In broadband antenna applications, the antenna's cavity is usually loaded with absorbers to eliminate the backward radiation, but in doing so the radiation efficiency of the antenna is decreased. To enhance the radiation efficiency of the antennas EBG structures are used, but they operate over a narrow band. Uniform electromagnetic band gap (EBG) structures are usually periodic structures consisting of metal patches that are separated by small gaps and vias that connect the patches to the ground plane. The electrical equivalent circuit consists of a resonant tank circuit, whose capacitance is represented by the gap between the patches and inductance represented by the via. EBG structures are equivalent to a magnetic surface at the frequency of resonance and thus have very high surface impedance; this makes the EBG structures useful when mounting an antenna close to conducting ground plane, provided the antenna's currents are parallel to the EBG structure. Because EBG structures are known to operate over a very narrow band, they are not useful when used with a broadband antenna. Mushroom-like uniform EBG structures (that use vias) are compact in size have low loss, can be integrated into an antenna to minimize coupling effects of ground planes and increase radiation efficiency of the antenna. The bandwidth of an EBG structure is defined as the band where the reflection-phase from the structure is between +900 to -900. In this dissertation analysis of EBG structures is established using circuit analysis and transmission line analysis. Methods of increasing the bandwidth of EBG structures are explored, by cascading uniform EBG structures of different sizes progressively and vertically (stacked), and applications with different types of antennas are presented. Analyses in this dissertation are compared with previously published results and with simulated results using 3D electromagnetic tools. Validation of applications with antennas is carried by manufacturing prototypes and comparing measured performance with analysis and 3D electromagnetic simulations. The improvements in performance by using wideband progressive EBG and wideband stacked EBG structures are noted. / Ph. D.

Millimeter wave (mm-Wave) has become increasingly popular frequency band for next-generation high-speed wireless communications. In mm-Wave, the wireless channel path loss is severe, demanding a high output power in transmitters (Tx) to meet a required SNR in receivers (Rx). Due to the intractable speed-power tradeoff ingrained in silicon processes, however, achieving a high power at mm-Wave, particularly over W-band (> 90 GHz), is challenging in silicon power amplifiers. To relieve the output power burden, phased-arrays are widely adopted in mm-Wave wireless communication systems -- namely, by leveraging a parallel power combining in the space domain, inherent in the phased arrays, the required output power per array element can be reduced significantly with increasing array size. In large arrays ( > 100's -- 1000's number of arrays), the required output power per element could be small, typically around several 10's mW or less in silicon-based phased arrays. In such small-to-medium scale output power level, the static power dissipations by transistor knee voltage and passive components could be a significant portion of the output power, decreasing power efficiency of power amplifiers drastically. This poses a significant concern on the power efficiency of the large-scale silicon-based phased arrays in mm-Wave. Another critical problem in mm-Wave wireless systems design is the increase of passive reactive components loss caused by worsening skin depth effect and increasing dielectric loss through silicon substrate. This essentially degrades the reactive components quality factor (Q) and limits frequency selectivity of the silicon-based mm-Wave systems. This thesis tackles these two major technical challenges to provide high frequency selectivity with maintaining high power efficiency for future mm-Wave wireless systems over W-band and beyond. First, various high-efficiency techniques such as impedance tuning with a reactive component at a cascoding stage in conventional stacked power amplifiers or load-pull based inter-stage matching technique, rather than conventional conjugate matching, have been applied to W-band CMOS and SiGe BiCMOS amplifiers to improve power efficiency with 5-10 dBm output power level, suitable for a large phased array applications, as detailed in Chapter 2 and 3. Second, a 4-tap finite impulse response (FIR) filter based receiver architecture is presented in Chapter 4. The FIR filtered receiver leverages a sinc-pulse type frequency nulls built-in in the transmission-line based FIR filter's frequency response to increase frequency selectivity. The proposed FIR filtered receiver achieves > 40-dB image rejection by placing an image signal at the null frequency at D-band, one of the largest image rejection performance at the highest frequency band reported so far. / Ph. D. / Due to recent advances in Silicon based solid-state technologies, the interest towards the millimeter wave (mm-Wave) frequency band has been emerging for next-generation high-speed wireless communication applications. One of the most significant parameters in a communication system would be the output power of a transmitter. However, the output power is limited especially at mm-wave frequencies. A phased array is one of the viable solutions to overcome this burden by utilizing a parallel power combing in the space domain. The required output power per element can be relieved, typically around several tens of mill watts or less. There are two major factors limiting the output power, which are the high loss of passive and active devices. This dissertation presents solutions to overcome these challenges. In addition, a 4-tap finite impulse response (FIR) filter based receiver architecture is introduced, which rejects unwanted image signals in heterodyne systems by utilizing sinc-pulse type frequency nulls. The proposed FIR filter achieves more than 40 dB of image rejection at D-band (110-170 GHz), which is one of the highest filtering performance in the millimeter-wave frequency band.

