By understanding and mimicking characteristics of postural control used by animals, scientist and engineers may develop standing autonomous robots that work safely within home environments, and treatment strategies that help people overcome postural impairments. To increase our understanding of postural control we developed physical and computational models of standing posture to explain the interrelation of stance width and feedback gain in controlling the stability and dynamics of the postural response. These models facilitated precise analysis of mechanical dynamics and their effects on compliant feedback control, and provided a physical implementation to verify predictions developed from simulation. We show that a scaling of active feedback gain is required to maintain postural stability. These results are consistent with previous studies that have shown that a correlation exists between increased stance width and decreased postural responses. However, these studies have not quantified the relation between stance and the active control of standing posture. This scaling of gains that we show is dependent on the changing kinematic relations of the mechanical structure as it undergoes stance width adjustments. Specifically, we show that increasing stance width increases the leverage of the mechanical system. Feedback gains must be reduced by the reciprocal of the increase in mechanical leverage in order to maintain a consistent postural response; otherwise, the system may become unstable with increasing oscillations. We also showed that increasing magnitudes of intrinsic stiffness increases postural stability by facilitating stable responses over larger ranges of active feedback gain and increasing the stability of responses by decreasing settling time, oscillations, and displacement magnitude. The conclusions of this study were that the variation of mechanical leverage is responsible for changing the dynamics of the response during stance width variation, and that scaling of feedback gains with the changing mechanical leverage of stance width variations is required to maintain consistent response dynamics across stance widths.

Measuring acoustic pressure gradients is critical in many applications such as directional microphones for hearing aids and sound intensity probes. This measurement is especially challenging with decreasing microphone size, which reduces the sensitivity due to small spacing between the pressure ports. Novel, micromachined biomimetic microphone diaphragms are shown to provide high sensitivity to pressure gradients on one side of the diaphragm with low thermal mechanical noise. These structures have a dominant mode shape with see-saw like motion in the audio band, responding to pressure gradients as well as spurious higher order modes sensitive to pressure. In this dissertation, integration of a diffraction based optical detection method with these novel diaphragm structures to implement a low noise optical pressure gradient microphone is described and experimental characterization results are presented, showing 36 dBA noise level with 1mm port spacing, nearly an order of magnitude better than the current gradient microphones. The optical detection scheme also provides electrostatic actuation capability from both sides of the diaphragm separately which can be used for active force feedback. A 4-port electromechanical equivalent circuit model of this microphone with optical readout is developed to predict the overall response of the device to different acoustic and electrostatic excitations. The model includes the damping due to complex motion of air around the microphone diaphragm, and it calculates the detected optical signal on each side of the diaphragm as a combination of two separate dominant vibration modes. This equivalent circuit model is verified by experiments and used to predict the microphone response with different force feedback schemes. Single sided force feedback is used for active damping to improve the linearity and the frequency response of the microphone. Furthermore, it is shown that using two sided force feedback one can significantly suppress or enhance the desired vibration modes of the diaphragm. This approach provides an electronic means to tailor the directional response of the microphones, with significant implications in device performance for various applications. As an example, the use of this device as a particle velocity sensor for sound intensity and sound power measurements is investigated. Without force feedback, the gradient microphone provides accurate particle velocity measurement for frequencies below 2 kHz, after which the pressure response of the second order mode becomes significant. With two-sided force feedback, the calculations show that this upper frequency limit may be increased to 10 kHz. This improves the pressure residual intensity index by more than 15 dB in the 50 Hz-10 kHz range, matching the Class I requirements of IEC 1043 standards for intensity probes without any need for multiple spacers.

