Η παρούσα διπλωματική εργασία ασχολείται με τη μελέτη των μη γραμμικών συστημάτων και τη σχεδίαση ενός προσαρμοστικού ελεγκτή για συστήματα δεύτερης τάξης με αβεβαιότητα. Αρχικά παρουσιάζονται η μορφή και οι ιδιότητες των μη γραμμικών συστημάτων τονίζοντας την ανάγκη για προσαρμοστικό έλεγχο. Στην συνέχεια προτείνεται ένας προσαρμοστικός ελεγκτής κατάλληλος για μη γραμμικά συστήματα δεύτερης τάξης που εγγυάται πλήρη ασυμπτωτική ευστάθεια για το συστημα κλειστού βρόγχου. Η μορφή του ελεγκτη βελτιώνεται σε περίπτωση που οι συναρτήσεις απόσβεσης και δυσκαμψίας είναι πολυωνυμικές και η αποδοτικότητα του επιβεβαιώνεται σε κάθε περίπτωση με εκτενείς προσομοιώσεις. Τέλος, ο προτεινόμενος ελεγκτής εφαρμόζεται σε δύο πραγματικά συστήματα (μηχανικός ταλαντωτής του duffing, απλό εκκρεμές) και συγκρίνεται με την τεχνική του feedback linearization, δίνοντας καλύτερες αποκρίσεις σε περιπτώσεις ύπαρξης αβεβαιότητας παραμέτρων. / This diploma thesis includes the analysis of non linear systems and the design of an aadaptive controller suitable for second order nonlinear systems. In the beginning we present the structure and the properties of nonlineat systems underlining the need for adaptive control. Futhermore, we propose an adaptive control law suitable for second order nonlinear systems which guarantees asympotic stability for the closed loop system. Finally, the proposed controller is applied in two real systems (Duffing's mechanical oscillator,pendulum) and compared with feedback linearization technique, resulting better responses in cases where the system includes uncertain parameters.

Είναι γνωστό ότι η ανάγκη για την ταυτοποίηση συστηµάτων εµφανίζεται σε πολλά προβλήµατα του ευρύτερου κλάδου της Επεξεργασίας Σηµάτων. Υπάρχουν πολλές εφαρµογές, στις οποίες είτε σήµατα προκύπτουν λόγω µη γραµµικών ϕαινοµένων, είτε σήµατα υπόκεινται σε µη γραµµική επεξεργασία. Παραδείγµατα τέτοιων εφαρµογών αποτελούν η µη γραµµική ακύρωση ηχούς, η εξίσωση καναλιών, η µοντελοποίηση ϕωνής, η προσαρµοστική ακύρωση ϑορύβου και η βελτίωση εικόνας. Σε τέτοιες περιπτώσεις, η ταυτοποίηση συστηµάτων χρησιµοποιώντας γραµµικά µοντέλα είναι εντελώς ακατάλληλη για τη µοντελοποίηση των αντιστοίχων ϕαινοµένων/διαδικασιών και τόσο η χρήση µη γραµµικών µοντέλων όσο και η ανάπτυξη σχετικών µεθόδων ταυτοποίησης είναι απαραίτητες. Στα πλαίσια της εν λόγω διπλωµατικής εργασίας, ϑα ασχοληθούµε µε το πρόβληµα της ταυτοποίησης µη γραµµικών συστηµάτων υψηλής τάξης µη γραµµικότητας, τα οποία επιδέχονται ανάπτυξη σε ισοδύναµες Σειρές Volterra. Αρχικά, ϑα ασχοληθούµε µε την ντετερµινιστική ταυτοποίηση µίας ειδικής µορφής διακριτού χρόνου πολυωνυµικών µη γραµµικών συστηµάτων, τα οποία περιγράφονται µέσω του λεγόµενου µοντέλου Volterra - Hammerstein. Πιο συγκεκριµένα, ϑα παρουσιάσουµε και ϑα αναλύσουµε µία σχετικά νέα µέθοδο ταυτοποίησης, η οποία βασίζεται στη διέγερση του υπό µελέτη συστήµατος χρησιµοποιώντας χρονικά µεταβαλλόµενα αρµονικά σήµατα πεπερασµένης διάρκειας, τα επονοµαζόµενα chirps, τα οποία διαθέτουν ορισµένες ενδιαφέρουσες χρονοσυχνοτικές ιδιότητες. Επιπλέον, παρουσιάζουµε καινούργια και ενδιαφέροντα αποτελέσµατα. Στη συνέχεια, εισάγουµε και αναλύουµε διεξοδικά το ϑεµελιώδες πρόβληµα της βέλτιστης στοχαστικής ταυτοποίησης συστηµάτων Volterra διακριτού χρόνου υπό την έννοια του ελάχιστου µέσου τετραγωνικού σφάλµατος, καθώς και το αντίστοιχο πρόβληµα της βέλτιστης ταυτοποίησης συστηµάτων Volterra διακριτού χρόνου υπό την έννοια των ελαχίστων τετραγώνων. Επιπλέον, ϑα επικεντρωθούµε στη διατύπωση των προαναφερθέντων προβληµάτων ελτιστοποίησης πάνω σε ισοδύναµους ορθοκανονικούς χώρους σηµάτων εισόδου. Ακόµα, εξετάζουµε αναλυτικά το πρόβληµα της βέλτιστης αναδροµικής και προσαρµοστικής ταυτοποίησης χρονικά µεταβλητών συστηµάτων Volterra διακριτού χρόνου υπό την έννοια του ελάχιστου µέσου τετραγωνικού σφάλµατος, µέσω του δηµοφιλούς ϕίλτρου Kalman. Χρησιµοποιώντας τη σχέση µίας ειδικής µορφής του ϕίλτρου Kalman µε τον εξίσου δηµοφιλή αλγόριθµο των εκθετικά σταθµισµένων αναδροµικών ελαχίστων τετραγώνων, αναπτύσσουµε προσαρµοστικά ϕίλτρα µε βελτιωµένες ικανότητες εντοπισµού. Τέλος, δοκιµάζουµε και συγκρίνουµε τις παραπάνω µεθόδους µέσω πειραµάτων και χρησιµοποιώντας δεδοµένα, τα οποία προέρχονται από περιβάλλοντα εξοµοίωσης. Επίσης, εξετάζουµε και συγκρίνουµε τη συµπεριφορά των µεθόδων υπό την επίδραση εξωτερικών διαταραχών µέτρησης στα αντίστοιχα σήµατα εξόδου. / -

