Peter M. Thompson, Gunter Stein, Michael Athans. / "July, 1983." Caption title. / Bibliography: p. 82. / NASA Grant NGL-22-009-124

CAPES / O objetivo desta dissertação é propor uma técnica para a síntese via realimentação de estados para sistemas lineares e invariantes no tempo, considerando que os estados realimentados são previamente quantizados. Para tanto, o erro de quantização é inicialmente modelado como um ruído externo. Assim, o problema de obter os ganhos de realimentação de estados, se torna um problema de projetar os ganhos que minimizem a norma H¥ do sistema controlado. Os ganhos de realimentação são calculados pela solução de um conjunto de condições descritas na forma de desigualdades matriciais lineares. A técnica é ilustrada pela aplicação da realimentação de estados quantizada em um servomecanismo. / The main objective of this dissertation is to propose a technique for synthesis by statefeedback for linear time-invariant systems, considering that the states are quantized before the feedback. To tackle such problem, the quantization error is initially modeled as an external noise. Therefore, the problem of getting the state-feedback gains, becomes a problem of designing the gains that minimize the H¥ norm of the system. The state-feedback gains are calculated by solving a set of conditions described in the form of linear matrix inequalities. The technique is illustrated by applying the of feedback of quantized states to a servo-mechanism.

CAPES / O objetivo desta dissertação é propor uma técnica para a síntese via realimentação de estados para sistemas lineares e invariantes no tempo, considerando que os estados realimentados são previamente quantizados. Para tanto, o erro de quantização é inicialmente modelado como um ruído externo. Assim, o problema de obter os ganhos de realimentação de estados, se torna um problema de projetar os ganhos que minimizem a norma H¥ do sistema controlado. Os ganhos de realimentação são calculados pela solução de um conjunto de condições descritas na forma de desigualdades matriciais lineares. A técnica é ilustrada pela aplicação da realimentação de estados quantizada em um servomecanismo. / The main objective of this dissertation is to propose a technique for synthesis by statefeedback for linear time-invariant systems, considering that the states are quantized before the feedback. To tackle such problem, the quantization error is initially modeled as an external noise. Therefore, the problem of getting the state-feedback gains, becomes a problem of designing the gains that minimize the H¥ norm of the system. The state-feedback gains are calculated by solving a set of conditions described in the form of linear matrix inequalities. The technique is illustrated by applying the of feedback of quantized states to a servo-mechanism.

Orientador: Jose Claudio Geromel / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Eletrica e de Computação / Made available in DSpace on 2018-08-06T08:12:23Z (GMT). No. of bitstreams: 1 Korogui_RubensHideo_M.pdf: 409144 bytes, checksum: 7291201982a90c16c32c29016cf26e33 (MD5) Previous issue date: 2006 / Resumo: Esta dissertação é dedicada ao estudo de métodos de análise e síntese de controladores para sistemas poli tópicos, tendo como ferramentas principais as desigualdades matriciais lineares (LMls) e o emprego de variáveis adicionais, chamadas multiplicadores ou matrizes de escalamento. Desenvolvemos os conceitos relacionados à análise de sistemas dinâmicos empregando multiplicadores, com os quais obtemos condições suficientes de estabilidade e desempenho robustos. Associado a essa perspectiva, introduzimos funções de Lyapunov dependentes de parâmetros para tratar de sistemas politópicos. De posse dos resultados de análise, propomos condições para o projeto de controladores por realimentação de estado, através de uma parametrização convexa. O texto é ilustrado com exemplos e tabelas / Abstract: This thesis is devoted to analysis and controllers design for polytopic systems in the Linear Matrix Inequalities (LMls) and Multipliers framework. We develop concepts related to the analysis of dynamic systems stability in terms of multipliers, so we are able to provide sufficient conditions for robust stability and performance. Moreover, we introduce parameter dependent Lyapunov functions in order to deal with polytopic systems. Based on the analysis results, we propose conditions for state feedback design using a convex parametrization. Throughout the text several examples and tables illustrate the development / 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 2018-08-13T17:01:09Z (GMT). No. of bitstreams: 1 CanahuireCabello_RuthVanessa_M.pdf: 2405747 bytes, checksum: 148678565951eb1c7507a6c9c63048fd (MD5) Previous issue date: 2009 / Resumo: Neste trabalho estudou-se o problema de controle H8 de vibrações levando em conta os efeitos da saturação dos atuadores. Duas situações de interesse foram exploradas. A primeira é aquela em que a saturação deve ser evitada para não causar danos no sistema, mesmo que nesta situação ocorra uma perda de desempenho. A segunda alternativa procuramodificar o controlador no sentido de obter um melhor desempenho, mesmo que ocorra a saturação. O projeto dos controladores H8 foi feito usando as formulações baseadas em desigualdades matriciais lineares (Linear Matrix Inequalities - LMI). Entre as técnicas que evitam a saturação, foram empregadas as funções de ponderação, escolha do peso do esforço de controle como uma das saídas de desempenho do problema de otimização, o escalonamento da matriz de saída do controlador e a inclusão de uma restrição adicional na forma de LMI para limitar o esforço de controle. Para as técnicas que procuram melhorar o desempenho mesmo na presença da saturação, o controlador foi modificado através da metodologia conhecida como Anti-windup. As técnicas estudadas foram avaliadas em dois problemas de controle de vibrações: um sistema massa-mola-amortecedor de dois graus de liberdade e em uma viga flexível. Estas técnicas foram implementadas usando o aplicativo MATLAB e ferramentas específicas para a solução de problemas de otimização com restrições na forma de LMI tais como o Yalmip, SeDuMi e SDPT3. Os resultados para estes dois exemplos são discutidos no trabalho. / Abstract: In this work, the H8 control of vibrations problem taking into account the effects of saturation of actuators is studied. Two situations of interest were explored. The first one considers that saturation should be avoided and does not cause damage to the system, even if this situation leads to loss of performance. The second alternative seeks to modify the controller to achieve better performance, even when the saturation occurs. The design of H8 controllers was done using the formulation based on linear matrix inequalities (LMI). The techniques that avoid saturation are the weighting functions selection, selecting of the weight of the control signal as a performance output of the optimization problem, the scaling of the output matrix of the controller and the inclusion of an additional restriction in the form of LMI to limit the control effort. A technique that seeks to improve performance, even in the presence of saturation, is the controller modification using the methodology known as Anti-windup. These techniques were evaluated in two problems of control of vibrations: a mass-springdamper system of two degrees of freedom and a flexible beam. These techniques were implemented using MATLAB and with the application of specific tools to solve problems of optimization with restrictions in the form of LMI, such as Yalmip, SeDuMi and SDPT3. The results for these two examples are discussed in the work. / Mestrado / Mecanica dos Sólidos e Projeto Mecanico / Mestre em Engenharia Mecânica

