MOVEMENT IS A fundamental characteristic of living
systems (see Figure 1). Plants and animals must move to
survive. Animals are distinguished from plants in that
they have to explore the world to feed. The carnivorous
plant remains at a fixed position to catch the imprudent
insect. Plants must make use of self-centered motions.
At the same time the cheetah goes out looking for food.
Feeding is a paragon of action. Any action in the
physical world requires self-centered movements,
exploration movements, or a combination of both. By
analogy, a manipulator robot makes use of self-centered
motions, a mobile robot moves to explore the world,
and a humanoid robot combines both types of motions.
Actions take place in the physical
space. Motions originate in the motor
control space. Robots—as any living
system—access the physical space only
indirectly through sensors and motors.
Robot motion planning and control explore the relationship between physical, sensory, and motor spaces; the
three spaces that are the foundations
32 How to translate actions
expressed in the physical space into
a motion expressed in motor coordinates? This is the fundamental robotics issue of inversion.
In life sciences, it is recognized that
optimality principles in sensorimotor
control explain quite well empirical
observations, or at least better than
40 The idea of expressing robot actions as motions to be optimized was first developed in robotics in the 1970s with the seminal work
41 It is now well developed
in classical robot control,
37 and also
along new paradigms jointly developed in multidisciplinary approaches.
35 Motion optimization appears to
be a natural principle for action selection. However, as we explained in the
24 optimality equations are intractable most of the time
and numerical optimization is notoriously slow in practice. The article aims
Robots move to act. While actions operate
in a physical space, motions begin in
a motor control space. So how do robots
express actions in terms of motions?
BY JEAN-PAUL LAUMOND, NICOLAS MANSARD,
AND JEAN BERNARD LASSERRE
˽ For robots and living beings, the link
between actions expressed in the physical
space and motions originated in the motor
space, turns to geometry in general
and, in particular, to linear algebra. In
life science the application of optimality
principles in sensorimotor control
unravels empirical observations. The idea
to express robot actions as motions to be
optimized has been developed in robotics
since the 1970s.
˽ Among all possible motions performing
a given action, optimization algorithms
tend to choose the best motion according
to a given performance criterion. More
than that, they also allow the realization
of secondary actions.
˽ Optimal motions are action signatures.
How to reveal what optimality criterion
underlies a given action? The question
opens challenging issues to inverse