Made available in DSpace on 2014-06-11T19:27:08Z (GMT). No. of bitstreams: 0 Previous issue date: 2012Bitstream added on 2014-06-13T18:30:56Z : No. of bitstreams: 1 scarabello_mc_me_prud.pdf: 7894728 bytes, checksum: 7780bf65e553d805c887201bc480e587 (MD5) / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) / Em um artigo publicado em maio de 2008 na revista Nature [17], um grupo de pesquisadores da Hewllet-Packard Company (HP) anunciou a fabricação de um componente eletrônico chamado memristor, uma contração para “memory resistor”. A existência teórica dos memristores havia sido prevista em 1971, pelo Engenheiro da Universidade da Califórnia em Berkeley, Leon Chua, com base em propriedades de simetria de certos circuitos elétricos, porém até 2008 sua existência física não havia sido comprovada. Tal componente é considerado o quarto componente eletrônico fundamental, ao lado do resistor, do capacitor e do indutor, pois possui propriedades que não podem ser duplicadas por nenhuma combinação desses três outros componentes. A construção física do memristor atraiu grande interesse no mundo todo, devido ao grande potencial de aplicações deste componente. No presente trabalho fazemos um estudo das bifurcações que ocorrem em um sistema de equações diferenciais ordinárias, que serve como modelo matemático de um circuito elétrico formado pelos quatro elementos fundamentais: um memristor, um capacitor, um indutor e um resistor. O circuito estudado foi proposto por Itoh e Chua em [9]... / In the present work we make a bifurcation analysis of a system of ordinary differential equations, which serves as a mathematical model of an electric circuit formed by the four fundamental elements: one memristor, one capacitor, one inductor and one resistor. The studied circuit was proposed by Itoh and Chua in [9] and was constructed based on the well-known Chua's oscillators. The studied model is given by a discontinuous piecewise-linear system, defined on three zones in R 3, determined by the following inequalities: |z|<1 (called central zone) and |z|>1 (called external zones). The z-axis is composed by equilibrium points of the system. The local normal stability of these equilibira in each zone is analyzed. We show that, due to the existence of this line of equilibria, the phase space R 3 is foliated by invariant planes transversal to the z-axis and parallel to each other, in each zone. The solutions of the system are contained in a combination of three of these invariant planes: one of them in the central zone and the other two in the external zones. We also show that the system may present nonlinear oscillations due to the existence of periodic orbits passing through two of the three zones or passing by three zones. The analysis developed here has analytical and numerical parts. The analytical part was developed based on the study of planar piecewise-linear systems with three zones presented by Freire et al. in [5]... (Complete abstract click electronic access below)

Probabilistic ideas and tools have recently begun to permeate into several fields where they had traditionally not played a major role, including fields such as numerical linear algebra and optimization. One of the key ways in which these ideas influence these fields is via the development and analysis of randomized algorithms for solving standard and new problems of these fields. Such methods are typically easier to analyze, and often lead to faster and/or more scalable and versatile methods in practice. This thesis explores the design and analysis of new randomized iterative methods for solving linear systems and