Automatic segmentation of speech is compulsory for building large speech databases to be used in speech processing applications. This study proposes a bimodal automatic speech segmentation system that uses either articulator motion information (AMI) or visual information obtained by a camera in collaboration with auditory information. The presence of visual modality is shown to be very beneficial in speech recognition applications, improving the performance and noise robustness of those systems. In this dissertation a significant increase in the performance of the automatic speech segmentation system is achieved by using a bimodal approach. Automatic speech segmentation systems have a tradeoff between precision and resulting number of gross errors. Boundary refinement techniques are used in order to increase precision of these systems without decreasing the system performance. Two novel boundary refinement techniques are proposed in this thesis / a hidden Markov model (HMM) based fine tuning system and an inverse filtering based fine tuning system. The segment boundaries obtained by the bimodal speech segmentation system are improved further by using these techniques. To fulfill these goals, a complete two-stage automatic speech segmentation system is produced and tested in two different databases. A phonetically rich Turkish audiovisual speech database, that contains acoustic data and camera recordings of 1600 Turkish sentences uttered by a male speaker, is build from scratch in order to be used in the experiments. The visual features of the recordings are extracted and manual phonetic alignment of the database is done to be used as a ground truth for the performance tests of the automatic speech segmentation systems.

This thesis includes the study of hybrid microwave integrated circuits (HMIC), miniaturization techniques applied on HMICs and its application on a frequency modulated continuous wave (FMCW) range sensor transceiver. In the scope of study, hybrid and monolithic microwave integrated circuits (HMIC and MMIC) are introduced, advantages and disadvantages of these two types are discussed. Large size of HMICs is the main disadvantage especially for military and civil applications requiring miniature volumes. This thesis is mainly devoted on miniaturization work of HMICs in order to cope with this problem. In this scope, miniaturization techniques of some HMICs such as 3 dB hybrid couplers and stubs are examined and analyzed. Their simulation and measurement results cohere with original circuit results. Nevertheless, considerable size reduction up to 80% is achieved. Moreover, planar interdigital capacitors (IDC), spiral inductors (SI) and their equivalent circuit models are introduced. Design technique is discussed with illustrative electromagnetic (EM) simulations. Furthermore, FMCW radar is introduced with its basic operation principles, brief history and usage areas. In addition, FMCW range sensor transceiver is designed with its sub&amp / #8208 / parts / power amplifier, low noise amplifier (LNA), coupler and front end. Multi technology based on chip transistors, interdigital capacitors, spiral inductors and hybrid couplers with wire&amp / #8208 / bond connections is used in the design. As the result of using hybrid miniaturized components small layout size is achieved for the transceiver system with its all components.

In this thesis, real time detection of traffic signs using FPGA hardware is presented. Traffic signs have distinctive color and shape properties. Therefore, color and shape based algorithms are chosen to implemented on FPGA. FPGA supports sufficient logic to implement complete systems and sub-systems. Color information of images/frames is used to minimize the search domain of detection process. Using FPGA, real time conversion of YUV space to RGB space is performed. Furthermore, color thresholding algorithm is used to localize the sign in the image/video depending on the color. Edges are the most important image/frame attributes that provide valuable information about the shape of the objects. Sobel edge detection algorithm is implemented on FPGA. After color segmentation, FPGA implementation of Sobel algorithm is used to find the edges of candidate traffic signs in real time. Later, radial symmetry based shape detection algorithm is used to determine circular traffic signs. Each FPGA implemented algorithm is tested by using video sequences and static images. In addition, combined implementation of color based and shape based algorithms are tested. Joint application of color and shape based algorithms are used in order to reduce search domain and the processing time of detection process. Designing architecture on FPGA makes traffic sign detection system portable as a final product and relatively more efficient than the computer based detection systems. The resulting hardware is suitable where cost and compactness constraints are important.

