<p>The major effort in this thesis has been directed towards the problem of short-term load demand forecasting. Special attention has been given to on-line load forecasting of several loading nodes simultaneously. A critical review of the state of the art in short-term load forecasting methods is presented. A new approach has been used to classify these methods according to the modelling approaches used for representing the load demand.</p> <p>Two multivariable, autoregressive and state space, models simultaneously have been proposed. The order of the autoregressive model is obtained without fitting coefficients to different models. Based on multivariable state space representation an efficient and completely automatic algorithm for on-line load forecasting is proposed. It has been shown that by utilizing the innovations representation an ordinary recursive least-square algorithm can be used to give unbiased estimates of the model parameters. The proposed algorithms obtain good forecasts even in case of loss of observations from one or more nodes.</p> <p>To obtain load forecasts for one-weak-ahead, the total load is divided into three components: nominal, weather sensitive, and stochastic. A special approach has been proposed to identify the weather-sensitive load component. Also, a study has been carried out to identify those weather variables which have a significant effect on the load demand.</p> <p>Actual load data from Ontario Hydro has been used to test the proposed methods.</p> / Doctor of Philosophy (PhD)

<p>This thesis explores two problems in target tracking and identification: (1) robust track state filtering, and (2) decision-level identity fusion. In the first part of the thesis, a novel finite-horizon, discrete-time, time-varying state estimation method based on on the robust semidefinite programming technique is proposed. The proposed method is robust to norm bounded parameter uncertainties in the system model as well as to uncertainties in the noise statistics. The robust performance of the proposed method is achieved by minimizing an upper bound on the worst case variance of the estimation error for all admissible systems. In the second part of the thesis, two decision-level identity fusion models are proposed: Similar Sensor Fusion (SSF) model and Dissimilar Sensor Fusion (DSF) model. In the SSF model, sensors provide reports on a set of common characteristics of a target, and the fusion objective is to find a fusion result which is most consistent with all the sensor reports. In comparison, sensors in the DSF model explore different characteristics of a target. Their reports are fused in a manner that leads to decreased uncertainty on teh arget identity. In other words, these reports reinforce each other to generate increased certainty on the target identity, rather than being averaged to minimize inconsistency. Furthermore, we propose several fundamental principles for identity fusion, based on which all existing and future identity fusion methods can be evaluated and compared. For the SSF model, two fusion methods are proposed: Convex quadratic fusion method and K-L fusion method. In the first method, inconsistencies between the fusion result and the sensor reports are measured by quadratic functions, and the problem is formulated as a convex quadratic programming problem. In the second method, Kullback-Leiber distance is used to measure the inconsistencies among the probabilistic sensor reports. The resulting formulation leads to a generalized analytic center problem. For the DSF model, we use a special objective function in the optimization formulation to accumulate the physical characteristics on a target explored by each sensor. The resulting fusion method involves solving an analytic center problem. Compared with the two classical decision-level identity fusion methods: Bayesian inference method and Dempster-Shafer evidential inference method, the three new fusion methods require no a priori information on the target, and enjoy small computation complexity. In addition, we show that the three new fusion methods, as well as the two classical methods, all satisfy the fundamental principles for identity fusion. The performance of the proposed fusion methods are illustrated in several numerical examples.</p> / Doctor of Philosophy (PhD)