inverting matrices. The methods are based on a novel sketch-and-project framework. By sketching we mean, to start with a difficult problem and then randomly generate a simple problem that contains all the solutions of the original problem. After sketching the problem, we calculate the next iterate by projecting our current iterate onto the solution space of the sketched problem. The starting point for this thesis is the development of an archetype randomized method for solving linear systems. Our method has six different but equivalent interpretations: sketch-and-project, constrain-and-approximate, random intersect, random linear solve, random update and random fixed point. By varying its two parameters – a positive definite matrix (defining geometry), and a random matrix (sampled in an i.i.d. fashion in each iteration) – we recover a comprehensive array of well known algorithms as special cases, including the randomized Kaczmarz method, randomized Newton method, randomized coordinate descent method and random Gaussian pursuit. We also naturally obtain variants of all these methods using blocks and importance sampling. However, our method allows for a much wider selection of these two parameters, which leads to a number of new specific methods. We prove exponential convergence of the expected norm of the error in a single theorem, from which existing complexity results for known variants can be obtained. However, we also give an exact formula for the evolution of the expected iterates, which allows us to give lower bounds on the convergence rate. We then extend our problem to that of finding the projection of given vector onto the solution space of a linear system. For this we develop a new randomized iterative algorithm: stochastic dual ascent (SDA). The method is dual in nature, and iteratively solves the dual of the projection problem. The dual problem is a non-strongly concave quadratic maximization problem without constraints. In each iteration of SDA, a dual variable is updated by a carefully chosen point in a subspace spanned by the columns of a random matrix drawn independently from a fixed distribution. The distribution plays the role of a parameter of the method. Our complexity results hold for a wide family of distributions of random matrices, which opens the possibility to fine-tune the stochasticity of the method to particular applications. We prove that primal iterates associated with the dual process converge to the projection exponentially fast in expectation, and give a formula and an insightful lower bound for the convergence rate. We also prove that the same rate applies to dual function values, primal function values and the duality gap. Unlike traditional iterative methods, SDA converges under virtually no additional assumptions on the system (e.g., rank, diagonal dominance) beyond consistency. In fact, our lower bound improves as the rank of the system matrix drops. By mapping our dual algorithm to a primal process, we uncover that the SDA method is the dual method with respect to the sketch-and-project method from the previous chapter. Thus our new more general convergence results for SDA carry over to the sketch-and-project method and all its specializations (randomized Kaczmarz, randomized coordinate descent ... etc.). When our method specializes to a known algorithm, we either recover the best known rates, or improve upon them. Finally, we show that the framework can be applied to the distributed average consensus problem to obtain an array of new algorithms. The randomized gossip algorithm arises as a special case. In the final chapter, we extend our method for solving linear system to inverting matrices, and develop a family of methods with specialized variants that maintain symmetry or positive definiteness of the iterates. All the methods in the family converge globally and exponentially, with explicit rates. In special cases, we obtain stochastic block variants of several quasi-Newton updates, including bad