Ce travail porte sur l'interaction avec des objets virtuels 3D, et plus particulièrement leur manipulation : sélection, translation et rotation. Les plates-formes les plus utilisées aujourd'hui pour ce type de tâche (ordinateur personnel et appareil mobile) ont été conçues avant tout pour l'interaction 2D, et sont donc peu adaptées à la manipulation 3D. Il existe pourtant une alternative plus efficace : les systèmes de réalité virtuelle. Cependant, les systèmes actuellement disponibles restent trop souvent encombrants, onéreux et/ou intrusifs pour l'utilisateur, et demeurent sous-utilisés pour cette raison. Dans cette thèse, nous étudions de nouvelles approches pour rendre ce type d'interaction moins contraignant, tout en conservant une bonne efficacité de manipulation. Les principaux axes explorés sont la portabilité de l'interface, sa non-intrusivité, et l'utilisation de dispositifs plus largement accessibles au grand public. Nous proposons une première approche visant à répondre simultanément aux critères ci-dessus. Celle-ci consiste à combiner un appareil mobile classique (pour sa portabilité et son accessibilité) avec des objets tangibles passifs (pour une manipulation 3D efficace et non-intrusive). Cette approche présente toutefois encore certaines contraintes : elle est constituée de multiples éléments qui doivent être transportés ensemble, et elle ne se prête pas à l'utilisation de dispositifs à retour de force. Nous introduisons donc dans un deuxième temps une nouvelle technique, appelée «découplage». Basée sur le principe du «pseudo-haptique», celle-ci permet à l'utilisateur de percevoir des forces virtuelles à travers la seule modalité visuelle. Il devient alors possible de se passer complètement de dispositifs haptiques, et des contraintes qui leur sont associées. Nous nous intéressons ensuite à une approche entièrement intégrée, visant à améliorer la portabilité par-rapport à une interface faite d'éléments séparés. Cette approche consiste à se servir des déplacements de l'appareil mobile lui-même dans l'espace réel pour manipuler des objets 3D affichés sur son propre écran. Cette configuration «localement couplée» présente cependant plusieurs particularités qui rendent la manipulation plus complexe. Nous examinons donc en détail les différentes questions que pose cette configuration spécifique. Nous proposons enfin une dernière approche, appelée le «volume tangible», visant à retrouver la simplicité de la manipulation par objets tangibles mais dans un unique dispositif intégré. Nous décrivons une première implémentation de ce dispositif, et discutons de sa faisabilité technique ainsi que de l'accessibilité de cette approche à court et moyen terme. / This work focuses on 3D interaction with virtual objects, more specifically on 3D manipulation: selection, translation, and rotation. The most commonly-used interaction platforms (personal computer, mobile device) were designed for 2D interaction and thus are not well suited to 3D manipulation. There is a more efficient alternative for this task: virtual reality. However, current virtual reality systems are too often bulky, expensive and/or intrusive to the user, and for these reasons remain underused. In this thesis, we propose and study new solutions to make this form of interaction more convenient without reducing its effectiveness. The main objectives we aim for are interface portability, non-intrusiveness, and the use of readily-available hardware. We propose a first approach to simultaneously meet the above criteria. It consists in combining a standard mobile device (for its portability and widespread availability) with passive tangible objects (for efficient 3D manipulation in a non-intrusive manner). Yet, this approach still has drawbacks: it is made of multiple parts that must always be carried together, and is not suitable for adding force-feedback using haptic devices. We thus introduce a new technique, called “decoupling”, that allows the user to perceive virtual forces through the visual modality alone. Based on the concept of “pseudo-haptics”, this technique makes it possible to provide force feedback without the constraints associated with actual haptic devices. We then study a different approach intended to be fully integrated, enhancing the portability aspect compared to an interface made of multiple parts. This approach consists in using the motion of the mobile device itself to manipulate 3D objects displayed on its own screen. However, this “locally-coupled” configuration presents several unique characteristics that make manipulation more complex. We thus investigate the questions raised by this specific configuration. Finally, we introduce a last approach, called the “tangible volume”, aimed at regaining the same directness of manipulation as with tangible objects but in a single integrated device. We describe an early implementation of such a device, and discuss its technical feasibility as well as its potential accessibility to end users in the short and medium term.