Orientador: Alberto Luiz Serpa / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Mecanica / Made available in DSpace on 2018-08-04T03:06:58Z (GMT). No. of bitstreams: 1 King_DiegoMelo_M.pdf: 7156086 bytes, checksum: e7d5891f788e5e71c720d6a82897b54d (MD5) Previous issue date: 2005 / Resumo: Este trabalho tem como objetivo verificar a ação de uma formulação obtida para o controle misto H2/ Hoo via realimentação de saída de uma estrutura modelada por elementos finÍtos. Para isso, também se verificaram as respectivas formulações via realimentação de saída dos controles H2 e Hoo isoladamente, para que os resultados obtidos em cada um deles pudessem validar o equacionamento descrito neste trabalho para o controle misto e servissem de parâmetro para balizar os seus resultados esperados. Os equacionamentos dos controladores seguiram a formulação da programação semi-definida, isto é, minimizam um objetivo linear sujeito a restrições em forma de LMI. Estas LMI representam as condições de estabilidade a que o sistema está sujeito, e as características da interação entre o modelo e a ação do controlador. Uma vez definidas todas estas condições, o problema de minimização, que é convexo, foi implementado para ser resolvido através do Matlab. Os resultados obtidos pela formulação para o controle misto também foram comparados com resultados produzidos pelas formulações clássicas do próprio Matlab, para que este último também servisse como um aferidor dos resultados da formulação apresentada neste trabalho. O modelo da estrutura, definida para uma viga engastada em balanço, também foi calculado através do Matlab, discretizada nos elementos finÍtos apropriados. As determinações das matrizes de massa, de rigidez elástica e de amortecimento, esta determinada pelo modelo de amortecimento proporcional, fazem a ligação com as matrizes de estado que representam o sistema na formulação utilizada para os controladores. Os resultados do controle misto mostram que a formulação convexa do controle via realimentação de saída é viável. Os valores apresentados pelas normas H2 e Hoominimizadas do sistema, e as respostas deste a uma entrada aleatória exógena, exibem a capacidade do controlador misto. Mesmo para um controle não-colocado, as vibrações da estrutura são atenuadas junto com a minimização do sinal de controle, e os picos da resposta em freqüência, vistos nas funções de resposta em freqüência, são reduzidos / Abstract: This work has the goal to verify the action of a mixed H2/ Hoo control formulation through output feedback for a structure modelIed in finite elements. Therefore, the pure H2 and Hoo control formulations through output feedback are also verified to validate the described equations in this work for the mixed control and show a reference to their expected results. These formulations are based on the semi-definite programming formulation, and consist of minimizing a linear objective subjected to LMI constraints. These LMI represent the system stability conditions and interaction characteristics between the model and the controller. Once all these conditions are settled, the minimization problem, which is convex, was solved through the Matlab software. The results obtained for the mixed control formulation were compared to the classical results obtained with the Matlab. The purpose is to validate the results ofthe formulation obtained in this work. The structure model, defined for a cantilever beam, was also obtained through the Matlab using the finite elements technique. The mass, stiffness and damping matrices, this last one obtained by the proportional damping model, make the connections to the state-space matrices representing the system for the controller formulation. The results obtained for the mixed control show the convex formulation to output feedback may offer good results. The minimized values of H2 and H00 norms and the behaviour of the controlIed systeIl! for the external random input show the capability of mixed control. Even for a noncollocated control problem the structure vibrations are reduced together with the control signal minimization, and the frequency response peaks, visible in the frequency response graphs, are also reduced. / Mestrado / Mecanica dos Sólidos e Projeto Mecanico / Mestre em Engenharia Mecânica