This thesis is focused on the analysis and development of Ensembles of Neural Networks. An ensemble is a system in which a set of heterogeneous Artificial Neural Networks are generated in order to outperform the Single network based classifiers. However, this proposed thesis differs from others related to ensembles of neural networks [1, 2, 3, 4, 5, 6, 7] since it is organized as follows. In this thesis, firstly, an ensemble methods comparison has been introduced in order to provide a rank-based list of the best ensemble methods existing in the bibliography. This comparison has been split into two researches which represents two chapters of the thesis. Moreover, there is another important step related to the ensembles of neural networks which is how to combine the information provided by the neural networks in the ensemble. In the bibliography, there are some alternatives to apply in order to get an accurate combination of the information provided by the heterogeneous set of networks. For this reason, a combiner comparison has also been introduced in this thesis. Furthermore, Ensembles of Neural Networks is only a kind of Multiple Classifier System based on neural networks. However, there are other alternatives to generate MCS based on neural networks which are quite different to Ensembles. The most important systems are Stacked Generalization and Mixture of Experts. These two systems will be also analysed in this thesis and new alternatives are proposed. One of the results of the comparative research developed is a deep understanding of the field of ensembles. So new ensemble methods and combiners can be designed after analyzing the results provided by the research performed. Concretely, two new ensemble methods, a new ensemble methodology called Cross-Validated Boosting and two reordering algorithms are proposed in this thesis. The best overall results are obtained by the ensemble methods proposed. Finally, all the experiments done have been carried out on a common experimental setup. The experiments have been repeated ten times on nineteen different datasets from the UCI repository in order to validate the results. Moreover, the procedure applied to set up specific parameters is quite similar in all the experiments performed. It is important to conclude by remarking that the main contributions are: 1) An experimental setup to prepare the experiments which can be applied for further comparisons. 2) A guide to select the most appropriate methods to build and combine ensembles and multiple classifiers systems. 3) New methods proposed to build ensembles and other multiple classifier systems.

X BAND TWO LAYER PRINTED REFLECTARRAY WITH SHAPED BEAM &Uuml / &ccedil / &uuml / nc&uuml / , G&ouml / khan MSc., Department of Electrical and Electronics Engineering Supervisor: Prof. Dr. H. &Ouml / zlem Aydin &Ccedil / ivi October, 2011, 110 pages X-band cosecant square shaped beam microstrip reflectarray is designed, fabricated and measured. Unit element of the reflectarray is in stacked patch configuration. With the aim of designing shaped beam pattern, phase-only synthesis method based on genetic algorithm is used. Phases of reflected electric field from antenna elements are adjusted by changing the dimensions of the patches. Unit cell simulations are performed using periodic boundary conditions and assuming infinite array approach to obtain reflection phase curves versus patch size. Then full reflectarray surface and its feed are designed and fabricated. Radiation patterns are measured in spherical near field range and results are compared with simulations. It is shown that the antenna is capable to operate in a band of 8.6 - 9.7 GHz.

