Dissertationen zum Thema „Quaternions space“

The term proximity operations has been widely used in recent years to describe a wide range of space missions that require a spacecraft to remain close to another space object. Such missions include, for example, the inspection, health monitoring, surveillance, servicing, and refueling of a space asset by another spacecraft. One of the biggest challenges in autonomous space proximity operations, either cooperative or uncooperative, is the need to autonomously and accurately track time-varying relative position and attitude references, i.e., pose references, with respect to a moving target, in order to avoid on-orbit collisions and achieve the overall mission goals. In addition, if the target spacecraft is uncooperative, the Guidance, Navigation, and Control (GNC) system of the chaser spacecraft must not rely on any help from the target spacecraft. In this case, vision-based sensors, such as cameras, are typically used to measure the relative pose between the spacecraft. Although vision-based sensors have several attractive properties, they introduce new challenges, such as no direct linear and angular velocity measurements, slow update rates, and high measurement noise. This dissertation investigates the problem of autonomously controlling and estimating the pose of a chaser spacecraft with respect to a moving target spacecraft, possibly uncooperative. Since this problem is inherently hard, the standard approach in the literature is to split the attitude-tracking problem from the position-tracking problem. Whereas the attitude-tracking problem is relatively simple, since the rotational motion is independent from the translational motion, the position-tracking problem is more complicated, as the translational motion depends on the rotational motion. Hence, whereas strong theoretical results exist for the attitude problem, the position problem typically requires additional assumptions. An alternative, more general approach to the pose control and estimation problems is to consider the fully coupled 6-DOF motion. However, fewer results exist that directly address this higher dimensional problem. The main contribution of this dissertation is to show that dual quaternions can be used to extend the theoretical results that exist for the attitude motion into analogous results for the combined position and attitude motion. Moreover, this dissertation shows that this can be accomplished by (almost) just replacing quaternions by dual quaternions in the original derivations. This is because dual quaternions are built on and are an extension of classical quaternions. Dual quaternions provide a compact representation of the pose of a frame with respect to another frame. Using this approach, three new results are presented in this dissertation. First, a pose-tracking controller that does not require relative linear and angular velocity measurements is derived with vision-based sensors in mind. Compared to existing literature, the proposed velocity-free pose-tracking controller guarantees that the pose of the chaser spacecraft will converge to the desired pose independently of the initial state, even if the reference motion is not sufficiently exciting. In addition, the convergence region does not depend on the gains of the controller. Second, a Dual Quaternion Multiplicative Extended Kalman Filter (DQ-MEKF) is developed from the highly successful Quaternion MEKF (Q-MEKF) as an alternative way to achieve pose-tracking without velocity measurements. Existing dual quaternion EKFs are additive, not multiplicative, and have two additional states. The DQ-MEKF is experimentally validated and compared with two conventional EKFs on the 5-DOF platform of the Autonomous Spacecraft Testing of Robotic Operations in Space (ASTROS) facility at the School of Aerospace Engineering at Georgia Tech. Finally, the velocity-free pose-tracking controller is compared qualitatively and quantitatively to a pose-tracking controller that uses the velocity estimates produced by the DQ-MEKF through a realistic proximity operations simulation. Third, a pose-tracking controller that does not require the mass and inertia matrix of the chaser satellite is suggested. This inertia-free controller takes into account the gravitational acceleration, the gravity-gradient torque, the perturbing acceleration due to Earth's oblateness, and constant -- but otherwise unknown -- disturbance forces and torques. Sufficient conditions on the reference pose are also given that guarantee the identification of the mass and inertia matrix of the satellite. Compared to the existing literature, this controller has only as many states as unknown elements and it does not require a priori known upper bounds on any states or parameters. Finally, the inertia-free pose-tracking controller and the DQ-MEKF are tested on a high-fidelity simulation of the 5-DOF platform of the ASTROS facility and also experimentally validated on the actual platform. The equations of motion of the 5-DOF platform, on which the high-fidelity simulation is based, are derived for three distinct cases: a 3-DOF case, a 5-DOF case, and a (2+1)-DOF case. Four real-time experiments were run on the platform. In the first, a sinusoidal reference attitude with respect to the inertial frame is tracked using VSCMGs. In the second, a constant reference attitude is maintained with respect to a target object using VSCMGs and measurements from a camera. In the third, the same sinusoidal reference attitude with respect to the inertial frame tracked in the first experiment is now tracked using cold-gas thrusters. Finally, in the fourth and last experiment, a time-varying 5-DOF reference pose with respect to the inertial frame is tracked using cold-gas thrusters.