In this dissertation, we present new approaches to improve standard designs in adaptive control theory, and novel adaptive control architectures. We first present a novel Kalman filter based approach for approximately enforcing a linear constraint in standard adaptive control design. One application is that this leads to alternative forms for well known modification terms such as e-modification. In addition, it leads to smaller tracking errors without incurring significant oscillations in the system response and without requiring high modification gain. We derive alternative forms of e- and adaptive loop recovery (ALR-) modifications. Next, we show how to use Kalman filter optimization to derive a novel adaptation law. This results in an optimization-based time-varying adaptation gain that reduces the need for adaptation gain tuning. A second major contribution of this dissertation is the development of a novel derivative-free, delayed weight update law for adaptive control. The assumption of constant unknown ideal weights is relaxed to the existence of time-varying weights, such that fast and possibly discontinuous variation in weights are allowed. This approach is particularly advantageous for applications to systems that can undergo a sudden change in dynamics, such as might be due to reconfiguration, deployment of a payload, docking, or structural damage, and for rejection of external disturbance processes. As a third and final contribution, we develop a novel approach for extending all the methods developed in this dissertation to the case of output feedback. The approach is developed only for the case of derivative-free adaptive control, and the extension of the other approaches developed previously for the state feedback case to output feedback is left as a future research topic. The proposed approaches of this dissertation are illustrated in both simulation and flight test.

Bibliography: p. 196-200. / M.I.T. Project DSR 76094. Contract no. NOW-66-0178-d.

O processo de organização e recuperação de imagens apresenta inúmeros problemas a serem abordados, compreender o significado subjetivo de uma consulta visual por meio de parâmetros numéricos que podem ser extraídos e comparados por meio de um computador é um dos maiores desafios. A disparidade entre a comparação realizada pela máquina e o que de fato o ser humano interpreta de uma consulta visual é denominado de lacuna semântica. Diversos trabalhos na literatura abordam técnicas para diminuir essa disparidade. Nesse contexto, a abordagem de realimentação de relevância apresenta-se como um meio eficaz para capturar a intenção do usuário e reduzir a diferença entre os conceitos semânticos oriundos da percepção visual do usuário, bem como das características visuais de baixo nível extraídas automaticamente de uma imagem. Entretanto, a lacuna semântica ainda é um desafio a ser vencido. Por tal motivo, o presente trabalho tem por objetivo o estudo, análise e proposta de um arcabouço para recuperação de imagens o qual combina a realimentação de relevância unida a algoritmos evolutivos visando aproximar a expectativa do usuário em relação aos resultados retornados pela recuperação de imagens, por meio da captura de intenção do mesmo e posterior definição dos parâmetros mais adequados. Para tanto, a ideia principal do trabalho é inserir memória de aprendizado ao processo de recuperação de imagens por conteúdo, armazenando os dados da interação do usuário com o sistema em perfis, os quais posteriormente serão utilizados para prover respostas ao usuário de maneira personalizada, consequentemente, contribuindo para diminuição da lacuna semântica. / The process of organization and retrieval of images presents numerous problems to be addressed, to understand the subjective meaning of a visual query through numerical parameters that can be extracted and compared by a computer is a challenge. The disparity between the comparison performed by the machine and that in fact the human being interprets a visual query is known as semantic gap. Several studies in the literature techniques to address this gap. In this context, relevance feedback (RF) is an effective approach to capture user intent and reduce the difference between semantic concepts and the visual characteristics of an image. However, the semantic gap is still a challenge to be overcome, therefore, the present work aims to study, analysis and proposal for a new framework for image retrieval through relevance feedback combined with evolutionary algorithms aiming to bring the user’s expectation througth the results returned by the retrieval of images through the his intention and definition of the most appropriate parameters. To this end, the main idea of the work is to insert learning memory in a content-based image retrieval system for this by storing the user interaction data with the system profiles so that later these data are able to provide parameters for the system “learn” and respond to personalized user and need. Thus, contributing to reduction of existing semantic gap between the results of a retrieval operation and indeed expected images for a specific context, improving the efficiency and effectiveness of the retrieval process.