Broyden (BB), good Broyden (GB), Powell-symmetric-Broyden (PSB), Davidon-Fletcher-Powell (DFP) and Broyden-Fletcher-Goldfarb-Shanno (BFGS). Ours are the first stochastic versions of these updates shown to converge to an inverse of a fixed matrix. Through a dual viewpoint we uncover a fundamental link between quasi-Newton updates and approximate inverse preconditioning. Further, we develop an adaptive variant of the randomized block BFGS (AdaRBFGS), where we modify the distribution underlying the stochasticity of the method throughout the iterative process to achieve faster convergence. By inverting several matrices from varied applications, we demonstrate that AdaRBFGS is highly competitive when compared to the well established Newton-Schulz and approximate preconditioning methods. In particular, on large-scale problems our method outperforms the standard methods by orders of magnitude. The development of efficient methods for estimating the inverse of very large matrices is a much needed tool for preconditioning and variable metric methods in the big data era.

[pt] Este trabalho apresenta um estudo sobre métodos de discretização aplicados à engenharia de Controle. É feita uma revisão dos sistemas lineares de controle discreto, ressaltando suas principais propriedades, assim como suas respostas no tempo e na freqüência. Também são apresentadas as figuras de mérito, que servirão como base de comparação de desempenho dos sistemas, e os tipos de conversores de sinais utilizados. O experimento é então apresentado e são discutidos todos os procedimentos para a sua realização através de um exemplo. Os controladores utilizados são apresentados e suas funções de transferências são fornecidas para sua posterior discretização. Finalmente são listados os resultados obtidos nas simulações utilizando tabelas e gráficos comparativos. / [en] This work presents a study of some discretization methods applied to Control Engineering. A review of discrete control linear systems, their properties and reponses in the and frequency domains are presented, as well as the figures, used to compare their performances, and the signal conversors used in this work. The experiment, the procedures, the controllers and their transfer functions are shown and the simulations results are presented in tables and graphics.

Orientador: Marcelo Messias / Banca: Cristiane Nespoli Morelato França / Banca: Paulo Ricardo da Silva / Resumo: Em um artigo publicado em maio de 2008 na revista Nature [17], um grupo de pesquisadores da Hewllet-Packard Company (HP) anunciou a fabricação de um componente eletrônico chamado memristor, uma contração para "memory resistor". A existência teórica dos memristores havia sido prevista em 1971, pelo Engenheiro da Universidade da Califórnia em Berkeley, Leon Chua, com base em propriedades de simetria de certos circuitos elétricos, porém até 2008 sua existência física não havia sido comprovada. Tal componente é considerado o quarto componente eletrônico fundamental, ao lado do resistor, do capacitor e do indutor, pois possui propriedades que não podem ser duplicadas por nenhuma combinação desses três outros componentes. A construção física do memristor atraiu grande interesse no mundo todo, devido ao grande potencial de aplicações deste componente. No presente trabalho fazemos um estudo das bifurcações que ocorrem em um sistema de equações diferenciais ordinárias, que serve como modelo matemático de um circuito elétrico formado pelos quatro elementos fundamentais: um memristor, um capacitor, um indutor e um resistor. O circuito estudado foi proposto por Itoh e Chua em [9]... