<p>A representative set of binary image processing techniques selected from the literature is described. The measurement of shape as the fundamental information contained in silhouettes is examined. Operations on digital binary images are demonstrated including smoothing, connectivity analysis and determination of position and orientation. The effects of digitizing errors at the boundary of a silhouette are discussed and examples of industrial vision systems which use binary images are presented.</p> <p>A binary image processing system has been designed and implemented. The apparatus is based on a General Electric TN2500 digital television camera and an Intel iSBC 86/12A microcomputer. Hardware for the acquisition of binary images from the camera is described followed by the software for calculating areas and centroids. The system is capable of "learning" a set of objects in a "Teach" mode and then making an identification based on their area in the "Run" mode.</p> / Master of Engineering (ME)

<p>Variable speed AC drives have become the new standard for high performance drive systems. Very high power traction drives present a particular challenge for AC drive technology because of the simultaneous need for high-torque vibration-free operation at standstill, traction-limited dynamic braking, and high speed operation.</p> <p>In this work, two independently controlled cycloconverters are used with a wound rotor machine to provide a double fed drive system that overcomes the frequency limitations of the cycloconverters and provides a stable high power drive, with potential for rapid torque response and power factor control.</p> <p>A new 'Jitter' control method is derived for the firing angle control of the cycloconverter. This method effectively spreads the cycloconverter output harmonics over a broader spectrum and thus minimizes the filtering requirements.</p> <p>The double fed motor (DFM) equations are applied to the DFM circle diagram and a simulation program has been written to plot the circle diagram and give a geometric interpretation of the developed torque. A new algorithm is proposed using δν, the pseudo torque angle, to give direct control of the torque using only position feed-back. A novel frequency hopping algorithm is also derived that allows the cycloconverters to be operated at high frequencies without generating harmonic torques.</p> <p>A dual cycloconverter drive was built to verify the stability and torque-speed performance of such a system. Reliable operation was achieved by using bank switching sensors based on detecting the reverse bias across a series connected diode, and by short circuit detection circuits with automatic reset facilities. The control software displays system status information on the screen, with on-line parameter modification, and provides a complete range of manual and automatic modes of operation to facilitate system development and testing.</p> / Doctor of Philosophy (PhD)

<p>Search and rescue satellite aided tracking (SARSAT) involves the use of satellites in low-polar orbits which relay the emergency signals of distressed vehicles to an earth station for signal analysis. A significant number of lives have been saved by using SARSAT. The program originated from the efforts of the three member countries of Canada, the United States and France who formed a joint venture with the Soviet Union in a program referred to as COSPAS/SARSAT. Since then, the United Kingdom, Norway and Sweden have joined the program and is expected that the system will be used worldwide.</p> <p>In this thesis, three different models for emergency locator transmitter (ELT) signals are proposed. Some basic concepts and a theoretical analysis of the spectrum produced by these signal models are studied. The first model, the Ideal Coherent Model ELT, produces a highly idealized spectrum which does not exist in practice. The second model, the Non-Ideal Coherent Model ELT, is capable of producing a wide range of spectra which can he closely related to actual ELT signals. Mathematical analysis identifies a design problem not previously recognized and provides the basis for a new design specification which should become mandatory. The third model, the non-Coherent Model ELT, represents a class of ELT signals having very poor spectral properties. This type of ELT design should be avoided.</p> <p>The performance of several different signal processors is evaluated using real ELT signals. The periodogram gives good performance for coherent ELT signals. However, the linear spectral estimation technique is not suitable for determining carrier frequency for non-coherent ELT signals. Use of a modified maximum entropy method in detecting these type of signals is demonstrated. In addition, spectrum ranking, sub-group averaging and spectrum levelling are found to benefit the spectral performance of the signal processors.</p> <p>In SARSAT signal environment, it is possible to receive many simultaneous emergency beacon signals combined with interference of various types. The thesis also examines the sources or interference which enter the 121.5/243 MHz frequency bands of the SARSAT system.</p> / Doctor of Philosophy (PhD)

<p>There has been an upsurge in interest recently in lower loss surface acoustic wave (SAW) devices for new signaI processing applications. In this thesis, a new generation of unidirectional transducers (UDTs) for the realization of low-loss SAW devices has been studied in which unidirectionaIity is effected by exploiting coherent internal reflections within a transducer. These single-phase UDTs (SPUDTs) offer many advantages and attractions that are unavaiIable in many previous UDT designs. CoupIed-mode theory is extensively used to analyze and characterize these structures. A general theory is proposed which will guide design and construction of all SPUDT devices. This includes specifications of the optimum conditions for achieving maximum directivity. An accurate and detailed model is also developed for SPUDTs with inter-IDT reflectors. The model accounts for all major electrode interaction effects and involves the cascade of transfer matrices of reflectors, transmission lines and transducers to realize the overall response. Experimental verifications have consistently shown that this model is very accurate in the analysis of SPUDTs.</p> <p>Based on a better understanding of the SPUDT concept, several new SPUDT designs have been developed. Special efforts are expended towards acquiring key properties never before obtained with SPUDTs, such as improved sidelobe suppression, wider bandwidths and higher operating frequencies. In these respects, the new designs represent a significant improvement over the earlier SPUDT designs. The final part of the thesis describes the theory and operation of a novel SPUDT-based multimode oscillator for frequency agile radar. An innovative injection-locking scheme involving FM chirp-mixing is also disclosed. Preliminary tests have indicated very encouraging results including low phase noise levels (120 dBc/Hz @ 10 kHz), wide tuning bandwidth (50%) and fast switching speeds (< 2 us).</p> / Doctor of Philosophy (PhD)

