Automatic object detection is a challenging field which has been evolving over decades. The application areas span many domains such as robotics inspection, medical imaging, military targeting, and reconnaissance. Some of the most concentrated efforts in automatic object detection have been in the military domain, where most of the problems deal with automatic target detection and scene analysis in the outdoors using a variety of sensors. One of the critical problems in Automatic Target Detection (ATD) systems is multiscenario adaptation. Most of the ATD systems developed until today perform unpredictably i.e. perform well in certain scenarios, and poorly in others. Unless ATD systems can be made adaptable, their utility in battlefield missions remains questionable. This thesis describes a methodology that adapts parameterized ATD systems with image metrics as the scenario changes so that ATD system can maintain better performance. The methodology uses experimentally obtained performance models, which are functions of image metrics and system parameters, to optimize performance measures of the ATD system. Optimization is achieved by adapting system parameters with incoming image metrics based on performance models as the system works in field. A simple ATD system is also proposed in this work to describe and test the methodology.

Making a computer &lsquo / see&rsquo / is certainly one of the greatest challanges for today. Apart from possible applications, the solution may also shed light or at least give some idea on how, actually, the biological vision works. Many problems faced en route to successful algorithms require finding corresponding tokens in different views, which is termed the correspondence problem. For instance, given two images of the same scene from different views, if the camera positions and their internal parameters are known, it is possible to obtain the 3-Dimensional coordinates of a point in space, relative to the cameras, if the same point may be located in both images. Interestingly, the camera positions and internal parameters may be extracted solely from the images if a sufficient number of corresponding tokens can be found. In this sense, two subproblems, as the choice of the tokens and how to match these tokens, are examined. Due to the arbitrariness of the image pairs, invariant schemes for extracting and matching interest points, which were taken as the tokens to be matched, are utilised. In order to appreciate the ideas of the mentioned schemes, topics as scale-space, rotational and affine invariants are introduced. The geometry of the problem is briefly reviewed and the epipolar constraint is imposed using statistical outlier rejection methods. Despite the satisfactory matching performance of simple correlation-based matching schemes on small-baseline pairs, the simulation results show the improvements when the mentioned invariants are used on the cases for which they are strictly necessary.

ABSTRACT SIMULATION BASED INVESTIGATION OF AN IMPROVEMENT FOR FASTER SIP RE-REGISTRATION TANRIVERDi, Eda M.Sc., Department of Electrical and Electronics Engineering Supervisor: Prof. Dr. Semih BiLGEN July 2004, 78 pages In this thesis, the Session Initiation Protocol (SIP) is studied and an improvement for faster re-registration is proposed. This proposal, namely the &ldquo / registration &ndash / activation&rdquo / , is investigated with a simulation prepared using OPNET. The literature about wireless mobile networks and SIP mobility is reviewed. Conditions for an effective mobile SIP network simulation are designed using message sequence charts. The testbed in [1] formed by Dutta et. al. that has been used to observe SIP handover performance is simulated and validated. The mobile nodes, SIP Proxy v servers, DHCP servers and network topology are simulated on &ldquo / OPNET Modeler Radio&rdquo / . Once the simulation is proven to be valid, the &ldquo / registration &ndash / activation&rdquo / is implemented. Different simulation scenarios are set up and run, with different mobile node speeds and different numbers of mobile nodes. The results show that the re-registration delay is improved by applying the &ldquo / registration &ndash / activation&rdquo / but the percentage of improvement depends on the improvement in the database access delay in the SIP Proxy server.

Phase control in microwave circuits is an impotant process. Especially, in certain applications such as phase array antennas, it is the main principle of opeation. In antenna arrays, each array element is fed by an individual oscillator. By controlling the phase of each oscillator, the radiation pattern and the RF power can be combined in space in certain directions. For such applications, phase shifters have been utilized extensively. However, their high costs, difficulties in design and efficiency are impotant disadvantages. More recently, another technique, &quot / Injection Locking&quot / or &quot / Phase Locking&quot / suggests to use a single reference signal injected into each oscilator element. Through this signal, the phase of the individual oscillators can be controlled and set to a desired value. Therefore, power combining in space or known as &quot / Spaial Power Combining&quot / is possible by using &quot / Phase Locking&quot / of individual oscillator elements. In this thesis, this new phase control technique is examined in theory and in application of a 1GHz oscillator system. A reference signal is injected into a voltage controlled oscillator, and the phase progression is obtained by tuning the oscilator&#039 / s free running frequency.

iv ABSTRACT DESIGN AND CONSTRUCTION OF REDUCED SIZE PLANAR SPIRAL ANTENNA IN THE 0.5-18 GHz FREQUENCY RANGE YILDIZ, 5nan&ccedil / M.S., Department of Electrical and Electronics Engineering Supervisor: Prof. Dr. Altunkan HIZAL October 2004, 106 pages In this thesis, theoretical and practical evaluation of usual spiral antenna is revised. Working principles of both types of planar spiral antennas as Equiangular and Archimedean are introduced. A predesigned microstrip tapered balun used for feeding section of a spiral antenna is simulated on Ansoft HFSS software. Successful simulation results are obtained and measurements of implemented balun structure are made by using an HP 8722 D vector network analyzer. Antenna measurement techniques used in this study are introduced. Measurement set-ups are defined and some preliminary knowledge is given on these. As the main matter of thesis, reduced size planar spiral antennas are designed and implemented. Return loss, gain / radiation and polarization patterns of antennas are measured. Datasets of measurements are compared with each other and with reference spiral antenna. Quite promising results are obtained and size reduction of spiral antenna is achieved in many aspects.

