La manipulation aérienne a été un domaine de recherche actif ces dernières années, principalement parce que les applications actives des véhicules aériens autonomes (UAV en anglais), augmente l'employabilité de ces véhicules pour diverses applications.Le développement récent de la manipulation aérienne a trouvé des applications potentielles dans les deux domaines, militaires et civils. Les applications militaires incluent le patrouilleur des frontières, la détection des mines, la reconnaissance, etc., tandis que les applications civiles sont en matière de gestion des catastrophes, d'inspection des ponts, de construction, de livraison de matériel, de recherche et de sauvetage, etc.La recherche sur la robotique aérienne implique principalement des hélicoptères et des architectures de décollage et d'atterrissage verticales (VTOL). Le principal avantage de ces plates-formes est leur maniabilité et la capacité d'effectuer des vols stationnaires, ce qui est essentiel pour les applications. Cette thèse porte sur les avions VTOL, où l'hélicoptère à quatre rotors ou quadrirotor est principalement étudié.En ce qui concerne le problème de la manipulation aérienne, la quantité d'applications augmente, mais en même temps, la complexité de la modélisation et du contrôle d'un tel système est également plus grande. L'un des plus grands défis réside dans leur charge utile limitée. Certaines approches ont essayé de résoudre le problème en utilisant plusieurs robots pour transporter des charges utiles avec des pinces ou des câbles, où leurs effecteurs et pinces doivent être légers eux-mêmes et capables de saisir des formes complexes. Un autre défi est que la dynamique du robot est considérablement modifiée par l'ajout de charges utiles. Cependant, pour le transport de la charge utile, il est nécessaire que les robots puissent estimer l'inertie de la charge utile et s'y adapter pour améliorer les performances de suivi.Selon les antécédents et les défis sur les véhicules VTOL portant des charges utiles ou des manipulateurs, la contribution du présent travail est centrée sur la modélisation et la conception d'une loi de commande non linéaire et une analyse de stabilité formelle pour la stabilisation asymptotique d'un véhicule VTOL portant un bras manipulateur. Pour cela, un modèle général d'un quadrirotor portant un bras manipulateur est proposé. Après cela, une loi de commande presque globalement asymptotique lisse pour la stabilisation de l'attitude qui prend en compte les effets de mouvement du bras est conçue. Une fois que le problème d'attitude est résolu, il est possible de concevoir un contrôleur non linéaire globalement asymptotique pour la dynamique de position basée sur l'utilisation de somme des fonctions saturés afin de prendre en compte les limitations des actionneurs. Enfin, certaines expériences pour valider les lois de commande proposées sont effectuées. / Aerial manipulation has been an active area of research in recent years, mainly because the active tasking of Unmanned Aerial Vehicles (UAV) increases the employability of these vehicles for various applications.The recent development of the aerial manipulation has found potential applications in both, military and civilian domains. Military applications include border patrolling, mine detection, reconnaissance, etc., while civilian applications are in disaster management, bridge inspection, construction, material delivery, search and rescue, etc.The research on aerial robotics has mainly involved helicopters and Vertical Take-off and Landing (VTOL) architectures. The main advantage of these platforms is theirmaneuverability and the capacity to perform hovers, which is essential for the applications. This thesis deals with VTOL aircrafts, where the four rotor helicopter, quadcopter or quadrotor is mainly studied.Regarding the problem of aerial manipulation, the amount of applications are increased, but at the same time the complexity of modeling and control of such a system are equally bigger. One of the biggest challenges arise from their limited payload. Some approaches have tried to solve the problem using multiple robots to carry payloads with grippers or with cables, where their end effectors and grippers have to be lightweight themselves and capable of grasping complex shapes. Another challenge is that the dynamics of the robot are significantly altered by the addition of payloads. However, for payload transport, it is necessary that the robots are able to estimate the inertia of the payload and adapt to it to improve tracking performance.According to the background and challenges on VTOL vehicles carrying payloads or manipulators, the contribution of the present work is centered on the modelling and the design of a nonlinear control and a formal stability analysis for the asymptotical stabilization of a VTOL vehicle carrying a manipulator arm. For this a general model of a quadcopter carrying a manipulator arm is proposed. After that, a smooth almost globally asymptotically control law for attitude stabilization which takes into account the arm motion effects is designed. Once the attitude problem is solved, it is possible to design a a globally asymptotically nonlinear controller for the translational dynamics based in the usage of nested and sum of saturation functions in order to take into account the actuators limitations. Finally, some experiments in order to validate the proposed control laws are carried out.