Au cours des dernières années, l’apprentissage profond est devenu l’approche privilégiée pour le développement d’une intelligence artificielle moderne (IA). L’augmentation importante de la puissance de calcul, ainsi que la quantité sans cesse croissante de données disponibles ont fait des réseaux de neurones profonds la solution la plus performante pour la resolution de problèmes complexes. Cependant, la capacité à parfaitement représenter la multidimensionalité des données réelles reste un défi majeur pour les architectures neuronales artificielles.Pour résoudre ce problème, les réseaux de neurones basés sur les algèbres des nombres complexes et hypercomplexes ont été développés. En particulier, les réseaux de neurones de quaternions (QNN) ont été proposés pour traiter les données tridi- mensionnelles et quadridimensionnelles, sur la base des quaternions représentant des rotations dans notre espace tridimensionnel. Malheureusement, et contrairement aux réseaux de neurones à valeurs complexes qui sont de nos jours acceptés comme une alternative aux réseaux de neurones réels, les QNNs souffrent de nombreuses lacunes qui sont en partie comblées par les différents travaux détaillés par ce manuscrit.Ainsi, la thèse se compose de trois parties qui introduisent progressivement les concepts manquants, afin de faire des QNNs une alternative aux réseaux neuronaux à valeurs réelles. La premiere partie présente et répertorie les précédentes découvertes relatives aux quaternions et aux réseaux de neurones de quaternions, afin de définir une base pour la construction des QNNs modernes.La deuxième partie introduit des réseaux neuronaux de quaternions état de l’art, afin de permettre une comparaison dans des contextes identiques avec les architectures modernes traditionnelles. Plus précisément, les QNNs étaient majoritairement limités par leurs architectures trop simples, souvent composées d’une seule couche cachée comportant peu de neurones. Premièrement, les paradigmes fondamentaux, tels que les autoencodeurs et les réseaux de neurones profonds sont présentés. Ensuite, les très répandus et étudiés réseaux de neurones convolutionnels et récurrents sont étendus à l’espace des quaternions. De nombreuses experiences sur différentes applications réelles, telles que la vision par ordinateur, la compréhension du langage parlé ainsi que la reconnaissance automatique de la parole sont menées pour comparer les modèles de quaternions introduits aux réseaux neuronaux conventionnels. Dans ces contextes bien spécifiques, les QNNs ont obtenus de meilleures performances ainsi qu’une réduction importante du nombre de paramètres neuronaux nécessaires à la phase d’apprentissage.Les QNNs sont ensuite étendus à des conditions d’entrainement permettant de traiter toutes les représentations en entrée des modèles de quaternions. Dans un scénario traditionnel impliquant des QNNs, les caractéristiques d’entrée sont manuellement segmentées en quatre composants, afin de correspondre à la representation induite par les quaternions. Malheureusement, il est difficile d’assurer qu’une telle segmentation est optimale pour résoudre le problème considéré. De plus, une segmentation manuelle réduit fondamentalement l’application des QNNs à des tâches naturellement définies dans un espace à au plus quatre dimensions. De ce fait, la troisième partie de cette thèse introduit un modèle supervisé et un modèle non supervisé permettant l’extraction de caractéristiques d’entrée désentrelacées et significatives dans l’espace des quaternions, à partir de n’importe quel type de signal réel uni-dimentionnel, permettant l’utilisation des QNNs indépendamment de la dimensionnalité des vecteurs d’entrée et de la tâche considérée. Les expériences menées sur la reconnaissance de la parole et la classification de documents parlés montrent que les approches proposées sont plus performantes que les représentations traditionnelles de quaternions
In the recent years, deep learning has become the leading approach to modern artificial intelligence (AI). The important improvement in terms of processing time required for learning AI based models alongside with the growing amount of available data made of deep neural networks (DNN) the strongest solution to solve complex real-world problems. However, a major challenge of artificial neural architectures lies on better considering the high-dimensionality of the data.To alleviate this issue, neural networks (NN) based on complex and hypercomplex algebras have been developped. The natural multidimensionality of the data is elegantly embedded within complex and hypercomplex neurons composing the model. In particular, quaternion neural networks (QNN) have been proposed to deal with up to four dimensional features, based on the quaternion representation of rotations and orientations. Unfortunately, and conversely to complex-valued neural networks that are nowadays known as a strong alternative to real-valued neural networks, QNNs suffer from numerous limitations that are carrefuly addressed in the different parts detailled in this thesis.The thesis consists in three parts that gradually introduce the missing concepts of QNNs, to make them a strong alternative to real-valued NNs. The first part introduces and list previous findings on quaternion numbers and quaternion neural networks to define the context and strong basics for building elaborated QNNs.The second part introduces state-of-the-art quaternion neural networks for a fair comparison with real-valued neural architectures. More precisely, QNNs were limited by their simple architectures that were mostly composed of a single and shallow hidden layer. In this part, we propose to bridge the gap between quaternion and real-valued models by presenting different quaternion architectures. First, basic paradigms such as autoencoders and deep fully-connected neural networks are introduced. Then, more elaborated convolutional and recurrent neural networks are extended to the quaternion domain. Experiments to compare QNNs over equivalents NNs have been conducted on real-world tasks across various domains, including computer vision, spoken language understanding and speech recognition. QNNs increase performances while reducing the needed number of neural parameters compared to real-valued neural networks.Then, QNNs are extended to unconventional settings. In a conventional QNN scenario, input features are manually segmented into three or four components, enabling further quaternion processing. Unfortunately, there is no evidence that such manual segmentation is the representation that suits the most to solve the considered task. Morevover, a manual segmentation drastically reduces the field of application of QNNs to four dimensional use-cases. Therefore the third part introduces a supervised and an unsupervised model to extract meaningful and disantengled quaternion input features, from any real-valued input signal, enabling the use of QNNs regardless of the dimensionality of the considered task. Conducted experiments on speech recognition and document classification show that the proposed approaches outperform traditional quaternion features

Les matériaux complexes sont aujourd'hui au cœur des enjeux sociétaux majeurs dans la plupart des grands domaines tels que l'énergie, le transport, l'environnement, la conservation/restauration du patrimoine, la santé ou la sécurité. En effet, de par les opportunités d'innovation offertes en matière de fonctionnalités, ces matériaux suscitent de nouvelles problématiques d'analyse et de compréhension multi-physiques et multi-échelles. Il en va de même pour l'instrumentation nécessaire à leur caractérisation.Répandues dans le domaine de la caractérisation non destructive des milieux complexes, les méthodes acoustiques utilisent les propriétés de propagation des ondes mécaniques dans ces matériaux pouvant être hétérogènes et anisotropes.Dans une approche multi-échelle, l'intérêt des méthodes ultrasonores est d'être particulièrement sensibles à leurs propriétés mécaniques, telles que l'élasticité, la rigidité et la viscosité. La nature hétérogène et multiphasique d'un milieu complexe conduit ainsi à la notion de milieu viscoélastique, caractérisé par les coefficients de Lamé généralisés complexes (��∗, ��∗) et leur variation en fonction de la fréquence.L'objectif de cette thèse est de développer une méthode de caractérisation de ces matériaux complexes viscoélastiques qui permette de mesurer simultanément la variation des deux coefficients de Lamé généralisés complexes (��∗, ��∗) en fonction de la fréquence. L'approche proposée est de suivre, dans l'espace et dans le temps, la propagation de l'onde de Rayleigh et d'extraire ses paramètres ellipsométriques (ellipticité χ et orientation θ) en complément des paramètres propagatifs (k' et k'') classiquement déterminés. Basée sur la détection de l'onde par vibrométrie laser 3D à la surface du matériau complexe, et au moyen de l'analyse de Gabor 2D dans l'espace des Quaternions, l'estimation de l'ensemble des paramètres - propagatifs et ellipsométriques - donne accès à la caractérisation complète du milieu avec cette seule onde de Rayleigh.Les développements théoriques proposés dans ce travail, ainsi que les résultats expérimentaux et issus de simulation, confirment l'intérêt de l'ellipsométrie acoustique pour la caractérisation de ces matériaux complexes
Complex materials are at the heart of major societal challenges in most major fields such as energy, transport, environment, heritage conservation/restoration, health and safety. Because of the opportunities for innovation offered in terms of features, these materials are giving rise to new problems of multi-physical and multi-scale analysis and understanding. The same applies to the instrumentation needed to characterize them.Acoustic methods, which are widely used in the non-destructive characterization of complex media, make use of the propagation properties of mechanical waves in these materials, which can be heterogeneous and anisotropic.In a multi-scale approach, the advantage of ultrasonic methods is that they are particularly sensitive to mechanical properties such as elasticity, rigidity and viscosity. The heterogeneous and multiphase nature of a complex medium thus leads to the notion of a viscoelastic medium, characterized by generalized complex Lamé coefficients (��∗, ��∗) and their variation as a function of frequency.The objective of this thesis is to develop a method for characterizing these complex viscoelastic materials that simultaneously measures the variation of the two generalized complex Lamé coefficients (��∗, ��∗) versus the frequency. The proposed approach is to follow, in space and in time, the propagation of the Rayleigh wave and to extract its ellipsometric parameters (ellipticity χ and orientation θ) in addition to the propagation parameters (k' and k'') conventionally determined. Based on the wave detection by 3D laser vibrometry at the surface of the complex material, and by means of 2D Gabor analysis in Quaternion space, the estimation of propagation and ellipsometric parameters gives access to the complete characterization of the complex material only by studying the interaction of a Rayleigh wave with the medium.The theoretical developments proposed in this work, together with experimental and simulation results, confirm the value of acoustic ellipsometry for characterizing these complex materials