The purpose of this research is to help increase the capacity of public-sector transportation agencies (such as state Departments of Transportation, Metropolitan Planning Organizations, and transit providers) to preserve and enhance transportation-related quality of life (QOL) outcomes in their jurisdictions. QOL is a multi-dimensional concept that is closely related to the concepts of livability and social sustainability. Public-sector agencies are charged with promoting the well-being (i.e. QOL) of the public, and they often must work within a complex inter-organizational context, with overlapping and intersecting jurisdictions and responsibilities, in order to influence QOL. Because of their responsibility to promote QOL, many public-sector transportation agencies mention QOL, livability, and/or sustainability in their vision statements, mission statements, and strategic planning documents. Furthermore, U.S. Federal guidance and regulations that govern the practice of transportation planning, engineering, and performance management have begun to refer to issues related to livability and sustainability. However, these complex concepts are still ambiguous in meaning and application for many transportation practitioners. In order to effectively preserve and enhance transportation-related QOL outcomes, practitioners need a clear conceptual framework that links concepts of livability and sustainability to practical performance management tools for an inter-jurisdictional context. The primary objective and contributions of this research are the development of such a conceptual framework - the stacked systems framework (SSF) - and a methodology for applying it to enhance transportation performance management in an inter-jurisdictional context. In order to develop the SSF, this research begins with an extensive literature review that clarifies the relationships among sustainability, livability, and transportation-related QOL outcomes; and integrates the concepts of social sustainability, soft systems methodologies, and the field of transportation performance management. To apply the SSF, this research includes a case study of public-sector transportation performance management processes in metropolitan Atlanta. The case study analyzes the influence of the regional inter-organizational system of public-sector transportation agencies on transportation-related QOL outcomes; identifies gaps in the current set of transportation performance measures used for decision making at the regional scale; and demonstrates the value to decision making of incorporating recommended performance measures that can more appropriately link organizational actions to broader QOL and livability outcomes via changes in transportation service quality. The case study methodology can be extended for future development of transportation performance management practices in metro Atlanta, and reproduced for other regions and geographic scales.

Ultra-broadband Monolithic Microwave Integrated Circuit (MMIC) amplifiers find applications in multi-gigabit communication systems for 5G and millimeter wave measurement instrumentation systems. The aim of the research was to achieve maximum bandwidth of operation of the amplifier from the foundry process used and high reverse isolation ( < -25.0 dB) across the whole bandwidth. To achieve this, several design variations of DC - 110 GHzMMIC Cascode TravellingWave Amplifier (TWA) on 100 nm AlGaAs/GaAs pHEMT process were done for application in next generation instrumentation and high data transfer rate (100 Gb/s) optical modulator systems. The foundry service and device models used for the design are of the WINPP10-10 process from WIN Semiconductor Corp., Taiwan, a commercial and highly stable process. The cut-off frequency ft and maximum frequency of oscillation fmax for this process are 135 GHz and 185 GHz respectively. Thus, the design was aimed at pushing the ultimate limits of operation for this process. The design specifications were targeted to have S21 = 9.0 to 10.0 ± 1.0 dB, S11 & S22 ≤ -10.0 dB and S12 ≤ -25.0 dB in the whole frequency range. In order to achieve the targeted RF performance, it is imperative to have accurate transistor models over the frequency range of operation, transistor configuration mode and operating bias points. Using smaller periphery transistors results in lower extrinsic & intrinsic input and output capacitances that lead to achieving very wide band performance. Thus, device sizes as small as 2x10 μm were used for the design. A cascode topology, which is a series connection of a common-source and common-gate field effect transistor (FET), was used to achieve large bandwidth of operation, high reverse isolation and high input and output impedance. Using very small periphery devices at cascode bias points posed limitation in the design in terms of accuracy of transistor models under these conditions, specifically at high frequencies i.e., above 50 GHz. One of the major systemrequirements for the application of MMIC ultra-broadband amplifiers in instrumentation is to achieve and maintain high reverse isolation (≤ -25.0 dB) over the whole frequency range of operation which cannot be achieved alone by the cascode topology and new design techniques have to be devised. These twomajor challenges, namely high frequency small periphery FET model modification & development and design technique to achieve high reverse isolation in ultra-broadband frequency range have been addressed in this research.

