Subject:
|
Re: Following the Line Follower
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Thu, 26 Sep 2002 14:36:47 GMT
|
Viewed:
|
994 times
|
| |
|
|
Very interesting (and good) work.
It may be useful for you to post the program used, because (as you said) it
can make a difference in the results.
However, I have to disagree with part of the conclusion:
"A line follower using a single light sensor will not be able to negotiate a
sharp 90 degree bend in both directions (clockwise and anticlockwise). You
need a line follower with 2 light sensors for that. "
This problem can be solved by either changing the construction of the robot
OR writing a slightly more complex program.
For example, on the robot you used, you had the one light sensor in front,
and between the wheels. If you move the light sensor to one side (I believe
the left side, for the program you used) so it is right in front of the
wheel, (like on your 2-sensor robot) I think you'll find it can actually
make it through the 90-degree turn in both directions.
I can elaborate, if anyone is interested.
Coincidently, this will shift the robot so it will drive more on-top of the
line, instead of on the edge. But that's a side effect.
There's a contest coming up in Seattle in a couple weeks, where this
information could be very useful.
http://www.workshop3d.com/rcx/index.htm
However, if you have a really good robot design or program, please DON'T
enter it. I'm hoping to win some more cool prizes.
Steve
In lugnet.robotics, Soh Chio Siong writes:
> Hi, people
>
> I have carried out an investigation into line following, a task that all
> respectable bots are expected to be able to do.
> =
>
> Diagrams often show the idealized zigzag path that the robot is supposed
> to take when following a line.=A0However, I was skeptical that the bot
> would actually travel in such idealized paths.
>
> Curious to know how robots actually follow a line, I stuck a felt pen to
> the front of the light sensor on my line follower. I tried with bots
> using a single light sensor as well as one equipped with 2 light
> sensors.
>
> The results are shown here =
>
> http://www.geocities.com/laosoh/robots/linefollow.htm
>
> I daresay that you can use this tracing method for comparing different
> mechanical designs of line following bots and for comparing different
> line following programs. This gives an objective way to decide whether
> the fine tuning (such as adjusting the threshold light values) you have
> made on your bot has an effect on its performance - important when you
> are preparing for competitive challenges. =
>
>
> One thing is certain, a bot with 2 light sensors is better than a bot
> with just a single light sensor for going round sharp bends. I haven't
> tried one with 3 light sensors, though but I should think 2 is enough. =
>
>
> Enjoy!
> -- =
>
> C S Soh
>
> CSSoh's Lego Pneumatics
> http://www.geocities.com/cssoh1
> =2E.. where air is power!
|
|
Message has 1 Reply:
 | | Re: Following the Line Follower
|
| (...) Thanks. I couldn't have done it without those ubiquitous felt pens. (...) The bots used fairly standard programs (written in ROBOLAB). For the single light sensor bot: turn right when it sees bright; turn left when it sees dark. For the bot (...) (22 years ago, 27-Sep-02, to lugnet.robotics)
|
Message is in Reply To:
| | Following the Line Follower
|
| Hi, people I have carried out an investigation into line following, a task that all respectable bots are expected to be able to do. Diagrams often show the idealized zigzag path that the robot is supposed to take when following a line. However, I (...) (22 years ago, 26-Sep-02, to lugnet.robotics)
|
5 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
|
|
|
|
