Subject:
|
Re: Drift Correction
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Fri, 3 Dec 1999 01:10:38 GMT
|
Original-From:
|
dave madden <DHM@MERSENNE.avoidspamCOM>
|
Viewed:
|
494 times
|
| |
|
|
=>From: Andy Gombos <gombos@ne.infi.net>
=>...
=>Does anyone know how to do drift correction? I have tried 4 different
=>ways, but they all don't work. Is it a difference in wheel
=>circumference? Or is it something with the programming? I drive for 5
=>seconds, then check the rotation sesnor values, If the are different, it
=>drives the wheels until the are the same(one wheel or the other). I
=>don't see why this will not work. If I am missing something, then please
=>tell me.
I don't think this method will work: while you're driving for 5
seconds, you're getting two motions (translation and rotation); but
when you drive only one motor at the end to fix the rotation sensor
values, you're only rotating. (Of course, in the physical world you
can't decompose motions into translation and rotation quite so simply,
but it gives a good way to begin thinking about it.) So I believe
that your method should leave the robot pointing more-or-less the same
direction as when it started, but not on the line between where you
started and where you were headed.
What you really need to do is watch the rotation sensors the whole
time you're moving. If one gets more than one tick ahead of the
other, idle the other motor until the counts are back together.
Perhaps you can achieve greater accuracy by running the two axles
through a differential driven backwards (with respect to the way it's
usually operated) and take the difference signal off the body of the
differential. Gear that up so that a small difference in wheel
rotation speeds turns into a large CW or CCW rotation on an axle, and
run that axle into a single rotation sensor. Then, use the sensor
output to determine when to adjust motor speeds. (I think this works
in theory, but suspect that LEGO(r) gears are too sloppy to make this
really practical.)
d.
|
|
Message has 1 Reply:
 | | Re: Drift Correction
|
| (...) I used this approach once - and with success. The gears are a bit sloppy and the rotation sensor only measures descrete angles. Anyway it worked for me. Jacob (25 years ago, 5-Dec-99, to lugnet.robotics)
|
Message is in Reply To:
| | Drift Correction
|
| Does anyone know how to do drift correction? I have tried 4 different ways, but they all don't work. Is it a difference in wheel circumference? Or is it something with the programming? I drive for 5 seconds, then check the rotation sesnor values, If (...) (25 years ago, 3-Dec-99, to lugnet.robotics)
|
6 Messages in This Thread:
  
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
This Message and its Replies on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