Presenting large amount of Multivariate Data is not a simple problem. When there are multiple correlated variables involved, it becomes difficult to comprehend data using traditional ways. Information Visualization techniques provide an interactive way to present and analyze such data. This thesis has been carried out at ABB Corporate Research, Västerås, Sweden. Use of Parallel Coordinates and Multiple Coordinated Views was has been suggested to realize interactive reporting and trending of Multivariate Data for ABB’s Network Manager SCADA system. A prototype was developed and an empirical study was conducted to evaluate the suggested design and test it for usability from an actual industry perspective. With the help of this prototype and the evaluations carried out, we are able to achieve stronger results regarding the effectiveness and efficiency of the visualization techniques used. The results confirm that such interfaces are more effective, efficient and intuitive for filtering and analyzing Multivariate Data.

Les tumeurs cérébrales sont une cause majeure de décès et d'invalidité dans le monde, ce qui représente 14,1 millions de nouveaux cas de cancer et 8,2 millions de décès en 2012. La radiothérapie et la radiochirurgie sont parmi l'arsenal de techniques disponibles pour les traiter. Ces deux techniques s’appuient sur une irradiation importante nécessitant une définition précise de la tumeur et des tissus sains environnants. Dans la pratique, cette délinéation est principalement réalisée manuellement par des experts avec éventuellement un faible support informatique d’aide à la segmentation. Il en découle que le processus est fastidieux et particulièrement chronophage avec une variabilité inter ou intra observateur significative. Une part importante du temps médical s’avère donc nécessaire à la segmentation de ces images médicales. L’automatisation du processus doit permettre d’obtenir des ensembles de contours plus rapidement, reproductibles et acceptés par la majorité des oncologues en vue d'améliorer la qualité du traitement. En outre, toute méthode permettant de réduire la part médicale nécessaire à la délinéation contribue à optimiser la prise en charge globale par une utilisation plus rationnelle et efficace des compétences de l'oncologue.De nos jours, les techniques de segmentation automatique sont rarement utilisées en routine clinique. Le cas échéant, elles s’appuient sur des étapes préalables de recalages d’images. Ces techniques sont basées sur l’exploitation d’informations anatomiques annotées en amont par des experts sur un « patient type ». Ces données annotées sont communément appelées « Atlas » et sont déformées afin de se conformer à la morphologie du patient en vue de l’extraction des contours par appariement des zones d’intérêt. La qualité des contours obtenus dépend directement de la qualité de l’algorithme de recalage. Néanmoins, ces techniques de recalage intègrent des modèles de régularisation du champ de déformations dont les paramètres restent complexes à régler et la qualité difficile à évaluer. L’intégration d’outils d’assistance à la délinéation reste donc aujourd’hui un enjeu important pour l’amélioration de la pratique clinique.L'objectif principal de cette thèse est de fournir aux spécialistes médicaux (radiothérapeute, neurochirurgien, radiologue) des outils automatiques pour segmenter les organes à risque des patients bénéficiant d’une prise en charge de tumeurs cérébrales par radiochirurgie ou radiothérapie.Pour réaliser cet objectif, les principales contributions de cette thèse sont présentées sur deux axes principaux. Tout d'abord, nous considérons l'utilisation de l'un des derniers sujets d'actualité dans l'intelligence artificielle pour résoudre le problème de la segmentation, à savoir le «deep learning ». Cet ensemble de techniques présente des avantages par rapport aux méthodes d'apprentissage statistiques classiques (Machine Learning en anglais). Le deuxième axe est dédié à l'étude des caractéristiques d’images utilisées pour la segmentation (principalement les textures et informations contextuelles des images IRM). Ces caractéristiques, absentes des méthodes classiques d'apprentissage statistique pour la segmentation des organes à risque, conduisent à des améliorations significatives des performances de segmentation. Nous proposons donc l'inclusion de ces fonctionnalités dans un algorithme de réseau de neurone profond (deep learning en anglais) pour segmenter les organes à risque du cerveau.Nous démontrons dans ce travail la possibilité d'utiliser un tel système de classification basée sur techniques de « deep learning » pour ce problème particulier. Finalement, la méthodologie développée conduit à des performances accrues tant sur le plan de la précision que de l’efficacité. / Brain cancer is a leading cause of death and disability worldwide, accounting for 14.1 million of new cancer cases and 8.2 million deaths only in 2012. Radiotherapy and radiosurgery are among the arsenal of available techniques to treat it. Because both techniques involve the delivery of a very high dose of radiation, tumor as well as surrounding healthy tissues must be precisely delineated. In practice, delineation is manually performed by experts, or with very few machine assistance. Thus, it is a highly time consuming process with significant variation between labels produced by different experts. Radiation oncologists, radiology technologists, and other medical specialists spend, therefore, a substantial portion of their time to medical image segmentation. If by automating this process it is possible to achieve a more repeatable set of contours that can be agreed upon by the majority of oncologists, this would improve the quality of treatment. Additionally, any method that can reduce the time taken to perform this step will increase patient throughput and make more effective use of the skills of the oncologist.Nowadays, automatic segmentation techniques are rarely employed in clinical routine. In case they are, they typically rely on registration approaches. In these techniques, anatomical information is exploited by means of images already annotated by experts, referred to as atlases, to be deformed and matched on the patient under examination. The quality of the deformed contours directly depends on the quality of the deformation. Nevertheless, registration techniques encompass regularization models of the deformation field, whose parameters are complex to adjust, and its quality is difficult to evaluate. Integration of tools that assist in the segmentation task is therefore highly expected in clinical practice.The main objective of this thesis is therefore to provide radio-oncology specialists with automatic tools to delineate organs at risk of patients undergoing brain radiotherapy or stereotactic radiosurgery. To achieve this goal, main contributions of this thesis are presented on two major axes. First, we consider the use of one of the latest hot topics in artificial intelligence to tackle the segmentation problem, i.e. deep learning. This set of techniques presents some advantages with respect to classical machine learning methods, which will be exploited throughout this thesis. The second axis is dedicated to the consideration of proposed image features mainly associated with texture and contextual information of MR images. These features, which are not present in classical machine learning based methods to segment brain structures, led to improvements on the segmentation performance. We therefore propose the inclusion of these features into a deep network.We demonstrate in this work the feasibility of using such deep learning based classification scheme for this particular problem. We show that the proposed method leads to high performance, both in accuracy and efficiency. We also show that automatic segmentations provided by our method lie on the variability of the experts. Results demonstrate that our method does not only outperform a state-of-the-art classifier, but also provides results that would be usable in the radiation treatment planning.