No description available.

[pt] Flexibilidade e liberdade são sempre desejados em ambientes de realidade virtual. Dispositivos de entrada tradicionais, como mouse ou teclado, dificultam as interações entre o usuário e o ambiente virtual. Para melhorar a interação em termos qualitativos em um ambiente virtual, a interação deve ser tão natural quanto possível, por isso, gestos com a mão se tornaram um meio popular para a interação humanocomputador. O desenvolvimento de dispositivos de imersão como os capacetes trouxeram a necessidade de uma nova forma de interação e um desafio para os desenvolvedores. O reconhecimento de gestos da mão usando sinais eletromiográficos (EMG) tem chamado a atenção devido ao surgimento de dispositivos mais baratos que conseguem gravar dados EMG precisos. Um dos dispositivos mais destacados nessa área é o bracelete Myo, equipado com oito sensores EMG e uma unidade de medição inercial (IMU). O objetivo deste trabalho é avaliar a usabilidade do bracelete Myo como um dispositivo de seleção e manipulação de objetos 3D em ambientes de realidade virtual, visando melhorar a experiência do usuário, aproveitando a possibilidade de medir a força aplicada a um gesto assim como de usar as vibrações do Myo como sistema de feedback. Este estudo pretende responder à seguinte pergunta: O bracelete Myo tem alto grau de usabilidade para a seleção/manipulação de objetos 3D em Ambientes de Realidade Virtual? Para atingir esse objetivo, foram propostas quatro subquestões para orientar essa pesquisa: I) Quais recursos do Myo podem ser usadas em Ambientes de Realidade Virtual (VRE)? II) Quais são as limitações do bracelete Myo? III) É possível realizar tarefas de seleção e manipulação usando o bracelete Myo? IV) Como o uso do bracelete Myo pode enriquecer as tarefas de seleção e manipulação? Para responder às duas primeiras subquestões foi realizada uma revisão da literatura que compreende a tecnologia do Myo, vantagens e limitações, e os trabalhos relacionados. Além disso, inclui conceitos básicos sobre Interações em VRE. Para responder às duas últimas subquestões foram propostas duas técnicas de seleção/manipulação usando o Myo e foram testadas com os usuários e os resultados foram comparados, avaliando sua usabilidade. / [en] Flexibility and freedom are always desired in virtual reality environments. Traditional inputs, like mouse or keyboard, hamper the interactions between the user and the virtual environment. To improve the interaction in qualitative terms in a virtual environment, the interaction must be as natural as possible, and because of that, hand gestures have become a popular means to the human-computer interaction. The development of immersion devices like head-mounted displays brought the need for a new way of interaction and a challenge to developers. Hand gestures recognition using electromyography signals (EMG) has increased the attention because the rise of cheaper wearable devices that can record accurate EMG data. One of the outstanding devices in this area is Myo armband, equipped with eight EMG sensors and a nineaxis inertial measurement unit (IMU). The objective of this work is to evaluate the usability of the Myo armband as a device for selection and manipulation of 3D objects in virtual reality environments, aiming to improve the user experience, taking advantage of the possibility to measure the force applied to a gesture and to use Myo vibrations as a feedback system. This study aims to answer the following question: Has Myo armband high grade of usability for selection/manipulation of 3D objects in Virtual Reality Environments? And to achieve that purpose, four sub-questions were proposed to guide this research: I) Which resources of Myo can be used in Virtual Reality Environments (VRE)? II) What are the limitations of the Myo armband? III) Can selection and manipulation tasks be performed using Myo armband? IV) How can Myo armband enrich the selection and manipulation tasks? To answer to the first two sub-questions, we conducted a literature review that covers Myo technology, its advantages and limitations, and related works. Also, it includes basic concepts about Interactions in VRE. To answer to the last two sub-questions, we proposed two selection/manipulation techniques using Myo, which were tested with users and the results were compared, evaluating their usability.