Section: New Results

Collaborative Virtual Environments

Exploration of large-scale 3D Virtual Environments (VEs) is often difficult because of lack of familiarity with complex virtual worlds, lack of spatial information that can be offered to users and lack of sensory details compared to the exploration of real environments. To address this problem, we presented a set of metaphors for assisting users in collaborative navigation to perform common exploration tasks in shared collaborative virtual environments [39] , [57] . Our propositions consist in three guiding techniques in the form of navigation aids to enable one or several users to help one main user (exploring user) to explore the VE efficiently. These three techniques consist in drawing directional arrows, lighting up path to follow, and orienting a compass to show a direction to the exploring user. Our experimental results could show that although the directional arrows and compass surpassed the light source in a navigation task, these three techniques are completely appropriate for guiding a user in 3D complex VEs.

When a user is fully immersed within a Virtual Environments (VE) through a large immersive display system, his feeling of presence can be altered because of disturbing interactions with his physical environment, such as collision with hardware parts of the system or loss of tracking. This alteration can be avoided by taking into account the physical features of the user and to embed them in the VE. In [20] we could present how we use the Immersive Interactive Virtual Cabin (IIVC) model to obtain such a virtual representation of the physical environment of the user and we illustrated how it can be used to guide efficiently a user for a navigation task in a VE. We also presented how we can add 3D representations of 2D interaction tools in order to cope with asymmetrical collaborative configurations, providing 3D cues for users in order to understand the actions of the other users even if they are not fully immersed in the shared virtual environment. Last, we explained how we could enhance 3D interaction and collaboration by embedding a symbolic 3D representation of the user that would give 3D information about his posture.

We introduced a collaborative virtual environment usable to conduct ergonomic design sessions, involving the worker, ergonomists and engineers [41] . We focused particularly on the representation of the ergonomic evaluation and the interaction between an ergonomist and the main user (worker). An ergonomic evaluation of the postures was presented. An interaction architecture between the main user and an ergonomist based on the combination of animation modes of two linked manikins was also proposed. Preliminary results and future developments of the CVE (e.g. additional ergonomic evaluation tools, graphical enhancement, interaction enhancement) were then presented.

In Collaborative Virtual Environments for Training (CVET), a group can learn and practice the completion of a task as a team using all the assets provided by Virtual Reality. We presented a novel mechanism that allows real and virtual humans to dynamically exchange the control of their embodiment in virtual environments [42] . Such a mechanism raises two important issues: the possibility of dynamic embodiment exchanges between real humans and virtual humans and the continuity of actions of the team members after an exchange. To address these issues we introduce a new entity, the Perceptive Puppet that abstracts real and virtual humans into one common entity containing its own knowledge.

In addition, in CVET different roles need to be played by actors, i.e. virtual agents or users. In order to abstract an actor from its embodiment in the virtual world, we have introduced a new entity, the Shell [37] . Through the Shell, users and virtual agents are able to collaborate in the same manner during the training. In addition to the embodiment's control, the Shell gathers and carries knowledge and provides interaction inputs. This knowledge and those inputs can be accessed and used homogeneously by both users and virtual agents to help them to perform the procedure.