Submitted by Viviane Lima da Cunha (viviane@biblioteca.ufpb.br) on 2015-05-19T14:20:06Z No. of bitstreams: 2 arquivototal.pdf: 9893623 bytes, checksum: 211bb6b2f2a20721abf5b2948352c518 (MD5) license_rdf: 22190 bytes, checksum: 19e8a2b57ef43c09f4d7071d2153c97d (MD5)
Approved for entry into archive by Maria Suzana Diniz (msuzanad@hotmail.com) on 2015-05-19T14:27:36Z (GMT) No. of bitstreams: 2 arquivototal.pdf: 9893623 bytes, checksum: 211bb6b2f2a20721abf5b2948352c518 (MD5) license_rdf: 22190 bytes, checksum: 19e8a2b57ef43c09f4d7071d2153c97d (MD5)
Made available in DSpace on 2015-05-19T14:27:36Z (GMT). No. of bitstreams: 2 arquivototal.pdf: 9893623 bytes, checksum: 211bb6b2f2a20721abf5b2948352c518 (MD5) license_rdf: 22190 bytes, checksum: 19e8a2b57ef43c09f4d7071d2153c97d (MD5) Previous issue date: 2013-08-14
Abstract: In this work we study some geometric transformations in the plane and the space. Initially, we present some special types of transformations in the plane and find the matrix of each of these transformations. In the second part we discourse the transformations in the space, emphasizing the rotations. We will use the angles of Euler to determine a rotation in the space around the Cartesian axes and define an equation which allows to rotate a vector around any axis. We also discuss the homogeneous spaces aiming the matrix representation of transformations of translation. Finally, we use the structure of the quaternions group to present a second form to rotation vectors and composition of rotations in the space. We emphasize that this study is essential to describe the motion of objects in the plane and in the space.
Neste trabalho estudamos algumas das transformações geométricas no Plano e no Espaço. Inicialmente, apresentamos alguns tipos de transformações especiais no Plano e encontramos a matriz de cada uma destas transformações. Na segunda parte abordamos as transformações no Espaço, dando ênfase as rotações. Utilizamos os ângulos de Euler para determinar uma rotação no espaço em torno dos eixos cartesianos e definimos uma equação que permite rotacionar um vetores em torno de um eixo qualquer. Também abordamos os espaços homogêneos objetivando a representa ção matricial da transformação de translação. Por último, usamos a estrutura do grupo dos Quatérnios para apresentar uma segunda forma de fazer rotações de vetores e composição de rotações no espaço. Ressaltamos que este estudo é fundamental para descrever o movimento de objetos no plano e no espaço.

The objects of study in this thesis are symmetric products and spaces of algebraic cycles. The first new result concerns symmetric products and it describes the geometry of truncated symmetric products (or, in other terminology, symmetric products modulo 2). We prove that if M is a closed compact connected triangulable manifold, a necessary and sufficient condition for its symmetric products modulo 2 to be manifolds is that M is a circle. We also show that the symmetric products of the circle modulo 2 are homeomorphic to real projective spaces and give an interpretation of this homeomorphism as a real topological analogue of Vieta's theorem. The second result concerns the spaces of real algebraic cycles, first studied by T.K. Lam. We describe a method of calculating the homotopy groups of the spaces of real cycles with integral coefficients on projective spaces; we give an explicit formula for the groups which lie in the "stable range". The third result (or, rather, a group of results) is the construction of a quaternionic analogue of Lawson's theory of algebraic cycles. We define quaternionic objects as those, which are invariant (in the case of varieties) or equivalent (in the case of polynomials) with respect to a free involution on CP²ⁿ⁺¹, induced by the action of the quaternion j on Hⁿ. Basic properties of quaternionic algebraic cycles are studied; a rational "quaternionic suspension theorem" is proved and the spaces of quaternionic cycles with rational coefficients on CP²ⁿ⁺¹ are described. We also present a method of calculating the Betti numbers of the spaces of quaternionic cycles of degree 2 and odd codimension on CP^∞. Some other results that are included in the thesis are a twisted version of the Dold-Thom theorem and an interpretation of the Kuiper-Massey theorem via symmetric products. After the main results on quaternionic cycles were proved, the author learned that similar results were obtained by Lawson, Lima-Filho and Michelson. Their version of the quaternionic suspension theorem is stronger and requires more sophisticated machinery for the proof.

