Submitted by RICARDO FERNANDO NUNES null (ricardofnes@gmail.com) on 2016-05-09T19:20:22Z No. of bitstreams: 1 Ricardo_Nunes_Pos_Defesa_FINAL.pdf: 4578185 bytes, checksum: 7d5f2601e5d33327aef16ab69587d77f (MD5) / Approved for entry into archive by Felipe Augusto Arakaki (arakaki@reitoria.unesp.br) on 2016-05-12T19:25:25Z (GMT) No. of bitstreams: 1 nunes_rf_me_ilha.pdf: 4578185 bytes, checksum: 7d5f2601e5d33327aef16ab69587d77f (MD5) / Made available in DSpace on 2016-05-12T19:25:25Z (GMT). No. of bitstreams: 1 nunes_rf_me_ilha.pdf: 4578185 bytes, checksum: 7d5f2601e5d33327aef16ab69587d77f (MD5) Previous issue date: 2016-03-31 / Neste trabalho apresenta-se uma abordagem para o mapeamento da cinemática inversa utilizando Redes Neurais Artificiais do tipo Perceptron Multicamadas na configuração em paralelo, tendo como referência o protótipo de um manipulador robótico de 5 graus de liberdade, composto por sete servomotores controlado pela plataforma de desenvolvimento Intel® Galileo Gen 2. As equações da cinemática inversa, normalmente apresentam múltiplas soluções, desta forma, uma solução interessante e frequentemente encontrada na literatura são as Redes Neurais Artificiais (RNA) em razão da sua flexibilidade e capacidade de aprendizado por meio do treinamento. As Redes Neurais são capazes de entender a relação cinemática entre o sistema de coordenadas das juntas e a posição final da ferramenta do manipulador. Para avaliar a eficiência do método proposto foram realizadas simulações no software MATLAB, as quais demostram pelos resultados obtidos e comparações a uma RNA do tipo MLP simples, aproximadamente redução das médias dos erros das juntas em até 87,8% quando aplicado à trajetória e 80% quando aplicado a pontos distribuídos no volume de trabalho. / This paper presents an approach to the mapping of inverse kinematics using Artificial Neural Networks Multilayer Perceptron in parallel configuration, in the prototype of a robotic manipulator 5 degrees of freedom, as reference, composed of seven servomotors controlled by development board Intel® Galileo Gen 2. The equations of inverse kinematics, usually have multiple solutions, therefore, an interesting solution and often found in the literature are the Artificial Neural Networks (ANN) because of their flexibility and learning capacity through training. Neural Networks are able to understand the kinematic relationship between the coordinate system of the joints and the final position of the manipulator tool. To evaluate the efficiency of the proposed, simulations in MATLAB software are performaded, that demonstrate by the results obtained and compared to a simple MLP type RNA, one reduction in mean errors of the joints by up to 87.8% when applied to the path and 80% when applied to points distributed in the work space.

Orientador: Suely Cunha Amaro Mantovani / Resumo: Neste trabalho apresenta-se uma abordagem para o mapeamento da cinemática inversa utilizando Redes Neurais Artificiais do tipo Perceptron Multicamadas na configuração em paralelo, tendo como referência o protótipo de um manipulador robótico de 5 graus de liberdade, composto por sete servomotores controlado pela plataforma de desenvolvimento Intel® Galileo Gen 2. As equações da cinemática inversa, normalmente apresentam múltiplas soluções, desta forma, uma solução interessante e frequentemente encontrada na literatura são as Redes Neurais Artificiais (RNA) em razão da sua flexibilidade e capacidade de aprendizado por meio do treinamento. As Redes Neurais são capazes de entender a relação cinemática entre o sistema de coordenadas das juntas e a posição final da ferramenta do manipulador. Para avaliar a eficiência do método proposto foram realizadas simulações no software MATLAB, as quais demostram pelos resultados obtidos e comparações a uma RNA do tipo MLP simples, aproximadamente redução das médias dos erros das juntas em até 87,8% quando aplicado à trajetória e 80% quando aplicado a pontos distribuídos no volume de trabalho. / Abstract: This paper presents an approach to the mapping of inverse kinematics using Artificial Neural Networks Multilayer Perceptron in parallel configuration, in the prototype of a robotic manipulator 5 degrees of freedom, as reference, composed of seven servomotors controlled by development board Intel® Galileo Gen 