteurs embarqués très performants tels que des télémètres laser, au coût relativement élevé. Les caméras numériques en revanche, de par leur coût largement inférieur, commencent à se démocratiser sur certains véhicules de grande série et assurent généralement des fonctions d'assistance à la conduite, pour l'aide au parking ou le freinage d'urgence, par exemple. En outre, cette implantation plus courante permet également d'envisager leur utilisation afin de reconstruire l'environnement dynamique proche des véhicules en trois dimensions. D'un point de vue scientifique, les techniques de SLAM visuel multi-objets existantes peuvent être regroupées en deux catégories de méthodes. La première catégorie et plus ancienne historiquement concerne les méthodes stéréo, faisant usage de plusieurs caméras à champs recouvrants afin de reconstruire la scène dynamique observée. La plupart reposent en général sur l'utilisation de paires stéréo identiques et placées à faible distance l'une de l'autre, ce qui permet un appariement dense des points d'intérêt dans les images et l'estimation de cartes de disparités utilisées lors de la segmentation du mouvement des points reconstruits. L'autre catégorie de méthodes, dites monoculaires, ne font usage que d'une unique caméra lors du processus de reconstruction. Cela implique la compensation du mouvement propre du système d'acquisition lors de l'estimation du mouvement des autres objets mobiles de la scène de manière indépendante. Plus difficiles, ces méthodes posent plusieurs problèmes, notamment le partitionnement de l'espace de départ en plusieurs sous-espaces représentant les mouvements individuels de chaque objet mobile, mais aussi le problème d'estimation de l'échelle relative de reconstruction de ces objets lors de leur agrégation au sein de la scène statique. La problématique industrielle de cette thèse, consistant en la réutilisation des systèmes multi-caméras déjà implantés au sein des véhicules, majoritairement composés d'un caméra frontale et de caméras surround équipées d'objectifs très grand angle, a donné lieu au développement d'une méthode de reconstruction multi-objets adaptée aux systèmes multi-caméras hétérogènes en stéréo wide-baseline. Cette méthode est incrémentale et permet la reconstruction de points mobiles éparses, grâce notamment à plusieurs contraintes géométriques de segmentation des points reconstruits ainsi que de leur trajectoire. Enfin, une évaluation quantitative et qualitative des performances de la méthode a été menée sur deux jeux de données distincts, dont un a été développé durant ces travaux afin de présenter des caractéristiques similaires aux systèmes hétérogènes existants
This Ph.D. thesis, which has been carried out in the automotive industry in association with Renault Group, mainly focuses on the development of advanced driver-assistance systems and autonomous vehicles. The progress made by the scientific community during the last decades in the fields of computer science and robotics has been so important that it now enables the implementation of complex embedded systems in vehicles. These systems, primarily designed to provide assistance in simple driving scenarios and emergencies, now aim to offer fully autonomous transport. Multibody SLAM methods currently used in autonomous vehicles often rely on high-performance and expensive onboard sensors such as LIDAR systems. On the other hand, digital video cameras are much cheaper, which has led to their increased use in newer vehicles to provide driving assistance functions, such as parking assistance or emergency braking. Furthermore, this relatively common implementation now allows to consider their use in order to reconstruct the dynamic environment surrounding a vehicle in three dimensions. From a scientific point of view, existing multibody visual SLAM techniques can be divided into two categories of methods. The first and oldest category concerns stereo methods, which use several cameras with overlapping fields of view in order to reconstruct the observed dynamic scene. Most of these methods use identical stereo pairs in short baseline, which allows for the dense matching of feature points to estimate disparity maps that are then used to compute the motions of the scene. The other category concerns monocular methods, which only use one camera during the reconstruction process, meaning that they have to compensate for the ego-motion of the acquisition system in order to estimate the motion of other objects. These methods are more difficult in that they have to address several additional problems, such as motion segmentation, which consists in clustering the initial data into separate subspaces representing the individual movement of each object, but also the problem of the relative scale estimation of these objects before their aggregation within the static scene. The industrial motive for this work lies in the use of existing multi-camera systems already present in actual vehicles to perform dynamic scene reconstruction. These systems, being mostly composed of a front camera accompanied by several surround fisheye cameras in wide-baseline stereo, has led to the development of a multibody reconstruction method dedicated to such heterogeneous systems. The proposed method is incremental and allows for the reconstruction of sparse mobile points as well as their trajectory using several geometric constraints. Finally, a quantitative and qualitative evaluation conducted on two separate datasets, one of which was developed during this thesis in order to present characteristics similar to existing multi-camera systems, is provided

Cette thèse a pour but d’étudier le principe d’un multiplexeur de sortie accordable pour la charge utile d’un satellite de télécommunication. La première étape consiste à analyser les principales topologies de multiplexeur de sortie utilisables et les comparer pour en retirer le meilleur candidat pour cette application. Par la suite, diverses études ont été menées sur des composants passifs imprimés en 3D plastique et métallisés de diverses manières, en cherchant à comprendre comment améliorer au mieux les résultats de ces prototypes. La dernière partie de ce manuscrit détaille la conception d’un multiplexeur à deux canaux pouvant être utilisé sur trois états différents en large bande à 19 GHz. Ces états sont créés par des bandes passantes relatives variant de 1,6 à 4,8 % et une réalisation de deux maquettes sert de preuves de concept et de performance (une en plastique métallisé et la seconde en aluminium usiné)
This thesis is about study of tunable output multiplexer principle for telecommunication satellite payload. The first step is analysis of the main topologies of output multiplexers and to compare them to conclude with the best candidate. Then, several studies were conducted on 3D plastic printed passive components with specific metallizations, searching how to improve results in term of losses and precision. Last part detailed two channels multiplexer design which can be used on three different states of relative bandwidth from 1.6 to 4.8 % at 19 GHz, and realization of two concept proof prototypes (one in metallized 3D printed plastic and another one in machined aluminum)