The main purpose of this thesis is to calculate the integral cohomology ring of the symmetric square of quaternionic projective space, which has been an open problem since computations with symmetric squares were first proposed in the 1930's. The geometry of this particular case forms an essential part of the thesis, and unexpected results concerning two universal Pin(4) bundles are also included. The cohomological computations involve a commutative ladder of long exact sequences, which arise by decomposing the symmetric square and the corresponding Borel space in compatible ways. The geometry and the cohomology of the configuration space of unordered pairs of distinct points in quaternionic projective space, and of the Thom space MPin(4), also feature, and seem to be of independent interest.

Let $S$ be a $2n$-dimensional complex manifold equipped with a line bundle with a real-analytic complex connection such that its curvature is of type $(1,1)$, and with a real analytic h-projective structure such that its h-projective curvature is of type $(1,1)$. For $n=1$ we assume that $S$ is equipped with a real-analytic M\"obius structure. Using the structure on $S$, we construct a twistor space of a quaternionic $4n$-manifold $M$. We show that $M$ can be identified locally with a neighbourhood of the zero section of the twisted (by a unitary line bundle) tangent bundle of $S$ and that $M$ admits a quaternionic $S^1$ action given by unit scalar multiplication in the fibres. We show that $S$ is a totally complex submanifold of $M$ and that a choice of a connection $D$ in the h-projective class on $S$ gives extensions of a complex structure from $S$ to $M$. For any such extension, using $D$, we construct a hyperplane distribution on $Z$ which corresponds to the unique quaternionic connection on $M$ preserving the extended complex structure. We show that, in a special case, the construction gives the Feix--Kaledin construction of hypercomplex manifolds, which includes the construction of hyperk\"ahler metrics on cotangent bundles. We also give an example in which the construction gives the quaternion-K\"ahler manifold $\mathbb{HP}^n$ which is not hyperk\"ahler. We show that the same construction and results can be obtained for $n=1$. By convention, in this case, $M$ is a self-dual conformal $4$-manifold and from Jones--Tod correspondence we know that the quotient $B$ of $M$ by an $S^1$ action is an asymptotically hyperbolic Einstein--Weyl manifold. Using a result of LeBrun \cite{Le}, we prove that $B$ is an asymptotically hyperbolic Einstein--Weyl manifold. We also give a natural construction of a minitwistor space $T$ of an asymptotically hyperbolic Einstein--Weyl manifold directly from $S$, such that $T$ is the Jones--Tod quotient of $Z$. As a consequence, we deduce that the Einstein--Weyl manifold constructed using $T$ is equipped with a distinguished Gauduchon gauge.

We present a uniform framework generalising and extending the classical theories of projective differential geometry, c-projective geometry, and almost quaternionic geometry. Such geometries, which we call \emph{projective parabolic geometries}, are abelian parabolic geometries whose flat model is an R-space $G\cdot\mathfrak{p}$ in the infinitesimal isotropy representation $\mathbb{W}$ of a larger self-dual symmetric R-space $H\cdot\mathfrak{q}$. We also give a classification of projective parabolic geometries with $H\cdot\mathfrak{q}$ irreducible which, in addition to the aforementioned classical geometries, includes a geometry modelled on the Cayley plane $\mathbb{OP}^2$ and conformal geometries of various signatures. The larger R-space $H\cdot\mathfrak{q}$ severely restricts the Lie-algebraic structure of a projective parabolic geometry. In particular, by exploiting a Jordan algebra structure on $\mathbb{W}$, we obtain a $\mathbb{Z}^2$-grading on the Lie algebra of $H$ in which we have tight control over Lie brackets between various summands. This allows us to generalise known results from the classical theories. For example, which riemannian metrics are compatible with the underlying geometry is controlled by the first BGG operator associated to $\mathbb{W}$. In the final chapter, we describe projective parabolic geometries admitting a $2$-dimensional family of compatible metrics. This is the usual setting for the classical projective structures; we find that many results which hold in these settings carry over with little to no changes in the general case.

The creation of anatomically correct three-dimensional joints for the simulation of humans is a complex process, a key difficulty being the correction of invalid joint configurations to the nearest valid alternative. Personalised models based on individual joint mobility are in demand in both animation and medicine [1]. Medical models need to be highly accurate animated models less so, however if either are to be used in a real time environment they must have a low temporal cost (high performance). This work briefly explores Support Vector Machine neural networks as joint configuration classifiers that group joint configurations into invalid and valid. A far more detailed investigation is carried out into the use of topologically evolved feed forward neural networks for the generation of appropriately proportioned corrective components which when applied to an invalid joint configuration result in a valid configuration and the same configuration if the original configuration was valid. Discontinuous vector fields were used to represent constraints of varying size, dimensionality and complexity. This culminated in the creation corrective quaternion constraints represented by discontinuous vector fields, learned by topologically evolved neural networks and trained via the resilient back propagation algorithm. Quaternion constraints are difficult to implement and although alternative methods exist [2-6] the method presented here is superior in many respects. This method of joint constraint forms the basis of the contribution to knowledge along with the discovery of relationships between the continuity and distribution of samples in quaternion space and neural network performance. The results of the experiments for constraints on the rotation of limb with regular boundaries show that 3.7 x lO'Vo of patterns resulted in errors greater than 2% of the maximum possible error while for irregular boundaries 0.032% of patterns resulted in errors greater than 7.5%.