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: In the present work we make a bifurcation analysis of a system of ordinary differential equations, which serves as a mathematical model of an electric circuit formed by the four fundamental elements: one memristor, one capacitor, one inductor and one resistor. The studied circuit was proposed by Itoh and Chua in [9] and was constructed based on the well-known Chua's oscillators. The studied model is given by a discontinuous piecewise-linear system, defined on three zones in R^3, determined by the following inequalities: |z|<1 (called central zone) and |z|>1 (called external zones). The z-axis is composed by equilibrium points of the system. The local normal stability of these equilibira in each zone is analyzed. We show that, due to the existence of this line of equilibria, the phase space R^3 is foliated by invariant planes transversal to the z-axis and parallel to each other, in each zone. The solutions of the system are contained in a combination of three of these invariant planes: one of them in the central zone and the other two in the external zones. We also show that the system may present nonlinear oscillations due to the existence of periodic orbits passing through two of the three zones or passing by three zones. The analysis developed here has analytical and numerical parts. The analytical part was developed based on the study of planar piecewise-linear systems with three zones presented by Freire et al. in [5]... (Complete abstract click electronic access below) / Mestre

CoordenaÃÃo de AperfeiÃoamento de Pessoal de NÃvel Superior / Sistemas de equaÃÃes lineares à um tema abordado no ensino mÃdio de forma limitada principalmente por conta do excesso de cÃlculos a serem feitos para encontrar seu conjunto soluÃÃo. A presente dissertaÃÃo tem como objetivo criar e disponibilizar uma ferramenta aos professores que tÃm interesse em ensinar a encontrar a soluÃÃo de sistemas lineares diversos aos seus alunos da educaÃÃo bÃsica utilizando a lÃgica computacional jà que este trabalho contÃm ferramentas para a introduÃÃo do estudo de lÃgica computacional aliada a uma linguagem de programaÃÃo simples e acessÃvel aos estudantes tanto para o ensino mÃdio regular como para a educaÃÃo tecnolÃgica. AlÃm disso, apresentamos uma interessante aplicaÃÃo em um dos problemas tradicionais da quÃmica, o balanceamento de reaÃÃes. Nesta dissertaÃÃo, tambÃm faremos uma abordagem de algumas definiÃÃes necessÃrias envolvendo matrizes e suas operaÃÃes. AlÃm disso, veremos como utilizar uma linguagem de programaÃÃo simples que à usada como ferramenta de aprendizagem inicial na maioria dos cursos bÃsicos de linguagem de programaÃÃo, o PortuguÃs Estruturado. Pontuaremos alguns elementos da sintaxe do VisualG 2.5, compilador gratuito e autoexecutÃvel que irà nos ajudar a atingir nosso objetivo. Veremos como encontrar a soluÃÃo de um sistema de equaÃÃes lineares com um nÃmero finito de variÃveis usando o mÃtodo do escalonamento de sistemas, a substituiÃÃo retroativa e seguindo um algoritmo computacional para chegar a tal soluÃÃo / Systems of linear equations are a topic covered in high school so limited mainly due to the excess calculations to be made to find a whole solution. This dissertation aims to create and provide a tool for teachers who are interested in teaching find the solution of several linear systems to their basic education students using computational logic as this work contains tools for introducing the study of computational logic ally a simple programming language, accessible to students both as a regular high school for technology education. Furthermore, we present an interesting application in one of the traditional problems of chemistry, balancin reactions. In this thesis we will also address some necessary definitions and operations involving matrices. Moreover, we will see how to use a simple programming language that is used as a tool for initial learning in most courses basic programming language, the Structured Portuguese. We will point some elements of syntax VisuAlg 2.5, free compiler and auto executable that will help us achieve our goal. We will see how to find the solution of a system of linear equations with a finite number of variables using the method of scheduling systems, the backward substitution and following an algorithm for finding such a solution.