<p>A new strategy for direct visual servoing for robotic position control is described. The approach does not rely on any external position or velocity sensors but directly sets motor current using visual feedback alone. The method is novel in that it can be implemented without specialized version hardware and is capable of processing visual feedback at high frame rates suitable for stable closed-loop position control of practical mechanical systems. A single RS-170 camera has a maximum sampling rate of 60Hz that significantly limits visual servoing performance. This limitation is overcome with multiple RS-170 cameras synchronized over a network in round-robin fashion to capture video fields at different instants in time. "Vision nodes", consisting of a camera and a dedicated computer, continuously process video at field rates to determine robot position. The vision algorithm, which is based on principal component analysis, is demonstrated to be suitable for accurate real-time position determination. Furthermore, the Euclidean distance in eigensepace in the presence of random occlusions is shown to be statistically related to the position measurement error variance. This leads to a novel approach for dealing with occlusions by considering them as "noise", the variance of which can be estimated directly from the Euclidean distance in eigenspace. A Kalman filter is then introduced to provide sensor fusion of the feedback from each vision node by weighting the position estimates from each camera to provide an improved overall position estimate. The Kalman filter also models the vision transport delays to reduce their effects on the visual feedback. Simulation results illustrate improvement in dynamic performance as the number of cameras are increased. Further simulations predict robustness to simulated occlusions. An experiment was designed and performed to experimentally verify the strategy for direct visual servoing. A 1 DOF servo drive equipped qith a rotating link constituted a simple "planar robot" testbed for demonstrating the distributed visioni and control techniques. The testbed was built using "off the shelf components" consisting of a network of four RS-170 cameras and computers connected to a master servo computer over a 100Mbps Ethernet network. An effective visual sampling rate of 240Hz was achieved. Several techniques were developed to achieve deterministic communications over Ethernet. The limitations on the number of cameras was analyzed and it was found that the Ethernet network could theoretically support up to 826 cameras. The main limitation was found to be the processor time on the master servo computer. Experimental results are shown for the system performing direct visual servoing under various different conditions. A direct visual servo emplying four cameras exhibits a step risetime of 190ms which closely matches the performance using traditional encoder feedback. Additional experimental results demonstrate the servo-hold performance and step responses with and without occlusions. Both full occlusions in a subset of cameras and partial occlusions in all cameras were investigated. The experment results validate the simulation results and verify that the strategy is capable of stable direct visual servoing and is robust to occlusions. Additional experiments are included that demonstrate performance under varying illumination conditions and various tele-robotic extensions.</p> / Doctor of Philosophy (PhD)

<p>The focus of this thesis is on blind identification techniques for multi-input, multi-output (MIMO) systems. In this respect we study three problems: 1. The joint diagonalization problem: Joint diagonalization is an efficient tool for blind identification techniques for MIMO systems. In this thesis we discuss new adaptive joint orthogonal diagonalization algorithms based on optimization methods over the stiefel manifold. 2. Blind identification of MIMO systems: We demonstrate that by using the second-oder statistics of the system outputs, by exploiting the non-stationary of sources, and some mild conditions on the sources and the system, the impulse response of the MIMO system can be identified up to an inherent scaling and permutation ambiguity. An efficient two-step frequency domain algorithm for identifying the MIMO system then has been proposed. Numerical simulations verify the theoretical results and the performance of the new algorithm. 3. Real room blind source separation problem: The final part of the thesis focuses on the practical problem of blind source separation of mixed audio signals in a real room. The new proposed algorithm exploits the non-stationarity of audio signals to separate them from their mixtures recorded in a reverberant environment. This method has successfully been applied to real data acquired during extensive recording experiments done in different office rooms on the McMaster campus.</p> / Doctor of Philosophy (PhD)

<p>A survey and analysis of visual measurement of camera and object position in visual sub-space (eigenspace) is provided leading to several improvements to existing methods, as well as new approaches. Specifically, novel techniques were developed to allow robust measurements in the presence of occlusions and other dynamic scene changes which is known to be a significant challenge in pose measurement methods for important applications such as visual servoing, autonomous robotics in manufacturing and tele-robotics including aerospace, medical operations and others. Local image information is shown to retain positional information in unoccluded regions that can then be used to determine position in the presence of significant dynamic occlusion. Local information is also shown to be more prone to ambiguity errors due to a lack of salient positional information. A subsectioning and recombination strategy is developed that features the advantages of local eigenspace independence for robustness to occlusion while maintaining the inherent resistance to ambiguity available from global eigenspace analysis. This is achieved by computing modified global projections, while excluding information from occluded sections. A new method for occlusion detection using an eigenspace reconstruction error measure is also developed and evaluated. A wide variety of experimental measurements are provided to demonstrate the performance of the new methods using an accurate XYZ platform and CCD cameras with metallic, machined parts. Experimental measurements are also performed to demonstrate improvements for eigenspace position accuracy through the use of multiple cameras. Several techniques are employed to combine and fuse multiple images from decoupled cameras whereby cameras are used for determining position in different directions to to improve accuracy. Subsequently, multiple cameras are applied to achieving three dimensional translational position measurements in visual subspace.</p> / Doctor of Philosophy (PhD)