In this research, a security architecture based on the feedback control theory has been proposed. The first loop has been designed, developed and tested. The architecture proposes a feedback model with many controllers located at different stages of network. The controller at each stage gives feedback to the one at higher level and a decision about network security is taken. The first loop implemented in this thesis detects one important anomaly of virus attack, rate of outgoing connection. Though there are other anomalies of a virus attack, rate of outgoing connection is an important one to contain the spread. Based on the feedback model, this symptom is fed back and a state model using queuing theory is developed to delay the connections and slow down the rate of outgoing connections. Upon implementation of this model, whenever an infected machine tries to make connections at a speed not considered safe, the controller kicks in and sends those connections to a delay queue. Because of delaying connections, rate of outgoing connections decrease. Also because of delaying, many connections timeout and get dropped, reducing the spread. PID controller is implemented to decide the number of connections going to safe or suspected queue. Multiple controllers can be implemented to control the parameters like delay and timeout. Control theory analysis is performed on the system to test for stability, controllability, observability. Sensitivity analysis is done to find out the sensitivity of the controller to the delay parameter. The first loop implemented gives feedback to the architecture proposed about symptoms of an attack at the node level. A controller needs to be developed to receive information from different controllers and decision about quarantining needs to be made. This research gives the basic information needed for the controller about what is going on at individual nodes of the network. This information can also be used to increase sensitivity of other loops to increase the effectiveness of feedback architecture.

Many practical problems in the field of engineering become so complex that they may be effectively solved only with the aid of a computer. An effective solution depends on the use of an efficient algorithm. Plotting root locus diagrams is such a problem. This thesis presents such an algorithm. Root locus design of feedback control systems is a very powerful tool. Stability of systems under the influence of variables can be easily determined from the root locus diagram. For even moderately complex systems of the type found in practical applications, determination of the locus is extremely difficult if accuracy is required. The difficulty lies in the classical method of graphically determining the location of points on the locus by trial and error. Such a method cannot be efficiently applied to a computer program. The text presents an original algorithm for plotting the root locus of a general system. The algorithm is derived using the combined methods of complex variable algebra and numerical analysis. For each abscissa desired a polynomial is generated. The real roots of this polynomial are the ordinate values for points on the root locus. Root finding methods from numerical analysis enable the solution of the problem to be one of convergent iteration rather than trial and error. Among the material presented is a computer program for solution of the general problem, an example of a completely analytic solution, and a table of solutions for more simple systems. The program inputs are the coefficients of the open loop transfer function and the range and increments of the real axis which are to be swept. The output lists the real and imaginary components of all solution points at each increment of the sweep. Also listed are the magnitude and angle components of the solution point and the value of system gain for which this is a solution. For less complex problems, the method can be applied analytically. This may result in an explicit relation between the real and imaginary components of all solution points or even in a single expression which can be analyzed using the methods of analytic geometry. As with any advance in the theory of problem solving, the ideas presented in the thesis are best applied in conjunction with previous solution methods. Specifically, an idea of the approximate location of the root locus can be obtained using sketching rules which are well known. The method presented here becomes much more efficient when even a rough approximation is known. Furthermore, the specific locations of system poles and zeros are not required, but can be helpful in planning areas in which to search for solutions.

No description available.

Electronic surveillance of employee performance is a common practice in the workplace today. Despite the fact that electronic monitoring is an inexpensive and objective technique for recording productivity figures, much resistance to the technique exists. The current study was conducted to determine if user acceptability and productivity could be improved in performance monitoring systems. It was hypothesized that frequent performance feedback would enhance productivity and task satisfaction as long as the feedback was not too frequent and the performance standard was not too difficult. The results of this experimental investigation did not support this hypothesis indicating that productivity (in terms of the amount of data entered into computer files) was actually higher when the performance standard was difficult and the performance feedback frequent. In addition, subjects assigned to the Difficult Goal/Frequent Feedback condition were more satisfied with the task and performance feedback than other subjects. Results are discussed in light of current viewpoints in applied psychology and human-computer interaction. / Ph. D.