En géométrie discrète, les objets euclidiens sont représentés par leurs approximations discrètes, telles que des sous-ensembles du réseau des points à coordonnées entières. Les déplacements de ces ensembles doivent être définis comme des applications depuis et sur un espace discret donné. Une façon de concevoir de telles transformations est de combiner des déplacements continus définis sur un espace euclidien avec un opérateur de discrétisation. Cependant, les déplacements discrétisés ne satisfont souvent plus les propriétés de leurs équivalents continus. En effet, en raison de la discrétisation, de telles transformations ne préservent pas les distances, et la bijectivité et la connexité entre les points sont généralement perdues. Dans le contexte des espaces discrets 2D, nous étudions des déplacements discrétisés sur les réseaux d'entiers de Gauss et d'Eisenstein. Nous caractérisons les déplacements discrétisés bijectifs sur le réseau carré, et les rotations bijectives discrétisées sur le réseau hexagonal régulier. En outre, nous comparons les pertes d'information induites par des déplacements discrétisés non bijectifs définis sur ces deux réseaux. Toutefois, pour des applications pratiques, l'information pertinente n'est pas la bijectivité globale, mais celle d'un déplacement discrétisé restreint à un sous-ensemble fini donné d'un réseau. Nous proposons deux algorithmes testant cette condition pour les sous-ensembles du réseau entier, ainsi qu'un troisième algorithme fournissant des intervalles d'angles optimaux qui préservent cette bijectivité restreinte. Nous nous concentrons ensuite sur les déplacements discrétisés sur le réseau cubique 3D. Tout d'abord, nous étudions à l'échelle locale des défauts géométriques et topologiques induits par des déplacements discrétisés. Une telle analyse consiste à générer toutes les images d'un ensemble du réseau fini sous des déplacements discrétisés. Un tel problème revient à calculer un arrangement d'hypersurfaces dans un espace de paramètres de dimension six. La dimensionnalité et les cas dégénérés rendent le problème insoluble, en pratique, par les techniques usuelles. Nous proposons une solution ad hoc reposant sur un découplage des paramètres, et un algorithme pour calculer des points d'échantillonnage de composantes connexes 3D dans un arrangement de polynômes du second degré. Enfin, nous nous concentrons sur le problème ouvert de déterminer si une rotation discrétisée 3D est bijective ou non. Dans notre approche, nous explorons les propriétés arithmétiques des quaternions de Lipschitz. Ceci conduit à un algorithme qui détermine si une rotation discrétisée donnée, associée à un quaternion de Lipschitz, est bijective ou non
In digital geometry, Euclidean objects are represented by their discrete approximations, e.g. subsets of the lattice of integers. Rigid motions of such sets have to be defined as maps from and onto a given discrete space. One way to design such motions is to combine continuous rigid motions defined on Euclidean space with a digitization operator. However, digitized rigid motions often no longer satisfy properties of their continuous siblings. Indeed, due to digitization, such transformations do not preserve distances, while bijectivity and point connectivity are generally lost. In the context of 2D discrete spaces, we study digitized rigid motions on the lattices of Gaussian and Eisenstein integers. We characterize bijective digitized rigid motions on the integer lattice, and bijective digitized rotations on the regular hexagonal lattice. Also, we compare the information loss induced by non-bijective digitized rigid motions defined on both lattices. Yet, for practical applications, the relevant information is not global bijectivity, but bijectivity of a digitized rigid motion restricted to a given finite subset of a lattice. We propose two algorithms testing that condition for subsets of the integer lattice, and a third algorithm providing optimal angle intervals that preserve this restricted bijectivity. We then focus on digitized rigid motions on 3D integer lattice. First, we study at a local scale geometric and topological defects induced by digitized rigid motions. Such an analysis consists of generating all the images of a finite digital set under digitized rigid motions. This problem amounts to computing an arrangement of hypersurfaces in a 6D parameter space. The dimensionality and degenerate cases make the problem practically unsolvable for state-of-the-art techniques. We propose an ad hoc solution, which mainly relies on parameter uncoupling, and an algorithm for computing sample points of 3D connected components in an arrangement of second degree polynomials. Finally, we focus on the open problem of determining whether a 3D digitized rotation is bijective or not. In our approach, we explore arithmetic properties of Lipschitz quaternions. This leads to an algorithm which answers whether a given digitized rotation—related to a Lipschitz quaternion—is bijective or not

Cette thèse est consacrée à l'étude de la manipulation et de la coordination robotique à deux bras ayant pour objectif le développement d'une approche unifiée dont différentes tâches seront décrites dans le même formalisme. Afin de fournir un cadre théorique compact et rigoureux, les techniques présentées utilisent les quaternions duaux afin de représenter les différents aspects de la modélisation cinématique ainsi que de la commande.Une nouvelle représentation de la manipulation à deux bras est proposée - l'espace dual des tâches de coopération - laquelle exploite l'algèbre des quaternions duaux afin d'unifier les précédentes approches présentées dans la littérature. La méthode est étendue pour prendre en compte l'ensemble des chaînes cinématiques couplées incluant la simulation d'un manipulateur mobile.Une application originale de l'espace dual des tâches de coopération est développée afin de représenter de manière intuitive les tâches principales impliquées dans une collaboration homme-robot. Plusieurs expérimentations sont réalisées pour valider les techniques proposées. De plus, cette thèse propose une nouvelle classe de tâches d'interaction homme-robot dans laquelle le robot contrôle tout les aspects de la coordination. Ainsi, au-delà du contrôle de son propre bras, le robot contrôle le bras de l'humain par le biais de la stimulation électrique fonctionnelle (FES) dans le cadre d'applications d'interaction robot / personne handicapée.Grâce à cette approche générique développée tout au long de cette thèse, les outils théoriques qui en résultent sont compacts et capables de décrire et de contrôler un large éventail de tâches de manipulations robotiques complexes
This thesis is devoted to the study of robotic two-arm coordination/manipulation from a unified perspective, and conceptually different bimanual tasks are thus described within the same formalism. In order to provide a consistent and compact theory, the techniques presented herein use dual quaternions to represent every single aspect of robot kinematic modeling and control.A novel representation for two-arm manipulation is proposed—the cooperative dual task-space—which exploits the dual quaternion algebra to unify the various approaches found in the literature. The method is further extended to take into account any serially coupled kinematic chain, and a case study is performed using a simulated mobile manipulator. An original application of the cooperative dual task-space is proposed to intuitively represent general human-robot collaboration (HRC) tasks, and several experiments were performed to validate the proposed techniques. Furthermore, the thesis proposes a novel class of HRC taskswherein the robot controls all the coordination aspects; that is, in addition to controlling its own arm, the robot controls the human arm by means of functional electrical stimulation (FES).Thanks to the holistic approach developed throughout the thesis, the resultant theory is compact, uses a small set of mathematical tools, and is capable of describing and controlling a broad range of robot manipulation tasks

This paper deals with studying skeletal animation techniques, and the subsequent design and implementation of skeletal animation extension for the GPUEngine library. The theoretical part describes the techniques of animation, skeletal animation and skinning. The following is an analysis of existing skeletal animation systems. The proposed solution seeks to reduce the data redundancy in the memory while rendering more skeletal models. According to the design a basic skeletal animation system has been implemented. Furthermore, a demonstration application has been created showing the skeletal system's use. The resulting skeletal system can be used in simple 3D applications and can serve as a basis for further works.