In this study, an automated building extraction system, which is capable of detecting buildings from satellite images using only RGB color band is implemented. The approach used in this work has four main steps: local feature extraction, feature selection, classification and post processing. There are many studies in literature that deal with the same problem. The main issue is to find the most suitable features to distinguish a building. This work presents a feature selection scheme that is connected with the classification framework of Adaboost. As well as Adaboost, four SVM kernels are used for classification. Detailed analysis regarding window type and size, feature type, feature selection, feature count and training set is done for determining the optimal parameters for the classifiers. A detailed comparison of SVM and Adaboost is done based on pixel and object performances and the results obtained are presented both numerically and visually. It is observed that SVM performs better if quadratic kernel is used than the cases using linear, RBF or polynomial kernels. SVM performance is better if features are selected either by Adaboost or by considering errors obtained on histograms of features. The performance obtained by quadratic kernel SVM operated on Adaboost selected features is found to be 38% in terms of pixel based performance criteria quality percentage and 48% in terms object based performance criteria correct detection with building detection threshold 0.4. Adaboost performed better than SVM resulting in 43% quality percentage and 67% correct detection with the same threshold.

This thesis presents fabrication of shunt, capacitive contact type RF MEMS switches which are designed according to given mm-wave performance specifications. The designed switches are modified for investigation in terms of reliability and lifetime. To observe the real-time performance of switches a time domain measurement setup is established and a CV (capacitance vs. voltage) curve measurement system is also included to measure CV curves, pull-in and hold-down voltages and the shifts of these due to actuations. By using the established setup reliability and lifetime measurements under different bias waveforms in different environments are performed. After investigation for the most suitable condition for improving lifetime long-term tests are performed and the outstanding result of more than 885 hours of operation under cycling bias waveform is obtained.

The aim of this study is to investigate the effects of changes in the physical parameters of tapered slot antennas (TSA) on their radiation characteristics and also to explore the possibility of reconfiguration in the radiation pattern of TSA by switching between two different types of tapering. There are mainly three physical parameters that affect the radiation pattern of TSAs. These are antenna length, aperture width and the ground extension. After designing a wideband microstrip line to slot line transition, the effect of antenna parameter variations on the beamwidth and sidelobe level of the antenna are investigated through the use of a commercially available electromagnetic simulation software HFSS by ANSOFT. The radiation characteristics of constant width slot antennas (CWSA) and linearly tapered slot antennas (LTSA) are compared. It is observed that CWSAs exhibit narrower beamwidth and higher sidelobe level whereas linearly tapered slot antennas (LTSA) have wider beamwidth with lower sidelobe level compared to each other. A novel switching architecture between CWSA and LTSA is proposed to obtain a reconfigurable antenna.

In this thesis work, minimization of platform-based coupling between the antennas of two VHF radios on an aircraft platform and two HF radios on a ship platform is aimed. For this purpose / an optimal antenna placement, which yields minimum average coupling between the antennas over the whole frequency band of operation is determined for each platform. Two important global optimization techniques, namely Genetic Algorithm Optimization and Particle Swarm Optimization, are used in determination of these optimal antenna placements. Aircraft &amp / ship platforms and antennas placed on them are modeled based on their real electrical and physical properties in CST &ndash / MWS (Microwave Studio) simulation tool. For each platform, antenna placements and coupling results determined by two different optimization techniques and performances of these optimization techniques are compared with each other. At the end of this thesis work / for each platform, far-field radiation pattern performances of the antennas at their optimal places are analyzed in terms of directivity and coverage.

