In this thesis, two separate problems are examines. First, sinusoidal signals are quite prevalent in practical applications. For example, any machine driven by a rotary shaft will exhibit periodic behaviour. For this reason, the estimation of sinusoidal parameters is studied extensively in the literature. Often in practical applications, there are unmodeled disturbances to the system, and the incoming measurements are noisy. Thus, estimation of the parameters of a sinusoidal signal in real-time for these conditions is of interest, calling for the use of a filter-based approach such as the Extended Kalman Filter. Considering the sinusoidal signal in its complex form, a novel approach is proposed resulting in a complex-valued filter. The resulting complex Extended Kalman Filter’s performance is evaluated in various test environments and is compared to standard approaches to the estimation problem using a Discrete Fourier Transform and standard Extended Kalman Filter. Results show that the complex Extended Kalman Filter outperforms the standard approaches in some cases in both accuracy and convergence rate. Second, research on hybrid systems has seen a large growth in interest in recent years. This is largely due to the increase of natural systems where discrete mode dynamics interact with continuous state dynamics. Switched systems are a subclass of hybrid systems that restrict their definition to continuous dynamic systems that interact with dis- crete switching events. Controller synthesis for such systems is no trivial task. Given the current trend in Artificial Intelligence and Machine Learning approaches, Dynamic Programming is explored as a means to approximate optimal control policies for switched systems. Discussions of discretization of the system’s state space are presented, followed by a high-level overview of an algorithm that leverages Dynamic Programming to find the approximated optimal control policies. Finally, the algorithm is applied to several examples to demonstrate its effectiveness. / Thesis / Master of Applied Science (MASc)

<p>Periodic signal modeling plays an important role in different fields. The unifying theme of this thesis is using nonlinear techniques to model periodic signals. The suggested techniques utilize the user pre-knowledge about the signal waveform. This gives these techniques an advantage as compared to others that do not consider such priors.</p><p>The technique of Part I relies on the fact that a sine wave that is passed through a static nonlinear function produces a harmonic spectrum of overtones. Consequently, the estimated signal model can be parameterized as a known periodic function (with unknown frequency) in cascade with an unknown static nonlinearity. The unknown frequency and the parameters of the static nonlinearity are estimated simultaneously using the recursive prediction error method (RPEM). A treatment of the local convergence properties of the RPEM is provided. Also, an adaptive grid point algorithm is introduced to estimate the unknown frequency and the parameters of the static nonlinearity in a number of adaptively estimated grid points. This gives the RPEM more freedom to select the grid points and hence reduces modeling errors.</p><p>Limit cycle oscillations problem are encountered in many applications. Therefore, mathematical modeling of limit cycles becomes an essential topic that helps to better understand and/or to avoid limit cycle oscillations in different fields. In Part II, a second-order nonlinear ODE is used to model the periodic signal as a limit cycle oscillation. The right hand side of the ODE model is parameterized using a polynomial function in the states, and then discretized to allow for the implementation of different identification algorithms. Hence, it is possible to obtain highly accurate models by only estimating a few parameters.</p><p>In Part III, different user aspects for the two nonlinear approaches of the thesis are discussed. Finally, topics for future research are presented. </p>

Periodic signal modeling plays an important role in different fields. The unifying theme of this thesis is using nonlinear techniques to model periodic signals. The suggested techniques utilize the user pre-knowledge about the signal waveform. This gives these techniques an advantage as compared to others that do not consider such priors. The technique of Part I relies on the fact that a sine wave that is passed through a static nonlinear function produces a harmonic spectrum of overtones. Consequently, the estimated signal model can be parameterized as a known periodic function (with unknown frequency) in cascade with an unknown static nonlinearity. The unknown frequency and the parameters of the static nonlinearity are estimated simultaneously using the recursive prediction error method (RPEM). A treatment of the local convergence properties of the RPEM is provided. Also, an adaptive grid point algorithm is introduced to estimate the unknown frequency and the parameters of the static nonlinearity in a number of adaptively estimated grid points. This gives the RPEM more freedom to select the grid points and hence reduces modeling errors. Limit cycle oscillations problem are encountered in many applications. Therefore, mathematical modeling of limit cycles becomes an essential topic that helps to better understand and/or to avoid limit cycle oscillations in different fields. In Part II, a second-order nonlinear ODE is used to model the periodic signal as a limit cycle oscillation. The right hand side of the ODE model is parameterized using a polynomial function in the states, and then discretized to allow for the implementation of different identification algorithms. Hence, it is possible to obtain highly accurate models by only estimating a few parameters. In Part III, different user aspects for the two nonlinear approaches of the thesis are discussed. Finally, topics for future research are presented.

