Spelling suggestions: "subject:"[een] NONLINEAR CONTROL"" "subject:"[enn] NONLINEAR CONTROL""
151 |
Bestärkendes Lernen zur Steuerung und Regelung nichtlinearer dynamischer SystemePritzkoleit, Max 21 January 2020 (has links)
In der vorliegenden Arbeit wird das bestärkende Lernen im Kontext der Steuerung und Regelung nichtlinearer dynamischer Systeme untersucht. Es werden zunächst die Grundlagen der stochastischen Optimalsteuerung sowie des maschinellen Lernens, die für die Betrachtungen dieser Arbeit relevant sind, erläutert. Anschließend werden die Methoden des bestärkenden Lernens im Kontext der datenbasierten Steuerung und Regelung dargelegt, um anschließend auf drei Methoden des tiefen bestärkenden Lernens näher einzugehen. Der Algorithmus Deep-Deterministic-Policy-Gradient (DDPG) wird zum Gegenstand intensiver Untersuchungen an verschiedenen mechanischen Beispielsystemen.
Weiterhin erfolgt der Vergleich mit einem klassischen Ansatz, bei dem die zu bewältigenden Steuerungsaufgaben mit einer modellbasierten Trajektorienberechnung, die auf dem iterativen linear-quadratischen Regler (iLQR) basiert, gelöst werden. Mit dem iLQR können zwar alle Steuerungsaufgaben erfolgreich bewältigt werden, aber für neue Anfangswerte muss das Problem erneut gelöst werden. Bei DDPG hingegen wird ein Regler erlernt, der das zu steuernde dynamische System – aus nahezu beliebigen Anfangswerten – in den gewünschten Zustand überführt. Nachteilig ist jedoch, dass der Algorithmus sich auf hochgradig nichtlineare Systeme bisher nicht anwenden lässt und eine geringe Dateneffizienz aufweist. / In this thesis, the application of reinforcement learning for the control of nonlinear dynamical systems is researched. At first, the relevant principles of stochastic optimal control and machine learning are explained. Afterwards, reinforcement learning is embedded in the context of optimal control. Three methods of deep reinforcement learning are analyzed. A particular algorithm, namely Deep-Deterministic-Policy-Gradient (DDPG), is chosen for further studies on a variety of mechanical systems. Furthermore, the reinforcement learning approach is compared to a model-based trajectory optimization method, called iterative linear-quadratic regulator (iLQR). All control problems can be successfully solved with the trajectory optimization approach, but for new initial conditions, the problem has to be solved again. In contrast, with DDPG a \emph{global} feedback controller is learned, that can drive the controlled system in the desired state. Disadvantageous is the poor data efficiency and the lack of applicability to highly nonlinear systems.
|
152 |
Fuel cell and intelligent power processing using nonlinear controlJanuary 2004 (has links)
This dissertation is a detailed scientific study concerning a proton exchange membrane fuel cell, which is coupled to a DC-to-DC converter as the power processor, serving as a power source. The novel aspect of the dissertation is the use of a new controller or nonlinear observer to predict parameter estimation of the fuel cell and the DC-to-DC converter as the load potential changes for the automated control system. Nonlinear control algorithms, which include nonlinear observers, were developed for such systems. / acase@tulane.edu
|
153 |
Control of Hyperbolic Heat Transfer Mechanisms Application to the Distributed Concentrated Solar CollectorsElmetennani, Shahrazed 04 1900 (has links)
This dissertation addresses the flow control problem in hyperbolic heat transfer mechanisms. It raises in concentrated distributed solar collectors to enhance their production efficiency under the unpredictable variations of the solar energy and the external disturbances. These factors which are either locally measured (the solar irradiance) or inaccessible for measurement (the collectors’ cleanliness) affect the source term of the distributed model and represent a major difficulty for the control design. Moreover, the temperature in the collector can only be measured at the boundaries. In this dissertation, we propose new adaptive control approaches to provide the adequate level of heat while coping with the unpredictable varying disturbances. First, we design model based control strategies for a better efficiency, in terms of accuracy and response time, with a relatively reduced complexity.
