| |
Abstract:
In this paper, we will present a number of empirical and
modeling studies that focus on the complexity of the information
the nervous system encodes concerning the mechanical periphery
and external loads. We will first summarize our earlier work that
suggests that patterns of coarticulation may arise as a
consequence of muscle properties and dynamics rather than
characteristics of the neural code. We will also present evidence
showing that in speech production, the nervous system is capable
of taking account of forces in motion planning. We report a study
that focuses on the forces on orofacial articulators that arise
during locomotion. Our results indicate a pattern of predictive
compensation for loads on the jaw suggesting that the nervous
system may adjust control signals to deal with the force
environment when planning speech movements. However, the
adjustment of control signals in anticipation of loads does not
necessitate direct specification of forces in movement planning.
A series of simulation studies, in this case using a model of
multi-joint arm movement, show that a simple equilibrium point
control scheme is capable of adaptive load compensation using a
direct adjustment of control signals based on positional error.
This scheme is demonstrated in the context of velocity dependent
force fields and forces arising due to joint interaction torques.
A clear advantage of this kind of positional control over force
control schemes is that it permits adaptation to load without the
need for coordinate transformation or inverse dynamics
calculations.
|
