Why study statically stable locomotion?
The current vogue in studying robot locomotion has been
dynamic locomotion: running, hopping etc. People take statically
stable locomotion for granted. It is arguably the most prevelant
form of transportation yet has almost no general framework
for basic research.
Locomotion in the broadest sense of the word (moving from place to
place) is seen everywhere. Not only in the obvious places like
cars, trains, horses walking, a baby crawling, earthworms digging,
but also in copy machines
(paper moving from one place to another), people eating (food moving from
the mouth to the stomach),
All of these examples are also statically stable. In some of these
examples, the duality between manipulation and locomotion becomes
apparent. In other words, a different way to view a person walking
on a globe is say the person is manipulating the globe with his feet.
The only important point is that the two objects are moving relative
to each other.
How can something move
from point A to point B? We can look for many specific examples in nature
but what does that tell us about locomotion in general? Looking at
locomotion in general would allow us to answer the following questions:
- How can we explore other ways of locomotion that may be more useful
or efficient that we have not already seen?
- How can we compare different classes of locomotion?
The first one has obvious value in furthering useful technology (assuming
that novel useful modes of locomotion are discovered). The second one
is done today by people without consciously thinking about it. For example,
if I want to transport 500 reams of paper, I think I need a truck. I wouldn't
use a motorcycle. It could be done in a car with many trips, but a truck
would be best. I have compared the class of trucks with that of cars and
motorcycles
etc.
If we have a broader understanding of locomotion, and we have for
example the question, what would be the best way of moving a
piece of paper from point A to point B, we have a frame work for
exploring the options.
One such framework is functional taxonomy.
Mark Yim, PhD
(mark@flamingo.stanford.edu )