La diminution de la longueur de grille des transistors a été le moteur essentiel de l’évolution des circuits intégrés microélectroniques ces dernières décennies. Toutefois, cette évolution des circuits microélectroniques a entrainé une densification des lignes d’interconnexion, donc la génération de fortes pertes, des ralentissements et de la diaphonie sur les signaux transmis, ainsi qu’une augmentation de l’impédance parasite des interconnexions. Cette dernière est néfaste pour l’intégrité de l’alimentation des composants actifs présents dans le circuit. Son augmentation multiplie le risque d’apparition d’erreurs numériques conduisant au dysfonctionnement d’un système. Il est donc nécessaire de réduire l’impédance sur le réseau d’alimentation des circuits intégrés. Pour ce faire, les condensateurs de découplage sont utilisés et placés hiérarchiquement à différents étages des circuits et dans leur intégralité (PCB, package, interposeur, puce).Ces travaux de doctorat s’inscrivent dans le cadre des développements récents des nouvelles solutions d’intégration 3D en microélectronique et ils portent sur l’étude de nouvelles architectures de capacités 3D, très intégrées et à fortes valeurs (>1 nF), élaborées en profondeur dans l’interposeur silicium. Ces composants, inspirés des architectures de via traversant le silicium (TSV, Through Silicon Via), sont nommées Through Silicon Capacitors (TSC). Ils constituent un élément clef pour l’amélioration des performances des alimentations des circuits intégrés car elles pourront réduire efficacement la consommation des circuits grâce à cette intégration directe de composants passifs dans l’interposeur silicium qui sert d’étage d’accueil des puces. Ces composants tridimensionnels permettent en effet d’atteindre de grandes densités de capacité de 35 nF/mm². Les enjeux sont stratégiques pour des applications embarquées et à haut débit et plus généralement dans un environnement économique et sociétal conscient de nos limites énergétiques. De plus ces condensateurs de découplage doivent fonctionner à des fréquences atteignant 2 GHz, voire 4 GHz, qui tendent à maximiser les effets parasites préjudiciables aux performances énergétiques des alimentations. Ceci est rendu possible par l’optimisation de leur intégration et l’utilisation de couches de cuivre avec, une bonne conductivité supérieure à 45 MS/m, comme électrodes.Les technologies d’élaboration des condensateurs TSC ont été développées au sein du CEA-LETI et de STMicroelectronics. Leur comportement électrique restait jusqu’alors mal connu et leurs performances difficiles à quantifier. Les études menées dans cette thèse consistaient à modéliser ces nouveaux composants en prenant en compte les paramètres matériaux et géométriques afin de connaitre les effets parasites. Les modèles électriques établis ont été confrontés à des caractérisations électriques effectuées sur une large bande de fréquence (du DC à 40 GHz). Ainsi ce travail a permis d’optimiser une architecture de capacité et leur intégration dans un réseau d’alimentation d’un circuit intégré 3D a pu montrer leur efficacité pour des opérations de découplage
The decrease of transistor’s gate length was the key driver of the development of microelectronic integrated circuits in recent decades. However, this development of microelectronic circuits has led to a greater density of interconnection lines, generating high losses, slowdowns and crosstalk on the transmitted signals, and an increase of the parasitic impedance of interconnections lines. The latter is detrimental to the power integrity of the active components in the circuit. Its increase increases the risk of developing numerical errors leading to a system’s malfunction. It is therefore necessary to reduce the impedance of the power distribution network of integrated circuits. To do this, the decoupling capacitors are used and placed hierarchically on different floors of the circuits and in their entirety (PCB, package, interposer, chip).These doctoral works are in the context of recent developments in new 3D integration solutions in microelectronics and they carry on studying new 3D capacitors, highly integrated, presenting high capacitance values (> 1 nF), and developed by using the depth of silicon interposeur level. Inspired from the Through Silicon Vias (TSV), these newly developed 3D capacitors are named Through Silicon Capacitors (TSC). They are a key element for improving the performance of the power integrated circuits because they can efficiently reduce the consumption of circuits thanks to their direct integration in silicon interposer which is used to stack chips. These 3D components allow tor reach high capacitance density up to 35 nF/mm². The issues are strategic for high speed embedded applications and more generally in an economic and societal environment aware of our energy limits. Moreover these decoupling capacitors must operate at frequencies up to 2 GHz or 4 GHz, which tend to maximize the parasitic effects which affect the energy efficiency of power distribution networks. This is made possible by optimizing their integration and by the use of copper layers with a good conductivity higher than 45 MS / m conductivity as electrodes.The technologies used to fabricate the TSC are developed by CEA-LETI and STMicroelectronics. The electrical behavior of those TSC remained hitherto little known and their performances difficult to quantify. The studies conducted in this thesis were to model these new components by taking into account the material and geometrical parameters in order to know the parasitic effects. The established electrical models have faced electrical characterizations carried out over a wide frequency range (DC to 40 GHz). This work allow to optimize the TSC architecture and their integration in a power distribution network (Power Distribution Network - NDS) prove that they are good candidate for decoupling operations

