Analyse de la forme 3D est un sujet de recherche extrêmement actif dans les deux l'infographie et vision par ordinateur. Dans la vision par ordinateur, l'acquisition de formes et de modélisation 3D sont généralement le résultat du traitement des données complexes et des méthodes d'analyse de données. Il existe de nombreuses situations concrètes où une forme visuelle est modélisé par un nuage de points observés avec une variété de capteurs 2D et 3D. Contrairement aux données graphiques, les données sensorielles ne sont pas, dans le cas général, uniformément répartie sur toute la surface des objets observés et ils sont souvent corrompus par le bruit du capteur, les valeurs aberrantes, les propriétés de surface (diffusion, spécularités, couleur, etc), l'auto occlusions, les conditions d'éclairage variables. Par ailleurs, le même objet que l'on observe par différents capteurs, à partir de points de vue légèrement différents, ou à des moments différents cas peuvent donner la répartition des points tout à fait différentes, des niveaux de bruit et, plus particulièrement, les différences topologiques, par exemple, la fusion des mains. Dans cette thèse, nous présentons une représentation de multi-échelle des formes articulés et concevoir de nouvelles méthodes d'analyse de forme, en gardant à l'esprit les défis posés par les données de forme visuelle. En particulier, nous analysons en détail le cadre de diffusion de chaleur pour représentation multi-échelle de formes 3D et proposer des solutions pour la segmentation et d'enregistrement en utilisant les méthodes spectrales graphique et divers algorithmes d'apprentissage automatique, à savoir, le modèle de mélange gaussien (GMM) et le Espérance-Maximisation (EM). Nous présentons d'abord l'arrière-plan mathématique sur la géométrie différentielle et l'isomorphisme graphique suivie par l'introduction de la représentation spectrale de formes 3D articulés. Ensuite, nous présentons une nouvelle méthode non supervisée pour la segmentation de la forme 3D par l'analyse des vecteurs propres Laplacien de graphe. Nous décrivons ensuite une solution semi-supervisé pour la segmentation de forme basée sur un nouveau paradigme d'apprendre, d'aligner et de transférer. Ensuite, nous étendre la représentation de forme 3D à une configuration multi-échelle en décrivant le noyau de la chaleur cadre. Enfin, nous présentons une méthode d'appariement dense grâce à la représentation multi-échelle de la chaleur du noyau qui peut gérer les changements topologiques dans des formes visuelles et de conclure par une discussion détaillée et l'orientation future des travaux. / 3D shape analysis is an extremely active research topic in both computer graphics and computer vision. In computer vision, 3D shape acquisition and modeling are generally the result of complex data processing and data analysis methods. There are many practical situations where a visual shape is modeled by a point cloud observed with a variety of 2D and 3D sensors. Unlike the graphical data, the sensory data are not, in the general case, uniformly distributed across the surfaces of the observed objects and they are often corrupted by sensor noise, outliers, surface properties (scattering, specularities, color, etc.), self occlusions, varying lighting conditions. Moreover, the same object that is observed by different sensors, from slightly different viewpoints, or at different time instances may yield completely different point distributions, noise levels and, most notably, topological differences, e.g., merging of hands. In this thesis we outline single and multi-scale representation of articulated 3D shapes and devise new shape analysis methods, keeping in mind the challenges posed by visual shape data. In particular, we discuss in detail the heat diffusion framework for multi-scale shape representation and propose solutions for shape segmentation and dense shape registration using the spectral graph methods and various other machine learning algorithms, namely, the Gaussian Mixture Model (GMM) and the Expectation Maximization (EM). We first introduce the mathematical background on differential geometry and graph isomorphism followed by the introduction of pose-invariant spectral embedding representation of 3D articulated shapes. Next we present a novel unsupervised method for visual shape segmentation by analyzing the Laplacian eigenvectors. We then outline a semi-supervised solution for shape segmentation based upon a new learn, align and transfer paradigm. Next we extend the shape representation to a multi-scale setup by outlining the heat-kernel framework. Finally, we present a topologically-robust dense shape matching method using the multi-scale heat kernel representation and conclude with a detailed discussion and future direction of work.