2. The equations of inverse kinematics, usually have multiple solutions, therefore, an interesting solution and often found in the literature are the Artificial Neural Networks (ANN) because of their flexibility and learning capacity through training. Neural Networks are able to understand the kinematic relationship between the coordinate system of the joints and the final position of the manipulator tool. To evaluate the efficiency of the proposed, simulations in MATLAB software are performaded, that demonstrate by the results obtained and compared to a simple MLP type RNA, one reduction in mean errors of the joints by up to 87.8% when applied to the path and 80% when applied to points distributed in the work space. / Mestre

This paper proposes the position control of an articulated manipulator of two degrees of freedom driven by induction motors with a controller in the current loop. The work includes the mechanical modeling of the handler that will be developed from the direct and inverse kinematics and position control of induction motors operating system and uncoupled allocated in the manipulator. Thus, we developed two strategies for control of manipulators with two degrees of freedom, one using the scheme for field-oriented vector control and other means of indirect sliding mode control (Sliding-Mode Control - SMC). These control strategies are applied to current loop of induction motors that drive the manipulator. The parameters of the position controller of manipulator are taken into account in calculating the controller parameters of the current control loop, in order to obtain satisfactory results in the positioning the degrees of freedom. In addition, we performed a comparative study between the indirect field-oriented vector control and sliding mode control applied to the current loop. The advantage of the SMC compared to the indirect field-oriented vector control due to the first had in its control law, developed in this thesis, the use of position degree of freedom manipulator applied directly in the control law, while the second acted only as a disturbed rejection controller for the position loop. The Proportional-Integral (PI) was used in the position and speed loops for both current control algorithms to provide a standard for comparison between. For the purpose of implementing the control system individually for each engine and motors coupled to the manipulator used a digital signal processor. / O presente trabalho propÃe o controle de posiÃÃo de um manipulador articulado de dois graus de liberdade acionado a partir de motores de induÃÃo trifÃsicos com um controlador na malha de corrente. O trabalho contempla a modelagem mecÃnica do manipulador que serà desenvolvida a partir das cinemÃticas direta e inversa e o controle de posiÃÃo dos motores de induÃÃo atuando desacoplados do sistema e alocados no manipulador. Deste modo, foram desenvolvidas duas estratÃgias de controle de manipuladores com dois graus de liberdade, uma utilizando o esquema por controle vetorial de campo orientado indireto e outra atravÃs de controle por modos deslizantes (Sliding Mode Control â SMC). Estas estratÃgias de controle sÃo aplicadas a malha de corrente dos motores de induÃÃo que acionam o manipulador. Os parÃmetros do controlador de posiÃÃo dos manipuladores sÃo levados em consideraÃÃo no cÃlculo dos parÃmetros do controlador da malha de controle de corrente, de modo a se obter resultados satisfatÃrios no posicionamento dos graus de liberdade. AlÃm disso, foi realizado um estudo comparativo entre o controle vetorial de campo orientado indireto e o controle de modos deslizantes aplicado na malha de corrente. A vantagem do SMC em relaÃÃo ao controle vetorial de campo orientado indireto deveu-se a que o primeiro possuÃa em sua lei de controle desenvolvida nesta tese a utilizaÃÃo direta da posiÃÃo do grau de liberdade do manipulador, enquanto que o segundo atuava somente como um controlador com rejeiÃÃo ao distÃrbio. O controlador Proporcional-Integral (PI) foi utilizado nas malhas de posiÃÃo e velocidade de modo a fornecer um padrÃo de comparaÃÃo confiÃvel entre os controladores de corrente. Com a finalidade de implementar o sistema de controle de cada motor individualmente e dos motores acoplados ao manipulador utilizou-se um processador digital de sinais.

