[Hans Blom, 930831], butting in on (Martin Taylor 930826 14:45) replying to
(Bruce Nevin 930826 12:08:03 EDT)
A control system can be perceived as a control system only from a point
of view outside the control system itself.
The claim is trivially true of an ECS, since what it perceives (and
controls) is a signal based on its perceptual input function. For it to
perceive itself as a control system demands that it "see" itself. If its
perceptual input function involved sensory input that included its own
perceptual signal, the opportunity for oscillation or chaotic behaviour
would be wide open; even if that were overcome by careful design, the ECS
in question would be a specialized one that controlled only for itself to
control itself to control.... Not very useful. So, of an ECS, the claim
is clearly true.
What is necessary for a control system, elementary or not, to be able to
"see" itself as a control system? I have experimented with such systems. A
requirement seems to be, that the system not only "see" its inputs (per-
ceptions), but also its outputs (how it effects the "outside" world). It
can then correlate its outputs with its inputs and hence determine their
effect: too much (overshoot), too little (too slow control), or just right?
This is the basis of adaptive control: the system does not only "know" that
it is a control system, it will - given the right internal mechanisms -
also be able to control the way it is controlling. If well designed, no
oscillation or chaotic behavior results, but optimization of the response
in the face of a changing world.
Greetings,
Hans Blom
[From Rick Marken (930901.0730)]
Hans Blom (930831) --
Thanks for the tip about Kalman filtering.
What is necessary for a control system, elementary or not, to be able to
"see" itself as a control system? I have experimented with such systems. A
requirement seems to be, that the system not only "see" its inputs (per-
ceptions), but also its outputs (how it effects the "outside" world). It
can then correlate its outputs with its inputs and hence determine their
effect: too much (overshoot), too little (too slow control), or just right?
Apparently "adaptive control" in control engineering is exactly the same
as "reorganization" in the reverse engineering of living systems done
by PCT. But I think the engineers are blowing it big time if they are basing
adaptive control on the correlation of outputs with inputs (as you described
them) since this correlation depends largely on the highly non-linear and
CHANGING environmental (feedback) function that connects output back to
input via the controlled variable. The PCT reorganization schemes are
based on the correlation between reference signal (what engineers
usually call "input") and perceptual signal (what engineers often call
"output"). The goal of reorganization (adaptive control) is to keep the
correlation between r and p VERY HIGH (which is the same as keeping
r-p very small). The problem with using your input-output correlation
as the basis of adaptive control is that sometimes this correlation
MUST be low (because of the feedback function) if control (perception-
reference correlation) is to be high. So an adaptive control scheme
that just tried to maximize the input-output correlation would actually
be REDUCING control in some cases. In PCT terms, isn't adaptive control
(in both control engineering and PCT) necessarily an attempt to maximize
the perception-reference, not the input-output, correlation?
Best
Rick