<p>A microcomputer based image processing and display system was constructed, consisting of an Intel 8080 microcomputer development system, a 128 kilobyte memory, a colour television monitor, and an image display controller and interface. This display system was used to compare images from airport and surveillance radar and meteorological satellites. the main purpose of the comparison was to detect and classify hazards to aircraft, particularly weather and migratory birds. The image display system is a relatively inexpensive means of performing basic processing of images for a variety of applications.</p> <p>Data from the Winnipeg airport radar and from the NOAA 5 meteorological satellite, of the same area, was simultaneously obtained during May 1978, the time of the bird migration. The two images were geometrically transformed into the same map projection and then compared. It was found that some clouds were common to both images, but many were not. Birds, on the other hand, are only detected by the radar. More data needs to be collected and analyzed, under different weather conditions and the radar optimized for detecting clouds.</p> / Master of Engineering (ME)

<p>The subject of this Thesis is the theoretical and experimental investigation of the capabilities and the limitations of Surface Acoustic Wave (SAW) Interdigitated Interdigital Transducer (IIDT) filters. These devices provide the means of achieving low-loss SAW filters with percentage bandwidths exceeding 10%. During the course of developing an analytical tool for IIDTs, a menu driven computer program based on a modified version of the Coupling of Modes (COM) theory was written to allow the frequency response of arbitrary SAW device structures, composed of transducers, reflector arrays and spaces, to be evaluated without further programming. Several IIDT filters were fabricated and tested with the aim of reducing the inherent ripple in the frequency response of the structures. Techniques are proposed for achieving losses of less than 6 dB with acceptable ripple suppression. In addition, some studies are performed on IIDT comb filters and their application to multimode oscillators.</p> / Doctor of Philosophy (PhD)

<p>A chain of 42 bidirectionally coupled relaxation oscillators was used to model electrical activity in the smooth muscle portion of the opossum esophagus. Each oscillator was represented by two first order nonlinear differential equations. The equations were programmed on a Nova 830 minicomputer to develop the model and to test the responses to various stimuli.</p> <p>The model simulated distally, proximally and bidirectionally propagated contractions (represented by oscillations at 2.5 Hz) when the pulse representing direct muscle stimulation was applied at the beginning, end or in the middle of the chain. Contractions were not present in the absence of a stimulus. The velocity of propagation increased from 3.6 cm/sec among oscillators representing the upper part of smooth muscle to 1.85 cm/sec among distal oscillators representing the lower part of the smooth esophageal muscle.</p> <p>The model showed "ON" and "OFF" responses upon depolarization or hyperpolarization of the oscillators. The velocity of propagation of "ON" responses was variable and depended upon the amplitude of the stimulus which was applied to all of the oscillators simultaneously. The velocity of propagation of the "OFF" responses was much faster than that of the "ON" responses when the stimulus was applied to all oscillators. The number of spikes in a burst of "OFF" responses was variable and depended upon the amplitude and the duration of the stimulus applied.</p> <p>In conclusion, the model simulated all of the characteristics observed in the smooth muscle portion of the opossum esophagus. This includes those which were observed on vagal stimulation and balloon distension. The model also elucidates the mechanism of functioning of the myogenic control system and its modulation by the neuronal control system.</p> / Master of Engineering (ME)