A l'aide de deux réactions de transfert (d,3He), la première avec MUST2 au GANIL et la deuxième avec le split-pole à Orsay, nous avons déterminé la position des états trous de protons dans les noyaux riches en neutrons : le 71Cu (N=42) et le 69Cu (N=40).A partir de la désintégration beta et de la spectroscopie laser, on sait que l'état excité f5/2 chute brutalement en énergie pour N>40 et devient même l'état fondamental dans le 75Cu. Cette chute en énergie a été expliqué par des travaux théoriques notamment à cause de la force tenseur entre les protons et les neutrons. La prédiction sur le partenaire spin-orbite f7/2 est qu'il devrait sentir également un effet à cause de cette force. Expérimentalement, les états trous de proton f7/2 ne sont pas connus pour N>40. Dans le 71Cu, deux états 7/2- sont connus autour de 1 MeV d'énergie d'excitation et sont des possibles candidats pour correspondre à cet état trou de proton. L'expérience au GANIL a eu lieu en mars 2011. Un faisceau secondaire de 72Zn à 38 AMeV a été produit par fragmentation et purifié à travers le spectromètre LISE. La réaction de transfert en cinématique inverse a été étudié à l'aide des télescopes MUST2 plus quatre détecteurs de silicium de 20 micromètres dans le but d'identifier les 3He de basse énergie. Le spectre en énergie d'excitation du 71Cu a été reconstruit grâce à la méthode de la masse manquante, les distributions angulaires ont été extraite et comparé avec un modèle de réaction utilisant les codes DWUCK4 et DWUCK5. A partir de ce travail au GANIL, aucun état n'a été peuplé autour de 1 MeV concluant que le centroid de la force réside à plus haute énergie d'excitation.Nous avalons également mesuré une nouvelle fois la les états trous de proton dans le 69Cu avec la même réaction de transfert mais cette fois en cinématique directe à Orsay dans le but d'étendre les données actuelles sur ce noyau où 60% de la force f7/2 est manquante et dans le but également d'avoir une analyse consistante des facteurs spectroscopiques entre les deux noyaux. Cette analyse consistante nous permet de pouvoir comparer l'évolution du centroid de la force entre les deux isotopes quand l'orbite neutron g9/2 commence à se remplir. Dans cette seconde expérience, nous avons utilisé un faisceau de deuton à 27 MeV produit par le tandem et une cible de 70Zn. Nous avons pu extraire trois nouvelles distributions angulaires et mesurer une nouvelle partie de la force f7/2.Enfin, pour pouvoir interpréter nos résultat, des calculs modèle en couche ont été effectué par le groupe de Strasbourg avec le code Antoine. L'espace de valence utilisé consiste en un coeur 48Ca avec les orbitales protons f7/2, p3/2, f5/2, p1/2 et les orbitales neutron p3/2, f5/2, p1/2, g9/2, d5/2. Les calculs inclus jusqu'à 8p-8h et montrent que la force se situe en effet à haute énergie d'excitation et qu'aucun état de trou de proton n'est calculé autour de 1 MeV. / In two (d,3He) transfer reactions with MUST2 at GANIL and the split-pole at Orsay, we have determined the position of the proton-hole states in the neutron-rich 71Cu (N=42) and 69Cu (N=40) isotopes. We have found that in 71Cu the hole strength of the f7/2 orbital lies at higher excitation energies than expected. From beta-decay and laser spectroscopy, the f5/2 first excited particle state in these isotopes was known to come down rapidly in energy when passing N=40 and even become the ground state in 75Cu. This sudden energy shift has been explained in a number of theoretical works. The prediction for the f7/2 spin-orbit partner was that it would change in energy too through a related effect. Experimentally, the f7/2 proton-hole state is not known for N>40. In 71Cu two 7/2- states around 1 MeV are candidates to be a proton-hole.The experiment at GANIL took place in March 2011. A secondary beam of 72Zn at 38 AMeV was produced by fragmentation and purified through the LISE spectrometer. The transfer reaction in inverse kinematics was studied with the MUST2 detectors plus four 20 micrometer silicon detector to identified the 3He of low kinetic energy. The excitation spectrum of 71Cu was reconstruct thanks to the missing mass method and the angular distributions were extracted and compared with a reaction model using the DWUCK4 and DWUCK5 code. From this work no states have been populated around 1 MeV concluding that the centroid of the f7/2 lies at higher excitation energy. We then remeasured the single-particle strength in 69Cu in the corresponding (d,3He) reaction at Orsay in March 2013 in order to extend the existing data where 60% of the f7/2 strength is missing and make sure that there is a consistent analysis of spectroscopic factors between both isotopes in order to well understood and well quantify the evolution of the f7/2 orbital when we start filling the g9/2 orbital. In this second experiment we have performed the reaction in direct kinematics using a deuteron beam at 27 MeV provided by the tandem and a target of 70Zn. In this work we were able to extract three new angular distributions and we have measured a new part of the f7/2 strength.Finally in order to interpret the results we have obtained from those two experiments, state-of-the-art shell-model calculations have been carried out in collaboration with the Strasbourg group using the Antoine code. The valence space consists in a core of 48Ca with the valence orbitals for protons f7/2, p3/2, f5/2, p1/2 and the orbitals p3/2, f5/2, p1/2, g9/2, d5/2 for neutrons. The calculations have been done allowing 8p-8h and show that the strength is indeed at high energy and no f7/2 proton-hole state lies around 1 MeV in 71Cu.