Les systèmes d’imagerie médicale ultrasonore ont considérablement amélioré le diagnostic clinique par une meilleure qualité des images grâce à des systèmes plus sensibles et des post-traitements. La communauté scientifique de l’imagerie ultrasonore a consenti à un très grand effort de recherche sur les posttraitements et sur le codage de l’excitation sans s’intéresser, outre mesure, aux méthodes de commandeoptimale. Ce travail s’est donc légitimement tourné vers les méthodes optimales basées sur l’utilisationd’une rétroaction de la sortie sur l’entrée. Pour rendre applicable ces méthodes, ce problème complexe decommande optimale a été transformé en un problème d’optimisation paramétrique sous-optimal et plussimple. Nous avons appliqué ce principe au domaine de l’imagerie ultrasonore : l’échographie, l’imagerieharmonique native et l’imagerie harmonique de contraste avec ou sans codage de la commande.La simplicité de l’approche nous a permis, par une modification de la fonction de coût, de l’adapter àl’imagerie harmonique. Cette adaptation montre que la méthode peut être appliquée à l’imagerie ultrasonoreen générale. / Medical ultrasound imaging systems have greatly improved the clinical diagnosis by improvingthe image quality thanks to more sensitive systems and post-processings. The scientific community has madea great effort of research on post-processing and on encoding the excitation. The methods of the optimalcontrol have been neglected. Our work has focused on the optimal methods based on the feedback fromoutput to input. We have transformed the complex problem of optimal control into an easier suboptimalparametric problem. We apply the principle of optimal control to the ultrasound imaging, the ultrasoundharmonic imaging and to the constrast harmonic imaging with or without encoding.The simplicity of the method has allowed us to adapt it to harmonic imaging by a change in the costfunction. This adaptation shows that our method can usually be applied to the ultrasound imaging.

State dependent Riccati equation (SDRE) plays an important role in nonlinear controller design. For autonomous nonlinear systems that can be expressed in linear form with state-dependent coefficients (SDC), SDRE-based controllers guarantee local asymptotic stability of the closed-loop system, under pointwise stabilizability and detectability conditions. Moreover, the optimal control for a quadratic cost function, when it exists, corresponds to an SDRE-based control design for a specific SDC parameterization of the associated nonlinear system. Unfortunately, the implementation of the SDRE-based controllers is computationally expensive. Various techniques have been developed for solving the SDRE, which are either computationally expensive or lack acceptable precision. In this dissertation, a modified state-dependent differential Riccati equation (MSDDRE) is proposed for approximating the solution of the SDRE, which is easy to implement with moderate computation power and its solution can be made arbitrarily close to that of the SDRE. Therefore, it can be used for real-time implementation of near-optimal controllers for nonlinear systems in state-dependent linear form. The proposed technique is then extended to SDRE-based filter design and its application to SDRE-based output feedback control technique. The proposed technique is also extended to state-dependent H-inf; robust control design for a constant noise attenuation bound, when the solution exists. To reduce the design conservativeness, the technique is further extended to state-dependent H-inf; robust control design with adaptive noise attenuation bound, using gain-scheduling technique and linear matrix inequality (LMI) optimization, to approximate H-inf; optimal control with state-dependent noise-attenuation bound. Local asymptotic stability of the closed-loop system is proven for all proposed techniques. Simulation results further confirm the validity of the development and demonstrate the efficiency of the proposed techniques.

[pt] É considerado o problema da estabilidade de sistemas dinâmicos bilineares em ambiente estocástico. Após a apresentação de uma coletânea de resultados já existentes, são propostas novas condições de estabilidade para sistemas discretos utilizando o método direto de Lyapunov. Tais resultados são comparados com os já existentes. / [en] The stability of bilinear dynamic systems in a stochastic environment is considered, including a survey on this field. New conditions for stability of discrete systems using the direct method of Lyapunov are proposed. The results developed here are compared with chose in the current literature.