Second, we enhance the controllers with on-line adaptation laws to continuously
update the efficient value of the external conditions. In this study, we approach the control problem using both, the infinite dimensional model (late lumping) and a finite dimensional approximate representation (early lumping). For the early lumping approach, we introduce a new reduced order bilinear approximate model for system analysis and control design. This approximate state representation is then used to derive a nonlinear state feedback resorting to Lyapunov stability theory. To compensate for the external disturbances and the approximation uncertainties, an adaptive controller is developed based on a phenomenological representation of the system dynamics. For the late lumping approach, we propose two PDE based controllers by stabilization of the reference tracking error distributed profile.
The control laws are explicitly defined as functions of the available measurement. The
first one is obtained using a direct approach for error stabilization while the second one
is derived through a nonlinear mapping. Furthermore, we endow the nonlinear controllers with an adaptation law to cope with the unpredictable unmeasured disturbances. The proposed adaptation law is based on a Proportional plus Integral correction feedback. We show that the control objectives with the required performance can be achieved following both approaches, but yet are conditioned with the physical limitations of the system.
|
154 |
Techniques Lyapunov pour une classe de systèmes hybrides et synthèses de contrôleurs à réinitialisation / Lyapunov techniques for a class of hybrid systems and reset controller syntheses for continuous-time plantsFichera, Francesco 11 October 2013 (has links)
Ce manuscrit présente des résultats de recherche concernant une certaine classe de systèmes hybrides. Les systèmes hybrides peuvent être utilises pour la modélisation de systèmes physiques complexes et hétérogènes dont l’évolution dans le temps présente des phénomènes discrets, tels que les commutations des convertisseurs ou les impacts des systèmes mécaniques. De la même manière, la théorie hybride peut être utilisée pour concevoir des contrôleurs hybrides, en général plus performants par rapport aux contrôleurs a temps continu.Dans ce cadre, les résultats de ce manuscrit peuvent être divises en trois parties. D'abord des résultats de stabilité par rapport à un indice de performance de type Hinfini sont présentes pour une classe plutôt large de systèmes hybrides. Ensuite, nous introduisons de nouvelles architectures de contrôleurs hybrides pour les systèmes à temps continu caractérisées par le fait que leur état peut être réinitialisé en fonction de la trajectoire. Enfin, nous présentons une technique de synthèse convexe pour la conception d'un contrôleur hybride multi-objectif. La comparaison avec les résultats classique met en évidence les avantages en termes de performance par rapport aux contrôleurs a temps continu classiques, tout en préservant la propriété de robustesse et la simplicité de conception.Bien que la théorie hybride soit en plein développement, ces travaux généralisent certains résultats existants, en améliorant la simplicité d’implémentation des solutions grâce à l'utilisation de la programmation semi-definie. En plus les architectures de contrôleurs hybrides présentées ont l'avantage de simplifier la généralisation de quelques résultats classiques concernant la synthèse optimale par rapport à des indices de performance communs. / This dissertation presents some results on hybrid systems. Hybrid systems can be used to model complex physical and heterogeneous systems whose time evolution experiences discrete phenomena, such as commutations in electronic converters or impacts in mechanical systems. In the meantime the hybrid theory can be used to design hybrid controllers which exhibit better performance than the classical continuous-time controllers.In this context, the results in this dissertation can be divided en three parts. First, some stability results with respect to the Hinfinity performance index are presented for a wide class of hybrid controllers. Second, we introduce new hybrid controller architectures for continuous-time systems, where the state of the hybrid controller can be reinitialized depending on the trajectory of the system. Finally, we present a convex synthesis of a multiobjective hybrid controller. The comparisons with the classical results show the improvements that can be achieved with hybrid controllers, maintaining the property of robustness and simplicity of design.Although the hybrid theory is in full development, this work generalizes some existing results by improving the simplicity of their usage by means of semidefinite programming tools. Moreover some hybrid architectures are able to generalize some classic results regarding the optimal synthesis with respect to popular performance indexes.