O processo de organização e recuperação de imagens apresenta inúmeros problemas a serem abordados, compreender o significado subjetivo de uma consulta visual por meio de parâmetros numéricos que podem ser extraídos e comparados por meio de um computador é um dos maiores desafios. A disparidade entre a comparação realizada pela máquina e o que de fato o ser humano interpreta de uma consulta visual é denominado de lacuna semântica. Diversos trabalhos na literatura abordam técnicas para diminuir essa disparidade. Nesse contexto, a abordagem de realimentação de relevância apresenta-se como um meio eficaz para capturar a intenção do usuário e reduzir a diferença entre os conceitos semânticos oriundos da percepção visual do usuário, bem como das características visuais de baixo nível extraídas automaticamente de uma imagem. Entretanto, a lacuna semântica ainda é um desafio a ser vencido. Por tal motivo, o presente trabalho tem por objetivo o estudo, análise e proposta de um arcabouço para recuperação de imagens o qual combina a realimentação de relevância unida a algoritmos evolutivos visando aproximar a expectativa do usuário em relação aos resultados retornados pela recuperação de imagens, por meio da captura de intenção do mesmo e posterior definição dos parâmetros mais adequados. Para tanto, a ideia principal do trabalho é inserir memória de aprendizado ao processo de recuperação de imagens por conteúdo, armazenando os dados da interação do usuário com o sistema em perfis, os quais posteriormente serão utilizados para prover respostas ao usuário de maneira personalizada, consequentemente, contribuindo para diminuição da lacuna semântica. / The process of organization and retrieval of images presents numerous problems to be addressed, to understand the subjective meaning of a visual query through numerical parameters that can be extracted and compared by a computer is a challenge. The disparity between the comparison performed by the machine and that in fact the human being interprets a visual query is known as semantic gap. Several studies in the literature techniques to address this gap. In this context, relevance feedback (RF) is an effective approach to capture user intent and reduce the difference between semantic concepts and the visual characteristics of an image. However, the semantic gap is still a challenge to be overcome, therefore, the present work aims to study, analysis and proposal for a new framework for image retrieval through relevance feedback combined with evolutionary algorithms aiming to bring the user’s expectation througth the results returned by the retrieval of images through the his intention and definition of the most appropriate parameters. To this end, the main idea of the work is to insert learning memory in a content-based image retrieval system for this by storing the user interaction data with the system profiles so that later these data are able to provide parameters for the system “learn” and respond to personalized user and need. Thus, contributing to reduction of existing semantic gap between the results of a retrieval operation and indeed expected images for a specific context, improving the efficiency and effectiveness of the retrieval process.

Orientador: José Cláudio Geromel / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Elétrica e de Computação / Made available in DSpace on 2018-08-16T14:10:01Z (GMT). No. of bitstreams: 1 Deaecto_GraceSilva_D.pdf: 2021318 bytes, checksum: 451e6647afe6d7beac67cf3717dff7da (MD5) Previous issue date: 2010 / Resumo: Esta tese trata do projeto de controladores dinâmicos H2 ou H? via realimentação de saída para sistemas lineares com comutação contínuos e discretos no tempo. Inicialmente, apresentamos a síntese via realimentação de estado de forma a colocar em evidência as principais dificuldades para a obtenção da solução do problema de projeto mais realista, correspondente ao controle via realimentação dinâmica de saída. As condições obtidas são necessárias e suficientes para a existência de solução com uma estrutura particular para as chamadas desigualdades de Lyapunov ou Riccati-Metzler. Tratamos também da síntese via realimentação de estado de sistemas sujeitos a incertezas limitadas em norma e politópicas. No que diz respeito às incertezas limitadas em norma, as condições são válidas inclusive se as mesmas apresentarem estrutura do tipo linear fracionária (LFT). No caso de incertezas politópicas, as condições podem ser empregadas como uma alternativa para o projeto de controle de sistemas com parâmetros lineares variantes no tempo (LPV). Aplicamos os resultados teóricos para a solução de três problemas de cunho essencialmente práticos, a saber, a regulação da tensão de saída de conversores de potência CCCC, a estimação da aceleração vertical de uma suspensão automotiva e o controle do ângulo de rolamento de uma aeronave / Abstract: This thesis is concerned to the output feedback H2 or H? control design for continuous-time and discrete-time switched linear systems. Initially, we present the state feedback synthesis in order to put in evidence the main difficulties we have to face towards the solution of the more realistic output feedback control design problem. The conditions are necessary and sufficient regarding the existence of a structured solution of the Lypunov or Riccati-Metzler inequalities. We also treat the state feedback synthesis for switched linear systems subject to norm bounded or polytopic uncertainties. In the former case, the conditions hold even if the uncertainty presents the linear fractional structure (LFT), and in the latter one, the conditions can be used as an alternative to the linear parameter varying control (LPV). The theoretical results are applied to three practical problems, namely, the output voltage regulation of DC-DC converters, the vertical stroke estimation in automotive suspensions and aircraft roll angle control / Doutorado / Automação / Doutor em Engenharia Elétrica

