Orientador: Pedro Luis Dias Peres / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Eletrica e de Computação / Made available in DSpace on 2018-08-05T03:23:54Z (GMT). No. of bitstreams: 1 Leite_ValterJuniordeSouza_D.pdf: 4310851 bytes, checksum: bcd0414d19eb46e02496290857fdf9bc (MD5) Previous issue date: 2005 / Resumo: Este trabalho trata da aplica¸c¿ao de funcionais de Lyapunov e Lyapunov-Krasovskii dependentes de parâmetro a alguns problemas selecionados da área de controle robusto, a saber: D-estabilidade robusta de polítipo de matrizes, D-estabilidade robusta de politopos de polinômios matriciais, estabilidade robusta de sistemas neutrais com atrasos variantes no tempo e controle robusto H8 de sistemas discretos no tempo com atraso nos estados. É utilizada a representação politópica para as incertezas dos sistemas estudados. São obtidas formulações convexas, na forma de desigualdades matriciais lineares, suficientes para a solução dos problemas selecionados. Essas condições podem ser resolvidas numericamente de maneira eficiente por meio de algoritmos especializados baseados em pontos interiores. Os resultados obtidos são menos conservadores que os encontrados na literatura, baseados em geral na estabilidade quadrática, isto é, as matrizes dos funcionais são fixas e independentes da incerteza / Abstract: This work deals with the application of parameter dependent Lyapunov and Lyapunov-Krasovskii functionals to some selected problems of robust control: robust D-stability of polytopes of matrices, robust D-stability of polytopes of polynomial matrices, robust stability of uncertain neutral systems with timevarying delays and robust H8 control of uncertain discrete time delay systems. The polytopic representation is used to describe the uncertainties. Convex formulations are obtained, in terms of inear matrix inequalities, that are sufficient for the solution of the selected problems. Those conditions can be solved in a efficient way through specialized interior point algorithms. The obtained results are less conservative than those from the literature, in general based on quadratic stability, i.e., the matrices in the functionals are fixed and do not depend on the uncertainty / Doutorado / Automação / Doutor em Engenharia Elétrica

Orientador: Pedro Luis Dias Peres / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Elétrica e de Computação / Made available in DSpace on 2018-08-22T18:08:10Z (GMT). No. of bitstreams: 1 Agulhari_CristianoMarcos_D.pdf: 2096468 bytes, checksum: 0e762392c4ee0ab5e249ccd096ab4acf (MD5) Previous issue date: 2013 / Resumo: As principais contribuições desta tese consistem no desenvolvimento de métodos para a síntese de controladores e para a análise de estabilidade de sistemas lineares, variantes ou invariantes no tempo. Com relação aos sistemas invariantes no tempo, o objetivo é a síntese de controladores robustos de ordem reduzida para sistemas a tempo contínuo com parâmetros incertos. O método apresentado para a síntese baseia-se em uma técnica de dois estágios, em que um ganho de realimentação de estados é construído no primeiro estágio e posteriormente utilizado no segundo estágio, que fornece o controlador robusto desejado. Cada etapa consiste na resolução de condições sob a forma de desigualdades matriciais lineares. No caso de sistemas variantes no tempo, em geral, dependendo das informações disponíveis, dois modelos matemáticos podem ser utilizados. Por um lado, para sistemas cujos elementos variantes no tempo são limitados em norma, mas não são completamente conhecidos, é possível utilizar modelos dependentes de parâmetros variantes no tempo, que levam a uma representação politópica. Nesse caso, a técnica de estabilização proposta é baseada no método de dois estágios, para gerar controladores dependentes dos parâmetros. Supõe-se que os parâmetros sejam mensuráveis em tempo real, e os controladores são sintetizados de forma a serem robustos a ruídos nas medições. Por outro lado, se a dinâmica variante no tempo é conhecida, o sistema pode ser tratado diretamente sem que seja utilizado nenhum tipo de parametrização. Duas técnicas de síntese são propostas para esse caso: a construção de ganhos estabilizantes utilizando diretamente a matriz de transição de estados, e uma técnica de síntese projetada a partir de um novo critério para a verificação da estabilidade do sistema. A validade dos métodos propostos é ilustrada por meio de exemplos numéricos, que mostram a qualidade dos resultados que podem ser obtidos / Abstract: The main contributions of this thesis concern the development of methods for the stability analysis and the synthesis of controllers for linear systems, either time varying or time-invariant. Concerning time-invariant systems, the objective is the synthesis of reduced-order robust controllers for continuous-time systems with uncertain parameters. The method presented for the synthesis is based on a two-stage technique, in which a stabilizing state-feedback gain is constructed in the first stage and then applied on the second stage to search for the desired controller. Each stage consists in the resolution of conditions based on linear matrix inequalities. In the case of time-varying systems, depending on the amount of available information, two mathematical models may be used. On one hand, if the time-varying elements of the system are not entirely known, one can model the system as function of time-varying parameters, resulting on a polytopic representation. In this case, the stabilization method proposed is based on the two-stage technique, which yields parameter-dependent controllers. The parameters are supposed to be real-time measurable, and the controllers are robust with respect to noises and uncertainties on the measures. On the other hand, if the time-varying dynamics are known, the system may be directly handled without using any parameterization. Two synthesis techniques are proposed in this case: the construction of stabilizing gains by using the state transition matrix, and a synthesis technique derived from a new stability criterion for time-varying systems. The validity of the proposed methods is illustrated through numerical examples that show the efficiency of the results that can be obtained / Doutorado / Automação / Doutor em Engenharia Elétrica