The goal of this thesis is to scan a ship hull with high 3D accuracy and resolution using an underwater stereo camera so as to enable the future autonomous detection of invasive biofouling organisms with autonomous underwater vehicles (AUVs). However, turbidity in most harbours necessitates being within a metre of the hull and thus requires ultra wide-angle camera lenses. But such ultra wide-angle lenses embedded in an underwater housing with a flat port lead to significant distance dependent image distortions. Prior research in this area has only considered narrower fields of view and so has not solved for the significant image distortions arising from wide-angle high resolution flat port underwater cameras. This thesis proposes a solution to modelling and calibrating the underwater camera for accurate 2D imaging and 3D reconstruction, and additionally demonstrates an accurate underwater real-time pose estimation system required for future ship hull relative AUV navigation. In this thesis an ultra wide-angle, short-baseline stereo camera is used, which is embedded in a flat port underwater housing. Flat port underwater housings represent a cost efficient way to use arbitrary in-air cameras underwater. However, the flat port of the underwater housing is subject to light refraction and causes distance dependent distortion, which is particularly visible at the large angles of the ultra wide-angle stereo camera used. To incorporate the effects of refraction, the thesis uses the well-known and accurate physics-based refractive underwater camera model. In contrast to the perspective camera-based underwater camera model, the refractive underwater camera model accurately describes the distance dependency of distortion. In the beginning of this thesis, the effects of refraction caused by a thick flat port underwater housing are summarised and extended. In this context, the fundamental magnification function is proposed, which enables the description of numerous known and also newly discovered effects. An additional quantitative analysis is carried out in which the importance to model the thickness of the port and the wavelength of light is revealed. In refractive geometry with a thick flat port, refractive forward projection represents a fundamental operation and describes where a 3D object point is observed in a 2D camera image. Refractive forward projection is required in numerous applications, such as refractive calibration, bundle adjustment, simultaneous localisation and mapping (SLAM) or image restoration. Unlike perspective projection in air, this operation is non-linear and computationally more expensive. This thesis compares existing and proposes new refractive forward projection methods and shows in contrast to previous research that refractive forward projection is efficient enough for real-time applications. The thesis also investigates the impact of the port and the impact of the indices of refraction on the camera's projection and reconstruction accuracy. A novel investigation shows that the water pressure, water salinity, water temperature, air pressure and the wavelength of light significantly affect the projection and reconstruction accuracy of wide-angle flat port underwater stereo cameras and should not be neglected by standard refractive indices. Moreover, this thesis proposes an accurate and efficient calibration method for thick flat port underwater stereo cameras. The proposed calibration method mainly achieves its high accuracy by the use of a significantly higher number of calibration images. In contrast to prior research, the computation of the reprojection error does not represent a bottleneck if the proposed refractive forward projection method is used. In this way, the calibration is similar to standard in-air camera calibration techniques and minimises the reprojection error. In combination with the proposed more accurate indices of refraction and refractive calibration, the underwater reconstruction accuracy of the novel configuration of a wide-angle flat port underwater short baseline stereo camera is evaluated under real-world conditions. In this context, a method is proposed, which enables the evaluation of the accuracy of the reconstructed 3D object space. Both chromatic aberration and pincushion distortion are effects of refraction and are particularly visible at the large angles of wide-angle underwater cameras. In order to obtained distortion-free images with minimised chromatic aberration to texturize reconstructed 3D ship hull surfaces, this thesis proposes accurate real-time methods to minimise chromatic aberration and to correct the distortion in the underwater camera images. The refractive camera model is based on image coordinates of images, which are distortion-free in air. But these in-air undistorted images are strongly distorted in-water by refraction, particularly at the large angles of wide-angle flat port underwater cameras. Image correspondence in these images is difficult. For that reason, this thesis proposes pseudo rectified images in which these distortions are minimised. Moreover, an accurate and efficient representation of epipolar curves is presented, which enables, for example, real-time constrained correspondence search or dense stereo. This thesis concludes with the demonstration of a pose estimation system for future ship hull relative navigation. The proposed pose estimation system is the first underwater SLAM and visual odometry system, which is based on the more accurate refractive underwater camera model. This thesis shows that the proposed pose estimation system is very accurate in a water tank experiment and efficiently works in real-time, and thus is superior to prior underwater SLAM research, which is based on the less accurate perspective camera-based underwater camera model.

Three-dimensional information is often used in robotics and 3D-mapping. There exist several ways to obtain a three-dimensional map. However, the time of flight used in the laser scanners or the structured light utilized by Kinect-like sensors sometimes are not sufficient. In this thesis, we investigate two CNN based stereo matching methods for obtaining 3D-information from a grayscaled pair of rectified images.While the state-of-the-art stereo matching method utilize a Siamese architecture, in this project a two-channel and a two stream network are trained in an attempt to outperform the state-of-the-art. A set of experiments were performed to achieve optimal hyperparameters. By changing one parameter at the time, the networks with architectures mentioned above are trained. After a completed training the networks are evaluated with two criteria, the error rate, and the runtime.Due to time limitations, we were not able to find optimal learning parameters. However, by using settings from [17] we train a two-channel network that performed almost on the same level as the state-of-the-art. The error rate on the test data for our best architecture is 2.64% while the error rate for the state-of-the-art Siamese network is 2.62%. We were not able to achieve better performance than the state-of-the-art, but we believe that it is possible to reduce the error rate further. On the other hand, the state-of-the-art Siamese stereo matching network is more efficient and faster during the disparity estimation. Therefore, if the time efficiency is prioritized, the Siamese based network should be considered.

