0
$\begingroup$

I am using MATLAB to identify a system. The problem with my method is that my result is perfect. I get a nearly 100% fit, which make me wonder if what I am doing is correct.

I feed a sine sweep with an amplitude being the max velocity the system is capable of moving. I am sweeping from $0.1 \mathrm{Hz}$ to $2 \mathrm{Hz}$ (and another set from $0.1 \mathrm{Hz}$ to $5 \mathrm{Hz}$).

I log how much the position has changed for each time I change the velocity. Which seems a bit incorrect to do, since a change in position is just a velocity. Aren't I just sending and measuring the same thing hence perfect fit or what?

The System is a Pan/tilt unit FLIR D48E The input to the Plant is a velocity, and the output of the plant is a position. the input for the controller should be summation of how much it has to angular displacement given from input source, and its present location.

$\endgroup$
  • $\begingroup$ Maybe you could provide the code that you're talking about, but it sounds like you are missing any element of measurement or dynamic noise. That is, if you assume that you have a perfect model for how a system behaves, perfect knowledge of what you put into it, and perfect knowledge of what comes out, then yes, I would expect that you get answers that are in great agreement with what you would expect. Only numerical precision effects are left, and in double-precision processing as in MATLAB, those will be small. In the real world, however, there are uncertainties in nearly every quantity. $\endgroup$ – Jason R May 22 '15 at 11:23
  • 1
    $\begingroup$ Would you mind posting some code (obviously by editing the original answer)? The way you are expressing it is very difficult to understand. I wouldn't be surprised with a 100% match, but let's first understand what you are trying to do. $\endgroup$ – jojek May 22 '15 at 11:25
  • $\begingroup$ I made a minor error, gave this high fit rate.. It is now 71 % which seems more realistic. But still as my question states, is the output i measure not the same as the input.. A change in position is as far i know a velocity? $\endgroup$ – sasd May 22 '15 at 15:23
  • $\begingroup$ Again. We don't know anything about your system, therefore we are unable to give you a satisfactory answer. $\endgroup$ – jojek May 22 '15 at 16:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.