Fichiers

Résumé

Despite recent technological improvements of immersive technologies, Virtual Reality suffers from severe intrinsic limitations, in particular the immateriality of the visible 3D environment. Typically, any simulation and manipulation in a cluttered environment would ideally require providing feedback of collisions to every body parts (arms, legs, trunk, etc.) and not only to the hands as has been originally explored with haptic feedback. This thesis addresses these limitations by relying on a cross modal perception and cognitive approach instead of haptic or force feedback. We base our design on scientific knowledge of bodily self-consciousness and embodiment. It is known that the instantaneous experience of embodiment emerges from the coherent multisensory integration of bodily signals taking place in the brain, and that altering this mechanism can temporarily change how one perceives properties of their own body. This mechanism is at stake during a VR simulation, and this thesis explores the new venues of interaction design based on these fundamental scientific findings about the embodied self. In particular, we explore the use of third person perspective (3PP) instead of permanently offering the traditional first person perspective (1PP), and we manipulate the user-avatar motor mapping to achieve a broader range of interactions while maintaining embodiment. We are guided by two principles, to explore the extent to which we can enhance VR interaction through the manipulation of bodily aspects, and to identify the extent to which a given manipulation affects the embodiment of a virtual body. Our results provide new evidence supporting strong embodiment of a virtual body even when viewed from 3PP, and in particular that voluntarily alternating point of view between 1PP and 3PP is not detrimental to the experience of ownership over the virtual body. Moreover, detailed analysis of movement quality show highly similar reaching behavior in both perspective conditions, and only obvious advantages or disadvantages of each perspective depending on the situation (e.g. occlusion of target by the body in 3PP, limited field of view in 1PP). We also show that subjects are insensitive to visuo-proprioceptive movement distortions when the nature of the distortion was not made explicit, and that subjects are biased toward self-attributing distorted movements that make the task easier.

Détails

Actions

Aperçu