Subject:
|
Re: Maze robot testing
|
Newsgroups:
|
lugnet.org.us.laflrc
|
Date:
|
Wed, 1 Feb 2006 15:54:36 GMT
|
Viewed:
|
9251 times
|
| |
|
|
In lugnet.org.us.laflrc, Steve Hassenplug wrote:
> > For those of you who have built maze running robots before
> > for NEIRG, what sort of test setup did you use?
The only time I ran it I used 1x4's and tested in a "U" shape (both clockwise
& counterclockwise) on my kitchen floor. This worked great, until event day when
the floor friction fouled me up (I may have to cobble up a test system like
Bryan suggested, with styrofoam... although isn't the NEIRG course painted black
as well, so the friction charecteristics would be due to the paint as well?).
> Greg pointed out the simplest shape which will test out most of the turn
> combinations is a '4' (with an open top).
I use a "U", and for simple turns an "L" would work... but turning around in
a dead-end (the "bottom" of the "U", or the "cup of the 4") seemed to be a
significant problem. Another that some robots might have is overshooting on
outside turns (less of a problem, but easy enough to test).
> > Does anybody have any hints about what to avoid... [to] keep me from
> > having surface-related problems in the maze?
Yes. The surface :-). Steve's suggestion is a good one, having nothing that
can drag or snag anywhere on the robot. Th three biggest problems I saw were (a)
some part of the robot getting snagged, (b), the robot getting turned around, or
(c) getting caught in a corner. The last caught me (twice, just a few feet away
from the exit).
> For reference, my maze robot runs on the small mag wheels,...
> A couple people have told me they have faster robots,...
Another point to remember for those in the "beat Steve" camp - keep in mind
larger wheels may not cut it. If you've not seen A-mazing run, it's *very*
smooth along the straights and wastes little time in turns. A robot that runs
faster along the straights, but does poorer in turns, will still have a tough
time. The NEIRG maze of years past has been heavy on turns, so turning time is
at least as important as "speed runs" down the (few) straightaways.
> remember the robot has to be able to make it through the maze, or
> it doesn't matter.
I entered a near-copy of A-mazing, and took third. I would have taken second
(mine was much fazster than the ultimate 2nd place), but my robot got caught in
a corner in one of the last dead-ends in the maze... *twice*... I'm not even
sure how I could have tested this out, since surprisingly it did fine on dozens
of dead-ends in the maze up to that point, always to somehow *just* hit wrong in
the same near-the-finish dead-end.
Oh, and who took first that year? Steve with A-mazing. It's a very elegant
robot.
--
Brian Davis
|
|
Message is in Reply To:
 | | Re: Maze robot testing
|
| (...) Greg pointed out the simplest shape which will test out most of the turn combinations is a '4' (with an open top). That's just 3 walls, two are 12"x12" and one is 12"x24". Remember, the walls are 12 inches high and 1/8 inch thick, and each (...) (21 years ago, 31-Jan-06, to lugnet.org.us.laflrc)
|
6 Messages in This Thread:
 
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