De type parabolique et soumis à l’action d’un couple de contrôles (h, k) où h et k jouent des rôles différents ; le contrôle k étant de type "contrôlabilité" et h de type "contrôle optimal".Il est alors naturel de considérer un problème d’optimisation multi-critères. Il existe plusieurs façons d’étudier de tels problèmes. Nous proposons, dans cette thèse, le contrôle de Stackelberg. Il s’agit d’une notion d’optimisation hiérarchique avec, ici, h qui est le "Leader" et k le "Follower". / In this thesis, we are interested in evolution problems governed by parabolic equations subjected to the action of a pair of controls (h, k) where h and k play different roles : the control k being of "controllability" type and h of "optimal control" type.It is then natural to consider a multi-criteria optimization problem. There are several ways to study such problems. We propose in this thesis, the Stackelberg control which is a notion of hierarchical optimization with here, h which is the "Leader" and k the "Follower".

Este trabalho desenvolveu uma metodologia de otimização da localização de material piezelétrico para avaliar vibrações estruturais. O principal objetivo foi estabelecer um procedimento para a determinação concomitante da localização mais adequada para atuadores e sensores piezelétricos através de uma formulação de um problema de otimização topológica. De acordo com a metodologia proposta, a localização desses atuadores e sensores é determinada através da maximização da controlabilidade e da observabilidade, ambas medidas por intermédio do seu gramiano, definindo onde o material deve ter propriedades piezelétricas. Os resultados do processo de otimização foram avaliados em malha fechada através do uso de dois controladores ótimos (LQR e LQG), utilizados em simulações para atenuar as oscilações estruturais resultantes da aplicação de perturbações externas. O desenvolvimento dos algoritmos de controle foi realizado com a utilização de um modelo modal truncado em seus primeiros modos de vibração. Os resultados mostram a eficácia do processo de otimização topológica quanto à localização de atuadores e sensores na estrutura. Além disso, verificou-se que a localização do material piezelétrico melhora o amortecimento estrutural, sendo importante para o desempenho das técnicas de controle utilizadas. / This work proposes a topology optimization methodology for the location of piezoelectric actuator/sensors for active vibration control. The main objective is to develop a procedure to determine the most suitable location for piezoelectric sensors and actuators using a topology optimization formulation. According to the proposed method, the location of these actuators and sensors is determined the maximization of the controllability and observability, both measured by the gramian matrix, defining where the material should have piezoelectric properties. The results of the optimization process are evaluated in closed loop by using two (LQR and LQG) active controllers, which are used in simulations to attenuate structural oscillations resulting from the application of external disturbances. The development of control algorithms was performed with the use of a modal model truncated to lowest modes. Results show the effectiveness of the topology optimization process as the location of actuators and sensors in the structure. Furthermore, it was found that location of the piezoelectric material improves the structural damping, which is important for the performance of control techniques.

This dissertation introduces a Dynamical Neural Network (DNN) model based adaptive inverse optimal control design for a class of nonlinear systems. A DNN structure is developed and stabilized based on a control Lyapunov function (CLF). The CLF must satisfy the partial Hamilton Jacobi-Bellman (HJB) equation to solve the cost function in order to prove the optimality. In other words, the control design is derived from the CLF and inversely achieves optimality when the given cost function variables are determined posterior. All the stability of the closed loop system is ensured using the Lyapunov-based analysis. In addition to structure stability, uncertainty/ disturbance presents a problem to a DNN in that it could degrade the system performance. Therefore, the DNN needs a robust control against uncertainty. Sliding mode control (SMC) is added to nominal control design based CLF in order to stabilize and counteract the effects of disturbance from uncertain DNN, also to achieve global asymptotic stability. In the next section, a DNN observer is considered for estimating states of a class of controllable and observable nonlinear systems. A DNN observer-based adaptive inverse optimal control (AIOC) is needed. With weight adaptations, an adaptive technique is introduced in the observer design and its stabilizing control. The AIOC is designed to control a DNN observer and nonlinear system simultaneously while the weight parameters are updated online. This control scheme guarantees the quality of a DNN's state and minimizes the cost function. In addition, a tracking problem is investigated. An inverse optimal adaptive tracking control based on a DNN observer for unknown nonlinear systems is proposed. Within this framework, a time-varying desired trajectory is investigated, which generates a desired trajectory based on the external inputs. The tracking control design forces system states to follow the desired trajectory, while the DNN observer estimates the states and identifies unknown system dynamics. The stability method based on Lyapunov-based analysis is guaranteed a global asymptotic stability. Numerical examples and simulation studies are presented and shown for each section to validate the effectiveness of the proposed methods.