Les systèmes physiques sont représentés par des modèles mathématiques qui peuvent être utilisés pour simuler, analyser ou contrôler ces systèmes. Selon la complexité du système qu’il est censé représenter, un modèle peut être plus ou moins complexe. Une complexité trop grande peut s’avérer problématique en pratique du fait des limitations de puissance de calcul et de mémoire des ordinateurs. L’une des façons de contourner ce problème consiste à utiliser l’approximation de modèles qui vise à remplacer le modèle complexe par un modèle simplifié dont le comportement est toujours représentatif de celui du système physique.Dans le cas des modèles dynamiques Linéaires et Invariants dans le Temps (LTI), la complexité se traduit par une dimension importante du vecteur d’état et on parle alors de modèles de grande dimension. L’approximation de modèle, encore appelée réduction de modèle dans ce cas, a pour but de trouver un modèle dont le vecteur d’état est plus petit que celui du modèle de grande dimension tel que les comportements entrée-sortie des deux modèles soient proches selon une certaine norme. La norme H2 a été largement considérée dans la littérature pour mesurer la qualité d’un modèle réduit. Cependant, la bande passante limitée des capteurs et des actionneurs ainsi que le fait qu’un modèle est généralement représentatif d’un système physique dans une certaine bande fréquentielle seulement, laissent penser qu’un modèle réduit dont le comportement est fidèle au modèle de grande dimension dans un intervalle de fréquences donné,peut être plus pertinent. C’est pourquoi, dans cette étude, la norme H2 limitée en fréquence,ou norme H2, Ω, qui est simplement la restriction de la norme H2 sur un intervalle de fréquences Ω, a été considérée. En particulier, le problème qui vise à trouver un modèle réduit minimisant la norme H2, Ω de l’erreur d’approximation avec le modèle de grande dimension a été traité.Deux approches ont été proposées dans cette optique. La première est une approche empirique basée sur la modification d’une méthode sous-optimale pour l’approximation H2. En pratique, ses performances s’avèrent intéressantes et rivalisent avec certaines méthodes connues pour l’approximation de modèles sur intervalles de fréquences limités.La seconde est une méthode d’optimisation basée sur la formulation pôles-résidus de la norme H2, Ω. Cette formulation généralise naturellement celle existante pour la norme H2 et permet également d’exprimer deux bornes supérieures sur la norme H∞ d’un modèle LTI, ce qui est particulièrement intéressant dans le cadre de la réduction de modèle. Les conditions d’optimalité du premier ordre pour le problème d’approximation optimale en norme H2, Ω ont été exprimées et utilisées pour créer un algorithme de descente visant à trouver un minimum local au problème d’approximation. Couplée aux bornes sur la norme H∞ de l’erreur d’approximation,cette méthode est utilisée pour le contrôle de modèle de grande dimension.D’un point de vue plus pratique, l’ensemble des méthodes proposées dans cette étude ont été appliquées, avec succès, dans un cadre industriel comme élément d’un processus global visant à contrôler un avion civil flexible. / Physical systems are represented by mathematical models in order to be simulated, analysed or controlled. Depending on the complexity of the physical system it is meant to represent and on the way it has been built, a model can be more or less complex. This complexity can become an issue in practice due to the limited computational power and memory of computers. One way to alleviate this issue consists in using model approximation which is aimed at finding a simpler model that still represents faithfully the physical system.In the case of Linear Time Invariant (LTI) dynamical models, complexity translates into a large dimension of the state vector and one talks about large-scale models. Model approximation is in this case also called model reduction and consists in finding a model with a smaller state vector such that the input-to-output behaviours of both models are close with respect to some measure. The H2-norm has been extensively used in the literature to evaluate the quality of a reduced-order model. Yet, due to the limited band width of actuators, sensors and the fact that models are generally representative on a bounded frequency interval only, a reduced-order model that faithfully reproduces the behaviour of the large-scale one over a bounded frequency interval only, may be morerelevant. That is why, in this study, the frequency-limited H2-norm, or H2,Ω-norm, which is the restriction of theH2-norm over a frequency interval, has been considered. In particular, the problem of finding a reduced-ordermodel that minimises the H2, Ω-norm of the approximation error with the large-scale model has been addressed here. For that purpose, two approaches have been developed. The first one is an empirical approach based on the modification of a sub-optimal H2 model approximation method. Its performances are interesting in practice and compete with some well-known frequency-limited approximation methods. The second one is an optimisationmethod relying on the poles-residues formulation of the H2,Ω-norm. This formulation naturally extends the oneexisting for the H2-norm and can also be used to derive two upper bounds on the H∞-norm of LTI dynamical models which is of particular interest in model reduction. The first-order optimality conditions of the optimal H2,Ω approximation problem are derived and used to built a complex-domain descent algorithm aimed at finding a local minimum of the problem. Together with the H∞ bounds on the approximation error, this approach isused to perform control of large-scale models. From a practical point of view, the methods proposed in this study have been successfully applied in an industrial context as a part of the global process aimed at controlling a flexible civilian aircraft.

