This thesis concerns blind source separation techniques using second order statistics and higher order statistics for reverberant environments. A focus of the thesis is algorithmic simplicity with a view to the algorithms being implemented in their online forms. The main challenge of blind source separation applications is to handle reverberant acoustic environments; a further complication is changes in the acoustic environment such as when human speakers physically move. A novel time-domain method which utilises a pair of finite impulse response filters is proposed. The method of principle angles is defined which exploits a singular value decomposition for their design. The pair of filters are implemented within a generalised sidelobe canceller structure, thus the method can be considered as a beamforming method which cancels one source. An adaptive filtering stage is then employed to recover the remaining source, by exploiting the output of the beamforming stage as a noise reference. A common approach to blind source separation is to use methods that use higher order statistics such as independent component analysis. When dealing with realistic convolutive audio and speech mixtures, processing in the frequency domain at each frequency bin is required. As a result this introduces the permutation problem, inherent in independent component analysis, across the frequency bins. Independent vector analysis directly addresses this issue by modeling the dependencies between frequency bins, namely making use of a source vector prior. An alternative source prior for real-time (online) natural gradient independent vector analysis is proposed. A Student's t probability density function is known to be more suited for speech sources, due to its heavier tails, and is incorporated into a real-time version of natural gradient independent vector analysis. The final algorithm is realised as a real-time embedded application on a floating point Texas Instruments digital signal processor platform. Moving sources, along with reverberant environments, cause significant problems in realistic source separation systems as mixing filters become time variant. A method which employs the pair of cancellation filters, is proposed to cancel one source coupled with an online natural gradient independent vector analysis technique to improve average separation performance in the context of step-wise moving sources. This addresses `dips' in performance when sources move. Results show the average convergence time of the performance parameters is improved. Online methods introduced in thesis are tested using impulse responses measured in reverberant environments, demonstrating their robustness and are shown to perform better than established methods in a variety of situations.

<p>This Master’s Thesis was part of the project called Multimodalanalysis at the Depart-ment of Biomedical Engineering and Informatics at the Ume˚ University Hospital inUme˚ Sweden. The aim of the project is to develop multivariate measurement anda,analysis methods of the skeletal muscle physiology. One of the methods used to scanthe muscle is functional ultrasound. In a study performed by the project group datawas aquired, where test subjects were instructed to follow a certain exercise scheme,which was measured. Since there currently is no superior method to analyze the result-ing data (in form of ultrasound video sequences) several methods are being looked at.One considered method is called Independent Vector Analysis (IVA). IVA is a statisticalmethod to find independent components in a mix of components. This Master’s Thesisis about segmenting and analyzing the ultrasound images with help of IVA, to validateif it is a suitable method for this kind of tasks.First the algorithm was tested on generated mixed data to find out how well itperformed. The results were very accurate, considering that the method only usesapproximations. Some expected variation from the true value occured though.When the algorithm was considered performing to satisfactory, it was tested on thedata gathered by the study and the result can very well reflect an approximation of truesolution, since the resulting segmented signals seem to move in a possible way. But themethod has weak sides (which have been tried to be minimized) and all error analysishas been done by human eye, which definitly is a week point. But for the time being itis more important to analyze trends in the signals, rather than analyze exact numbers.So as long as the signals behave in a realistic way the result can not be said to becompletley wrong. So the overall results of the method were deemed adequate for the application at hand.</p> / Multimodalanalys

Documenting temporal changes to coastal zones is an essen­tial part of understanding and managing these environ­ments. The exclusive use of traditional surveying tools may not be practical for monitoring large, remote, or rapidly changing areas. This paper investigates the utility of multispectral Landsat Thematic Mapper satellite data for docu­menting changes to a Caribbean coastal zone using the change vector analysis processing technique. The area of study was the coastal region near the village of Buen Hombre on the north coast of the Dominican Republic. The primary habitats of interest were the intertidal mangrove for­ ests, and the shallow water seagrasses, macroalgae, and coral reefs. The change vector analysis technique uses any number of spectral bands from multidate satellite data to produce change images that yield information about both the magnitude and direction of differences in pixel values (which are proportional to radiance). The final products were created by appending color-coded change pixels onto a black-and-white base map. The advantages and limitations of the technique for coastal inventories are discussed.

Promotion of agroforestry practices in sub-Sahara Africa may help sustain subsistent food and wood production by integrating trees and crops on farmlands to replenish soil fertility and improve crop yield. Using rotational woodlot and pigeonpea intercropping systems in semi-arid Tanzania as case studies, my research screened suitable tree species to increase fuelwood supply and examined mechanisms for reducing tree-crop competition. By adopting nutrient use efficiency (the ratio of biomass yield to nutrient uptake) as a criterion, I found that selecting tree species of low wood nutrient concentrations would minimize nutrient exports by 42 – 60 %, thus reducing soil nutrient depletion while concurrently sustaining local fuelwood supply harvested from rotational woodlots. Currently smallholder farmers cannot afford to replenish soil fertility because of high fertilizer costs. However, 5-year tree fallowing raised soil N and P levels for maize culture as high as those from recommended fertilizer applications. Post-fallow maize yield was also increased significantly over natural fallow practices. Apparently there is a trade-off between yields of maize and fuelwood under rotational woodlot culture providing farmers the choice to proportion tree and crop composition based on priority demands. An alternative practice of intercropping pigeonpea with maize may also rapidly replenish soil fertility as well as enhance maize yield when competitive interactions between trees and crops are controlled. Vector analysis revealed that such interactions suppressed biomass yields of maize and pigeonpea by 30 % and 60 %, respectively, due to limited soil nutrients and/or moisture. Optimizing yields of both crops would require prescribed fertilizer addition when intercropped, but dose rates can be lowered by half under the improved fallow system due to alleviating interspecific competition. My findings form the basis of a plea for greater use of rotational woodlot and pigeonpea intercropping systems in semi-arid areas. I conclude that smallholder farm management of rotational agroforestry systems can be significantly improved by refining tree selection criteria and mitigating nutrient competition between trees and crops to maintain food and fuelwood production.