La present tesi doctoral s’ha centrat en el repte d’aconseguir, mitjançant Fabricació Additiva (FA), models per a assaig quirúrgic, sota la premissa que els equips per fer-los haurien de ser accessibles a l’àmbit hospitalari. L’objectiu és facilitar l’extensió de l’ús dels prototips com a eina de preparació d’operacions quirúrgiques, transformant la pràctica mèdica actual de la mateixa manera que en el seu moment ho van fer tecnologies com les que van facilitar l’ús de radiografies. El motiu d’utilitzar FA, en lloc de tecnologies més tradicionals, és la seva capacitat de materialitzar de forma directa les dades digitals obtingudes de l’anatomia del pacient mitjançant sistemes d’escanejat tridimensional, fent possible l’obtenció de models personalitzats. Els resultats es centren en la generació de nou coneixement sobre com aconseguir equipaments d’impressió 3D multimaterials accessibles que permetin l’obtenció de models mimètics respecte als teixits vius. Per facilitar aquesta buscada extensió de la tecnologia, s’ha focalitzat en les tecnologies de codi obert com la Fabricació per Filament Fos (FFF) i similars basades en líquids catalitzables. La recerca s’alinea dins l’activitat de desenvolupament de la FA al CIM UPC, i en aquest àmbit concret amb la col·laboració amb l’Hospital Sant Joan de Déu de Barcelona (HSJD). El primer bloc de la tesi inclou la descripció de l’estat de l’art, detallant les tecnologies existents i la seva aplicació a l’entorn mèdic. S’han establert per primer cop unes bases de caracterització dels teixits vius -sobretot tous- per donar suport a la selecció de materials que els puguin mimetitzar en un procés de FA, a efectes de millorar l’experiència d’assaig dels cirurgians. El caràcter rígid dels materials majoritàriament usats en impressió 3D els fa poc útils per simular tumors i altres referències anatòmiques. De forma successiva, es tracten paràmetres com la densitat, la viscoelasticitat, la caracterització dels materials tous a la indústria, l’estudi del mòdul elàstic de teixits tous i vasos, la duresa d’aquests, i requeriments com l’esterilització dels models. El segon bloc comença explorant la impressió 3D mitjançant FFF. Es classifiquen les variants del procés des del punt de vista de la multimaterialitat, essencial per fer models d’assaig quirúrgic, diferenciant entre solucions multibroquet i de barreja al capçal. S’ha inclòs l’estudi de materials (filaments i líquids) que serien més útils per mimetitzar teixits tous. Es constata com en els líquids, en comparació amb els filaments, la complexitat del treball en processos de FA és més elevada, i es determinen formes d’imprimir materials molt tous. Per acabar, s’exposen sis casos reals de col·laboració amb l’HJSD, una selecció d’aquells en els que el doctorand ha intervingut en els darrers anys. L’origen es troba en la dificultat de l’abordatge d’operacions de resecció de tumors infantils com el neuroblastoma, i a la iniciativa del Dr. Lucas Krauel. Finalment, el Bloc 3 té per objecte explorar nombrosos conceptes (fins a 8), activitat completada al llarg dels darrers cinc anys amb el suport dels mitjans del CIM UPC i de l’activitat associada a treballs finals d’estudis d’estudiants de la UPC, arribant-se a materialitzar equipaments experimentals per validar-los. La recerca ampla i sistemàtica al respecte fa que s’estigui més a prop de disposar d’una solució d’impressió 3D multimaterial de sobretaula. Es determina que la millor via de progrés és la de disposar d’una pluralitat de capçals independents a fi de capacitar la impressora 3D per integrar diversos conceptes estudiats, materialitzant-se una possible solució. Cloent la tesi, es planteja com seria un equipament d’impressió 3D per a models d’assaig quirúrgic, a fi de servir de base per a futurs desenvolupaments.
La presente tesis doctoral se ha centrado en el reto de conseguir, mediante Fabricación Aditiva (FA), modelos para ensayo quirúrgico, bajo la premisa que los equipos para obtenerlos tendrían que ser accesibles al ámbito hospitalario. El objetivo es facilitar la extensión del uso de modelos como herramienta de preparación de operaciones quirúrgicas, transformando la práctica médica actual de la misma manera que, en su momento, lo hicieron tecnologías como las que facilitaron el uso de radiografías. El motivo de utilizar FA, en lugar de tecnologías más tradicionales, es su capacidad de materializar de forma directa los datos digitales obtenidos de la anatomía del paciente mediante sistemas de escaneado tridimensional, haciendo posible la obtención de modelos personalizados. Los resultados se centran en la generación de nuevo conocimiento para conseguir equipamientos de impresión 3D multimateriales accesibles que permitan la obtención de modelos miméticos respecto a los tejidos vivos. Para facilitar la buscada extensión de la tecnología, se ha focalizado en las tecnologías de código abierto como la Fabricación por Hilo Fundido (FFF) y similares basadas en líquidos catalizables. Esta investigación se alinea dentro de la actividad de desarrollo de la FA en el CIM UPC, y en este ámbito concreto con la colaboración con el Hospital Sant Joan de Déu de Barcelona (HSJD). El primer bloque de la tesis incluye la descripción del estado del arte, detallando las tecnologías existentes y su aplicación al entorno médico. Se han establecido por primera vez unas bases de caracterización de los tejidos vivos – principalmente blandos – para dar apoyo a la selección de materiales que los puedan mimetizar en un proceso de FA, a efectos de mejorar la experiencia de ensayo de los cirujanos. El carácter rígido de los materiales mayoritariamente usados en impresión 3D los hace poco útiles para simular tumores y otras referencias anatómicas. De forma sucesiva, se tratan parámetros como la densidad, la viscoelasticidad, la caracterización de materiales blandos en la industria, el estudio del módulo elástico de tejidos blandos y vasos, la dureza de los mismos, y requerimientos como la esterilización de los modelos. El segundo bloque empieza explorando la impresión 3D mediante FFF. Se clasifican las variantes del proceso desde el punto de vista de la multimaterialidad, esencial para hacer modelos de ensayo quirúrgico, diferenciando entre soluciones multiboquilla y de mezcla en el cabezal. Se ha incluido el estudio de materiales (filamentos y líquidos) que serían más útiles para mimetizar tejidos blandos. Se constata como en los líquidos, en comparación con los filamentos, la complejidad del trabajo en procesos de FA es más elevada, y se determinan formas de imprimir materiales muy blandos. Para acabar, se exponen seis casos reales de colaboración con el HJSD, una selección de aquellos en los que el doctorando ha intervenido en los últimos años. El origen se encuentra en la dificultad del abordaje de operaciones de resección de tumores infantiles como el neuroblastoma, y en la iniciativa del Dr. Lucas Krauel. Finalmente, el Bloque 3 desarrolla numerosos conceptos (hasta 8), actividad completada a lo largo de los últimos cinco años con el apoyo de los medios del CIM UPC y de la actividad asociada a trabajos finales de estudios de estudiantes de la UPC, llegándose a materializar equipamientos experimentales para validarlos. La investigación amplia y sistemática al respecto hace que se esté más cerca de disponer de una solución de impresión 3D multimaterial de sobremesa. Se determina que la mejor vía de progreso es la de disponer de una pluralidad de cabezales independientes, a fin de capacitar la impresora 3D para integrar diversos conceptos estudiados, materializándose una posible solución. Para cerrar la tesis, se plantea cómo sería un equipamiento de impresión 3D para modelos de ensayo quirúrgico, a fin de servir de base para futuros desarrollos.

