Made available in DSpace on 2014-12-17T14:54:48Z (GMT). No. of bitstreams: 1 JoseAP.pdf: 3661698 bytes, checksum: ce1565573ad07f2677ac0b9d8cde09d8 (MD5) Previous issue date: 2007-12-07 / Conselho Nacional de Desenvolvimento Cient?fico e Tecnol?gico / This work describes the study and the implementation of the vector speed control for a three-phase Bearingless induction machine with divided winding of 4 poles and 1,1 kW using the neural rotor flux estimation. The vector speed control operates together with the radial positioning controllers and with the winding currents controllers of the stator phases. For the radial positioning, the forces controlled by the internal machine magnetic fields are used. For the radial forces optimization , a special rotor winding with independent circuits which allows a low rotational torque influence was used. The neural flux estimation applied to the vector speed controls has the objective of compensating the parameter dependences of the conventional estimators in relation to the parameter machine s variations due to the temperature increases or due to the rotor magnetic saturation. The implemented control system allows a direct comparison between the respective responses of the speed and radial positioning controllers to the machine oriented by the neural rotor flux estimator in relation to the conventional flux estimator. All the system control is executed by a program developed in the ANSI C language. The DSP resources used by the system are: the Analog/Digital channels converters, the PWM outputs and the parallel and RS-232 serial interfaces, which are responsible, respectively, by the DSP programming and the data capture through the supervisory system / Este trabalho descreve o estudo e a implementa??o do controle vetorial de velocidade de uma m?quina de indu??o sem mancais trif?sica com bobinado dividido de 4 p?los e 1.1kW utilizando estima??o neural de fluxo do rotor. O controle vetorial de velocidade opera em conjunto com os controles de posicionamento radial e das correntes nos enrolamentos de cada fase do estator. Para o posicionamento radial utilizam-se as for?as controladas pelos campos magn?ticos no interior da m?quina. Para a otimiza??o das for?as radiais operando com influ?ncia m?nima do torque rotacional, foi utilizado um modelo especial de bobinado do rotor com circuitos independentes. A estima??o neural de fluxo aplicada ao controle vetorial de velocidade tem o objetivo de compensar a depend?ncia dos estimadores convencionais em rela??o ?s varia??es nos par?metros da m?quina devido a aumentos de temperatura ou satura??o magn?tica do rotor. O sistema de controle implementado possibilita uma compara??o direta dos respectivos desempenhos de velocidade e posi??o radial da m?quina sob orienta??o do estimador neural em rela??o ao estimador convencional de fluxo. Todo o controle do sistema ? realizado por um programa desenvolvido em linguagem padr?o ANSI C. Os recursos do DSP utilizados pelo sistema s?o: os canais de convers?o A/D, as sa?das PWM e as interfaces paralela e serial RS-232, as quais s?o respons?veis, respectivamente, pela programa??o do DSP e a captura de dados atrav?s de um sistema de supervis?o

Made available in DSpace on 2014-12-17T14:55:05Z (GMT). No. of bitstreams: 1 FranciscoCHQ.pdf: 791868 bytes, checksum: bd94a6e450520ec9b64d043384db8ccb (MD5) Previous issue date: 2008-03-17 / Coordena??o de Aperfei?oamento de Pessoal de N?vel Superior / This work describes the study and the implementation of the speed control for a three-phase induction motor of 1,1 kW and 4 poles using the neural rotor flux estimation. The vector speed control operates together with the winding currents controller of the stator phasis. The neural flux estimation applied to the vector speed controls has the objective of compensating the parameter dependences of the conventional estimators in relation to the parameter machine s variations due to the temperature increases or due to the rotor magnetic saturation. The implemented control system allows a direct comparison between the respective responses of the speed controls to the machine oriented by the neural rotor flux estimator in relation to the conventional flux estimator. All the system control is executed by a program developed in the ANSI C language. The main DSP recources used by the system are, respectively, the Analog/Digital channels converters, the PWM outputs and the parallel and RS-232 serial interfaces, which are responsible, respectively, by the DSP programming and the data capture through the supervisory system / Este trabalho descreve o estudo e a implementa??o de um controle vetorial de velocidade para um motor de indu??o trif?sico de 1.1 kW / 4 p?los utilizando estima??o neural de fluxo do rotor. O controle vetorial de velocidade opera em conjunto com o controle das correntes nos enrolamentos de cada fase do estator. A estima??o neural de fluxo aplicada ao controle vetorial de velocidade tem como objetivo compensar a depend?ncia dos estimadores convencionais em rela??o ?s varia??es nos par?metros da m?quina devido a aumentos de temperatura ou satura??o magn?tica do rotor. O sistema de controle implementado possibilita uma compara??o direta dos respectivos desempenhos de velocidade sob orienta??o do estimador neural em rela??o ao estimador convencional de fluxo. Todo o controle do sistema ? realizado por um programa desenvolvido em linguagem padr?o ANSI C. Os principais recursos do DSP utilizados pelo sistema s?o, respectivamente, os canais de convers?o A/D, as sa?das PWM e as interfaces paralela e serial RS-232, as quais s?o respons?veis, respectivamente, pela programa??o do DSP e a captura de dados atrav?s de um sistema de supervis?o