<p>The Transient Energy Function (TEF) method represents a powerful technique to analyze the transient stability of large-scale power systems. Currently, in the applications of the TEF method, the power network is reduced by eliminating all buses and retaining only the internal nodes of the generators. This Reduced Network Formulation (RNF) yields dense (non-sparse) matrices in the computations and consumes significant computational time. This represents a serious drawback of the RNF, especially in applications to large power networks. Also, all system loads are modeled as constant impedance loads in order to use conventional techniques to reduce the network to the internal nodes of the generators. Many loads in practical power systems can be represented as constant power loads. Such loads are conventionally approximated as constant impedance type based on the pre-fault conditions. Consequently, accurate results may not be obtained. Moreover, the TEF is not applicable to very large-scale power systems due to the computer storage related problems (e.g, file paging) and excessive computational time.</p> <p>A novel formulation of the TEF method, retaining the original structure of the system network, is presented and the associated computerized algorithm is described. All the above mentioned problems are solved using the proposed Sparse Formulation (SF).</p> <p>The sparse formulation avoids network reduction completely. All matrices used in the calculation of both the Stable Equilibrium Point (SEP) and the Unstable Equilibrium Point (UEP), for which the computational times are dominant in the calculation process of the energy margin (the stability index), are very sparse. This leads to a significant saving in computational time, i.e. the sparse formulation is more efficient as compared with the RNF approach.</p> <p>The sparse formulation is applied to different (realistic) utility systems of up to 300 generators and 1724 buses. The results prove the superiority of the sparse formulation in contrast with other current methods.</p> <p>In addition, either constant impedance or constant power load models, or any combination thereof, can be handled explicitly. Considering these actual load models, the stability indices (the critical clearing time and the calculated more accurately.</p> <p>The proposed technique can handle very large scale power systems which are beyond the scope of RNF approach. Consquently, it enables an improved design methodology of transmission networks by including provision for modeling the network in more detail. Using the sparse formulation, it is possible to perform a transient stability analysis on a microcomputer. This will render coat-effective the use of such analysis throughout the world. Also, a very powerful and robust numerical technique to deal with ill-conditioned power systems is described. Therefore, practical (stressed) power systems can be handled, i.e. the sparse formulation is more reliable than other techniques such as RNF.</p> / Doctor of Philosophy (PhD)

<p>This thesis offers a unified and integrated treatment of three essential aspects of computer-aided circuit design: effective use of the state-of-the-art optimization tools, efficient calculation of exact and approximate gradients, and adequate mathematical representation of the engineering problems.</p> <p>The recent advances in gradient-based ℓp optimization are reviewed. The essence of the trust region Gauss-Newton method and the quasi-Newton solution to optimality equations is described. A new algorithm for linearly constrained one-sided ℓ₁ optimization is presented.</p> <p>Efficient approaches to network sensitivity analysis are addressed. Useful formulas are derived for general multi-ports, especially two-ports. Novel proofs of an important result for lossless two-ports are given.</p> <p>The basic formulations of nominal circuit optimization are introduced through a hierarchy of simulation models. Variables, error functions and ℓp objectives are identified. Optimization of multi-coupled cavity filters is described and illustrated by examples of elliptic, self-equalized and asymmetric designs. Large-scale optimization of multiplexers is also discussed.</p> <p>Realistic consideration of tolerance and uncertainties is of prominent interest to circuit, especially integrated circuit designers. A multi-circuit approach to design centering, tolerancing, tuning and yield enhancement is presented. Techniques for statistical design are reviewed. A generalized ℓp centering algorithm is developed.</p> <p>A novel approach to device modeling which utilizes multiple circuits and exploits the theoretical properties of the ℓp norm is described. It emphasizes the uniqueness and consistency of an equivalent circuit model. Practical applications are formulated and illustrated through industrial examples.</p> <p>A new algorithm for optimization with integrated gradient approximations is offered. Implementations for the minimax and ℓp problems are shown. The efficiency and usefulness are demonstrated by a large variety of examples.</p> / Doctor of Philosophy (PhD)

<p>This thesis addresses itself to computer oriented techniques for large scale optimization of analog circuits. New techniques for simulation and sensitivity analysis are described and are used to improve the performance of circuit optimization. A powerful automatic decomposition technique is developed directly enabling a normal optimizer to solve large circuit problems. Our theory is applied to the design of microwave circuits.</p> <p>The status of large scale circuit optimization and the state-of-the-art of microwave CAD are reviewed. The necessity of circuit oriented optimization techniques is demonstrated by formulating design, modelling, diagnosis and tuning into optimization problems.</p> <p>A comprehensive treatment of large change sensitivity computation for linearized circuits using generalized Householder formulus is presented. A technique for circuit response updating via a minimum order reduced system is developed. By avoiding re-analysis of the complete circuit, our method is responsible for efficient simulation of large circuits when a subset of the circuit parameters is frequently perturbed.</p> <p>An elegant theory for simulation and exact sensitivity analysis of branched cascaded networks is described. Our approach explicitly takes the circuit-structure into consideration and does not deteriorate as the overall network becomes large. The practicality of the theory is illustrated by efficient optimization of microwave multiplexers consisting of multi-cavity filters distributed along a waveguide manifold. Examples of optimizing 12- and 16-channel multiplexers are provided.</p> <p>A novel and general automatic decomposition technique for large scale optimization of microwave circuits is presented. The partitioning approach proposed by Kondoh for FET modelling problems is verified. The application of our technique is demonstrated by the large scale optimization of a 16-channel multiplexer involving 399 nonlinear functions and 240 variables.</p> / Doctor of Philosophy (PhD)