Dumb bots do it better
|
Steven Skaar, a robotics professor at Notre Dame, thinks there's a kind of Emperor's New Clothes world where academicians won't admit to themselves or others how little progress they are actually making. Skaar made a wheelchair that can take its occupant among just a handful of destinations -- the toilet, the kitchen, the bed. And it makes this journey very slowly. Skaar's students taught the chair its course by pushing it along all the permutations of possible routes and methodically saving them in the robot's memory. This simple robot, Skaar argues, is about as good as a robot gets.
Skaar believes that since robotic technology is at a "dead end", we should make pretty dumb robots that do what they do well. The future, he contends, is going to belong to androids with robotic arms spot-welding the same joint in the same car at the same spot on the assembly line. Via The New York Times. |
1 TrackBacks
Listed below are links to blogs that reference this entry: Dumb bots do it better.
TrackBack URL for this entry: http://www.we-make-money-not-art.com/cgi-bin/mt-tb.cgi/2585
Julianne moore nude angelina jolie nude, lindsay lohan pics jennifer aniston naked. Christina aguilera naked mariah carey pics, jennifer aniston pictures anna nicole smith naked. Katie holmes nude avril lavigne wallpaper, alyssa milano nude sienn... Read More
Leave a comment |
|
Um, that web page describes research done in 1994.
How does this get reported as news?
to answer your question:
- because this is big news to me, the opinion of that guy really surprised me as i keep writing about robots and new technologies and utopias, etc. and sudenly, for me, there's this guy claiming that none of that is really trustworthy.
- because it is an ongoing research, so it's not so old.
but your question is very interesting and it makes me reflect once again about what i'm writing.
you see, i keep finding projects, researches, etc. that i find absolutely awesome. I just leave them aside, thinking "mmmh too old for nearnear future", and sometimes (like inthis case), i just don't bother if it's not too fresh coz i find it interesting.
Hi Regine.
I found Prof. Skaar's sentiment interesting too, since I'm actually helping to develop a prototype autonomous wheelchair -- not experimental research but as a potential product.
Even though robots these days do all of the things that he says (said) can't be done easily (they prettyl reliably navigate autonomously, avoid obstacles), I think that his idea that simple things done well is often better than complex things done poorly is quite true. Our system is very complex and requires extensive configuration, lots of computing power, expensive hardware, and is still buggy.
It was really quite odd reading his website with the assumption that it was written recently. In the past 15-20 years there have been several important research advances in mobile robot navigation and lots of people have started working on it, at that time it seems it was still pretty rough. It's cool to see the "early days" when it was just another part of general robotics, now it's forked off into it's own specialty.
The future belongs to welding arms? Arms have been assembling our cars for thirty years! :)
On the other hand, though in the pages linked and in the NYT mag article it says robotics is at a "dead end" and is too hard, Prof. Skaar's primary research seems to be in computer vision, which is full of truly difficult problems, and really is a general purpose approach to sensing rather than a specialized and "dumb" approach.
reed
PS. I was not upset that you blogged about it (as "news") but there was no mention of the time factor in the NYT article and therefore not here either.