Ces dernières années, le domaine de l’apprentissage profond a connu des progrès énormes dans des applications allant de la génération d’images, détection d’objets, modélisation du langage à la réponse aux questions visuelles. Les approches classiques telles que l’apprentissage supervisé nécessitent de grandes quantités de données étiquetées et spécifiques à la tâches. Cependant, celles-ci sont parfois coûteuses, peu pratiques, ou trop longues à collecter. La modélisation efficace en données, qui comprend des techniques comme l’apprentissage few-shot (à partir de peu d’exemples) et l’apprentissage self-supervised (auto-supervisé), tentent de remédier au manque de données spécifiques à la tâche en exploitant de grandes quantités de données plus “générales”. Les progrès de l’apprentissage profond, et en particulier de l’apprentissage few-shot, s’appuient sur les benchmarks (suites d’évaluation), les métriques d’évaluation et les jeux de données, car ceux-ci sont utilisés pour tester et départager différentes méthodes sur des tâches précises, et identifier l’état de l’art. Cependant, du fait qu’il s’agit de versions idéalisées de la tâche à résoudre, les benchmarks sont rarement équivalents à la tâche originelle, et peuvent avoir plusieurs limitations qui entravent leur rôle de sélection des directions de recherche les plus prometteuses. De plus, la définition de métriques d’évaluation pertinentes peut être difficile, en particulier dans le cas de sorties structurées et en haute dimension, telles que des images, de l’audio, de la parole ou encore du texte. Cette thèse discute des limites et des perspectives des benchmarks existants, des fonctions de coût (training losses) et des métriques d’évaluation (evaluation metrics), en mettant l’accent sur la modélisation générative - les Réseaux Antagonistes Génératifs (GANs) en particulier - et la modélisation efficace des données, qui comprend l’apprentissage few-shot et self-supervised. La première contribution est une discussion de la tâche de modélisation générative, suivie d’une exploration des propriétés théoriques et empiriques des fonctions de coût des GANs. La deuxième contribution est une discussion sur la limitation des few-shot classification benchmarks, certains ne nécessitant pas de généralisation à de nouvelles sémantiques de classe pour être résolus, et la proposition d’une méthode de base pour les résoudre sans étiquettes en phase de testing. La troisième contribution est une revue sur les méthodes few-shot et self-supervised de détection d’objets , qui souligne les limites et directions de recherche prometteuses. Enfin, la quatrième contribution est une méthode efficace en données pour la description de vidéo qui exploite des jeux de données texte et vidéo non supervisés. / In recent years, the field of deep learning has seen tremendous progress for applications ranging from image generation, object detection, language modeling, to visual question answering. Classic approaches such as supervised learning require large amounts of task-specific and labeled data, which may be too expensive, time-consuming, or impractical to collect. Data-efficient methods, such as few-shot and self-supervised learning, attempt to deal with the limited availability of task-specific data by leveraging large amounts of general data. Progress in deep learning, and in particular, few-shot learning, is largely driven by the relevant benchmarks, evaluation metrics, and datasets. They are used to test and compare different methods on a given task, and determine the state-of-the-art. However, due to being idealized versions of the task to solve, benchmarks are rarely equivalent to the original task, and can have several limitations which hinder their role of identifying the most promising research directions. Moreover, defining meaningful evaluation metrics can be challenging, especially in the case of high-dimensional and structured outputs, such as images, audio, speech, or text. This thesis discusses the limitations and perspectives of existing benchmarks, training losses, and evaluation metrics, with a focus on generative modeling—Generative Adversarial Networks (GANs) in particular—and data-efficient modeling, which includes few-shot and self-supervised learning. The first contribution is a discussion of the generative modeling task, followed by an exploration of theoretical and empirical properties of the GAN loss. The second contribution is a discussion of a limitation of few-shot classification benchmarks, which is that they may not require class semantic generalization to be solved, and the proposal of a baseline method for solving them without test-time labels. The third contribution is a survey of few-shot and self-supervised object detection, which points out the limitations and promising future research for the field. Finally, the fourth contribution is a data-efficient method for video captioning, which leverages unsupervised text and video datasets, and explores several multimodal pretraining strategies.