The recent evolution of radio communications combined with innovative manufacturing techniques, such as 3D printing, has driven antennas development and implementation of new structures made of unusual materials. An example of this type of evolution are the lens antennas. Lens antennas are always associated with a source antenna (usually a microstrip patch antenna) and allow changing the source antenna’s radiation characteristics (varying the gain or directivity). Thus, lenses can improve the performance of some types of communication systems, such as phased arrays, which are used for beamforming. However, they have some limitations in coverage due to the array elements having low directivity. The use of a lens antenna changes the radiation diagram to obtain a wider beamwidth and is a potential solution to the problem of phased arrays. Throughout this dissertation, was studied the possibility of using lens antennas to change the radiation beam and increase the beamwidth of a simple microstrip patch antenna. For this purpose, simulations of several lens antenna structures were performed with a patch antenna (calibrated for 7.8GHz) to determine the array’s behavior and verify if it is possible to increase the beamwidth. One of the requirements to produce prototypes of lens antennas with 3D printing is knowing the electrical characteristics of the manufacturing materials (PLA, PETG, and nylon), more precisely, their dielectric constant. For that several samples of these materials were characterized considering different manufacturing conditions. The last step is the fabrication, by 3D printing, of prototype antennas using different materials and fabrication conditions. Nine lenses (six with a single material structure and three with several different materials) and seven patch antennas (five linearly polarized and two circularly polarized) were fabricated. Finally, was made a comparative study of the results obtained by simulation with the measurements performed in an anechoic chamber for both the patch antennas and the lens antenna array.
A recente evolução das radiocomunicações combinada com as inovadoras técnicas de fabrico, como a impressão 3D, impulsionaram o desenvolvimento e implementação de antenas com novas estruturas fabricadas com materiais incomuns. Um exemplo deste tipo de evolução são as antenas lente. As antenas lente estão sempre associadas a uma antena fonte (usualmente uma antena microstrip patch) e permitem alterar as caraterísticas de radiação (variar o ganho ou a directividade) da antena fonte. Assim, as lentes podem ser usadas para melhorar o desempenho de alguns tipos de sistema radiantes, como por exemplo o caso dos phased arrays, utilizados para fazer beamforming. No entanto, estes apresentam algumas limitações de cobertura, devido aos seus elementos do array terem ganho diretivo variável na zona de interesse. A utilização duma antena lente faz com que ocorra a alteração do diagrama de radiação de modo a obter uma maior largura de feixe podendo ser uma solução para referida limitação. Ao longo desta dissertação foi estudada a possibilidade de se utilizarem lentes para aumentar a largura de feixe de uma simples antena microstrip patch. Para isso, foram estudadas e realizadas simulações de várias estruturas de antenas lente com uma antena patch (calibrada para os 7.8GHz) com o intuito de determinar qual o comportamento do conjunto e verificar a possibilidade de tornar mais uniforme o diagrama de radiação no semi-espaço pretendido . A produção de protótipos de antenas lente com a impressão 3D requer o conhecimento das características elétricas dos materiais de fabrico (PLA, PETG e nylon), mais precisamente sua constate dielétrica. Para tal, foi feita uma caracterização de várias amostras desses materiais tendo em conta diferentes condições de fabrico. A última etapa foi a fabricação, por impressão 3D, de protótipos de antenas e lentes, utilizando diferentes materiais e condições de fabrico. No total foram fabricadas nove lentes (seis com uma estrutura de um único material e três com vários materiais distintos) e sete antenas patch (cinco de polarização linear e duas de polarização circular). Finalmente foi feito um estudo comparativo dos resultados obtidos por simulação com as medidas realizadas em câmara anecoica tanto para as antenas patch, como para o conjunto antena lente.
Mestrado em Engenharia Eletrónica e Telecomunicações

碩士
國立交通大學
資訊學院資訊科技產業專班
97
In this thesis, the use of an open graphics architecture CWT(CYC Window Toolkit) among the implementation of the OpenGL version. Design a system for mixing widgets of the 2D and 3D graphics. And by the experiment to verify whether the developer can provide a better choice of 3D Game UI toolkit.

We describe an algorithm that takes as inputs a coarse3D model of an environment, and a video sequence acquiredwithin the environment, and produces as output an estimateof the camera’s 6-DOF egomotion expressed in the coordinatesof the 3D model. Our method has several novelaspects: it performs line-based structure-from-motion; italigns the local line constellation to the known model; andit uses off-line visibility analysis to dramatically acceleratethe alignment process.We present simulation results demonstrating themethod’s operation in a multi-room environment. We showthat the method can estimate metric egomotion accuratelyand could be used for for many minutes of operation andthousands of video frames.

碩士
國立交通大學
電子研究所
101
In this thesis, we present a wide-baseline stereo system for 3D scene reconstruction. We implement our system with multiple un-calibrated cameras which are set widely. The main challenge of the system lies on how to match image pairs at wide-baseline, in which there appear large perspective distortions and large occlusion areas between images. In this research, we attempt to tackle the problem based on machine learning and optimization techniques. In order to match image more accurately, we apply random forest to overcome large perspective distortions, and add Conditional Random Field (CRF) with modified Histogram of Oriented Gradients (HOG) to solve the matching problem. Combining conditional random field with random forest can not only correct error correspondences but handle some occlusions. After getting matched points, we use these correspondences to find a 3D point set and camera matric by bundle adjustment (BA) that minimizes re-projection error. Then, we use the idea of spectral matting to refine the 3D point set. Finally, we build a 3D model with the refined point set.

碩士
國立清華大學
電機工程學系
101
The three-dimensional integrated circuit (3D IC) is considered a promising approach that can obtain high data band-width and low power consumption for future electronic systems that require high integration level. One of the popular drivers for 3D IC is the integration of a memory stack and a logic die. Because the yield of a 3D IC is the product of respective yields of the mounted dies, the yields of the memory dies and logic die must be high enough, or the 3D IC will be too expensive to be manufactured. It is generally agreed that the yield of large memories that are manufactured with advanced technologies is lower than the logic die. As a consequence, to obtain a high yield of 3D ICs, efficient test and repair methodologies for memories are necessary. In this thesis, we target the wide-I/O dynamic random access memory (DRAM) and propose two 3D redundancy architectures, i.e., Cubical Redundancy Architectures 1 and 2 (CRA1 and CRA2). In CRA1, spares are associated with each DRAM die as in a conventional 2D architecture. In CRA2, we use a static random access memory (SRAM) on the logic die as spares. We implemented both CRA1 and CRA2, and compared their repair rates as well as area overhead with the traditional redundancy architecture (TRA). Experimental results show that the CRA1 can obtain up to 3% higher stack yield than the TRA with the same area overhead. On the other hand, the CRA2 can obtain the same yield as the CRA1 with 40% less spares, but 1.27% higher area overhead than the CRA1.

