DPRG mailing list archives: [DPRG] DPRG monthly meeting & talk this Saturday December 10: David Anderson

Mon Dec 12 17:32:53 CST 2011

Howdy,

Thanks to Ron and Paul for the kind words and to Glen for recording and 
posting the video.  I had a great time Saturday.  Hope this stuff is 
useful.

Watching the video I see lots of mistakes... (base is isosceles 
triangle, not equilateral) so maybe I need to make an addenda.  I'll 
make some notes and post them.  Isosceles.

I was unable at the warehouse to demonstrate the robot perimeter 
following due to a mechanical problem, but here are a couple of videos 
that might be helpful in that regard.  The first is the LegoBot 
perimeter following an actual wall, using front-looking IR avoidance to 
stay off the wall, and side looking SHARP IR distance detectors to 
approach the wall:

<http://www.geology.smu.edu/dpa-www/robo/lego/lego-04.mpg>

I use the term "perimeter following" rather than "wall following" 
because the perimeter my be loosely defined as the boundary around a 
group of objects, rather than an actual wall.  Here is LegoBot perimeter 
following around the irregular perimeter of my office:

<http://www.geology.smu.edu/dpa-www/robo/lego/lego-03.mpg>

Here is the jBot perimeter following around part of the law school at 
SMU, using forward looking Sonar to stay off the wall, foliage, etc , 
and side looking Sonar to approach the wall:

<http://www.geology.smu.edu/dpa-www/robo/jbot/jbot_wf1.mpg>

Perimeter following is a very powerful tool for solving all sorts of 
problems.  The last link is an annotated video of jBot switching between 
navigation mode and perimeter following mode in order to navigate around 
a large building complex that includes dead-end cul de sacs and other 
similar "concave singularity" problems for the robot to solve.

The algorithm tries to attempt to acquire the target waypoint in 
navigation mode, and it switches to perimeter following mode if it finds 
that the target waypoint is directly behind the robot, switching back to 
navigation mode when the waypoint is directly ahead of the robot.  This 
works very robustly to solve the waypoint trap problem we talked about 
Saturday.

<http://www.geology.smu.edu/dpa-www/robo/jbot/jbot2/jbot_ti2.mpg>

Thanks again to everyone that came out Saturday.

happy roboting,
David

On 12/12/2011 07:57 AM, Glenn Pipe wrote:
> David's presentation video is now online at:
>
> http://www.youtube.com/watch?v=8CXReb7f0Eo&feature=youtu.be 
> <http://www.youtube.com/watch?v=8CXReb7f0Eo&feature=youtu.be>
>
> Thanks,
>
> Glenn
>
> ------------------------------------------------------------------------
> *From:* Ron Grant <deltagraph at aol.com>
> *To:* davida at smu.edu; dprglist at dprg.org
> *Sent:* Sunday, December 11, 2011 6:53 AM
> *Subject:* Re: [DPRG] DPRG monthly meeting & talk this Saturday 
> December 10: David Anderson
>
> Thank You, David, for the great presentation! I promise to read your 
> published information too, rather than pretend I can invent everything 
> myself. Ha!
> Thanks to Glenn and Doug P. for doing video recording. I look forward 
> to watching the video as I am sure many people across the world will.
> Thanks also to all who set up the meeting space in the Makerspace 
> warehouse area.
> I recopied David's links, as a reminder to read!
> <http://geology.heroy.smu.edu/dpa-www/robo/subsumption/>
> <http://geology.heroy.smu.edu/dpa-www/robo/challenge/>
> <http://geology.heroy.smu.edu/dpa-www/robo/Encoder/imu_odo/>
> Ron
>
>
> _______________________________________________
> DPRGlist mailing list
> DPRGlist at dprg.org <mailto:DPRGlist at dprg.org>
> http://list.dprg.org/mailman/listinfo/dprglist
>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://list.dprg.org/pipermail/dprglist/attachments/20111212/c62a4c10/attachment.html 