Subject:
|
Re: Science Central
|
Newsgroups:
|
lugnet.org.us.lrgoaa
|
Date:
|
Thu, 4 Mar 2004 00:14:51 GMT
|
Viewed:
|
2840 times
|
| |
|
|
Steve Hassenplug wrote:
> Did anyone take pictures of the stuff at Science Central on Saturday?
Some, but I haven't dug through them yet either (just got back from a short
vacation). I spent more time talking and photographing.
> Actually, I was wanting to see some pictures of Brian's maze robot,
> and find out what problems the robot had.
Well, that I might be able to arrange, as I've not even unpacked from NEIRG
as yet. But, honestly, it looked an awful lot like A-Mazing. Back-to-back motors
under the rear edge of a sideways-mounted RCX, with a front non-steering skid
wheel (medium pulley) about 3 studs in front of the front edge of the RCX. Front
sensors were a light sensor "color wheel" (using black, white, and two (old)
shades of grey 1x1 round plates, two staggered rows trapped between medium
pulleys) geared up 3:1 from a standoff arm with an undriven "feeler" wheel, and
a NC touch sensor for front wall detection.
Honestly, the major problem was overshoot. If it was too close to the wall,
it turned away, but by the time the standoff sensor indicated a "good" distance
from the wall, the 'bot was angled away from a wall-parallel trajectory, and
thus while the code did no trajectory corrections, the 'bot moved away from the
wall. As the distance increased too much, it would start turning back towards
the wall, but initally this would result in no decrease in the 'bot-wall
distance because first it had to turn back towards the wall (wasting time while
it drifted further away), then it was still too far away and would dive into the
wall.
I had this partially tuned out of the design at home, but the funky flooring
in the NEIRG maze threw all that out the window. I had a second (actually
faster) 'bot built, but not programmed for the event. Oh well, perhaps next
time. I really think it's possible to come close to A-mazings performance with
standard firmware, but the surface calibration issue is a real one. And you saw
how Trellis (the line follower) calibrated, I *LIKE* designs that are as
enviroment-independant as possible.
What part were you most curious about?
> Did everyone else have fun?
Absolutely!
--
Brian Davis
|
|
Message is in Reply To:
 | | Science Central
|
| Did anyone take pictures of the stuff at Science Central on Saturday? I haven't had a chance to look at mine, yet. I spent yesterday at a Children's Expo in Indy, loading chips into my Connect-Four robot, and trying to convince people that it (...) (21 years ago, 1-Mar-04, to lugnet.org.us.lrgoaa)
|
6 Messages in This Thread:
 
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