In dieser Arbeit werden lineare zeitinvariante endlichdimensionale Systeme (LTI-Systeme) mit m > 1 Eingängen und p > 1 Ausgängen untersucht (MIMO-Systeme). Diese lassen sich darstellen durch lineare Gleichungen mit Matrizen, deren Einträge Polynome im Ableitungsoperator d/dt sind. Bei Nutzung der Laplace-Transformation handelt es sich um Polynome in s. Algebraisch bilden diese einen Euklidischen Ring. Durch Überführung der Matrizen in die Hermitesche Normalform werden m Basisgrößen definiert. Die Verläufe oder Trajektorien der Basisgrößen lassen sich frei vorgegeben. Damit werden die Trajektorien sämtlicher übrigen Signale, insbesondere die der erforderlichen Eingangssignale, festgelegt und können ohne Integration berechnet werden. Ein linksteilerfremdes (auch steuerbar genanntes) Modell ist dabei nicht zwingend erforderlich. Damit eignen sich die Basisgrößen besonders zur Planung von Trajektorien. Genauer untersucht wird die Planung mit Polynomen in der Zeit als Ansatzfunktionen und die Planung von Trajektorien, die ein quadratisches Kostenfunktional minimieren. In der technischen Praxis werden die Systeme stets von den geplanten Trajektorien abweichen. Insbesondere bei instabilen Regelstrecken ist deshalb ein stabilisierender Folgeregler unentbehrlich. Die Struktur der Folgeregelung wird eingeführt und es wird deutlich gemacht, dass jede Methode zum Entwurf linearer Regler angewendet werden kann. Die Nullstellenzuweisung durch dynamische Ausgangsrückführung mit Reglern vorgegebener möglichst geringer dynamischer Ordnung wird detailliert untersucht und eine neue Lösungsmöglichkeit aufgezeigt. Durch Nutzung der modifizierten z-Transformation lässt sich die Theorie auf ein hybrides System, bestehend aus einer zeitkontinuierlichen Regelstrecke und einer zeitdiskreten digitalen Steuerung und Regelung, ausdehnen. Dabei werden die Verläufe der Signale zwischen den Abtastzeitpunkten in die Planung einbezogen. Zum Schluss werden die linearen Beobachter im Licht der polynomialen Matrizendarstellung neu untersucht. Es wird gezeigt, dass die polynomiale Matrizendarstellung einen theoretischen Rahmen bietet, in dem sich sämtliche linearen Beobachter mit einer Methode entwerfen lassen. - (Die Dissertation ist veröffentlicht in der Reihe Fortschritt-Berichte VDI, Reihe 8 - Mess-, Steuerungs- und Regelungstechnik, Band 1164 im VDI Verlag GmbH, Düsseldorf, ISBN 978-3-18-516408-8) / In this thesis linear time invariant lumped systems (LTI-systems) with m>1 inputs and p > 1 outputs (MIMO-systems) are investigated. These systems can be represented by linear equations with matrices, whose entries are polynomials in the differential operator d/dt. If Laplace-transform is employed, the polynomials are in s. Algebraically polynomials form a Euclidean ring. The conversion of the matrices to the Hermite form leads to defining m basic variables. The trajectories of the basis variables may be chosen arbitrarily. With that choice the trajectories of all remaining variables and especially the input variables are determined and can be calculated without integration. A left coprime (also called controllable) model is not required. Hence basis variables are particularly useful for planning trajectories. Special attention is paid to planning trajectories with polynomials in time as basic functions and planning trajectories which minimise a quadratic functional of costs. In engineering practice the systems will always differ from the planed trajectories. Especially with unstable plants a stabilising tracking controller is compulsory. The structure of the tracking control is introduced. It becomes apparent that every linear theory for the design of closed loop controllers is suitable. Pole assignment by dynamic output feedback with low order controllers of a fixed structure is looked at in more detail. A new approach to this problem is presented. Using the modified z-transform the theory is extended to hybrid systems consisting of a digital or discrete time controller and a plant in continuous time. Thereby the course of the signals between the sampling moments is taken into account. Finally linear observers are reinvestigated using the polynomial matrix representation. It is shown that the polynomial matrix representation provides a theoretical framework in which all linear observers can be designed.

