<p>The identification of discrete linear time-invariant multivariable systems has been investigated. This thesis consists of three interrelated parts. In the first part, linear systems have been described in canonical forms which are mostly suitable for system identification. Inter-relationships between the input-output difference canonical form due to Bonivento, Guidorzi and Marro (1973) and the state-space canonical form due to Luenberger (1967) have been studied. In the second part structures of linear multivariable systems have been studied. An efficient algorithm has been developed to identify a set of structural indices directly from the input-output data of a system. The order of the system is, then, determined in terms of these structural indices. In the last part, estimation of system parameters from the input-output data has been studied. Recursive algorithms have been developed for estimating the system parameters sequentially from noise-free and noise-corrupted data. Digital computer simulations with noise-free as well as noise-corrupted data have been utilized to illustrate the usefulness of the proposed procedures.</p> / Master of Engineering (ME)

<p>Based on Altman's model, a state model and a measurement model are obtained. Both models are linearized and the linear state variable feedback method is then applied to decouple the linearized system.</p> <p>Computer programmes for decoupling are written and the results are then discussed.</p> <p>Suggestions for future work are proposed.</p> / Master of Engineering (ME)

<p>By making use of measurements of the input and output of an unknown system, the characterizing parameters are found by using matrix pseudoinverse method. For noise-corrupted measurement, least squares estimation of the parameters are obtained.</p> <p>Stochastic approximation is employed to improve the estimation.</p> <p>Both methods are tested on-line in real time using the PDP-11/45 minicomputer while the system is simulated on the TR-20 analog computer.</p> / Master of Engineering (ME)

<p>This thesis covers the design and fabrication of a Surface Acoustic Wave (SAW) oscillator. The oscillator was designed to operate at 124 MHz in the fundamental mode and was constructed on an ST-X quartz substrate. The type of SAW oscillator incorporated was of the delay line type where a SAW delay line is inserted in the feedback loop of an amplifier with excess gain. The delay line exhibited a bandpass characteristic with an insertion loss of 32 dB. The oscillator has a short-term stability of 2.0 parts in 10⁷over 1 second and a medium-term stability of less than 8 ppm over 12 hours. The oscillator had an effective loaded Q of 628 and power output was -10 dBm into 50 Ω with the second harmonic down 22 dB, and the third down 33 dB. The theory of operation of the SAW delay line oscillator is covered along with fabrication procedures and experimental results.</p> / Master of Engineering (ME)

<p>The concepts of the analogue channel vocoder and digital FFT vocoder are utilized in the analysis of a Walsh vocoder. The FFT and Walsh vocoders are simulated in Fortran on a CDC6400 computer. The simulations contain a simplified cepstrum pitch detector to simulate a hardware pitch detector. The Walsh vocoder simulation is identical to the FFT vocoder simulation with the Fast Walsh Transform operation replacing the Fast Fourier Transform operations (FWT and FFT respectively) in the channel signal calculations and transforms into the time domain.</p> <p>Sentences of different lengths and speakers are processed by the simulated vocoders and the resultant synthesized speech is analyzed for comparative quality and intelligibility.</p> <p>An unsuccessful attempt is made to effect a vocoder which does not require a pitch detector with the calculation of the time invariant sequency power spectrum replacing the channel signal calculations in the vocoder simulations. Long term frequency and sequency enveloped are used to shape the frequency and sequency power spectra derived from the transmitted time invariant sequency power spectrum in the FFT and Walsh vocoder simulations respectively.</p> / Master of Engineering (ME)

<p>The purpose of this thesis is to present a model for magnetic domains wholly situated inside an ion implanted channel. In this model, ion implantation is assumed to change the saturation magnetization, the wall energy density, and the in-plane susceptibility. As a result the channel and the domain induce magnetic pole distributions on the channel walls which interact with the bubble domain, altering its geometry and energy. An analysis for a circular domain with variable penetration and location in the channel, as well as a variational formulation for a generally deformable domain of fixed penetration in the implanted channel, are presented. The latter case has been programmed and the results obtained are discussed. The case of a domain penetrating into the unimplanted region outside the channel is also considered.</p> / Master of Engineering (ME)

<p>In this work, a computer simulation for ion implantation has been carried out based on a simplified random model. This simulation has been used to obtain range profiles and deposited damaging energy distributions, for any system and for a wide energy range. The obtained distributions show good agreement with the third order Edgeworth expansion approximation of the standard transport theory solution (the WSS theory). By taking into account the motion of the first knock on atoms, as well as the incident ion itself, a good agreement between the simulation and the WSS theory is found for the damage distribution. A random model for dealing with poly-atomic targets is presented. Results obtained show good agreement with the available experimental results for range profiles. The simulation includes obtaining a complete picture of the collision cascade (in space and time). The complete picture of the collision is discussed in relation to the recently observed "molecular effect" and explains some available experimental results. The examination of typical cascades suggests that high energy density regions are responsible primarily for the obtained departures from "Linear Cascade Theory".</p> / Master of Engineering (ME)

<p>The problem of adaptive state estimation which involves the identification of the Kalman gain matrix without a priori information on the noise statistics is presented. A scheme incorporating an identification algorithm and a tracking algorithm is proposed. This scheme provides a powerful approach for adaptive state estimation.</p> <p>An ARMA model for system description is derived for preliminary analysis of the noise transition matrix when the observation noise is sequentially correlated.</p> <p>The innovations process for systems with coloured observation noise is shown to be white for optimum filtering.</p> <p>Simulations are performed on an inertial navigation system for both white and coloured observation noise. Numerical results indicate the superiority of the filter with tracking over one without. Performance of the filter for coloured observation noise confirms the theoretical derivation of the ARMA model.</p> / Master of Engineering (ME)

<p>Recently the problem of dynamic instability has been increased due to the recent trends in synchronous machine design and the growth in the relative size of the power stations. In this thesis such a problem has been examined by analysing the damping and synchronizing torque coefficients. Different methods for calculating those torque coefficients have been reviewed, and a new method based on a more accurate representation for the machine has been developed. Because of the more detailed representation the new resultant torque coefficients permit the investigation of additional properties of those coefficients, as well as more precise analysis of the properties previously studied. The problem of weak or negative damping associated with the use of high gain and quick response static exciters has been given special attention. Different schemes for producing a stabilizing signal have been discussed. A design procedure for a stabilizer which assures successful operation over the whole generation range has been outlined.</p> / Master of Engineering (ME)

<p>Theory, design and fabrication of a Surface Acoustic Wave (SAW) TV-IF filter is presented in this thesis. The filter is a linear phase bandpass filter with a passband extending from 41.25 to 47.25 MHz. Optimization techniques were used for designing the filter impulse response. Conventional photolithography was used for the device fabrication. The results obtained correspond very well to the theoretical design.</p> / Master of Arts (MA)