Biomedical Engineering Reference
In-Depth Information
say, pushing the left arrow key. It also differentially tenses muscles in the left and
right legs to indicate leftward movement and engages the posture that one might
adopt when leaning into a leftward turn to counteract the impending inertia. Spatial
knowledge acquisition may or may not be helped, however, by other physical actions
such as waving in a defined pattern to call-up a menu or even flicking one's wrist
upward to jump—an action typically performedwith one's legs. Indeed, this is an area
ripe for further research. It is well known that spatial congruence between controls
and actions can facilitate users' behavior (e.g., [
79
,
93
]). In the previous example,
flicking one's wrist upward to jump would certainly be preferable to some other
arbitrary motion—say, rotating one's wrist leftward—given that the upward motion
is congruent with an upward jumping action. Whereas the research is unequivocal in
favoring spatially mapped controls relative to arbitrary controls, it is unclear whether
well-mapped controls are as beneficial to users as the actual motions they replace. In
other words, flicking one's wrist upward to jump is certainly better than rotating it
leftward, but is it as good as actually jumping? From a perceptual standpoint, it seems
unlikely. Users of these systems are still largely stationary or confined to a small area
with a fixed toward-the-screen orientation, so some deficits in spatial knowledge are
still likely.
Uses for Desktop VEs
While navigation in a desktop VE may be the most limiting from a multi-sensory
perspective, there are compelling reasons to use desktop systems, and perhaps even
several uses for which it is ideal. Desktop VEs are clearly superior to CAVEs and
HMDs on many non-spatial dimensions, including system cost, convenience, and
availability to a wide audience. Thus, one might be willing to trade off the fidelity of
spatial perception and learning in favor of other factors. Additionally, many computer
simulations require little navigation or environmental knowledge. Specifically, when
simulating a single virtual object or a small number of objects—a newproduct design,
for example—it may more intuitive to present the VE as a desktop simulation. Users
can sit at a desk or table and interact with the virtual object on that table in a similar
manner to which they might sit and interact with a physical object on the table. If a
VE does not require users to travel, or it is relatively unimportant for them to stay
oriented or update their movements, then a desktop VE may be best suited to that
particular task.
Conversely, for simulations in which navigation is crucial or in which accurate
spatial perception and learning is desired, a desktop VE is likely to be a poor choice.
For example, a training simulation designed to teach elite police forces to raid a
building would be best implemented in a system that enabled users to physically
move, crouch, look around, and remain keenly aware of their position and heading.