Dynamical structure functions were developed as a partial structure representation of linear time-invariant systems to be used in the reconstruction of biological networks. Dynamical structure functions contain more information about structure than a system's transfer function, while requiring less a priori information for reconstruction than the complete computational structure associated with the state space realization. Early sufficient conditions for network reconstruction with dynamical structure functions severely restricted the possible applications of the reconstruction process to networks where each input independently controls a measured state. The first contribution of this thesis is to extend the previously established sufficient conditions to incorporate both necessary and sufficient conditions for reconstruction. These new conditions allow for the reconstruction of a larger number of networks, even networks where independent control of measured states is not possible. The second contribution of this thesis is to extend the robust reconstruction algorithm to all reconstructible networks. This extension is important because it allows for the reconstruction of networks from real data, where noise is present in the measurements of the system. The third contribution of this thesis is a Matlab toolbox that implements the robust reconstruction algorithm discussed above. The Matlab toolbox takes in input-output data from simulations or real-life perturbation experiments and returns the proposed Boolean structure of the network. The final contribution of this thesis is to increase the applicability of dynamical structure functions to more than just biological networks by applying our reconstruction method to wireless communication networks. The reconstruction of wireless networks produces a dynamic interference map that can be used to improve network performance or interpret changes of link rates in terms of changes in network structure, enabling novel anomaly detection and security schemes.

No description available.

Approved for public release, distribution is unlimited / In this thesis, the ocean was modeled as a waveguide with an ideal pressure - release surface, and an ideal rigid bottom. The ocean waveguide was then treated as a linear, time - invariant, space - variant (TISV) filter or communication channel. The filter is time - invariant because no motion was modeled and because the properties of the ocean were assumed to be constant. The filter is space - variant because of the presence of the two boundaries, that is, the ocean surface and ocean bottom. This thesis investigates the ocean as a linear TISV filter by evaluating 1) the complex frequency response, 2) the impulse response, and 3) the transfer function of the ocean with respect to depth. It is shown that the TISV impulse response of the ocean contains information that can be used to help localize a target in range and whether the target is above or below the receiver. Computer simulation results were obtained by evaluating the three filter functions for several different test cases. / Ensign, United States Navy