碩士
國立中正大學
物理所
98
Wide-field optical microscopy is one of the most popular techniques for observing biological specimens. However, the optical resolution of microscopy is limited to be larger than ~ 0.5??laterally and 1.5? in depth due to diffraction effect. It is shown that the wide-field super-resolution imaging with the lateral resolution on the order of 100 nm is achievable if the sample is illuminated with periodic structured light. In order to retrieve a super-resolution image, several images illuminated with space-shifted patterned light must be taken and processed numerically. This will greatly reduce the imaging frame rate as mechanical movements are involved. In this thesis, we show that optically sectioned super-resolution microscopic imaging can be achieved by using a single SLM to modulate the illumination light. The 2D sinusoidal modulation mesh pattern projected on the sample is shifted by rapidly varying the driving signal on the LCoS panel at a frame-refresh rate as high as 60 Hz. The sectioned and super-resolved image is generated simultaneously from the same set of patterned excitation images, and the image acquisition rate of the image set can be as high as one set per second. By varying the period of 2D sinusoidal pattern on SLM, we can control the enhancement factor of the spatial resolution while keeping the field of view of the image fixed. With the SLM based super-resolution sectioning microscope, we improve the lateral resolution to 0.25 wavelengths and achieve a depth resolution of 0.38 wavelengths simultaneously. By taking the advantage of our advanced imaging system, 3D sectioning images of BPAE cells with mouse anti-?-tubulin and f-actin have been successfully acquired. These results demonstrate the ability of observing the sub-cell structure. Furthermore, this simple and high-resolution wide-field optical microscopy can be easily implemented on conventional fluorescence microscopes and has the potential to be applied to dynamical analyses of the organelles inside a live cell.

In this work, Paper-based inkjet-printed Ultra-wide band (UWB) fractal antennas are presented. Three new designs, a combined UWB fractal monopole based on the fourth order Koch Snowflake fractal which utilizes a Sierpinski Gasket fractal for ink reduction, a Cantor-based fractal antenna which performs a larger bandwidth compared to previously published UWB Cantor fractal monopole antenna, and a 3D loop fractal antenna which attains miniaturization, impedance matching and multiband characteristics. It is shown that fractals prove to be a successful method of reducing fabrication cost in inkjet printed antennas while retaining or enhancing printed antenna performance.

碩士
逢甲大學
光電學系
106
In this study, we have successfully extended the capability of traditional scanning white-light interference microscopy (SWIM) to wide-field 3D color surface measurement. The new SWIM system uses a color CCD to record interference images. The outputs of the CCD are three RGB interferograms. In order to maintain high axial resolution, the RGB interference images are combined to generate one white-light interferogram. The surface profile of a tested sample is restored via vertical scanning and centroid methods. The colors of the sample are the average values of RGB interference signals near zero optical path difference. The field of view is broadened by stitching many surface-profiles together. Every two adjacent fields overlap partially. The heights of the overlapped regions are corrected according to plane-fitting results. Finally, we use least root-mean-square error of the height difference between the overlapped areas for determining the offset along X and Y directions of two adjacent fields. 3×3-field measurements with this new SWIM were performed on multiple samples, the system is proved very effective in measuring wide-field 3D color surface. The measurement area is increased by 6 times compared with the conventional white-light interference microscope.

碩士
國立交通大學
電子研究所
105
Nowadays, big data becomes one of most popular topics in the world. Analyzing these data needs large amount of memory accessing. For the requests of multi users, the memory need high bandwidth and high density. The power of moving data also needs to be considered in the big data generation. High density 3D-stacked DRAM is the potential solution for the big data storage. By applying the through-silicon-vias (TSVs) technology in 3D-stacked DRAM, the I/O pins between logic tiers and DRAM are at least 32x larger comparing the conventional DRAMs. In order to utilize the 3D- DRAM lager capacity and memory bandwidth, a new advanced 3-D memory controller is needed. In this paper, we present an energy-efficient DRAM controller to fully utilize the benefit offer for 3D-Stacked DRAM. The controller uses command rescheduling and rank interleaving to parallel the commands from multi users, and the self-refresh can save the power for 3D-stacked DRAM refresh. Using near-data processing short the data path of transfer and makes the energy consumption decreased. The controller can realize the bandwidth improvement by 66.8%, and the execution time improvement by 40.08%. The energy consumption also decrease 27.18%.

碩士
國立交通大學
資訊科學與工程研究所
101
Due to the gap between memory and CPU speed, memory has become a bottleneck in computing systems. Improving memory access latency will improve system performance. As TSV technology matures, chips stacked in different stratums can reduce access latency. However, as SoC development moves towards 3D, it becomes increasingly difficult to evaluate complex systems designs. In this thesis an ESL platform is implemented which can support JEDEC wide I/O interface to evaluate memory performance. The simulator supports multi-threaded modeling and speedups the simulation time. After analyzing address mapping methods and properties of wide I/O, this thesis proposes two mechanisms to improve the performance of 3D architecture. Adaptive-prefetching will analyze memory intensive blocks and reference command queue status to prefetch data and improve RLP. Latency overlapping scheduling executes precharge command by beforehand analyzing TSV bus utilizaion.

