[Avery Andrews 931011.1238]
It might be helpful for someone who's up on such things to say about
about the characteristics of the `slowed proportional' control method
typically used in PCT models, vs. the integral component in the PID
controllers normally found in control theory and robotics textbooks.
From just naively staring at the PID control law, it looks to me like
there might be a problem if there are both steady disturbances and
fluctuating ones that can produce temporary episodes of large error,
the problem being that such errors can pile up in the integral term,
perhaps producing bad control when their causes go away (perhaps this
is one of the problems that adaptive control is supposed to solve?).
Slowed Proportional doesn't have this problem, but on the other hand it
can't *completely* cancel out a constant disturbance (but perhaps this
doesn't really matter, since the can be set as high as you like). At
any rate, a quick this aspect of PCT vs. engineering control theory
would help define the relationship of PCT to the mainland, as it were.
Avery.Andrews@anu.edu.au