Este trabalho aborda o problema de seguimento e rejeição de sinais periódicos em sistemas lineares sujeitos a saturação nos atuadores. Para garantir o seguimento/rejeição, dois controladores baseados no princípio do modelo interno são considerados: o primeiro baseia-se no modelo interno em sua formulação clássica, isto é, um controlador dinâmico contendo um número finito de modos (marginalmente) instáveis da referência/perturbação é introduzido na malha de controle, em uma abordagem chamada de controladores ressonantes; a segunda abordagem considera o controlador repetitivo, onde um elemento de atraso é inserido namalha de controle emumlaço de realimentação positiva, fazendo o papel do modelo interno de ordem infinita. Nos dois casos, o objetivo principal é a obtenção de condições na forma de inequações matriciais lineares (do inglês, Linear Matrix Inequalities - LMIs) para a síntese simultânea de uma realimentação de estados estabilizante e do ganho do laço estático de anti-windup. Partindo do pressuposto que as referências e perturbações pertencem a um certo conjunto admissível, estes ganhos garantem que as trajetórias do sistema em malha fechada inciadas em um certo conjunto elipsoidal convergem para outro conjunto elipsoidal invariante contido na região de operação linear do sistema. Nesta região, a presença do modelo interno na malha de controle garante o seguimento e a rejeição dos sinais de interesse. Nas duas abordagens são propostos problemas de otimização visando a maximização do conjunto invariante de estados admissíveis e/ou a maximização do conjunto de referências/perturbações admissíveis. Extensões da metodologia para sistemas de tempo discreto também são apresentadas. / This work addresses the tracking/rejection problem of periodic signals for linear systems subject to control saturation. To ensure the tracking/rejection, two internal model based controllers are considered: the first one considers the internal model in a classical framework, i.e. a dynamic controller containing a finite number of (marginally) unstable modes of the reference/disturbance signal is introduced in the control loop. In this work, this approach is called resonant controller. The second approach considers the repetitive controller, where a delay element is introduced in the control loop in a positive feedback loop, playing the role of an infinite order internal model. In both cases, the main objective is to obtain conditions in the form of LMIs to simultaneously compute a stabilizing state feedback gain and an anti-windup gain. Assuming that the references and disturbances signals belong to a certain admissible set, these gains guarantee that the trajectories of the closed-loop system starting in a certain ellipsoidal set contract to another invariant ellipsoidal set inside the linearity region of the closed-loop system. In this region the presence of the internal model ensures tracking/rejection of the considered periodic signals. In both frameworks, optimization problems aiming at the maximization of the invariant set of admissible states and/or the maximization of the set of admissible references/disturbances are proposed. Extensions of the proposed framework to discrete-time systems are also presented.

