Game Development Reference
In-Depth Information
Proxy
Proxy
Target
Target
Figure 16.6 Principle of Virtual Proxy
Mendoza & Laugier, 2000), but they are restricted to the calculation of the position
of a single point with its three translation parameters. We would also like to mention
that this system of constraints can be adapted to the interactions with deformable bod-
ies (Mendoza & Laugier, 2001). Position of the proxy is often calculated by a static
method which makes it possible to find a position close to the interface position and
compatible with the obstacles of the environment, using the minimisation technique.
To sum up, the position of the avatar is controlled through positioning, which makes
us take k s
0 in figure 16.4. However, another more general approach is to
control the proxy by forces by simulating it as a mechanical object (which follows static
or dynamic equations). This physical approach lets us generalise the proxy approach
to the manipulation of any type of complex physical bodies. The initial systems sim-
ulated the proxy as a rigid body (McNeely et al., 1999), and the physical approach
proved its usefulness in processing interactions (Zhuang & Canny, 2000). A dynamic
solution is also useful for proxies formed by an articulation of rigid bodies. Several
works focus on controlling the end of an articulated arm (Ruspini & Khatib, 1998;
Constantinescu et al., 2003). However, by associating the degrees of freedom of the
haptic interface with all or one sub-unit of degrees of freedom of the avatar, the total
state of the articulated rigid proxy can be controlled by the dynamics (Meseure et al.,
2004). The differences between the degrees of freedom of the interface and that of the
proxy create forces and momentums, necessary for simulation and haptic feedback.
However, we must also mention that certain studies criticise the methods based on the
dynamics of creating inertia in the movements of the proxy. We have already men-
the position of the avatar is then controlled by a 2 degree differential equation. The
damping parameter that controls this movement is sometimes difficult to calibrate.
A very high value reduces the speed of the proxy and makes it less reactive, whereas a
small value makes it swing around the target position.
=
1 and γ s
=
16.2.3.3 Benefits of virtual proxy
Other than the fact that the virtual proxy method helps the impedance devices to adhere
to the constraints of the environment, there are some other advantages that we would
like to discuss. First, we can observe that the forces returned to the simulation and
the device need not be identical at all. Naturally, they should be absolute. However,
this flexibility can be used when the forces of the simulation and those of the device
are not subject to the same constraints of scale. We thus operate a decoupling, as