<p>Magnetic nerve stimulation has proven to be an effective, non-invasive technique to excite peripheral and central nervous systems. In this technique, the excitement of the neural tissue depends on exposure to a transient magnetic field generated by passing a high pulse of current through a coil. By positioning the coil in a specific orientation over the targete tissue, the transient magnetic field will induce an electric field in the conductive milieu of the body. If this field reaches a certain threshold within a specific time period, neural depolarization is then evident.</p> <p>Recently, transcranial magnetic stimulation showed promise as a novel treatment for mood disorder and other psychiatric illnesses. Considering that these sicknesses are currently one of the larges causes of disability (with a cost sometimes exceeding that of diabetes, heart disease, and hypertension) it is evident that pursuing this therapeutic approach is beneficial. However, as magnetic stimulation is relatively new, many limitations and obstacle are to be addressed before this innovative technology is approved for clinical applications.</p> <p>The main limitations in applying magnetic nerve stimulation are poor field focality and inadequate strength in deep tissues (targeted area). This, the excitation of these regions requires very high coil currents to achieve a strong field that is capable of penetrating deep tissues. However, these fields might activate adjacent tissues as well as the targeted area. Further, the high currents will result in coil heating, especially for a high rate of repetition.</p> <p>Another limitation, especially in transcranial stimulation, is the lack of knowledge of precisely which area of neural tissue is being affected since no immediate behavioral effects of current reversal in the coils and the virtual cathode of the coil suggests conflicting results.</p> <p>The primary objective of this thesis is the development and testing of new coil designs that can focus the magnetic field more effectively. Two such coils have been built. The first coil has an air core, while the other has a magnetic core. The magnetic fields of these coils, applied to the human upper limb, have been determined theoretically, and the results compared to the field generated by the most generated by the most common commercial coil, the Figure-8 coil. To design these coils and to test them experimentally, a current pulse generator has been designed and built. Further, a novel meaurement system using surface mount inductances and a computer based data acquisition system has been designed and built. The experimental results confirm the theoretical findings, that the air core coil is slightly better than the Figure-8, as far as field strength and focality are concerned. In addition, the experimental results, prove that the coil with the ferromagnetic core, is superior.</p> <p>The second objective is to investigate the effect of stimulus waveforms theoretically, experimentally, and through in vivo study. The goals of the study are to establish a quantitative relationship among various waveforms and to investigate the effect of these waveforms in determining the site of stimulation. Accordingly, a multi subject trail was conducted: a Figure-8 coil was applied to the median nerve of ten subjects at the upper limb. The motor responses of the thenar muscle were then recorded. The study results show that the biphasic stimulating pulse is more effective than the monophasic pulse, as has been concluded by other researchers.</p> <p>However, the effective stimulating point, or virtual cathode of the coil, was found to be not simply 3 to 4 cm from the coil center as had been reported. In fact, the study shows that the site of the virtual cathode is affected by the current amplitude and the degree of inhomogeneity of the tissues surrounding the nerve. Furthermore, reversing the coil current direction results in a different level of stimulation but does not affect the virtual cathode position.</p> <p>In summary, the research presented in this thesis covers theoretical concepts, experimental aspects, and human studies related to neuro magnetic stimulation. The results of the experiment and the study are consistent with the theoretical analysis. The proposed coil design is novel and offers promise for a better coil system for magnetic nerve stimulation</p> / Doctor of Philosophy (PhD)

<p>This thesis presents a continuation in the process of rationalizing, unifying and improving existing model reduction techniques. Thus a method of reduction is developed which combines the method of aggregation and partial Pade approximation in such a way as to maintain their separate advantages while simultaneously removing their disadvantages. The important aspects associated with the reduced-order models obtained are: guaranteeing the stability of the reduced-order models saving computation time, retaining the invariance property under state variable feedback conditions and matching some of the original system time moments.</p> <p>Also, a criterion is proposed for selecting the state variables of the original system to be retained in the reduced-order model. This criterion leads to developing a reduction technique which can be regarded as a combination of the methods of aggregation and singular perturbation. Therefore, the reduced-order model obtained retains the physical significance of the state variables and the dominant eigenvalues of the original system.</p> <p>Furthermore, a procedure is developed for obtaining dynamic equivalents of multimachine systems. This procedure utilizes the concept of component cost analysis for identifying the coherent groups of generators.</p> <p>Verification of the methods developed in the thesis is established using a variety of realistic power system models including a single synchronous machine connected to an infinite bus, a three-machine system and a 10-machine system. These applications include simulation, analysis and simple controller design.</p> / Doctor of Philosophy (PhD)