La détermination des structures tri-dimensionnelles (3D) des complexes protéiques est cruciale pour l’avancement des recherches sur les processus biologiques qui permettent, par exemple, de comprendre le développement de certaines maladies et, si possible, de les prévenir ou de les traiter. Face à l’intérêt des complexes protéiques pour la recherche, les difficultés et le coût élevé des méthodes expérimentales de détermination des structures 3D des protéines ont encouragé l’utilisation de l’informatique pour développer des outils capables de combler le fossé, comme par exemple les algorithmes d’amarrage protéiques. Le problème de l’amarrage protéique a été étudié depuis plus de 40 ans. Cependant, le développement d’algorithmes d’amarrages précis et efficaces demeure un défi à cause de la taille de l’espace de recherche, de la nature approximée des fonctions de score utilisées, et souvent de la flexibilité inhérente aux structures de protéines à amarrer. Cette thèse présente un algorithme pour l’amarrage rigide des protéines, qui utilise une série de recherches exhaustives rotationnelles au cours desquelles seules les orientations sans clash sont quantifiées par ATTRACT. L’espace rotationnel est représenté par une hyper-sphère à quaternion, qui est systématiquement subdivisée par séparation et évaluation, ce qui permet un élagage efficace des rotations qui donneraient des clashs stériques entre les deux protéines. Les contributions de cette thèse peuvent être décrites en trois parties principales comme suit. 1) L’algorithme appelé EROS-DOCK, qui permet d’amarrer deux protéines. Il a été testé sur 173 complexes du jeu de données “Docking Benchmark”. Selon les critères de qualité CAPRI, EROS-DOCK renvoie typiquement plus de solutions de qualité acceptable ou moyenne que ATTRACT et ZDOCK. 2) L’extension de l’algorithme EROS-DOCK pour permettre d’utiliser les contraintes de distance entre atomes ou entre résidus. Les résultats montrent que le fait d’utiliser une seule contrainte inter-résidus dans chaque interface d’interaction est suffisant pour faire passer de 51 à 121 le nombre de cas présentant une solution dans le top-10, sur 173 cas d’amarrages protéine-protéine. 3) L’extension de EROSDOCK à l’amarrage de complexes trimériques. Ici, la méthode proposée s’appuie sur l’hypothèse selon laquelle chacune des trois interfaces de la solution finale doit être similaire à au moins l’une des interfaces trouvées dans les solutions des amarrages pris deux-à-deux. L’algorithme a été testé sur un benchmark de 11 complexes à 3 protéines. Sept complexes ont obtenu au moins une solution de qualité acceptable dans le top-50 des solutions. À l’avenir, l’algorithme EROS-DOCK pourra encore évoluer en intégrant des fonctions de score améliorées et d’autres types de contraintes. De plus il pourra être utilisé en tant que composant dans des workflows élaborés pour résoudre des problèmes complexes d’assemblage multi-protéiques
Determination of tri-dimensional (3D) structures of protein complexes is crucial to increase research advances on biological processes that help, for instance, to understand the development of diseases and their possible prevention or treatment. The difficulties and high costs of experimental methods to determine protein 3D structures and the importance of protein complexes for research have encouraged the use of computer science for developing tools to help filling this gap, such as protein docking algorithms. The protein docking problem has been studied for over 40 years. However, developing accurate and efficient protein docking algorithms remains a challenging problem due to the size of the search space, the approximate nature of the scoring functions used, and often the inherent flexibility of the protein structures to be docked. This thesis presents an algorithm to rigidly dock proteins using a series of exhaustive 3D branch-and-bound rotational searches in which non-clashing orientations are scored using ATTRACT. The rotational space is represented as a quaternion “π-ball”, which is systematically sub-divided in a “branch-and-bound” manner, allowing efficient pruning of rotations that will give steric clashes. The contribution of this thesis can be described in three main parts as follows. 1) The algorithm called EROS-DOCK to assemble two proteins. It was tested on 173 Docking Benchmark complexes. According to the CAPRI quality criteria, EROS-DOCK typically gives more acceptable or medium quality solutions than ATTRACT and ZDOCK. 2)The extension of the EROS-DOCK algorithm to allow the use of atom-atom or residue-residue distance restraints. The results show that using even just one residue-residue restraint in each interaction interface is sufficient to increase the number of cases with acceptable solutions within the top-10 from 51 to 121 out of 173 pairwise docking cases. Hence, EROS-DOCK offers a new improved search strategy to incorporate experimental data, of which a proof-of-principle using data-driven computational restraints is demonstrated in this thesis, and this might be especially important for multi-body complexes. 3)The extension of the algorithm to dock trimeric complexes. Here, the proposed method is based on the premise that all of the interfaces in a multi-body docking solution should be similar to at least one interface in each of the lists of pairwise docking solutions. The algorithm was tested on a home-made benchmark of 11 three-body cases. Seven complexes obtained at least one acceptable quality solution in the top-50. In future, the EROS-DOCK algorithm can evolve by integrating improved scoring functions and other types of restraints. Moreover, it can be used as a component in elaborate workflows to efficiently solve complex problems of multi-protein assemblies

Dans une première partie de cette thèse, nous donnons des minorations universelles ne dépendant que de la dimension – explicites, de trois invariants globaux des quotients des espaces hyperboliques quaternioniques : leur rayon maximal, leur volume, ainsi que leur caractéristique d’Euler. Nous donnons également une majoration de leur constante de Margulis, montrant que celle-ci décroit au moins comme une puissance négative de la dimension. Dans une seconde partie, nous étudions un réseau remarquable des isométries du plan hyperbolique quaternionique, le groupe modulaire d’Hurwitz. Nous montrons en particulier qu’il est engendré par quatres éléments, et construisons un domaine fondamental pour le sous-groupe des isométries de ce réseau qui stabilisent un point à l’infini
In the first part of this thesis, we derive explicit universal – that is, depending only on the dimension – lower bounds on three global invariants of quaternionic hyperbolic sapces : their maximal radius, their volume, and their Euler caracteristic. We also exhibit an upper bound on their Margulis constant, showing that this last quantity decreases at least like a negative power of the dimension. In the second part, we study a specific lattice of isometries of the quaternionic hyperbolic plane : the Hurwitz modular group. In particular, we show that this group is generated by four elements, and we construct a fundamental domain for the subgroup of isometries of this lattice stabilising a point on the boundary of the quaternionic hyperbolic plane