碩士
大同工學院
工業設計研究所
85
AbstractFor the design department students, the basic design instruction is most important, and it influence students'' design ability。Under present teaching method, the design concept of the student and educate all come from the guidance of the department from the school. Very few an organisms would and other school alternating current and the works of the student completeness from the lesson teacher with the personal standpoint to adjudicate: like this of mode apparently than is sealing off and also have to owe objective. In addition, painting of student at inform the hour of stereo modeling concept, often look through planar come express, than can''t right of expressive space inside of modeling relation and take a lot of time at manufacture and whole grass molds of then inform the space concept. Present network system very of prosperous, is to obtain and deliver the tool of very convenient information. However, the main purpose of the this research just to hope by the integration of the network system, and blend virtual reality of real - time and with each other to move the characteristic, can fleetness of respond design of design concept. The real - time of inform the correlation of the modeling in the environment of the conjecture, make the teaching environment more comprehensiveness.Basic 3D modeling education emphasize to train in the modeling of two times of dollars and three dollar, it educates at the whole to educate the most basic modeling perception of the student in process, to a student that design department however talk equal importance. Because the fundamental capability of a designer is here phase created. For this reason this research then educate with the foundation stereo modeling course is main range, design one in the network system ascend, and regarding VR as the teaching mode of the Virtual Environment. This system can provide the student differ from tradition teaching method of designing and studying interface, and can search the related modeling theories to make up the tradition teaching to lose.This research completion of network teaching structure. Its inside includes curricular of the study environment of the basic information and teaching unit. Look through experiment come inspect and examine the student during the use condition of the design, at satisfaction degree, accept degree, data search, use time and creativity degree top, all more traditional teaching method is good, experiment result display. By the network support, it will surpass the mode that tradition check the references material on the hunt of the data, and look through the conjecture make environment of lend support to can attain following of achievement. First, student''s design concept of inform the aspect, not because of self-painting of capability however have the restrict. Second, developments of student at proceed design concept and inform the hour, can manufacture in the three dimensions of the conjecture. And inform design the space of the concept directly to relate to, need not the flower take a lot of time in the profile chart of the different visual angle come inform design concept. Third, hours of correction of student at proceed concept, need not again a graphics face to need only direct revise conjecture article, can save a flood of time. Above conclusion lend support to the tradition with the new type teaching system of tuition network to design educational of possibility, advantage and its existent value, and providing other researcher in design education inside the others part of a guide of the application.Keywords：Basic instruction in industrial design、Virtual Reality、Virtual Environment、Network （World Wide Web）。

碩士
國立清華大學
動力機械工程學系
96
Human see this world stereoscopically, we are trying to reproduce the visual experiences from what we see from the real world. Basic principles of most 3D displays technologies were discovered a hundred years ago, but the 3D displays have not yet been put into mass production due to several problems, such as the low 3D resolution, complicated mechanism, and inconvenience of wearing glasses, etc. Nowadays many researchers are getting interested in 3D displays again because the performance of flat panel display, especially liquid crystal display (LCD) has been largely improved. This mature display technology accompanying with the new-development method can overcome the problem of the previous 3D technology. This research proposes a solution for high efficiency high resolution liquid crystal (LC) panel for autostereoscopic 3D display with wide-view and high 3D resolution. The slanted barrier is applied on a LC panel for sharing the resolution reduction in both horizontal and vertical direction. In order to reach a high resolution with high efficiency, a new design of the LC pixel layout is necessary. The pixel design is based on the slanted barrier. The LC pixel with high directional emission profile can not only improve efficiency but also eliminate some deficiencies that affect the image quality in conventional parallax barrier design. By applying the slanted barrier and the pixel layout, we can overcome the issue of low resolution and serious resolution reduction in horizontal direction on conventional straight barrier design and have a LC panel which is more suitable for the application of high light efficiency parallax barrier type 3D display.

碩士
國立臺灣科技大學
機械工程系
107
This study is the early research on the development of page-wide piezoelectric printhead module high-speed sintering 3D printing machine. At first, the design and development of the heating process was carried out with a modified single thermal bubble printhead powder bed sintering 3D printing machine. At the same time, the material selection and the properties of the material were analyzed to establish the reaction mechanism and print the finished product. Organize a better printing strategy and apply this experience to machine process development. In addition, the nozzle is assembled into a page-wide module to simplify the printing process becoming single pass to achieve high speed process. Thermoplastic polyurethane, TPU, is widely used in current footwear manufacturing industry. As is a kind of amorphous materials, TPU has no clear melting point, which makes it difficult to define the morphology in the powder bed sintering 3D printing process. In this research, detecting of temperature in different area is used to make products and then calculate the difference between dimensional accuracy and mechanical property as a reference for tuning the process parameters. Finally, printability test of the piezoelectric printhead and ink can be carried out as the basis for material selection and adjustment, and using high speed camera to observe the droplets during printing to determine whether the design of the piezoelectric waveform is appropriate.

Σήμερα υπάρχουν πολυάριθμα εργαλεία για την κατασκευή και παραμετροποίηση τρισδιάστατων μοντέλων με το καθένα από αυτά να χρησιμοποιεί το δικό του πρωτόκολλο περιγραφής της τρισδιάστατης πληροφορίας που αποθηκεύει και επεξεργάζεται. Εντούτοις δεν είναι λίγες η φορές που απαιτείται η μεταφορά περιεχομένου απο το ένα εργαλείο στο άλλο. Το παραπάνω οδήγησε στην ανάγκη δημιουργίας κοινώς αποδεκτών πρωτοκόλλων περιγραφής της τρισδιάστατης πληροφορίας για την διευκόλυνση της διαχείρισής της. Απο τα πρωτόκολλα αυτά, το COLLADA φαίνεται να επικρατεί λόγω του οτι είναι ανοιχτό, επεκτάσιμο και ευρέως διαδεδομένο. Στην παρούσα εργασία γίνεται παρουσίαση αυτού και δίνεται έμφαση στους λόγους που οδηγούν στην ολοένα και ταχύτερη ανάπτυξή του και στις τελευταίες λειτουργείες που έχουν προστεθεί. Στη συνέχεια, με βάση το πρωτοκολλο αυτό, υλοποιείται μια πιλοτική εφαρμογή για την απεικόνιση και την αλληλεπίδραση με τρισδιάτατα αντικείμενα, χρησιμοποιώντας σύγχρονες τεχνικές και τεχνολογίες για διαδικτυακή παρουσίαση τρισδιάστατου περιεχομένου. Η εφαρμογή αυτή στοχεύει σε φυλλομετρητές προσωπικών υπολογιστών, ενδέχεται ωστόσο να επεκταθεί ώστε να είναι δυνατή η χρήση της ακόμα και από φυλλομετρητές τελευταίας τεχνολογίας κινητών τηλεφώνων (iPhone). Τα συμπεράσματα που θα προκύψουν από την χρήση της εφαρμογής αυτής ενδέχεται να οδηγήσουν στην δημιουργία ενός μετέπειτα ολοκληρωμένου προϊόντος.
Today there are numerous tools for the construction and configuration of three-dimensional models, each of them uses its own protocol describes the three-dimensional information stored and processed. However, there are few times when the need to transfer content from one tool to another. The above led to the need for commonly accepted protocol describes the three-dimensional information to facilitate management. Of these protocols, the COLLADA seems to prevail because it is open, scalable and ubiquitous. In the present study it and focus on the reasons leading to ever more rapid development in recent operations that have been added. Then, based on this protocol, implemented a pilot application to display and interact with trisdiatata items using modern techniques and technologies for web-dimensional presentation of content. This application is targeted at PC browsers, but may be extended to allow the use of browsers and even the art mobile phones (iPhone). The conclusions arising from the use of this application may result in the creation of a finished product later.