Submitted by Maria Aparecida Cézar (aparecida.czar@gmail.com) on 2015-12-11T19:14:27Z No. of bitstreams: 1 DISSERTAÇÃO FINAL de Maria Aparecida Cézar.pdf: 2063662 bytes, checksum: c1bc99295f1fd2b116bbb23d22eb160b (MD5) / Approved for entry into archive by Maria Auxiliadora da Silva Lopes (silopes@ufba.br) on 2015-12-17T17:18:57Z (GMT) No. of bitstreams: 1 DISSERTAÇÃO FINAL de Maria Aparecida Cézar.pdf: 2063662 bytes, checksum: c1bc99295f1fd2b116bbb23d22eb160b (MD5) / Made available in DSpace on 2015-12-17T17:18:57Z (GMT). No. of bitstreams: 1 DISSERTAÇÃO FINAL de Maria Aparecida Cézar.pdf: 2063662 bytes, checksum: c1bc99295f1fd2b116bbb23d22eb160b (MD5) / A presente dissertação faz uma análise da gestão da educação a partir das orientações do Plano de Ações Articuladas (PAR) para a definição e o desenvolvimento da política educacional do município. A partir da utilização de conceitos como descentralização, participação e autonomia procuramos evidenciar as diversas percepções dos atores sociais para empreender a gestão democrática no processo de implementação do PAR no município de Serrinha do Estado da Bahia. A abordagem proposta situa a gestão educacional no processo de mudanças políticas e econômicas influenciadas pelo neoliberalismo a partir dos anos de 1990, destacando a reforma do Estado. Trata-se de uma pesquisa qualitativa do tipo estudo de caso, no qual a coleta de dados foi realizada através de entrevistas, bem como a consulta a documentos. A pesquisa fundamentou-se na vertente do neoinstitucionalismo histórico com o propósito de apropriar-se de seu quadro metodológico aplicado à análise de políticas públicas, a fim de construir uma interpretação da política educacional direcionada à educação básica. Com isso, refletimos sobre a conjuntura que reivindica mudanças na cultura e na organização das instituições educacionais, confrontando as referências teóricas e documentais com os dados coletados no trabalho de campo; problematizando os efeitos das ações do PAR na gestão do sistema de educação municipal e suas escolas ao tentar construir práticas e instrumentos norteadores de suas políticas educacionais. As análises empreendidas apontaram as fragilidades desse processo para a superação do centralismo e controle, característicos de uma gestão por resultados. Constatamos que a adoção desse tipo de gestão no sistema de ensino como estratégia para melhoria da qualidade da educação, tem restringido em muitos aspectos a possibilidade da gestão da educação municipal, com a colaboração dos vários segmentos sociais, em efetivar a gestão democrática partindo da definição das próprias políticas públicas educacionais, de forma a atender as reais necessidades em prol da melhoria da educação. / ABSTRACT This dissertation is a education management analysis from the guidelines of Articulated Action Plan (Plano de Ações Articuladas - PAR) for the definition and development of educational policy of the city. From the use of concepts as decentralization, participation and autonomy, we seek to highlight the different perceptions of social actors to undertake the democratic management in the PAR‟s implementation process at Serrinha city of Bahia. The proposed approach places the educational management in the political and economic changes influenced by neoliberalism from the 1990s, highlighting the reform of the state. It is a qualitative research of a case study, in which data collection was carried out through interviews and consultation documents. The research was based on the historical aspect of neo-institutionalism in order to take ownership of its methodological framework applied to the analysis of public policies, in order to build an interpretation of educational policy directed to basic education. With this, we reflect on the situation claiming changes in the culture and organization of educational institutions, comparing the theoretical and documentary references to the data collected in the field work; questioning the effects of the actions of PAR in the management of municipal education system and their schools to try to build practices and guiding instruments of its educational policies. The current analysis showed the weaknesses of this process to overcome the centralization and control, characteristic of a results-based management. We found that the adoption of this type of management in the education system as a strategy for improving the quality of education, is restricted in many ways the possibility of management of municipal education, with the collaboration of various social segments, in effect the democratic management begins with the identification d their own educational policies, to meet the real needs for the sake of improving education.