Orientador: Jose Claudio Geromel / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Eletrica e de Computação / Made available in DSpace on 2018-08-13T09:58:44Z (GMT). No. of bitstreams: 1 Korogui_RubensHideo_D.pdf: 1047054 bytes, checksum: 1a066f84d7ea84703434bdb049a62a20 (MD5) Previous issue date: 2009 / Resumo: Uma das temáticas abordadas neste trabalho é o problema de filtragem robusta em norma H2 para sistemas lineares a tempo contínuo e a tempo discreto. Consideramos que os sistemas estão sujeitos a incertezas paramétricas inicialmente do tipo politópicas e, em seguida, do tipo que admite representação linear fracionária. Calculamos limitantes inferior e superior para a norma H2 do erro de estimação, resolvendo-se problemas minimax, escrevendo-os em formato de desigualdades matriciais lineares LMI. Dessa maneira calculamos um intervalo de otimalidade que permite certificar o desempenho do filtro robusto projetado. Em seguida, aplicamos a metodologia proposta em problemas de estimação envolvendo um motor de indução trifásico e uma linha de transmissão com derivação. O presente trabalho também apresenta o que chamamos de sistema linear de comparação, cuja proposta é servir como alternativa para estudar sistemas com atraso. Utilizando a substituição de Rekasius construímos um sistema linear invariante no tempo que permite obter informações sobre a estabilidade e o cálculo de norma H8 para sistemas desta classe. Em vista da formulação através de matrizes de representação de estado é possível estender sem maiores dificuldades os resultados para o projeto de controladores via realimentação de estado e de filtros para sistemas com atraso. / Abstract: One of the themes considered in this work is the robust H2 filtering design problem for linear time invariant continuous and discrete time systems. We assume that the systems are subject to parametric uncertainty, initially of the polytopic type and later as linear fractional transformation parametric uncertainties. We calculate lower and upper bounds to the H2 squared norm of the estimation error by means of the equilibrium solution of a minimax problem, that can be formulated in a linear matrix inequality framework. Under this approach we provide an optimality gap that allows us to certify the performance of the designed robust filter. Afterwards, we apply the proposed methodology to estimation problems involving a three-phase induction motor and a transmission line with a stub. This work also considers what we call a linear comparison system, whose goal is to serve as an alternative to study time-delay systems. Using the Rekasius substitution we construct a linear time invariant system that allows us to get information about stability and H8 norm of this class of systems. Based on this approach in terms of state space matrices it is possible to extend the results to state feedback design and filtering design without any further assumptions. / Doutorado / Automação / Mestre em Engenharia Elétrica