We consider two different problems in quantization theory. During the first part we discuss the so called Bennett's White Noise Hypothesis, introduced to study quantization errors of different schemes. Under this hypothesis, one assumes that the reconstruction errors of different channels can be considered as uniform, independent and identically distributed random variables. We prove that in the case of uniform quantization errors for frame expansions, this hypothesis is in fact false. Nevertheless, we also prove that in the case of fine quantization, the errors of different channels are asymptotically uncorrelated, validating, at least partially, results on the computation of the mean square error of reconstructions that were obtained through the assumption of Bennett's hypothesis. On the second part, we will introduced a new scalar quantization scheme, called a Beta Alpha Encoder. We analyze its robustness with respect to the quantizer imperfections. This scheme also induces a challenging dynamical system. We give partial results dealing with the ergodicity of this system.

Thesis (Ph.D.)--University of Nebraska-Lincoln, 2008. / Title from title screen (site viewed Oct. 31, 2008). PDF text: 134 p. : col. ill. ; 2 Mb. UMI publication number: AAT 3307115. Includes bibliographical references. Also available in microfilm and microfiche formats.

Thesis (M.Sc.) - Carleton University, 2006. / Includes bibliographical references (p. 91-93). Also available in electronic format on the Internet.

Thesis (Ph. D.)--University of Missouri-Columbia, 1998. / Typescript. Vita. Includes bibliographical references (leaves 122-126). Also available on the Internet.

Thesis (Ph. D.)--University of Missouri-Columbia, 1998. / Typescript. Vita. Includes bibliographical references (leaves 122-126). Also available on the Internet.

Este trabalho desenvolve um novo método de controle de eficiência energética no acionamento de motores de indução. Esta técnica utiliza um modelo de perdas e uma estrutura de controle baseada no mínimo consumo de energia elétrica. Em um primeiro momento, são apresentados os estudos com os principais controladores de eficiência energética da última década. Depois, são apresentados os modelos elétricos e mecânicos do motor de indução trifásico, a partir do modelo clássico, com o eixo de referência qd0. O controlador utilizado, é o controle por campo orientado indireto (IFOC - Indirect Field Oriented Control), composto por uma malha externa de controle de velocidade e fluxo, e uma malha interna de controle de correntes, que demonstram um bom desempenho de torque e velocidade. Por se tratar de um controlador IFOC, onde os parâmetros do motor são fundamentais para obter bons resultados, o ensaio do motor de indução é realizado e apresentado. Sabendo-se das principais fontes de perdas do motor, o controlador de eficiência energética pode ser elaborado. O controlador baseado no modelo de perdas (LMC -Loss-Model-Based Control) é descrito, assim como o controlador de busca recursiva (SC - Search Control), ambos muito utilizados na literatura. O LMC proposto é implementado através de curvas padrão de fluxo ótimo baseado na potência absorvida, dependendo da carga no eixo do motor, uma nova curva padrão é utilizada. O controlador de eficiência proposto chamado de (MBC - Model Based Control), utiliza o método LMC com um controlador IFOC, este é simulado via software Matlab e verificado experimentalmente através de uma plataforma de acionamento projetada e montada, conforme o projeto descrito. O controlador de eficiência proposto, é comparado com o acionamento de fluxo constante e com o controlador SC, para diferentes perfis de carga no eixo. Para todos os casos o controlador proposto, apresentou o melhor rendimento da máquina. / This paper develops a new method of controlling the drive energy efficiency of induction motors . This technique uses a model of losses and a control structure based on minimum energy consumption . At first , studies with the main drivers of energy efficiency in the last decade are presented . Then the electrical and mechanical models of three-phase induction motor are presented , from the classic model , with the reference axis qd0 . The controller used is an indirect field-oriented control (IFOC), composed of an outer control loop speed and flow , and an inner current control loop , showing a good performance of torque and speed. Because it is a IFOC controller, where the engine parameters are critical to obtain good results , the testing of induction motor is carried out and presented . Knowing the main sources of losses of the engine, the controller of energy efficiency can be drawn . The loss model based (LMC) controller is described , as well as the controller recursive search (SC), both widely used in the literature . The LMC proposed is implemented through optimum standard curves of flow based on power absorbed. Depending on the load on the motor shaft , a new standard curve is used . The proposed controller called efficiency ( MBC - Model Based Control ) uses the LMC method with a IFOC controller , this is simulated via textit software emph Matlab and verified experimentally through a platform drive designed and assembled , as the project described . The controller proposed efficiency is compared with the drive and the constant flow controller SC to different profiles of the axle. For all cases, the proposed controller showed the best performance of the machine.