This master’s thesis describes the design, implementation and testing of an active antenna intended for use with global navigation satellite systems. The active antenna is composed of two major parts, a dual-band circular patch antenna and a dual-band low-noise amplifier. To streamline the design process, a modular solution was adopted. This enabled the functionality of every part in the signal path to be verified before the final active antenna was designed. A practical method to develop dual-band stacked circular patch antennas, along with a systematic way to tune the resonant frequencies and impedance of the antenna, is also presented. Testing of the antenna in realistic scenarios shows that the active antenna performs as expected and predicted by simulations. / <p>Examensarbetet är utfört vid Institutionen för teknik och naturvetenskap (ITN) vid Tekniska fakulteten, Linköpings universitet</p>

Stacking, also known as stacked generalization, is a method of ensemble learning where multiple base models are trained on the same dataset, and their predictions are used as input for one or more metamodels in an extra layer. This technique can lead to improved performance compared to single layer ensembles, but often requires a time-consuming trial-and-error process. Therefore, the previously developed Visual Analytics system, StackGenVis, was designed to help users select the set of the most effective and diverse models and measure their predictive performance. However, StackGenVis was developed with only one metamodel: Logistic Regression. The focus of this Bachelor's thesis is to examine how alternative metamodels affect the performance of stacked ensembles through the use of a visualization tool called MetaStackVis. Our interactive tool facilitates visual examination of individual metamodels and metamodels' pairs based on their predictive probabilities (or confidence), various supported validation metrics, and their accuracy in predicting specific problematic data instances. The efficiency and effectiveness of MetaStackVis are demonstrated with an example based on a real healthcare dataset. The tool has also been evaluated through semi-structured interview sessions with Machine Learning and Visual Analytics experts. In addition to this thesis, we have written a short research paper explaining the design and implementation of MetaStackVis. However, this thesis  provides further insights into the topic explored in the paper by offering additional findings and in-depth analysis. Thus, it can be considered a supplementary source of information for readers who are interested in diving deeper into the subject.