Este trabalho aborda o problema de seguimento e rejeição de sinais periódicos em sistemas lineares sujeitos a saturação nos atuadores. Para garantir o seguimento/rejeição, dois controladores baseados no princípio do modelo interno são considerados: o primeiro baseia-se no modelo interno em sua formulação clássica, isto é, um controlador dinâmico contendo um número finito de modos (marginalmente) instáveis da referência/perturbação é introduzido na malha de controle, em uma abordagem chamada de controladores ressonantes; a segunda abordagem considera o controlador repetitivo, onde um elemento de atraso é inserido namalha de controle emumlaço de realimentação positiva, fazendo o papel do modelo interno de ordem infinita. Nos dois casos, o objetivo principal é a obtenção de condições na forma de inequações matriciais lineares (do inglês, Linear Matrix Inequalities - LMIs) para a síntese simultânea de uma realimentação de estados estabilizante e do ganho do laço estático de anti-windup. Partindo do pressuposto que as referências e perturbações pertencem a um certo conjunto admissível, estes ganhos garantem que as trajetórias do sistema em malha fechada inciadas em um certo conjunto elipsoidal convergem para outro conjunto elipsoidal invariante contido na região de operação linear do sistema. Nesta região, a presença do modelo interno na malha de controle garante o seguimento e a rejeição dos sinais de interesse. Nas duas abordagens são propostos problemas de otimização visando a maximização do conjunto invariante de estados admissíveis e/ou a maximização do conjunto de referências/perturbações admissíveis. Extensões da metodologia para sistemas de tempo discreto também são apresentadas. / This work addresses the tracking/rejection problem of periodic signals for linear systems subject to control saturation. To ensure the tracking/rejection, two internal model based controllers are considered: the first one considers the internal model in a classical framework, i.e. a dynamic controller containing a finite number of (marginally) unstable modes of the reference/disturbance signal is introduced in the control loop. In this work, this approach is called resonant controller. The second approach considers the repetitive controller, where a delay element is introduced in the control loop in a positive feedback loop, playing the role of an infinite order internal model. In both cases, the main objective is to obtain conditions in the form of LMIs to simultaneously compute a stabilizing state feedback gain and an anti-windup gain. Assuming that the references and disturbances signals belong to a certain admissible set, these gains guarantee that the trajectories of the closed-loop system starting in a certain ellipsoidal set contract to another invariant ellipsoidal set inside the linearity region of the closed-loop system. In this region the presence of the internal model ensures tracking/rejection of the considered periodic signals. In both frameworks, optimization problems aiming at the maximization of the invariant set of admissible states and/or the maximization of the set of admissible references/disturbances are proposed. Extensions of the proposed framework to discrete-time systems are also presented.

Este trabalho aborda o problema de seguimento e rejeição de sinais periódicos em sistemas lineares sujeitos a saturação nos atuadores. Para garantir o seguimento/rejeição, dois controladores baseados no princípio do modelo interno são considerados: o primeiro baseia-se no modelo interno em sua formulação clássica, isto é, um controlador dinâmico contendo um número finito de modos (marginalmente) instáveis da referência/perturbação é introduzido na malha de controle, em uma abordagem chamada de controladores ressonantes; a segunda abordagem considera o controlador repetitivo, onde um elemento de atraso é inserido namalha de controle emumlaço de realimentação positiva, fazendo o papel do modelo interno de ordem infinita. Nos dois casos, o objetivo principal é a obtenção de condições na forma de inequações matriciais lineares (do inglês, Linear Matrix Inequalities - LMIs) para a síntese simultânea de uma realimentação de estados estabilizante e do ganho do laço estático de anti-windup. Partindo do pressuposto que as referências e perturbações pertencem a um certo conjunto admissível, estes ganhos garantem que as trajetórias do sistema em malha fechada inciadas em um certo conjunto elipsoidal convergem para outro conjunto elipsoidal invariante contido na região de operação linear do sistema. Nesta região, a presença do modelo interno na malha de controle garante o seguimento e a rejeição dos sinais de interesse. Nas duas abordagens são propostos problemas de otimização visando a maximização do conjunto invariante de estados admissíveis e/ou a maximização do conjunto de referências/perturbações admissíveis. Extensões da metodologia para sistemas de tempo discreto também são apresentadas. / This work addresses the tracking/rejection problem of periodic signals for linear systems subject to control saturation. To ensure the tracking/rejection, two internal model based controllers are considered: the first one considers the internal model in a classical framework, i.e. a dynamic controller containing a finite number of (marginally) unstable modes of the reference/disturbance signal is introduced in the control loop. In this work, this approach is called resonant controller. The second approach considers the repetitive controller, where a delay element is introduced in the control loop in a positive feedback loop, playing the role of an infinite order internal model. In both cases, the main objective is to obtain conditions in the form of LMIs to simultaneously compute a stabilizing state feedback gain and an anti-windup gain. Assuming that the references and disturbances signals belong to a certain admissible set, these gains guarantee that the trajectories of the closed-loop system starting in a certain ellipsoidal set contract to another invariant ellipsoidal set inside the linearity region of the closed-loop system. In this region the presence of the internal model ensures tracking/rejection of the considered periodic signals. In both frameworks, optimization problems aiming at the maximization of the invariant set of admissible states and/or the maximization of the set of admissible references/disturbances are proposed. Extensions of the proposed framework to discrete-time systems are also presented.