Physically-rendered Remote Collaborators. When a user
reaches out and places a hand on the table, its shape is physically
rendered at the remote site. This conveys presence and
enables remote pointing and manipulation. Figure 10 shows
an example application, where a remote expert supports a local
user.
Unidirectional Shape Output
The shape output link described in the above scenarios can
be uni- or bidirectional. An ideal bidirectional configuration,
where both sites have a shape display, may not always be feasible
due to size, cost and required infrastructure. A unidirectional
link is still advantageous, as it allows remote participants
to be more present with physically rendered body representations.
It makes it possible to also capture and transmit
physical objects to the remote site.
Collaborative 3D Modeling
A specific example of a telepresence workspace with a shared
digital model is collaborative 3D modeling. In this case, users
can deform the model, with any change reflected through the
model on the connected site (see Figure 11). However, rendering
the remote person’s hands may obstruct the changes
to the model. Therefore, our application allows switching to
rendering the model geometry without the user’s hands.
Another consideration is the potential conflict that may occur
when two users try to simultaneously deform a model. We
therefore use a turn-taking protocol for manipulation. When
the system detects deformation, control is passed to the active
user. The changes will be reflected on the remote connected
model, while concurrent remote input is ignored. When the
user stops modifying the model, the collaborator at the remote
site may take control by starting a new deformation. Passing
control between users can be applied to the whole model or
to each pin individually.
Shared Presence: Connected Membranes
This application conveys the presence of remote participants
through a shared digital membrane that uses two connected
shape displays, each rendering the inverse shape of the other
(see Figure 12). When a users pushes down a pin, the corresponding
remote pin pops out. This is conceptually similar
Figure 11. Collaborative 3D modeling of two linked landscape models.
Figure 12. Connected Membrane: When a user pushes into the membrane
on one site, the inverse shape appears at the other.
to a physical link between the pins. We implemented two
modes:
Pin screen. Pins remain in their position after users push or
pull them. This leaves behind a trace of past deformations,
until it is erased.
Elastic. Pins always spring back to a neutral flat state when
they are not being deformed. This mode conveys whether
someone on the other end is pushing at that very moment.