Motor skills, such as experts' skills are hard to learn. It is not only due
to musclar strength but due to lack of efficient instruction method.
Motor skill is difficlut to explain in words. Most of the case, we tend to say
"do it in this way" in demonstration but learners may not understand.
In visual instruction, the rough shape of a motion could be taught but still
it is impossible to teach how to reproduce the motion. For human, crutial information
to it is when, which muscle to activate. Still we cannot extract this information
but it may be possible to extract good hints to it from experts' motion.
We performed motion capture experiment of a kneading motion in ceramic art (in
fact, kneading a cray is very similar to mixing process of bakery, etc.). We
may think that expert's motion is hard to analyse. However, expert's motion
is highly periodic and no obvious critical point in the kneading, it was possible.
We should note that professionals must make highly similar products as ordered.
On the other hand, novice's motions are disorgnised and difficult to analyse.
(In this respect, baby's motion is not necessarily easy target).
We compared motion of an expert (10 years of professional experience) and experienced
person (1 year of experience at a hobby school). Within highly periodic (i.e.
convergence is seen in the phase space) trajectories, we found differentiation
of motion patterns and phases between joints. The former is seen as differentiation
of torso's oscillation pattern from single pendulum (experienced) to double
pendulum (exptert), where the latter pivotting its motion at the hip. This may
correspond to differentiation of DOFs. In the latter, phase difference is stabilised
in the experts; push the cray at the pushing back, rather than pushing down.
We think these features may correspond to hierarchical orgnisation of phase
space of human movement dynamics but it is still to be examined.
We also begun a study of pitching motion. This is an explosive motion but what
we have seen in the kneading motion may helpful. We are also planning to make
a machine which can perform explotive movement based on the resuts we obtain
in this study.
publications:
Mamiko Abe, Tomoyuki Yamamoto and Tsutomu Fujinami
A Dynamical Analysis of Kneading Using a Motion Capture Device
proceedings of Epigenetic Robotics 2003, pp 41-48