|
155 |
Trajectory Design Based on Robust Optimal Control and Path Following Control / ロバスト最適制御と経路追従制御に基づく軌道設計Okura, Yuki 25 March 2019 (has links)
付記する学位プログラム名: デザイン学大学院連携プログラム / 京都大学 / 0048 / 新制・課程博士 / 博士(工学) / 甲第21761号 / 工博第4578号 / 新制||工||1713(附属図書館) / 京都大学大学院工学研究科航空宇宙工学専攻 / (主査)教授 藤本 健治, 教授 泉田 啓, 教授 太田 快人 / 学位規則第4条第1項該当 / Doctor of Philosophy (Engineering) / Kyoto University / DGAM
|
156 |
Friction compensation in the swing-up control of viscously damped underactuated roboticsDe Almeida, Ricardo Galhardo January 2018 (has links)
A dissertation submitted to the Faculty of Engineering and the Built Environment,
University of the Witwatersrand, Johannesburg, in fulfilment of the requirements
for the degree of Master of Science in Engineering in the Control Research Group
School of Electrical and Information Engineering, Johannesburg, 2017 / In this research, we observed a torque-related limitation in the swing-up control
of underactuated mechanical systems which had been integrated with viscous
damping in the unactuated joint. The objective of this research project was thus to
develop a practical work-around solution to this limitation.
The nth order underactuated robotic system is represented in this research as a
collection of compounded pendulums with n-1 actuators placed at each joint with
the exception of the first joint. This system is referred to as the PAn-1 robot (Passive
first joint, followed by n-1 Active joints), with the Acrobot (PA1 robot) and the PAA
robot (or PA2 robot) being among the most well-known examples. A number of friction
models exist in literature, which include, and are not exclusive to, the Coulomb
and the Stribeck effect models, but the viscous damping model was selected for
this research since it is more extensively covered in existing literature. The effectiveness
of swing-up control using Lyapunov’s direct method when applied on the
undamped PAn-1 robot has been vigorously demonstrated in existing literature, but
there is no literature that discusses the swing-up control of viscously damped systems.
We show, however, that the application of satisfactory swing-up control using
Lyapunov’s direct method is constrained to underactuated systems that are either
undamped or actively damped (viscous damping integrated into the actuated joints
only). The violation of this constraint results in the derivation of a torque expression
that cannot be solved for (invertibility problem, for systems described by n > 2) or a
torque expression which contains a conditional singularity (singularity problem, for
systems with n = 2). This constraint is formally summarised as the matched damping
condition, and highlights a clear limitation in the Lyapunov-related swing-up control
of underactuated mechanical systems. This condition has significant implications
on the practical realisation of the swing-up control of underactuated mechanical
systems, which justifies the investigation into the possibility of a work-around. We
thus show that the limitation highlighted by the matched damping condition can be
overcome through the implementation of the partial feedback linearisation (PFL)
technique. Two key contributions are generated from this research as a result, which
iii
include the gain selection criterion (for Traditional Collocated PFL), and the convergence
algorithm (for noncollocated PFL).
The gain selection criterion is an analytical solution that is composed of a set of
inequalities that map out a geometric region of appropriate gains in the swing-up
gain space. Selecting a gain combination within this region will ensure that the
fully-pendent equilibrium point (FPEP) is unstable, which is a necessary condition
for swing-up control when the system is initialised near the FPEP. The convergence
algorithm is an experimental solution that, once executed, will provide information
about the distal pendulum’s angular initial condition that is required to swing-up a
robot with a particular angular initial condition for the proximal pendulum, along
with the minimum gain that is required to execute the swing-up control in this
particular configuration. Significant future contributions on this topic may result
from the inclusion of more complex friction models. Additionally, the degree of
actuation of the system may be reduced through the implementation of energy
storing components, such as torsional springs, at the joint.