La capture de mouvement ambulatoire est un sujet en plein essor pour des applications aussi diverses que le suivi des personnes âgées, l'assistance des sportifs de haut niveau, la réhabilitation fonctionnelle, etc. Ces applications exigent que le mouvement ne soit pas contraint par un système externe, qu’il puisse être réalisé dans différentes situations, y compris en extérieur, que l’équipement soit léger et à un faible coût, qu’il soit réellement ambulatoire et sans procédure complexe de calibration.Actuellement, seuls les systèmes utilisant un exosquelette ou bien des modules inertiels (souvent combinés avec des modules magnétiques) permettent d'effectuer de la capture de mouvement de façon ambulatoire. Le poids de l’exosquelette est très important et il impose des contraintes sur les mouvements de la personne, ce qui le rend inutilisable pour certaines applications telles que le suivi de personnes âgées. La technologie inertielle est plus légère. Elle permet d'effectuer la capture du mouvement sans contrainte sur l’espace de mesure ou sur les mouvements réalisés. Par contre, elle souffre de dérives des gyromètres, et le système doit être recalibré.L'objectif de cette thèse est de développer un système de capture de mouvement de chaînes articulées, bas-coût et temps réel, réellement ambulatoire, ne nécessitant pas d'infrastructure de capture spécifique, permettant une utilisation dans de nombreux domaines applicatifs (rééducation, sport, loisirs, etc.).On s'intéresse plus particulièrement à des mesures intra-corporelles. Ainsi, tous les capteurs sont placés sur le corps et aucun dispositif externe n'est utilisé. Outre un démonstrateur final permettant de valider l'approche proposée, on s'astreint à développer également des outils qui permettent de dimensionner le système en termes de technologie, nombre et position des capteurs, mais également à évaluer différents algorithmes de fusion des données. Pour ce faire, on utilise la borne de Cramer-Rao.Le sujet est donc pluridisciplinaire. Il traite des aspects de modélisation et de dimensionnement de systèmes hybrides entièrement ambulatoires. Il étudie des algorithmes d'estimation adaptés au domaine de la capture de mouvement corps entier en traitant les problématiques d'observabilité de l'état et en tenant compte des contraintes biomécaniques pouvant être appliquées. Ainsi, un traitement adapté permet de reconstruire en temps réel la posture du sujet à partir de mesures intra-corporelles, la source étant également placée sur le corps. / Ambulatory motion capture is of great interest for applications ranging for the monitoring of elderly people, sporty performances monitoring, functional rehabilitation, etc. These applications require that the movement is not constrained by an external system, that it can be performed in different situations, including outdoor environment. It requires lightweight and low cost equipment; it must be truly ambulatory without complex process of calibration.Currently, only systems using an exoskeleton or inertial modules (often combined with magnetic modules) can be used in such situations. Unfortunately, the exoskeleton weight is not affordable and it imposes constraints on the movements of the person, which makes it unusable for certain applications such as monitoring of the elderly.Inertial technology is lighter. Itcan be used for the capture of movement without constraints on the capture space or on the movements. However, it suffers from gyros drift, and the system must be recalibrated.The objective of this thesis is to develop a system of motion capture for an articulated chain, low-cost, real-time truly ambulatory that does not require specific capture infrastructure, that can be used in many application fields (rehabilitation, sport, leisure, etc.).We focus on intra-corporal measurements. Thus, all sensors are placed on the body and no external device is used. In addition to a final demonstrator to validate the proposed approach, we also develop tools to evaluate the system in terms of technology, number and position of sensors, but also to evaluate different algorithms for data fusion. To do this, we use the Cramer-Rao lower bound. \\The subject is multidisciplinary. It addresses aspects of modelling and design of fully ambulatory hybrid systems. It studies estimation algorithms adapted to the field of motion capture of a whole body by considering the problem of observability of the state and taking into account the biomechanical constraints that can be taken into account. Thus, with an appropriate treatment, the pose of a subject can be reconstructed in real time from intra-body measurements.