In this dissertation, new approaches in the design of a voice transformation (VT) system for Turkish are proposed. Objectives in this thesis are two-fold. The first objective is to develop standard speech corpora and segmentation tools for Turkish speech research. The second objective is to consider new approaches for VT. A triphone-balanced set of 2462 Turkish sentences is prepared for analysis. An audio corpus of 100 speakers, each uttering 40 sentences out of the 2462-sentence set, is used to train a speech recognition system designed for English. This system is ported to Turkish to obtain a phonetic aligner and a phoneme recognizer. The triphone-balanced sentence set and the phonetic aligner are used to develop a speech corpus for VT. A new voice transformation approach based on Mixed Excitation Linear Prediction (MELP) speech coding framework is proposed. Multi-stage vector quantization of MELP is used to obtain speaker-specific line-spectral frequency (LSF) codebooks for source and target speakers. Histograms mapping the LSF spaces of source and target speakers are used for transformation in the baseline system. The baseline system is improved by a dynamic programming approach to estimate the target LSFs. As a second approach to the VT problem, quantizing the LSFs using k-means clustering algorithm is applied with dimension reduction of LSFs using principle component analysis. This approach provides speaker-specific codebooks out of the speech corpus instead of using MELP&#039 / s pre-trained LSF codebook. Evaluations show that both dimension reduction and dynamic programming improve the transformation performance.

The introduction of broadband systems to communication and radar technologies has demanded the design of broadband antennas. In this thesis, broadband log-periodic dipole antennas and reflector antennas are investigated. A dual polarized log-periodic dipole antenna and a reflector antenna are designed and analyzed within the frequency band of 0.4-18 GHz. Basic theory and calculations about mutual coupling between the linear antenna elements are given. The currents at the bases of dipoles of the log-periodic antenna are found and these currents are used as inputs for the design of the reflector antenna. After the simulations made with MATLAB&Ograve / and FORTRAN, a prototype single polarized log-periodic antenna is produced. Developments on the prototype antenna are performed, both to improve the electrical characteristics of the antenna and to make the final design realizable. After the performance of the single polarized log-periodic antenna is found satisfactory, dual polarized antenna is produced. Performances of both single and dual polarized log-periodic antennas are measured in the anechoic chamber at ASELSAN&reg / Inc. After the measurements of the antenna, using the currents at the bases of the dipoles, a reflector antenna is designed using &ldquo / MIR&rdquo / program written in FORTRAN, which is based on physical optics method. A method to calculate the aperture blockage of the reflector is developed. Simulation results of reflector antenna are given.

In this thesis, performances of some Mobile IP handoff schemes have been compared. The comparison has been based on simulation results. Simulation study has been carried out with a MIP model developed using OMNeT++ and available model frameworks. The literature on Mobile IP and several improvements including handoff management schemes have been surveyed. A MIP model has been constructed and then validated with the help of some scenarios in the literature, especially the one found in [7]. The model has been then used to investigate performances of FMIP. The study also included performance of FMIP under local traffic where mobile hosts communicate with each other in the same domain. Simulations are carried out under several scenarios involving UDP and TCP transfers. Mobile host speed and base station buffer size variables have been changed throughout the simulations. The result show that use of L2 triggers reduces handoff latency as both FMIP Post-Reg and Pre-Reg have better performance than HMIP without L2 triggers. The results also show that FMIP Post-Reg is a good candidate for future MIP infrastructures with its low latency handoff characteristics due to bidirectional tunneling between old and new points of attachment. Moreover, the results suggest that FMIP Post-Reg is also the best handoff scheme under local traffic where mobile hosts communicate among each other in the same foreign network.

Congestion control is concerned with allocating the resources in a network such that the network can operate at an acceptable performance level when the demand exceeds or is near the capacity of the network resources. These resources include bandwidths of links, buffer space (memory) and processing capacity at intermediate nodes. Although resource allocation is necessary even at low load, the problem becomes more important as the load increases. Without proper congestion control mechanisms, the throughput may be reduced considerably under heavy load. Future applications are expected to require increasingly higher bandwidth and generate a heterogeneous mix of network traffic. ATM network is potentially capable of supporting all classes of traffic (e.g., voice, video, and data) and have multiple service classes allow audio, video and data to share the same network. Of these, the Available Bit Rate (ABR) service class is designed to efficiently support data traffic. Switch algorithms have been the most investigated topic of ABR. This has happened because the specification of ABR given by the ATM Forum allows a diversity of switch algorithms to be implemented. These range from the simplest binary switches to the more complex ER switches. The major part of this thesis has been devoted to ABR. First an introduction to the concept of congestion control and a brief literature survey of congestion control for ABR service of ATM networks are presented. Then two proposed congestion control mechanisms for the ABR service class in ATM networks are examined by means of simulation, showing the different degree of performance and complexity. The simulation results presented in this thesis were obtained using a network simulator written in C++. This network simulator is a small event driven program. Analytical results were derived for different network configurations and different scenarios using this program.

The plasma antennas offer a new solution to new requirements that are imposed on antenna systems with the advancing communication technology and increasing demand on wider frequency bands. In this thesis the plasma antennas are investigated for the radar and communication applications. The interaction of gas and semiconductor plasma with electromagnetic waves is inspected theoretically, and several experiments on the interaction of microwaves with gas plasma are conducted. Results of these experiments show that a relatively simple setup can produce plasma dense enough to interact with microwaves of frequency about 8 GHz. The previous studies of other institutes on plasma antennas are surveyed, emphasizing the results important for the use in radar and communication applications. Finally, semiconductor plasma is introduced, and an antenna system utilizing the semiconductor plasma generated by optical excitation is proposed.