碩士
國立清華大學
電機工程學系
103
With the evolution of MOS technology based on Moore’s Law, we have to face the difficulties on designing and the bottleneck on physics and materials including increasing leakage current, RC delays on wire routing and yield issue. 3D integration has the most potential to solve these problems by using Though Silicon Via (TSV) technology and providing outstanding performance and high density advantage at the same time. However, there are many challenges for TSV-based 3D IC such as bad energy efficiency due to large loading, multi layer addressing and large TSV pitch. Therefore, the capability of achieving low energy efficiency by circuit design for 3D IC is the main target of our works. In the thesis, we propose a Data-Aware Charge-Sharing Low Voltage Transmission Scheme with Asymmetric Sense Amplifier to reduce the power consumption and solve the Vref selecting problems for conventional symmetric sense amplifier. A 1kb I/O macro has been fabricated in 65nm CMOS technology to verify the ideas of this works. The measurement results demonstrate the functionality of this works and the energy efficiency can achieve to 0.12 mw/Gbps.

博士
國立清華大學
電機工程學系
103
The fast growing demand of thin and compact mobile and wearable devices has driven the efforts to reduce the size of camera module. CMOS image sensor (CIS) with small pixel dimension is an effective solution to implement a small size camera module. The design challenges of CMOS image sensor with small pixel dimension are low dynamic range, low full well capacity (FWC) and low sensitivity. In this thesis, three new techniques are proposed to address the problems. Firstly, a dual-exposure single-capture wide dynamic range CMOS image sensor for mobile devices is proposed. The proposed sensor achieves column-wise highly/lowly-illuminated pixel detection, and only the “adequate” voltage signal (long- or short-exposure signal) is digitized. With an integrated highly/lowly-illuminated pixel detection function in the column-wise single slope ADC, each pixel is read out only once with highly- or lowly-illuminated pixel index for synthesis of a wide DR frame. This approach can dramatically reduce the power dissipation compared to existing multi-frame-readout solutions. The dynamic range expansion ratio is programmable, and depends on the time ratio of long-exposure to short-exposure period. Secondly, a novel single-slope ADC design and operation is proposed to expand full well capacity of CMOS image sensor with small pixel dimension. With the proposed technique, charges stored in the photodiode and floating diffusion of 4T active pixel sensor are all read out and accumulated by the proposed SS ADC to improve the FWC. Only one A/D conversion is required for each pixel, which decreases chip power consumption compared to the general double A/D conversion operation. Finally, because 3D IC is an emerging solution to reduce chip size, a 3D-integrated IV CMOS image sensor layer with built-in self-test function for 3-layer stacking CMOS imager is proposed. A modular CIS sub-array is proposed with new readout and control scheme. The proposed readout structure with in-pixel two-dimensional (2D) decoding function achieves high spatial resolution, without degrading the frame rate. A BIST circuit is also proposed to filter out unqualified CIS layer before chip stacking, improving the yield performance of the final 3D integrated imagers, without adding extra transistor in the pixel. The proposed 3D-integrated CIS layer is very suitable for small size camera module applications.

In recent years, many researchers have investigated automated progress tracking for construction projects. These efforts range from 2D photo-feature extraction to 3D laser scanners and radio frequency identification (RFID) tags. A multi-sensor data fusion model that utilizes multiple sources of information would provide a better alternative than a single-source model for tracking project progress. However, many existing fusion models are based on data fusion at the sensor and object levels and are therefore incapable of capturing critical information regarding a number of activities and processes on a construction site, particularly those related to non-structural trades such as welding, inspection, and installation activities. In this research, a workflow based data fusion framework is developed for construction progress, quality and productivity assessment. The developed model is based on tracking construction activities as well as objects, in contrast to the existing sensor-based models that are focussed on tracking objects. Data sources include high frequency automated technologies including 3D imaging and ultra-wide band (UWB) positioning. Foreman reports, schedule information, and other data sources are included as well. Data fusion and management process workflow implementation via a distributed computing network and archiving using a cloud-based architecture are both illustrated. Validation was achieved using a detailed laboratory experimental program as well as an extensive field implementation project. The field implementation was conducted using five months of data acquired on the University of Waterloo Engineering VI construction project, yielding promising results. The data fusion processes of this research provide more accurate and more reliable progress and earned value estimates for construction project activities, while the developed data management processes enable the secure sharing and management of construction research data with the construction industry stakeholders as well as with researchers from other institutions.

Thesis (Ph.D.)--University of Alberta, 2009.
Title from PDF file main screen (viewed on Apr. 1, 2010). A thesis submitted to the Faculty of Graduate Studies and Research in partial fulfillment of the requirements for the degree of Doctor of Philosophy, Department of Electrical and Computer Engineering, University of Alberta. Includes bibliographical references.