O aumento do custo e da demanda de energia, juntamente com a escassez de recursos não renováveis e o crescente apelo para a sustentabilidade dos processos, incentiva o desenvolvimento de equipamentos e sistemas mais eficientes mediante a otimização e o controle dos recursos energéticos disponíveis. A proposta deste trabalho consiste em estabelecer uma abordagem termodinâmica para o controle e a otimização de processos visando o aumento da eficiência na operação dos mesmos. Para tanto, foi desenvolvido um problema de controle ótimo formulado com base no critério de eficiência termodinâmica de acordo com a 2ª Lei, com o objetivo de otimizar a transição entre estados com respeito a taxa de geração de entropia do sistema, mantendo as variáveis controladas como restrições de desigualdade dependentes do tempo. A estratégia proposta, chamada COBET (controle baseado na eficiência termodinâmica), foi aplicada a uma coluna de destilação binária para controlar a pureza do destilado e do produto de fundo, manipulando a razão de refluxo e a taxa de transferência de calor do refervedor. O COBET apresentou ajuste simples, tempo computacional razoável e desempenho de controle aceitável, quando comparado com controladores convencionais. Foi proposta uma estratégia alternativa, de modo a aprimorar o COBET, chamada R-COBET, que consiste no uso de restrições adicionais que garantem que as variáveis de controle permaneçam entre o estado inicial e final. O R-COBET obteve um desempenho aceitável, menor integral do erro quadrático das variáveis de controle e maior lucro, quando comparado ao COBET. A fim de demonstrar a aplicabilidade da metodologia proposta a outras funções objetivo, uma função empírica de eficiência baseada na 1ª Lei da Termodinâmica foi determinada e utilizada no problema de controle ótimo, sujeito às mesmas restrições consideradas no controlador R-COBET. Também se aplicou a metodologia proposta a um reator CSTR com o intuito de controlar a temperatura e a concentração do produto, considerando-se o controlador na forma R-COBET, e obteve-se um desempenho similar ao do estudo da coluna de destilação. Com estas análises, concluiu-se que a metodologia proposta pode ser aplicada com sucesso a problemas de controle ótimo com estados final e inicial definidos. / The increasing demand of energy and its rising costs, in addition to the non-renewable resources depletion and to the growing interest in the sustainability in industrial processes, stimulate the development of equipment and systems that are more efficient through the control and optimization of available energy resources. The purpose of this work is to establish a thermodynamic approach to process control and optimization, in order to improve operational efficiency. To this end, it is developed in this work an optimal control problem formulated on the basis of a thermodynamic efficiency criterion, as dictated by the 2nd Law. The principle is to optimize the transition between states with respect to thermodynamic efficiency – regarded as the square of entropy generation rate – while maintaining the controlled variables as time dependent inequality constraints. The proposed strategy, called here TEBC (Thermodynamic Efficiency Based Control) was applied to a binary distillation column in order to control distillate and bottom product purities within specifications while manipulating the reflux ratio and reboiler heat transfer rate. When compared to other controllers, TEBC presented practical tuning, reasonable computational time and acceptable control performance. An alternative strategy was also proposed in order to improve TEBC, called R-TEBC, which considers additional constraints to ensure that control variables remain between the initial and final states. The R-TEBC achieved acceptable performance, smaller integral squared error in the control variables and larger profit, when compared to TEBC. In order to demonstrate the applicability of the proposed methodology to other objective functions, an empirical efficiency function was proposed for the distillation column example and used in the optimal control problem, in the form of a R-TEBC controller. Moreover, an exothermic CSTR reactor with product concentration and reactor temperature as controlled variables was also studied, using additional constraints (R-TEBC), and the results was similar to those obtained in the distillation column. With this analysis, it was concluded that the proposed methodology can be applied to optimal control problems of this kind with fixed initial and final states.

