|  | Bumper Logic Explored | Posted 25 Oct 2006 at 16:54 UTC by Rog-a-matic  |
Chris Schur reasons through bumper sensor design and programming
for mobile robots in this
excellent
article.
Complete with pseudo-code for better obstacle detection and
responses.
I don't think I've seen bumper sensors this well explained before. Nice piece of work. Definitely worth a view. Good job.
Thanks!, posted 25 Oct 2006 at 20:07 UTC by cschur »
(Master)
Thanks Dave, were going to do one on IR prox navigation next...
Chris
Hope you don't mind, I posted a link to the article on our club website. Outstanding explanantion! Even better, I found myself nodding at a number of your examples and thinking, "Boy I've seen THAT before." Now I have new tools to go back and remedy some of them!
Thanks for the work and energy you put into that.
Tom
Good indeed, posted 26 Oct 2006 at 02:04 UTC by marcin »
(Journeyer)
I was wondering though, Chris, whether you consider the bumber design used in your vacbot to be obsolete? I mention this as I have just finished building a bumper very much derived from vacbot's bumper.
I also found the article to be very informative, but then I've been a fan since the explanation of subsumption architecture.
Marcin
Thanks Chris, your work will save me a lot of time when I resume working
on my own robotics projects. I posted some notes and lisp code about
bumpers and finite state automata on my site a while ago
http://asmith.id.au/robotics.html but have not added to it since.
Your article is excellent and is the kind of thorough treatment of the
subject that we all need. Thanks again.
Andrew
Vacbots bumper, posted 26 Oct 2006 at 12:47 UTC by cschur »
(Master)
Marcin, That robots bumper is still certainly a valid type, following the guidelines of designing a mechanical system first, then adding the switches second. I use my vacbot up at our cabin in the mountains, where the floor surfaces are not so cluttered with household items like at home!
Chris
We started last night on our "IR Prox Logic" work, and quickly came to the conclusion that the IR beams MUST face forward looking to be able to navigate the robot through tight household confines. By pointing the beams radially, like most people do, you effectively increase the virtual diameter of ther robot substantially.
I've previously thought about going overboard with bumpers, along the lines of a honey-comb style pattern of buttons/bumpers densely stacked around the perimeter of the robot, but more than likely the 6-bumper design, together with some thoughtful logic does the trick just as well with a lot less hookup wire.
M
|