[From Rick Marken (930309.1500)]
Bill Powers said:
The only aspect of a control loop that is under reliable control,
therefore, is the sensor signal.
Hans Blom (930309) answers --
In practice, this is true only if the sensor signal is ever-present and
noise-free. If the sensor signal contains (much) noise, as is often true in
engineering applications, some means is required to separate the signal
from the noise. You often call such means an input function. There are two
general approaches to getting rid of the noise: averaging over multiple
sensors (redundancy) or averaging over time (filtering).
But then it is this filtered sensory signal that is controlled, right?
If the sensor signal may be absent for shorter or longer periods, a model
(in the sense of a built-in or acquired approximation of the object to be
controlled) is required that temporarily provides an alternative means of
pseudo-feedback. Such models are usually not highly accurate and therefore
drift will occur away from the optimal operating point.
And again, it is these model produced sensor signals that are controlled.
I don't get you point here, Hans. When a system is controlling (holding
some aspect of the environment in a fixed or variable reference state
while resisting the effects of disturbances) then it is the sensory
(perceptual) signal (whatever its cause) is really controlled. If there
is lots of noise or intermittancy in the system that there will be
NO CONTROL. In this case the sensory signal is, indeed, NOT CONTROLLED
because NONE OF THE VARIABLES IN THE LOOP ARE CONTROLLED.
The central insight of PCT is that WHEN THERE IS CONTROL THEN IT IS THE
FUNCTIONAL EQUIVALENT OF THE SENSORY SIGNAL (PERCEPTION) THAT IS CONTROLLED.
This is not just true of living systems; it is true of ALL NEGATIVE
FEEDBACK CONTROL ORGANIZATIONS when they are IN CONTROL (that is, when
their loop gain and dynamics are set appropriately so that the system
is STABLE and keeping a variable at a fixed or varying reference against
disturbance). The behavior of every single control system IS THE CONTROL
OF PERCEPTION. The evidence for this fact about control systems is
1) mathematical; solving the simultaneous equations for a negative
feedback control system we find that
p = r
(the perceptual signal, AND ONLY THE PERCEPTUAL SIGNAL, depends on
the setting of the reference signal -- or setpoint for the controlled
variable).
2) experimental; varying the sensory representation of the controlled
variable changes the environmental level at which it is maintained.
For example, wearing prism glasses changes the arm position that is
counted as pointing at a target.
This fact about control system operation must be of little interest to
a control engineer. But it is of FUNDEMENTAL significance a student of
the behavior of living systems.
Your statement above gives the impression that you don't completely
agree with the idea that control ALWAYS and ONLY involves the control
of a sensory (perceptual) variables. It sounds like you are suggesting
that other variables, besides sensory variables, can be controlled
in a control loop. I would really like to know what these variables
are because, if they exist, it would take an awful lot of the wind
out of the PCT sails.
Best
Rick