[from Jeff Vancouver (2003.08.07.1700 EST()]
Hi all,
I have been unable to post to CSGnet for some time, but presumably I
have resubscribed (if you are reading this, then it was successful).
I am working on a revision to my response to Bandura and Locke's (2003)
critique of control theory. I have a quick question for Bill and perhaps
others. Specifically, I have been reading Jagacinski and Flech book on
Control Theory for Humans. They talk a lot about Kalman filters. It
seems these would be very similar to the kinds of subsystems involved in
imagination mode processes. Is there any connection between Kalman
filters and PCT or HPCT?
Jeff Vancouver
[From Bill Powers (2003.08.08.0731 MDT)]
Jeff Vancouver (2003.08.07.1700 EST)--
I have been reading Jagacinski and Flech book on
Control Theory for Humans. They talk a lot about Kalman filters. It
seems these would be very similar to the kinds of subsystems involved in
imagination mode processes. Is there any connection between Kalman
filters and PCT or HPCT?
Not really, or at least not yet. Kalman filters as discussed in Jagacinzki
and Flach (not Flech) are a way of estimating the outputs that must be
produced in order to make the controlled part of the environment behave in
a predetermined way. They're part of "modern control theory." This is the
model that requires computing inverse kinematics and dynamics and having a
complete and accurate model of the system's own actuators and their effects
in the outside world, in order to compute the required command signals.
They require enormous computing power in comparison to a PCT model and in
my opinion are totally unbelievable as a model of a living system.
Best,
Bill P.
[From Jeff Vancouver (2003.08.08.1315 EST)]
Bill Powers (2003.08.08.0731
<Kalman filters as discussed in Jagacinzki and Flach (not Flech) . . .
<require enormous computing power in comparison to a PCT model and in my
<opinion are totally unbelievable as a model of a living system.
Yes, the high computational requirements of the more advanced
descriptions have this problem, but simpler versions are described
(though it appears for pedagogical reasons). As it turns out, these
simpler versions are very much like some of my modeling. I was just
wondering if others in the group had gotten to where I am via simplified
Kalman filters.
Jeff
[From Bill Powers (2003.08.08.1142 MDT)]
Jeff Vancouver (2003.08.08.1315 EST) --
I was just wondering if others in the group had gotten to where I am via
simplified
Kalman filters.
I don't knowe -- how does a "simplified Kalman Filter" work? And where are you?
Best,
Bill P.