This thesis presents the development and microfabrication of capacitive micromachined ultrasonic transducers (CMUT) with diamond membranes for the first time in the literature. Although silicon and silicon nitride (Si3N4) membranes have been generally used as the membrane material in CMUTs. These membrane materials have moderate properties that can cause damage during the operation of CMUTs. In this thesis, a new material for the membrane is introduced for CMUTs. Diamond has exceptional potential in the area of micro-nano technologies due to unrivalled stiffness and hardness, excellent tribological performance, highly tailorable and stable surface chemistry, high thermal conductivity and low thermal expansion, high acoustic velocity of propagating waves, and biocompatibility. Based on these excellent material properties, diamond is employed in the new generation CMUT structures for more robust and reliable operations. The microfabrication process of CMUT has been generally performed with either sacrificial release process or wafer bonding technique. High yield and low cost features of wafer bonding process makes it preferable for CMUT devices. In this thesis, plasma-activated direct wafer bonding process was developed for the microfabrication of 16-element 1-D CMUT arrays with diamond membranes. They were designed to operate at different resonance frequencies in the range of 1 MHz and 10 MHz with different cell diameters (120, 88, 72, 54, 44 &mu / m) and element spacing (250, 375 &mu / m). 1-D CMUT array devices can be used for focusing ultrasound applications. The electronic circuit for 1-D CMUT devices with diamond membranes was designed and implemented on PCB for the ultrasound focusing experiment. This electronic circuit generates continuous or burst AC signals of &plusmn / 15 V with different and adjustable phase shifting options at 3 MHz frequency. 16 elements of 72 &mu / m 1-D CMUT array were successfully tested. Fully functional 7 elements of 1-D CMUT array are focused at an axial distance of 5.81 mm on the normal to the CMUT center plane. The CMUT array was excited using 10 Vp&minus / p with 10 cycles sinusoidal signals at 3 MHz. The microfabrication process and focusing ultrasound of 1-D CMUT devices with diamond membranes are done successfully in this thesis.

Potential applications such as high intensity focused ultrasound in medical therapeutics require larger output pressures. To offer unprecedented acoustic output pressure in transmit without the limitations, Capacitive Micromachined Ultrasonic Transducer operation modes of collapse and collapse-snapback are introduced in literature. Both operation modes require the membrane to contact the substrate surface, which poses a problem on the durability of the membrane in terms of structural integrity and tribological property. Based on the additional requirements of these modes, diamond is proposed as the ultimate solution to be used as the membrane material. Mechanical, thermal, and electrical properties of diamond are all in favor of its use in the microfabrication of CMUTs. This thesis introduces the design and test results of the first diamond-based CMUTs as an alternative to silicon and silicon nitride based CMUTs. Simulations are performed using Finite Element Methods (FEM) using a commercially available software package, ANSYS. The diamond-based CMUT is operated successfully both in air and immersion for the first time. Fully customizable in-house software is developed to command and control the test setup equipments for current dissertation and future work. Fresnel and Fraunhofer regions of the CMUT are characterized in sunflower oil using a combination of advanced hardware and software. The experimental results of radiation and diffraction for the diamond-based circular CMUT are verified by the theoretical calculations for a circular piston transducer. The results obtained from the first generation diamond-based CMUTs presented the diamond as a promising material for membranes in CMUTs.

In this thesis, the synchronization in the arrays of identical and non-identical coupled harmonic oscillators is studied. Both linear and nonlinear coupling is considered. The study consists of two main parts. The first part concentrates on theoretical analysis and the second part contains the simulation results. The first part begins with introducing the harmonic oscillators and the basics of synchronization. Then some theoretical aspects of synchronization of linearly and nonlinearly coupled harmonic oscillators are presented. The theoretical results say that linearly coupled identical harmonic oscillators synchronize for any frequency of oscillation. For nonlinearly coupled identical harmonic oscillators, synchronization is shown to occur at large enough frequency values. In the second part, the simulator and simulation results are presented. A GUI is designed in MATLAB to run the simulations. In the simulations, synchronization of coupled harmonic oscillators are studied according to different coupling strength values, different frequency values, different coupling graph types (e.g. all-to-all, ring, tree) and different coupling function types (e.g. linear, saturation, cubic). The simulation results do not only support the theoretical part of the thesis but also give some idea about the part of the synchronization of coupled harmonic oscillators uncovered by theory.