Subject:
|
Re: line following
|
Newsgroups:
|
lugnet.robotics.handyboard
|
Date:
|
Wed, 16 Apr 1997 21:25:24 GMT
|
Original-From:
|
Chuck McManis <CMCMANIS@NETCOM.COMantispam>
|
Reply-To:
|
CMCMANIS@NETCOM.COMspamless
|
Viewed:
|
1447 times
|
| |
|
|
Katt97@aol.com wrote:
> In my rather limited experience with reflective photosensors,
> I've found that black on white is not always the best of
> choices -- the white can come up as different numbers depending
> on how clean/dirty the floor is and the texture of the white
> (is it shiny and smooth, dull and rough, patterned, etc).
The experience is good, I'm just not sure you took away the right
lesson from it. :-) Everything Katt say's is true, when you are
following lines detecting the line is sometimes not as straight
forward as the description would have you believe.
What Katt has discovered is that the 'real world' introduces a
lot of variables that can confuse a simple line following robot.
What he has suggested (modifying the environment) is not as
applicable as "defensive" sensor management.
One of the goals here is to learn how to build reliable robots,
and to do that you have to take into account these variables.
One of the scouts I coached on building a line-following robot
came up with the following scheme that worked well for many
applications:
"Build into the robot a 'learning' mode where the
robot gets to learn what the line looks like. The
learning mode consists of the robot moving forward
and backward over the line. The samples taken are
plotted and the lowest and highest number are saved
off. The robot divides these numbers into three ranges
by dividing the total difference by 3. When the value
is in the lowest range, that sensor is considered to
not be over the line, when the value is in the upper
range the sensor is over the line, if the reading is
in the middle range the sensor is probably moving
on to the line or off the line."
Now what this person discovered was that when both sensors were
in the 'middle' range, the robot was straddling the line, when
one of them was over the line the other was usually either in
the middle range or the lower range. Using this knowledge the
speeds of the motors were adjusted to keep the motor on track.
This person also figured out that three states were better than
just dividing the range in half. It should be pretty obvious why
that was.
Interestingly enough this robot followed black on white or white
on black lines.
--Chuck
--
cmcmanis@netcom.com http://www.professionals.com/~cmcmanis
All opinions in the non-included text above are the sole opinions of
the author.
|
|
Message is in Reply To:
 | | Re: line following
|
| (...) <snip> (...) <snip> In my rather limited experience with reflective photosensors, I've found that black on white is not always the best of choices -- the white can come up as different numbers depending on how clean/dirty the floor is and the (...) (27 years ago, 16-Apr-97, to lugnet.robotics.handyboard)
|
3 Messages in This Thread:
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