Made available in DSpace on 2014-08-20T13:48:05Z (GMT). No. of bitstreams: 1 Dulce Mari da Silva Voss_Tese.pdf: 1207388 bytes, checksum: f12856361c468b422ea9b714c252befa (MD5) Previous issue date: 2012-07-25 / The Thesis analyzed the reconstruction movements of the All for education policy on the Articulated Action Plan and the effects in scholar and teaching work, through a case study in the town of Pinheiro Machado, Rio Grande do Sul. This policy, created by the MEC, is constituted strategy to decrease the scholar failure and evasion and increase the IDEB, putting Brazil in the our country in the global education reformulation that support their discourse in the market neoliberal logic. The study was made on the Policies Cycle, proposed by Ball, and the discourse analyzes under a Foucault orientation, searching for an understanding of the inter-relations possible among the global and national reformulation movements in the education area and the discursive technologies that had an influence on the All for Education Commitment policy. On the searched context, the official discourse was recontextualized in a hybrid way, in a conflicted game of negotiation, exclusion and punishment. The local subjects were seduced by the possibilities that PAR has created of receiving more resources and increase the teaching conditions on the cities. The PAR action was marked by the adoption of control technologies and scholar and teaching work, mixed to client political practices on the decision processes, conditions inequality among the municipal net schools of education workers as well as teachers, who were engaged on performing conducts that come possible to reach the expected results. / A Tese aborda os movimentos de recontextualização da política Compromisso Todos pela Educação na gestão do Plano de Ações Articuladas (PAR) e seus efeitos no trabalho escolar e docente, através de um Estudo de Caso no município de Pinheiro Machado, Rio Grande do Sul. Essa política, criada pelo MEC, constitui-se numa estratégia para alcançar a qualidade da educação, diminuir a reprovação e a evasão escolar e elevar o IDEB, visando inserir o Brasil no contexto global das reformas educacionais, cujos discursos estão fundamentados na lógica neoliberal do mercado. A Abordagem do Ciclo de Políticas, proposta por Ball, e a Análise do Discurso de orientação foucaultiana, serviram de ferramentas para a interpretação dos movimentos e das relações de poder-saber produzidas no contexto local com a gestão do PAR, no estudo das possíveis inter-relações entre os movimentos das reformas globais e nacionais, produzidos no campo da educação, e as tecnologias discursivas que influenciam a política Compromisso Todos pela Educação. No contexto pesquisado, o discurso oficial foi recontextualizado de forma híbrida, num jogo conflituoso de negociações, inclusões/exclusões e penalizações. Os sujeitos locais foram seduzidos pelas possibilidades que o PAR cria de ampliar os recursos e melhorar as condições do ensino no município. A gestão do PAR foi marcada pela adoção de tecnologias de controle e regulação do trabalho escolar e docente, mescladas com práticas políticas clientelistas nos processos decisórios. Na gestão local da política, a desigualdade de condições entre as escolas da rede municipal foi acentuada e a preocupação das professoras em atingir os resultados esperados, produziu identidades profissionais performativas.

The automation of the tramming or load, haul and dump (LHD) procedure, performed by a LHD vehicle, holds the potential to improve productivity, efficiency and safety in the mining environment. Productivity is mainly increased by longer working hours; efficiency is improved by repetitive, faultless and predictable work; and safety is improved by removing the human operator from the harsh environment. However, before the automation of the process can be addressed, a thorough understanding of the process and its duty in the overall mining method is required. Therefore, the current applicable mining methods and their areas of potential automation are given. Since the automation of the LHD vehicle is at the core of this project, its implementation in the tramming process is also justified. Also, the current underground navigation methods are given and their shortcomings are named. It is concluded that infrastructure-free navigation is the only viable solution in the ever-changing mining environment. With that in mind, the feasibility of various navigation sensors is discussed and conclusions are drawn. Both kinematic and dynamic modelling of LHD vehicles are introduced. Various forms of kinematic models are given and their underlying modelling assumptions are named. The most prominent assumptions concern the vehicle’s half-length and the inclusion of a wheel-slip factor. Dynamic modelling techniques, with a strong emphasis on tyre modelling, are also stated. In order to evaluate the modelling techniques, field tests are performed on the articulated vehicles, namely the Wright 365 LHD and the Bell 1706C loader. The test on the Wright 365 LHD gives a good impression of the harsh ergonomics under which the operator has to work. A more thorough test is performed on the Bell 1706C articulated loader. The test results are then compared to simulation results obtained from the kinematic models. Also, the above-named assumptions are tested, evaluated and discussed. A dynamic model is also simulated and discussed. Lastly, two localization and control methods are given and evaluated. The first method is an open-loop nonlinear optimal control strategy with periodic position resetting and the second method is a pathtracking controller. AFRIKAANS : Automatisering van die laai-, vervoer- en dompel- (LVD) prosedure het die potensiaal om die produktiwiteit, effektiwiteit en veiligheid van die mynbedryf te verbeter. Produktiwiteit word hoofsaaklik deur langer werksure verhoog, effektiwiteit word deur herhalende, foutlose en voorspelbare werk verbeter en veiligheid word verbeter omdat menslike operateurs uit die gevaarlike ondergrondse omgewing verwyder word. Voordat aandag aan die automatisering van die prosedure geskenk kan word, moet die prosedure en die algemene mynbedrywighede rakende die prosedure deeglik bestudeer en verstaan word. As gevolg hiervan word die huidige, toepaslike mynboumetodes hier gedokumenteer. Die implementering van ʼn gekoppelde LVD-voertuig in die LVD-prosesword ook geregverdig. Verder word die huidige metodes van ondergrondse navigasie genoem en hulle tekortkominge aangedui. Die gevolgtrekking dat infrastruktuur-vrye navigasie die enigste lewensvatbare navigasiemetode in die immer veranderende ondergrondsemynbouomgewing is, word ook gemaak. In die lig daarvan word ʼn verskeidenheid sensors genoem en bespreek. Kinematiese en dinamiese modellering van ʼn LVD-voertuig word bekendgestel. Verskeie kinematiese modelle en hulle onderliggende aannames word genoem. Die mees prominente aannames is die lengte van die gekoppelde voertuig se hoofdele en die insluiting van ʼn wielglipfaktor. Die tegnieke van dinamiese modellering, met die klem op bandmodellering, word ook gegee. Praktyktoetse op gekoppelde voertuie is ook gedoen om die verskillende modelle te evalueer. Die toets op die Wright 365-LVD bied goeie insig in die strawwe ergonomiese toestande waaronder die operateurs moet werk. ʼn Deeglike toets is op ʼn BELL 1706C- gekoppelde laaier, wat kinematies identies aan ʼn LVD-voertuig is, uitgevoer. Die bevindinge van die toets word met bogenoemde modelsimulasies vergelyk en gevolgtrekkings word gemaak. Laastens word lokalisiering en beheer van ʼn LVDvoertuig behandel. Twee beheermetodes, opelus- nie-lineêre optimale beheer met periodieke herstel en padvolgingbeheer word geëvalueer en bespreek. Copyright / Dissertation (MEng)--University of Pretoria, 2012. / Electrical, Electronic and Computer Engineering / unrestricted

