FROM BOB CLARK
This post was a response to a suggestion from Gary Cziko.
Your suggestion about Cruise Control is interesting. It can lead to
a discussion of several aspects of control system operation:
sensitivity, response time, offset (or "dead zone"), relation to
other systems, etc. These are interesting and important in
themselves.
However I selected Thermostatic Systems because they are familiar to
many people and they include the basic elements of Negative Feedback
Control Systems. That is, they include: 1) a means for detecting a
variable, 2) a means for affecting that same variable, 3) a means for
subtracting the magnitude of that variable from some "preset" value,
with a resulting positive difference acting to produce a positive
output from the second "means". This is the usual combination of
components composing a Negative Feedback Control System.
Such a system need not have a continuous output to achieve its
result. It is interesting to observe that "continuity" is, in part,
a matter of "viewpoint." Thus, if the Thermal System is observed over
a period of several hours, its control approximates continuity. And
the Cruise Control, observed in milliseconds, reveals various
limitations.
Also, it is true that the Thermal System is a "one-way" system as
usually presented with a furnace, etc, that is only one of several
limitations it suffers. Another "one-way" system is a living muscle
fiber! It can only pull, not push.
The Thermostatic System can also be used to illustrate other aspects
of control systems -- and other forms of control system. Thus:
Before Thermostatic Systems were developed, people kept warm in the
winter. My father had a coal fired furnace that had a damper that
adjusted its operation. Too cold -- open the damper; too ward --
close the damper. And it was a fairly continuous operation.
Where was the Control System? Clearly the situation was livable,
although not as convenient as one would like. Obviously, there was a
control system in operation where the temperature (where?) was the
controlled variable even though the control was accomplished by
adjusting the rate of heating.
But where was the Control System? Without a person, the temperature
was not controlled -- but also without a damper and a fire box,the
temperature also was not controlled. Some person - DECIDES - whether
action is needed, and in which direction. He then uses his (lower
order) muscle systems to affect his - ENVIRONMENT - according to his -
UNDERSTANDING - of his environment.
What about "his - ENVIRONMENT - ?" This usually refers, perhaps
vaguely, to his physical surroundings outside his skin.
But someone else might be available and asked to "open the damper,"
"turn up the furnace," etc.
Where now is the Control System? The person could "do it himself" or
"ask someone else." Having made his - DECISION - , he used his lower
order systems to get his desired result. Were there two (or more?)
levels of control involved: the "other person" and the "furnace."
Thermostatic Systems now take the place of the "other person," much
more efficient and convenient.
For those who are familiar with thermostats, most of this is
unnecessary. But what about those whose - ENVIRONMENT - does not
include - UNDERSTANDING - of control systems? You must have seen
people turn the setting up higher and higher when the room doesn't
warm up fast enough? The furnace was already at full speed, so
raising the setting has no immediate effect. Later, however, the
room is too warm and the setting is reduced. This is "over-control"
and the system is oscillating!
Notice the importance of the - TIME - scale of the person vs the
"response time" of the System.
This illustrates the difference between regarding the assembly of
parts as a Control System rather simply a group of connected parts.
Such a difference in - VIEWPOINT - can result in a difference in
behavior. Often these differences have little effect, but sometimes
they are very important!
Both viewpoints are valid and result in the same mathematical
representation. However one is more useful for using the system, the
other for modifying the system.
Several words have been noted above: DECIDES, UNDERSTANDING,
ENVIRONMENT, CHOICE, DECISION, VIEWPOINT, TIME. These words and
their associated concepts are used routinely and seem to be readily
accepted. However each of them is very important, and merits closer
examination.
And how does each relate to a Hierarchy of Control Systems?
Regards, Bob Clark