Subject:
|
Re: Robot navigation
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Mon, 1 Nov 1999 22:37:18 GMT
|
Original-From:
|
Andy Gombos <GOMBOS@NE.INFI.NETsaynotospam>
|
Reply-To:
|
gombos@ne.infi.STOPSPAMMERSnet
|
Viewed:
|
947 times
|
| |
|
|
If you started with the wheels pointed in an arbitrary direction (say north) ,and you knew how many degrees it took to change a compass from (north) to (north east), then you could do crude
navigation.
1. Compass and robot say north.
2. Robot travels, and it thinks it is going north, but compass says it is going northeast.
3. Robot turns enough to bring wheels (and robot) back to facing north.
4. Compass now says north.
5. Go to step 1.
This way, you would be able to follow a 'straight' line, straight being ending up in a pretty close line to from the start.
Line that should be followed: Line that is actually traveled:
Start Start
\ \
\ \
\ /
\ \
\ /
\ \
End \ End
The picture shows that the robot ends up pretty close to the place that it would have ended up if it had traveled straight. This would probably be only a few inches( a foot at the most) from where
the robot really ended up.
The accuracy of the robot could be adjusted by changing the rate at which you sample the compass. The higher the rate, the more chance you will notice right when the compass changes.
If you want an accurate non electronic 1 position compass, then do this.
Take a compass that is cheap, but not filled with water.
Drill a hole right through the heading that you want to measure. (Both the top and bottom) The hole needs to be straight.
Take a light source (laser module?) and mount it on the bottom of the compass over the hole.
Then take a light sensor, and mount it over the hole in the top.
If the needle is in the way, then the reading will be low, but if the needle moves out of the way, then the reading will go high, and you will know you have left the heading (although not which
way...)
That is how you know you left the heading.
Andy
|
|
Message is in Reply To:
 | | RE: Robot navigation
|
| I have been working on this too, and I decided that I am going to live with a certain amount of uncertainty. If I can get to within a few inches of where I think I am, that's going to be good enough. Until it isn't anymore, and then back to the (...) (25 years ago, 29-Oct-99, to lugnet.robotics)
|
2 Messages in This Thread:
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
