[Avery Andrews 930428.2028]
One of my coordinative structures programs is in a semi-usable state,
for C-hackers. It's a sort of lip closure simulation, wherein
the two `lips' have equilibrium points with zero passive restoring
points, & the control system produces a force on each lip which
tends to effect closure. The control system compares actual
separation with the reference separation, with the reference
separation varying as a square pulse. You can freeze either
lip and watch the other go further, and also watch it not have
much effect on the separation achieved. At the moment the parameters
are hard-coded, so it has to be recompiled when they are changed,
but they should be readable from a file in a day or two.
Avery.Andrews@anu.edu.au
From Tom Bourbon (930428.0950)
[Avery Andrews 930428.2028]
One of my coordinative structures programs is in a semi-usable state,
for C-hackers. It's a sort of lip closure simulation, wherein
the two `lips' have equilibrium points with zero passive restoring
points, & the control system produces a force on each lip which
tends to effect closure. The control system compares actual
separation with the reference separation, with the reference
separation varying as a square pulse. You can freeze either
lip and watch the other go further, and also watch it not have
much effect on the separation achieved. At the moment the parameters
are hard-coded, so it has to be recompiled when they are changed,
but they should be readable from a file in a day or two.
Avery, this project is starting to sound interesting. Have you ever seen
one of the devices used in some speech and hearing facilities to monitor
accelerations and forces produced by the two lips and the tongue? Signals
from the transducers are used (typically one signal at a time) to drive the
position of a cursor on a screen. The client-patient-subject is asked to
make the cursor move to and remain at a static target position on the
screen.
The device would be a natural for replicating all of the interactive
tracking and control tasks Rick and I have done (the ones we admit do
not solve the world's most pressing social problems). I would model the
tasks with interacting PCT models. When you have pushed them a little
farther along, it would be interesting to see how your models handle the data.
Currently, most clinicians and researchers extract only low-grade data from
the devices -- mean latency to target, variances, etc., and many group
statistics. What a shame, they are sitting just above a vein of high-grade
ore! And I suspect this is a clinical area in which a PCT interpretation of
the clinical problems, and a control-theoretic design for research and
clinical procedures, could make a real difference.
I'll be watching for more news on your project.
Until later,
Tom Bourbon