<p>This thesis contributes to the development of novel absorbing boundary conditions (ABCs) for two finite-difference time-domain (FDTD) methods in computational electromagnetics; the well-known FDTD method to the solution of Maxwell's equations and the wave equation in the time-domain (WTD) method. The conventional approach to create a perfectly matched layer (PML) ABC for the FDTD solution to Maxell's equations is reviewed, as well several other state-of-the-art techniques to define a PML medium. The novel WETD technique is described for applications in numerical algorithms both for optical waveguides and structures, and in the microwave and millimetre-wave structure analysis. The new algorithm requires a reliabe efficient ABC, which can handle both open problems (i.e., radiation and scattering) and problems involving port terminations (such as high-frequency circuit problems and optical guiding structures). A new degree of freedom is introduced in the definition of the PML variable profiles and thus improved algorithms for PML ABCs are formulated and developed both for the WETD method and for the FDTD solution to Maxwell's equations. The proposed modified PML profiles handle equally well both port terminations in guide-wave problems and truncations of the computational domain of open problems. The performance of the proposed PML absorber is further improved by new types of PML termination walls using single-layer ABCs. For that, a lossy version of Mur's second order ABC and a lossy version of the second order dispersive boundary condition (DBC) have been developed on orthogonal non-uniform grids. It handles inhomogeneous dielectrics intersecting the PML boundaries. Various numerical simulations have been carried out to validate the theoretical models at microwave and optical frequencies, as well as in depth detailed comparison with commonly used PML ABCs is presented. Suggestions for further research are provided.</p> / Doctor of Philosophy (PhD)

<p>Most intriguing and complicated scientific problems are solved with the aid of high-speed computers. Advancements in computer technology in conjunction with lower cost of computers have tremendously contributed to this effort. In this thesis, adopting the same philosophy, we explore the solutions to electromagnetic problems using a numerical technique, namely the Finite Difference Time Domain technique. This technique requires large computer resources: specifically, fast computing and large amount of computer memory. Therefore, our developments of this numerical technique focus on reducing both the computational time and computer memory requirements. A unique line-of-sight (LoS) approximation to the equivalence principle, which is a far more efficient alternative to the standard equivalence principle, is introduced. This includes the theory associated with the LoS equivalence, the examination of the error due to the approximation, and the verification with simulations and measurements. The FDTD method is formulated by discretizing Maxwell's curl equations over a finite volume and approximating the derivatives with central difference approximations. The error due to this approximation depends directly on the method of implementation such as the non-uniform mesh. An experimental investigation of the numerical error due to the non-uniform FDTD technique is presented. In the area of active and passive microwave devices, a unique method of incorporating the device equations into the FDTD algorithm is presented. The measurement and simulation results verify this method. In addition to device analysis, new antennas for the use in Personal Communication Systems are presented. The finale to the thesis is an electromagnetic simulation software package that incorporates all the new techniques developed in the course of this work.</p> / Doctor of Philosophy (PhD)

<p>The rapid growth in mobile and satellite communications has intensified the requirements for good performance, compact structure, high quality and low cost waveguide filters and diplexers. This dissertation is devoted to the full wave analysis and modeling of various circular-rectangular (C-R) coaxial waveguide structures which are commonly used to develop combline filters and diplexers. Specifically, models that can be cascaded to simulate the system performance of the filters and diplexers are being sought in the dissertation. The research includes three parts: (1) modal analysis of the higher-order modes in the C-R waveguide; (2) modal analysis of the TEM mode in the CR waveguide; and (3) the scattering characteristics of the right-angle bend and the T waveguide junctions loaded with a generic post. A rigorous analysis, which combines the orthogonal expansion method and the Galerkin method, is performed to obtain the higher-order eigenmodes in the C-R waveguide. The Bessel-Fourier series is employed to merge the circular and rectangular coordinate systems used in the analysis. The cutoff frequencies of the higher-order modes are determined using the singular value decomposition (SVD) technique. The modal solution of the TEM mode in the C-R waveguide is obtained by superposition of the resonant modes in an equivalent rectangular cavity loaded with a conducting post. The characteristic impedance and attenuation coefficient of the waveguide are derived from the solution of the TEM mode. Analytic models of the right-angle bend and T -junctions loaded with posts of varying heights are derived. A novel technique of the extended eigen mode functions is developed to deal with the complex boundary conditions in the junction structures. The general scattering matrices of the right-angle bend and T - junctions are obtained.</p> / Doctor of Philosophy (PhD)