Orientador: Prof. Dr. Marat Rafikov / Dissertação (mestrado) - Universidade Federal do ABC, Programa de Pós-Graduação em Engenharia Mecânica, 2016. / O objetivo desta dissertação é o controle de trajetória e navegação de um robô móvel autônomo subaquático. O robô abordado nesta dissertação faz parte de uma nova classe de AUV¿s (Autonomous Underwater Vehicles) chamados de planadores subaquáticos. Os planadores subaquáticos se destacam por substituírem os propulsores por um mecanismo de translação e rotação de massas internas e variação de flutuabilidade, para se locomoverem. Essa nova classe de veículos visa superar o problema de autonomia, encontrado em outros tipos de AUV¿s, devido à limitada duração das baterias. O controle de posicionamento é realizado através da linearização do modelo não ¿linear e aplicada a estratégia de controle ótimo conhecida por LQR (Regulador Linear Quadrático). Essa estratégia é aplicada ao problema de regulação do robô a uma referência e a eficácia do método de controle do robô é verificada através de simulações numéricas. Nas simulações verificou ¿ se que o controle LQR foi capaz de fazer com que o veículo convergisse para a trajetória de referência desejada, demonstrando assim que a aplicação de um controlador linear para o modelo não ¿ linear do veículo é uma opção de aplicação em missões reais. / Within the structures optimization study area, one of the extensively explored methods is the TO objetivo desta dissertação é o controle de trajetória e navegação de um robô móvel autônomo subaquático. O robô abordado nesta dissertação faz parte de uma nova classe de AUV¿s (Autonomous Underwater Vehicles) chamados de planadores subaquáticos. Os planadores subaquáticos se destacam por substituírem os propulsores por um mecanismo de translação e rotação de massas internas e variação de flutuabilidade, para se locomoverem. Essa nova classe de veículos visa superar o problema de autonomia, encontrado em outros tipos de AUV¿s, devido à limitada duração das baterias. O controle de posicionamento é realizado através da linearização do modelo não ¿linear e aplicada a estratégia de controle ótimo conhecida por LQR (Regulador Linear Quadrático). Essa estratégia é aplicada ao problema de regulação do robô a uma referência e a eficácia do método de controle do robô é verificada através de simulações numéricas. Nas simulações verificou ¿ se que o controle LQR foi capaz de fazer com que o veículo convergisse para a trajetória de referência desejada, demonstrando assim que a aplicação de um controlador linear para o modelo não ¿ linear do veículo é uma opção de aplicação em missões reais.

abstract: This thesis considers two problems in the control of robotic swarms. Firstly, it addresses a trajectory planning and task allocation problem for a swarm of resource-constrained robots that cannot localize or communicate with each other and that exhibit stochasticity in their motion and task switching policies. We model the population dynamics of the robotic swarm as a set of advection-diffusion- reaction (ADR) partial differential equations (PDEs). Specifically, we consider a linear parabolic PDE model that is bilinear in the robots' velocity and task-switching rates. These parameters constitute a set of time-dependent control variables that can be optimized and transmitted to the robots prior to their deployment or broadcasted in real time. The planning and allocation problem can then be formulated as a PDE-constrained optimization problem, which we solve using techniques from optimal control. Simulations of a commercial pollination scenario validate the ability of our control approach to drive a robotic swarm to achieve predefined spatial distributions of activity over a closed domain, which may contain obstacles. Secondly, we consider a mapping problem wherein a robotic swarm is deployed over a closed domain and it is necessary to reconstruct the unknown spatial distribution of a feature of interest. The ADR-based primitives result in a coefficient identification problem for the corresponding system of PDEs. To deal with the inherent ill-posedness of the problem, we frame it as an optimization problem. We validate our approach through simulations and show that reconstruction of the spatially-dependent coefficient can be achieved with considerable accuracy using temporal information alone. / Dissertation/Thesis / Masters Thesis Mechanical Engineering 2014