The theory of regular functions over the quaternions introduced by Gentili and Struppa in 2006, already quite rich, is in continuous development. Despite their diverse peculiarities, regular functions reproduce numerous properties of holomorphic functions of one complex variable. This Thesis is devoted to investigate properties of regular functions defined on the unit ball B of the quaternions H. As it happens in the complex case, this particular subset of H represents a special domain for the class of regular function. It is the simplest example of the most natural set of definition for a regular function, namely of a "symmetric slice domain". Furthermore, on open balls centred at the origin, regular functions are characterized by having a power series expansion, hence they behave very nicely. The first Chapter, starting from the very first definitions, includes all the preliminary results that will be used in the sequel. The second Chapter discusses some properties of the modulus of regular functions, in particular how it is related with the modulus of the "regular conjugate" of a regular function. The main result presented is an analogue of the Borel-Carathéodory Theorem, a tool useful to bound the modulus of a regular function by means of the modulus of its real part. The central part of the Thesis contains geometric theory results. The third Chapter contains the analogue of the Bohr Theorem concerning power series, together with a weaker version, that follows as in the complex case from the Borel-Carathéodory Theorem. In the fourth Chapter we prove a Bloch-Landau type theorem, showing that in some sense the image of a ball under a regular function can not be too much thin. The fifth Chapter is dedicated to Landau-Toeplitz type theorems, that study the possible shapes that the image of a regular function can assume. The last Chapter is devoted to the study of the quaternionic Hardy spaces. We begin by the definition of the spaces H^p(B) and H^{\infty}(B), then we prove some of their basic properties. We introduce in conclusion the Corona Problem in the quaternionic setting, proving a partial statement of the Corona Theorem.

A simplicial cell complex K is the face poset of a regular CW complex W such that the boundary complex of each cell is isomorphic to the boundary complex of a simplex of same dimension. If a topological space X is homeomorphic to W then we say that K is a pseudotriangulation of X. For d 1, a (d + 1)-colored graph is a graph = (V; E) with a proper edge coloring : E ! f0; : : : ; dg. Such a graph is called contracted if (V; E n 1(i)) is connected for each color A contracted graph = (V; E) with an edge coloring : E ! f0; : : : ; dg determines a d-dimensional simplicial cell complex K( ) whose vertices have one to one correspondence with the colors 0; : : : ; d and the facets (d-cells) have one to one correspondence with the vertices in V . If K( ) is a pseudotriangulation of a manifold M then ( ; ) is called a crystallization of M. In [71], Pezzana proved that every connected closed PL manifold admits a crystallization. This thesis addresses many important results of crystallization theory in combinatorial topology. The main contributions in this thesis are the followings. We have introduced the weight of a group which has a presentation with number of relations is at most the number of generators. We have shown that the number of vertices of any crystallization of a connected closed 3-manifold M is at least the weight of the fundamental group of M. This lower bound is sharp for the 3-manifolds RP3, L(3; 1), L(5; 2), S1 S1 S1, S2 S1, S2 S1 and S3=Q8, where Q8 is the quaternion group. Moreover, there is a unique such vertex minimal crystallization in each of these seven cases. We have also constructed crystallizations of L(kq 1; q) with 4(q + k 1) vertices for q 3, k 2 and L(kq +1; q) with 4(q + k) vertices for q 4, k 1. In [22], Casali and Cristofori found similar crystallizations of lens spaces. By a recent result of Swartz [76], our crystallizations of L(kq + 1; q) are vertex minimal when kq + 1 are even. In [47], Gagliardi found presentations of the fundamental group of a manifold M in terms of a crystallization of M. Our construction is the converse of this, namely, given a presentation of the fundamental group of a 3-manifold M, we have constructed a crystallization of M. These results are in Chapter 3. We have de ned the weight of the pair (hS j Ri; R) for a given presentation hS j R of a group, where the number of generators is equal to the number of relations. We present an algorithm to construct crystallizations of 3-manifolds whose fundamental group has a presentation with two generators and two relations. If the weight of (hS j Ri; R) is n then our algorithm constructs all the n-vertex crystallizations which yield (hS j Ri; R). As an application, we have constructed some new crystallization of 3-manifolds. We have generalized our algorithm for presentations with three generators and a certain class of relations. For m 3 and m n k 2, our generalized algorithm gives a 2(2m + 2n + 2k 6 + n2 + k2)-vertex crystallization of the closed connected orientable 3-manifold Mhm; n; ki having fundamental group hx1; x2; x3 j xm1 = xn2 = xk3 = x1x2x3i. These crystallizations are minimal and unique with respect to the given presentations. If `n = 2' or `k 3 and m 4' then our crystallization of Mhm; n; ki is vertex-minimal for all the known cases. These results are in Chapter 4. We have constructed a minimal crystallization of the standard PL K3 surface. The corresponding simplicial cell complex has face vector (5; 10; 230; 335; 134). In combination with known results, this yields minimal crystallizations of all simply connected PL 4-manifolds of \standard" type, i.e., all connected sums of CP2, CP2, S2 S2, and the K3 surface. In particular, we obtain minimal crystallizations of a pair 4-manifolds which are homeomorphic but not PL-homeomorphic. We have also presented an elementary proof of the uniqueness of the 8-vertex crystallization of CP2. These results are in Chapter 5. For any crystallization ( ; ) the number f1(K( )) of 1-simplices in K( ) is at least d+1 . It is easy to see that f1(K( )) = d+1 if and only if (V; 1(A)) is connected for each d 2 2 1)-set A called simple. All the crystallization in Chapter 5 (. Such a crystallization is are simple. Let ( ; ) be a crystallization of M, where = (V; E) and : E ! f0; : : : ; dg. We say that ( ; ) is semi-simple if (V; 1(A)) has m + 1 connected components for each (d 1)-set A, where m is the rank of the fundamental group of M. Let ( ; ) be a connected (d +1)-regular (d +1)-colored graph, where = (V; E) and : E ! f0; : : : ; dg. An embedding i : ,! S of into a closed surface S is called regular if there exists a cyclic permutation ("0; "1; : : : ; "d) (of the color set) such that the boundary of each face of i( ) is a bi-color cycle with colors "j; "j+1 for some j (addition is modulo d+1). Then the regular genus of ( ; ) is the least genus (resp., half of genus) of the orientable (resp., non-orientable) surface into which embeds regularly. The regular genus of a closed connected PL 4-manifold M is the minimum regular genus of its crystallizations. For a closed connected PL 4-manifold M, we have provided the following: (i) a lower bound for the regular genus of M and (ii) a lower bound of the number of vertices of any crystallization of M. We have proved that all PL 4-manifolds admitting semi-simple crystallizations, attain our bounds. We have also characterized the class of PL 4-manifolds which admit semi-simple crystallizations. These results are in Chapter 6.