Orientador: Jose Claudio Geromel / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Eletrica e de Computação / Made available in DSpace on 2018-08-09T08:47:02Z (GMT). No. of bitstreams: 1 Deaecto_GraceSilva_M.pdf: 5350126 bytes, checksum: 88f05b9392a9876bfe07fae6a8f95202 (MD5) Previous issue date: 2007 / Resumo: Neste trabalho, são propostas condições suficientes para a estabilidade assintótica de sistemas lineares, contínuos no tempo, com comutação que asseguram um custo garantido de desempenho. Estas condições dependem da solução de um conjunto de desigualdades de Lyapunov-Metzler, definidas em [Geromel & Colaneri, 2006], que possuem natureza não-convexa, sendo portanto, de difícil solução. Para contornar este problema, apresentamos condições de estabilidade mais conservadoras baseadas em uma subclasse de matrizes de Metzler com elementos iguais na diagonal principal, que podem ser resolvidas através de desigualdades matriciais lineares e busca unidimensional. Os resultados apresentados em [Geromel & Colaneri, 2006], que fornecem condições para a estabilidade de sistemas dinâmicos lineares com comutação, são generalizados para lidar com sistemas mais gerais, a saber, sistemas dinâmicos sujeitos a perturbações impulsivas. A proposta elaborada assegura a estabilidade assintótica inclusive na presença de possíveis modos deslizantes / Abstract: In this work, sufficient conditions are proposed for the study of asymptotic stability of continuous time linear systems with commutation, which assure a guaranteed cost of performance. These conditions depend on the solution of a set of Lyapunov-Metzler inequalities, defined in [Geromel & Colaneri, 2006], which are difficult to solve due to their non-convex nature. However, to circumvent this difficulty, we present more conservative stability conditions based on a subclass of Metzler matrices characterized by having equal elements on the main diagonal. Although these conditions are more conservative, they can be solved by Linear Matrix Inequalities (LMls) and unidimensional search. The analysis done in [Geromel & Colaneri, 2006], which provides the stability conditions for linear dynamic systems with commutation is expanded to cover a more general class of systems with commutation and subject to impulsive disturbances. The conditions presented in this work assure the stability even in the presence of possible sliding modes / Mestrado / Automação / Mestre em Engenharia Elétrica

Orientador: Alberto Luiz Serpa / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Mecanica / Made available in DSpace on 2018-08-11T04:04:36Z (GMT). No. of bitstreams: 1 Mazoni_AlyssonFernandes_M.pdf: 2883879 bytes, checksum: a514d74c83605535d0c2eaf4a060356f (MD5) Previous issue date: 2008 / Resumo: Esta dissertação aborda técnicas modernas de controle robusto H8 para sistemas dinâmicos lineares. Com isso pretende-se dizer que são usados como ferramentas matemáticas os resultados da teoria de controle de sistemas lineares para o caso com incertezas de vários tipos admitidas sobre o modelo. Os modelos são primariamente estruturas flexíveis e os métodos de projeto são implementados usando exclusivamente a solução de problemas sujeitos a desigualdades matriciais lineares. São abordadas as incertezas paramétrica, dinâmica e politópica com o objetivo de apresentar métodos matemáticos de projeto de controladores para os sistemas incertos. Para o caso de incerteza dinâmica, apresenta-se a técnica de filtros de ponderação. Em contraposição a essa abordagem, os resultados recentes da literatura sobre o lema generalizado de Kalman-Yakubovi?c-Popov e o H8 restrito na freqüência também são usados como métodos de controle independentes de filtros de ponderação. Os métodos são comparados usando modelos de simulação e experimentos no âmbito de estruturas flexíveis. Palavras-chave: Teoria dos sistemas dinâmicos, Programação Convexa, Sistema de controle por realimentação / Abstract: This dissertation deals with modern techniques from the Robust H8 Control of Linear Dynamic Systems. By this it is meant that the results from linear control systems theory are used as mathemathical tools when considering several kinds of uncertainty on the models. These models are mostly of flexible structures and the design methods are implemented using solely the solution of problems subjected to linear matrix inequalities. The types of uncertainty approached are: parametric, dynamic and polytopic; this is done aiming to present mathematical design methods for the uncertain systems considered. When dealing with dynamic uncertainty, the weighting functions are introduced. In contrast with this approach, recent results from literature on the generalised Kalman-Yakubovi?c-Popov lemma and frequency restricted H8 are also used as control design methods whose application is independent of weighting functions. All methods are compared using simple simulation models and experiments with flexible structures . Keywords: Theory of dynamical systems, Convex programming, Feedback control systems / Mestrado / Mecanica dos Sólidos e Projeto Mecanico / Mestre em Engenharia Mecânica