Subject:
|
Re: Lego or Biscuit pc/Linux (for path following robot)
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Thu, 6 Mar 2003 04:03:10 GMT
|
Viewed:
|
793 times
|
| |
|
|
Steve Baker wrote:
> Teoman Naskali wrote:
>
> > I have to build a robot for my thesis projet that will be able to
> > follow a line and inetrpret some graphical commands on the line • >> (like left triangle turn left, big circle wait 10 secs etc)
>
>
> Well - if you want to see a 'graphic' of (say) a circle and intend that
> to tell
> the robot to go around in a circle, you'll need a camera. The RCX isn't
> powerful
> enough to interface to a camera (although some dubious Lego adverts seem to
> indicate otherwise!). You can interface to the Lego camera only via a
> long cable
> going back to the PC.
>
> HOWEVER, apart from the task of recognising those icons, it's all very
> easy.
Steve, for the most part I agree with you. However, I was thinking that
a robot with two or three light sensors should be able to accomplish
this task. See the following description (untested as of yet)
One or two light sensors for following the line (2 sensors would help
keep the robot straight with respect to the line) The third light
sensor could be mounted on a arm that could detect an image (graphic),
then be motor controlled to scan the image into an array. The array
could then be analyzed.
I would guess that the robot would start by detecting the edge of the
graphic, then back up a small amount, then begin a scanning run. Move
forward a tiny bit, scan one line of the graphic (expected to be no
larger than an inch or two), then move forward a bit, scan another line,
so on and so forth.
Now, the graphics would have to be pretty distinct, otherwise comparison
could be tricky.
// Joe
|
|
Message is in Reply To:
4 Messages in This Thread:
  
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