A simplicial cell complex K is the face poset of a regular CW complex W such that the boundary complex of each cell is isomorphic to the boundary complex of a simplex of same dimension. If a topological space X is homeomorphic to W then we say that K is a pseudotriangulation of X. For d 1, a (d + 1)-colored graph is a graph = (V; E) with a proper edge coloring : E ! f0; : : : ; dg. Such a graph is called contracted if (V; E n 1(i)) is connected for each color A contracted graph = (V; E) with an edge coloring : E ! f0; : : : ; dg determines a d-dimensional simplicial cell complex K( ) whose vertices have one to one correspondence with the colors 0; : : : ; d and the facets (d-cells) have one to one correspondence with the vertices in V . If K( ) is a pseudotriangulation of a manifold M then ( ; ) is called a crystallization of M. In [71], Pezzana proved that every connected closed PL manifold admits a crystallization. This thesis addresses many important results of crystallization theory in combinatorial topology. The main contributions in this thesis are the followings. We have introduced the weight of a group which has a presentation with number of relations is at most the number of generators. We have shown that the number of vertices of any crystallization of a connected closed 3-manifold M is at least the weight of the fundamental group of M. This lower bound is sharp for the 3-manifolds RP3, L(3; 1), L(5; 2), S1 S1 S1, S2 S1, S2 S1 and S3=Q8, where Q8 is the quaternion group. Moreover, there is a unique such vertex minimal crystallization in each of these seven cases. We have also constructed crystallizations of L(kq 1; q) with 4(q + k 1) vertices for q 3, k 2 and L(kq +1; q) with 4(q + k) vertices for q 4, k 1. In [22], Casali and Cristofori found similar crystallizations of lens spaces. By a recent result of Swartz [76], our crystallizations of L(kq + 1; q) are vertex minimal when kq + 1 are even. In [47], Gagliardi found presentations of the fundamental group of a manifold M in terms of a crystallization of M. Our construction is the converse of this, namely, given a presentation of the fundamental group of a 3-manifold M, we have constructed a crystallization of M. These results are in Chapter 3. We have de ned the weight of the pair (hS j Ri; R) for a given presentation hS j R of a group, where the number of generators is equal to the number of relations. We present an algorithm to construct crystallizations of 3-manifolds whose fundamental group has a presentation with two generators and two relations. If the weight of (hS j Ri; R) is n then our algorithm constructs all the n-vertex crystallizations which yield (hS j Ri; R). As an application, we have constructed some new crystallization of 3-manifolds. We have generalized our algorithm for presentations with three generators and a certain class of relations. For m 3 and m n k 2, our generalized algorithm gives a 2(2m + 2n + 2k 6 + n2 + k2)-vertex crystallization of the closed connected orientable 3-manifold Mhm; n; ki having fundamental group hx1; x2; x3 j xm1 = xn2 = xk3 = x1x2x3i. These crystallizations are minimal and unique with respect to the given presentations. If `n = 2' or `k 3 and m 4' then our crystallization of Mhm; n; ki is vertex-minimal for all the known cases. These results are in Chapter 4. We have constructed a minimal crystallization of the standard PL K3 surface. The corresponding simplicial cell complex has face vector (5; 10; 230; 335; 134). In combination with known results, this yields minimal crystallizations of all simply connected PL 4-manifolds of \standard" type, i.e., all connected sums of CP2, CP2, S2 S2, and the K3 surface. In particular, we obtain minimal crystallizations of a pair 4-manifolds which are homeomorphic but not PL-homeomorphic. We have also presented an elementary proof of the uniqueness of the 8-vertex crystallization of CP2. These results are in Chapter 5. For any crystallization ( ; ) the number f1(K( )) of 1-simplices in K( ) is at least d+1 . It is easy to see that f1(K( )) = d+1 if and only if (V; 1(A)) is connected for each d 2 2 1)-set A called simple. All the crystallization in Chapter 5 (. Such a crystallization is are simple. Let ( ; ) be a crystallization of M, where = (V; E) and : E ! f0; : : : ; dg. We say that ( ; ) is semi-simple if (V; 1(A)) has m + 1 connected components for each (d 1)-set A, where m is the rank of the fundamental group of M. Let ( ; ) be a connected (d +1)-regular (d +1)-colored graph, where = (V; E) and : E ! f0; : : : ; dg. An embedding i : ,! S of into a closed surface S is called regular if there exists a cyclic permutation ("0; "1; : : : ; "d) (of the color set) such that the boundary of each face of i( ) is a bi-color cycle with colors "j; "j+1 for some j (addition is modulo d+1). Then the regular genus of ( ; ) is the least genus (resp., half of genus) of the orientable (resp., non-orientable) surface into which embeds regularly. The regular genus of a closed connected PL 4-manifold M is the minimum regular genus of its crystallizations. For a closed connected PL 4-manifold M, we have provided the following: (i) a lower bound for the regular genus of M and (ii) a lower bound of the number of vertices of any crystallization of M. We have proved that all PL 4-manifolds admitting semi-simple crystallizations, attain our bounds. We have also characterized the class of PL 4-manifolds which admit semi-simple crystallizations. These results are in Chapter 6.