Alois Knoll
Autonomous Robots, No.12, 2002
Summary
Since the 1920, when Karel Capek published “R.U.R.
– Rossum’s Universal Robots”, man-made robot of human-like shape inspired
people’s imagination, making robots viewed as humans partners, as friends or
potential enemy. Reality so far has actually being very far from fiction and
Robotics and AI are not yet at the point intended, technologies are not at a
that point and therefore the vision of autonomous robots has shifted more at
the basics, trying to obtain improvements in perception, communication and
manipulation and integrating these characteristics in order to finally obtain a
humanoid.
Originally the first
humanoids are dated back in the 70’s with the “Wabots”, when in Waseda University (Japan), humanoids where build
mainly with the intent of doing practical feasibility studies in mechanical
engineering, at that time computing technology was not at the point of
guaranteeing any kind of possible self-control or self-guidance. In the 80’s
the research in humanoids slowed down due to the fact that no industrial
application was found, today possible applications can be imagined easily in
service, prosthetics, education, entertainment and so on. Of course humanoids
open different fields of research which are related to the fact that it’s a
matter of Cognitive Science, even
studies of possible inter-humanoid relations appear to be interesting an
challenging.
Different are the
challenges in humanoids research, first of all the humanoids brain, which will
be strictly dependent on its creator, the structure of it’s brain can be
comparable with the one of human at birth.
Adult human brain operated
in two different ways (Christaller ,. Al, 1999), fast parallel processing and
slower sequent processing, not always it’s easy to fully distinguish the two
and often repetitive sequent activities become fast parallel activities.
Therefore a humanoid will have to reach this level of complexity in thought and
action.
A next important step is
the capacity of the humanoid of generating categories, attempts have been done
in the field based on sensory stimuli, but still the subsequent step of self
building a tailored non trivial reasoning has yet to be done. Also aspects
which are usually considered as more achievable, such as planning and language,
my not be that easy, for example there are studies which consider language
genetically related (Chomsky).
Interaction with the
humans of course is another important point which has to be kept into consideration,
in fact also aspect as body-language are extremely important, specially in
regard of “teaching by demonstration”, which still today hasn’t been achieved
due to the fact that monomodal fixed cameras are mostly used, so there are no
redundant multidimensional views available and the instruction is not in the
form of a dialog between the instructor (human) and the robot. Better
integration of certain class of actions must be obtained, such as vision,
speech and touch, so that the concept of imitation would be more affordable,
encompassing both motor and cognitive aspects.
Interaction with human
covers also physical aspects, there is evidence of connection between human
motor skill an cognitive development (Lakoff, 1990). Therefore there several aspects
which have to be studies and analyzed: whole body cross and fine motion,
body-centered behaviors and fine motion, learning arm motion, grasping and
sensorimotor and posture stability, planning of movement and collision,
locomotion and simulation of body limb dynamics for designer and as a mental
tool for the humanoid.
Progress therefore is on
the way and must be performed in the following three sectors: mechatronics,
sensor design and evolutionary hardware.
The authors conclude
discussing the future in humanoids’ research, asking for benchmarking and international
competitions and contests (such as the “Robocup”), which would ensure
improvements in the field. If a standard platform will be found, and for a
reasonable price, then humanoid would enter in the sphere of interest of many
people and investors, this will give a complete redesign of actuator and will
require special power supply systems, where technologies can be taken easily
from other sectors (such as the automotive one).
Key
Concepts
Humanoids, Human Robot
Interaction
No comments:
Post a Comment