In summary, we present two contributions in the form of the gain selection criterion
and the convergence algorithm which accommodate the circumnavigation of the
limitation formalised as the matched damping condition. This condition pertains to the
Lyapunov-related swing-up control of underactuated mechanical systems that have
been integrated with viscous damping in the unactuated joint. / CK2018
|
157 |
Low-cost control of discontinuous systems including impacts and frictionSvahn, Fredrik January 2007 (has links)
For a successful design of an engineering system it is essential to pay careful attention to its dynamic response. This is particularly true, in the case of nonlinear systems, since they can exhibit very complex dynamic behaviour, including multiple co-existing stable solutions and chaotic motions, characterized by large sensitivity to initial conditions. In some systems nonlinear characteristics are desired and designed for, but in other cases they are unwanted and can cause fatigue and failure. A type of dynamical system which is highly nonlinear is discontinuous or non-smooth systems. In this work, systems with impacts are primarily investigated, and this is a typical example of a discontinuous system. To enhance or optimize the performance of dynamical systems, some kind of control can be implemented. This thesis concerns implementation of low-cost control strategies for discontinuous systems. Low-cost control means that a minimum amount of energy is used when performing the control actions, which is a desirable situation regardless of the application. The disadvantage of such a method is that the performance might be limited as compared with a control strategy with no restrictions on energy consumption. In this work, the control objective is to enforce a continuous or discontinuous grazing bifurcation of the system, whichever is desirable. In Paper A, the dynamic response and bifurcation behaviour of an impactoscillator with dry friction is investigated. For a one-degree-of-freedom model of the system, analytical solutions are found in separate regions of state space. These are then used to perform a perturbation analysis around a grazing trajectory. Through the analysis, a condition on the parameters of the system is derived, which assures a continuous grazing bifurcation. It is also shown that the result has bearing on the dynamic response of a two-degree-of-freedom model of the system. A low-cost active control strategy for a class of impact oscillators is proposed in Paper B. The idea of the control method is to introduce small adjustments in the position of the impact surface, at discrete moments in time, to assure a continuous bifurcation. A proof is given for what control parameters assures the stabilization. In Paper C, the proposed low-cost control method is implemented in a quarter-car model of a vehicle suspension, in order to minimize impact velocities with the bumpstop in case of high amplitude excitation. It is shown that the control method is effective for harmonic road excitation. / QC 20101118
|
158 |
Attitude control on manifolds via optimization and contractions with automatic gain tuningVang, Bee 27 September 2021 (has links)
The attitude (or orientation) of an object is often crucial in its ability to perform a task, whether the task is driving a car, flying an aircraft, or focusing a satellite. In traditional control approaches, the attitude is often parameterized by Euler angles or unit quaternions which exhibit problems such as gimbal lock or ambiguity in representation, respectively. These complications prevent the controllers from achieving global stability and worse they may cause real physical harm due to unexpected large motions. More recent works have achieved global stability and avoided these system failures by working directly on the configuration manifold, but these approaches are generally complex or lack automatic, user-friendly ways to tune them.
The goal of this dissertation is to develop simple geometric attitude controllers that are globally, exponentially stable and can be automatically tuned. By simple, we mean that the controllers are computationally efficient for real time implementation on embedded computers and the tuning parameters have geometric interpretations. These properties make the controllers user friendly and practical for real hardware implementation even on fast dynamical systems. Furthermore, we aim to obtain an automatic tuning procedure that ensures convergence, and can also quantify and optimize performance guarantees.
We achieve our goal through four major contributions. The first is a substantial generalization on the theory of classical Riemannian metrics for tangent bundles which provides the ability to compare and combine attitude and velocity terms in the stability analysis, allowing us to consider a larger set of feasible controller gains. The second contribution is a framework to study the stability of attitude systems on manifolds and to automatically tune the controller gains by combining Riemannian geometry, contraction theory, and offline optimization. The third contribution is the development of a globally, exponentially stable attitude controller. This controller overcomes the topological limitation that prevents continuous, time-invariant controllers from achieving global stability by using a time-varying intermediate reference trajectory. The fourth contribution is the improvement of the proposed controllers by way of point-wise-in-time quadratic programming.
|
159 |
Studies on sparse optimal control and passivity-based control for nonlinear mechanical systems / 非線形機械系を対象としたスパース最適制御と受動性に基づく制御に関する研究Hamada, Kiyoshi 23 March 2022 (has links)
京都大学 / 新制・課程博士 / 博士(工学) / 甲第23887号 / 工博第4974号 / 新制||工||1777(附属図書館) / 京都大学大学院工学研究科航空宇宙工学専攻 / (主査)教授 藤本 健治, 教授 泉田 啓, 教授 大塚 敏之 / 学位規則第4条第1項該当 / Doctor of Philosophy (Engineering) / Kyoto University / DFAM
|
160 |
Predictive Control for Linear and Nonlinear Systems Subject to Exogenous DisturbancesParry, Adam Christopher 20 December 2022 (has links)
No description available.
|
Page generated in 0.0475 seconds