Fundação Carlos Chagas Filho de Amparo a Pesquisa do Estado do Rio de Janeiro / Nesta dissertação é apresentado um estudo sobre a recuperação de sinais modelados por somas ponderadas de exponenciais complexas. Para tal, são introduzidos conceitos elementares em teoria de sinais e sistemas, em particular, os sistemas lineares invariantes no tempo, SLITs, que podem ser representados matematicamente por equações diferenciais, ou equações de diferenças, para sinais analógicos ou digitais, respectivamente. Equações deste tipo apresentam como solução somas ponderadas de exponenciais complexas, e assim fica estabelecida a relação entre os sistemas de tipo SLIT e o modelo em estudo. Além disso, são apresentadas duas combinações de métodos utilizadas na recuperação dos parâmetros dos sinais: métodos de Prony e mínimos quadrados, e métodos de Kung e mínimos quadrados, onde os métodos de Prony e Kung recuperam os expoentes das exponenciais e o método dos mínimos quadrados recupera os coeficientes lineares do modelo. Finalmente, são realizadas cinco simulações de recuperação de sinais, sendo a última, uma aplicação na área de modelos de qualidade de água. / A study on the recovery of signals modeled by weighted sums of complex exponentials complex is presented. For this, basic concepts of signals and systems theory are introduced. In particular, the linear time invariant systems (LTI Systems) are considered, which can be mathematically represented by differential equations or difference equations, respectively, for analog or digital signals. The solution of these types of equations is given by a weighted sum of complex exponentials, so the relationship between the LTI Systems and the model of study is established. Furthermore, two combinations of methods are used to recover the parameters of the signals: Prony and least squares methods, and Kung and least squares methods, where Prony and Kung methods are used to recover the exponents of the exponentials and the least square method is used to recover the linear coefficients of the model. Finally, five simulations are performed for the recovery of signals, the last one being an application in the area of water quality models.

Fundação Carlos Chagas Filho de Amparo a Pesquisa do Estado do Rio de Janeiro / Nesta dissertação é apresentado um estudo sobre a recuperação de sinais modelados por somas ponderadas de exponenciais complexas. Para tal, são introduzidos conceitos elementares em teoria de sinais e sistemas, em particular, os sistemas lineares invariantes no tempo, SLITs, que podem ser representados matematicamente por equações diferenciais, ou equações de diferenças, para sinais analógicos ou digitais, respectivamente. Equações deste tipo apresentam como solução somas ponderadas de exponenciais complexas, e assim fica estabelecida a relação entre os sistemas de tipo SLIT e o modelo em estudo. Além disso, são apresentadas duas combinações de métodos utilizadas na recuperação dos parâmetros dos sinais: métodos de Prony e mínimos quadrados, e métodos de Kung e mínimos quadrados, onde os métodos de Prony e Kung recuperam os expoentes das exponenciais e o método dos mínimos quadrados recupera os coeficientes lineares do modelo. Finalmente, são realizadas cinco simulações de recuperação de sinais, sendo a última, uma aplicação na área de modelos de qualidade de água. / A study on the recovery of signals modeled by weighted sums of complex exponentials complex is presented. For this, basic concepts of signals and systems theory are introduced. In particular, the linear time invariant systems (LTI Systems) are considered, which can be mathematically represented by differential equations or difference equations, respectively, for analog or digital signals. The solution of these types of equations is given by a weighted sum of complex exponentials, so the relationship between the LTI Systems and the model of study is established. Furthermore, two combinations of methods are used to recover the parameters of the signals: Prony and least squares methods, and Kung and least squares methods, where Prony and Kung methods are used to recover the exponents of the exponentials and the least square method is used to recover the linear coefficients of the model. Finally, five simulations are performed for the recovery of signals, the last one being an application in the area of water quality models.

Linear time-invariant (LTI) dynamic networks are described by their dynamical structure function, and generally, they have many possible state space realizations. This work characterizes the necessary and sufficient conditions on a state transformation that preserves the dynamical structure function, thereby generating the entire set of realizations of a given order for a specific dynamic network.