DPRG
DPRG List  



[DPRG] RE: How much do our choices effect our results?

Subject: [DPRG] RE: How much do our choices effect our results?
From: Jeff Koenig koenig.jeff at gmail.com
Date: Fri Sep 7 19:21:13 CDT 2007

On 8/19/07, Randy M. Dumse <rmd at newmicros.com> wrote:

> Maybe we should turn this thread into a brainstorming discussion
> of what sensors we'd like.

I'd like for the Sharp IR rangefinders to have a longer range (maybe
they do, or they have ones with a "further" range  - I haven't checked
them in quite a while.)

I used several GP2D12's on one robot, and was able to use their
knife-like width to accurately detect cans, but only to a maximum of
about 4 feet.

I found that I could separate two of these sensors by a little more
than the width of a soda can, and angle them outwards very slightly so
that the inner edge of each "cone" of sensing stayed at about the
width of a can.

I then added a third sensor, right between the two, and wrote some
code to compare the three sensor's readings.

If the middle sensor's detection was closer than both of the outer
sensors by at least the width of a can, then the robot assumed there
was one dead ahead. (and the "Charge!" behaviour kicked in)  I also
had other, similar rules to determine if the "can" was to the right or
the left.

I wrote some "wander - spin - wander" code, and the robot would
consistently accelerate and run into randomly placed cans.  It worked
very well.  (I didn't test it much with edges - mostly open field).  I
showed this at one RoboRama.

But, it only to about 4', the range of the GP2D12.

More information about the DPRG mailing list