Orientador: Ricardo Coração de Leão Fontoura de Oliveira / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Elétrica e de Computação / Made available in DSpace on 2018-08-26T15:54:55Z (GMT). No. of bitstreams: 1 Vieira_HenriquedeSouza_M.pdf: 1027517 bytes, checksum: 9dcae7f42b15c7b6a4659c57bea547f7 (MD5) Previous issue date: 2015 / Resumo: Os problemas de estabilização e controle H? robustos por realimentação de estados para sistemas lineares incertos em domínios politópicos são investigados nesta dissertação. São propostas novas técnicas de síntese baseadas em condições LMIs com busca em escalares, abordando de maneira unificada sistemas contínuos e discretos no tempo. A principal novidade da técnica proposta é o uso efetivo de matrizes de Lyapunov polinomiais de grau arbitrário para certificar a estabilidade robusta do sistema em malha fechada. A segunda vantagem da abordagem proposta é que as melhores condições de síntese por realimentação de estados atualmente disponíveis na literatura podem ser reproduzidas por meio de escolhas particulares dos parâmetros escalares. Para o problema de controle H? também é proposto um procedimento iterativo como alternativa à busca dos escalares. Experimentos numéricos ilustram o potencial e a eficácia da técnica proposta / Abstract: The problems of robust stabilization and H? control by state-feedback for uncertain linear systems in polytopic domains are investigated in this dissertation. New synthesis techniques based on LMI conditions with scalar searches, addressing in a unified way continuous and discrete-time systems, are proposed. The main novelty of the proposed method is the effective use of polynomial Lyapunov matrices of arbitrary degree to certify the robust stability of the closed-loop system. The second advantage of the proposed approach is that the best currently available state-feedback synthesis conditions in the literature can be reproduced by particular choices of the scalars. Regarding the H? control problem, an iterative procedure is also proposed as an alternative to the scalar searches. Numerical experiments illustrate the potential and efficacy of the proposed methods / Mestrado / Automação / Mestre em Engenharia Elétrica

Orientador: Rafael Santos Mendes / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Eletrica / Made available in DSpace on 2018-07-18T13:49:42Z (GMT). No. of bitstreams: 1 Paiva_ElyCarneirode_M.pdf: 4160343 bytes, checksum: 3e6c7b7d3a3f81cf6c8ca7027018dfe1 (MD5) Previous issue date: 1993 / Resumo: Sistemas lineares discretos, invariantes no tempo, sujeitos a incertezas de parâmetros do processo, são considerados neste trabalho. Para um controlador dado, a maior região de incerteza hiperretangular no espaço de parâmetros do processo é calculada, tal que os pólos do sistema em malha fechada estejam contidos em uma região conexa desejada no círculo unitário. Isto é equivalente a determinar os intervalos máximos para os parâmetros incertos do processo, de modo que a estabilidade relativa do sistema seja assegurada. Uma medida de robustez é definida a partir desta região de incerteza, para um controlador dado. Além do problema da robustez, considera-se também neste trabalho, a presença de perturbações estocásticas, sendo um dos objetivos do controle, a minimização da variância, denotada por ¿J IND. 2¿ dos sinais de saída. Um procedimento de projeto é proposto, para a obtenção do controlador que minimiza a maior variância ¿J IND. 2¿ (dentre todos os parâmetros do processo considerados), ao mesmo tempo em que assegura a robustez diante das incertezas nominais / Abstract: Linear time-invariant discrete-time systems subject to uncertainties of plant parameters are considered in this work. For a given controller the greatest hyperrectangular region of uncertainty in the plant parameter space is calculated, so that the cIosed-loop system poles stay confined to a desired connected region in the unit circle. This is equivalent to determining the maximal interval bounds on the uncertainties of the plant parameters such that the relative stability of the system be invariant. A robustness measure of a given controller is defined from the structure adopted to the uncertainties. Also the presence of stochastic perturbations acting in the plant are considered and the output/control signal variances are the performance index (called ¿J IND. 2¿) to be minimized. Finally, a design procedure, based on gradient directions, that Iteratively modifies the controller parameters such that the ¿J IND. 2¿ performance index is decreased under the restriction of robust D-stability, is presented. / Mestrado / Mestre em Engenharia Elétrica

The numerical solutions of many types of problems are generally obtained by solving approximating linear algebraic systems. Moreover, in solving a nonlinear problem, one may replace it by a sequence of linear systems providing progressively improved approximations. For the study of these linear systems of equations a geometric terminology with the compact symbolism of vectors and matrices is useful. A resume of the basic principles of higher algebra necessary for the development of the material to follow is therefore included. / "A Paper." / "Submitted to the Graduate Council of Florida State University in partial fulfillment of the requirements for the degree of Master of Science." / Advisor: Paul J. McCarthy, Professor Directing Paper. / "May, 1958." / At head of title: Florida State University. / Typescript. / Includes bibliographical references.