Control Law Development

(aka PID stuff)

I wrote a paper a year ago which was sort of a tutorial on PID control equations. Over the year since, I've become aware that there is a lot of confusion in this area; and, I really have to go back and expand on the paper.

There seems to be an assumption that the classic PID block diagram is a good set of equations for our robotic purposes. The classic PID is good for control of a linear system with a relatively stable setpoint (command). It's also good for people who do analysis. It is not particularly good for working with non-linear systems (which most of ours are) or for dealing with large command changes. I don't think anyone would claim it is an optimal general purpose system. Hence, there is a lot more that can be done to write good equations for motor speed control, steering, etc. The Proportional, Derivative and Integral terms of PID can be useful in these equations, but often won't do the job well by themselves.

In the meantime, my paper is available here in two forms (sorry, I can't do PDF):

And, the paper includes a QBasic program which demonstrates the effects of the PID equations on control. That program is listed in the paper, or you can download it here.