This thesis deals with the design of a low-cost robot used for educational purposes at robotic group Robotárna and at the secondary technical school SPŠ and VOŠ Brno, Sokolská, příspěvková organizace. The main intent of the robot is in teaching students programming. The research part deals with collaborational robots phenomenon and their practical usage in present production plants. The thesis covers the mechanical design of the robot, calculations of the required torques for each joints, design of the robot gripper and also a theoretical solution to robot control.

The subject of the master’s thesis is static design and assessment of the static load-bearing steel structure. The structure is designed for the site Brno. Specifically this is an eight-storey office building with a 54 x 12 m floor plan and a total height is 30,2 m. The steel structure is lovated betwrrn teo reinforced concrete towers secure for vertical transport. The structure is designed in three variants. In the one of them a steel frame with hingee attached elements. The ceiling structure is designed as composite steel and concrete structure.

This thesis deals with configuration of radial five-cylinder engine with attention to crank mechanism. Covered topics are motion of crank mechanism with articulated connecting rod, force analysis and balancing of inertial forces. The second part discusses FEM analysis and calculation of safety factor of selected component.

The main intent of the thesis is design of articulated mining dumper. The contet of own designed mining dumper respects technical, aesthetic and ergonomic requirements. The thesis brings a new perspective of design solution in category of currently the most powerful articulated mining dumpers.

Being able to safely control autonomous heavy machinery is of uttermost importance for the conversion of traditional machines to autonomous machines. With the continuous growth of autonomous vehicles around the globe, an increasing effort has been put into certifying autonomous vehicles in terms of reliability and safety. In this thesis, we will investigate the problem with a deviation from the planned path for an autonomous hauler from Volvo Construction Equipment. The autonomous hauler has an error within the kinematic model, the feed-forward curvature-steering controller, due to a slip-effect that comes with the third wheel-axle. The deviation can especially be seen in sharp curves, where the deviation needs to be decreased in order to make the autonomous hauler more dependable and achieve an increased accuracy when following any given path. The aim of the thesis is to develop a fully functional Artificial Neural Network that has a new steering angle as output. The hypothesis for this thesis is to use an ANN to mimic the steering of a human driver, since a real driver compensates for the slipping behavior; both because the operator knows where on the road the machine is and also in the way that a human thinks many steps ahead whilst driving. This proposed ANN will have a monitor function which ensures that the steering angle command operates within its boundaries. Hence this thesis implies that it is indeed possible to ensure that the ANN performs reliably with the help of a monitor function in a simulated environment and can thus be used in dependable systems.