Slashdot Mirror
Hobbyist One-Ups Sandia Labs
An anonymous reader writes "A robotics hobbyist has created what he claims is the world's smallest autonomous robot. The robot is half the volume of the robots produced by Sandia National Labs in 2001, moves quite a bit faster, and was made using techniques and supplies accessible to anyone." While Sandia Labs has had some time to improve on their original designs, it's still pretty cool to see what one can do at home as well.
Why not place it in a box? I understand he doesn't have the sensor working to stop it from going off the table, why would he not then place it in a box as it looks like the device recognizes an obstacle and changes courses.
but it's not cute enough. Must add some more cute. Once there is enough cute (and the cowbell) they just may take over the world.
You can't handle the truth.
The GeekBot.
I for one would like to welcome our new minute robotic overlords!
But functionally useless...
.. but colour me impressed when someone comes up with one which can handle stairs. And if anyone's thinking of mentioning Asimo, I present Exhibit A
Why do researchers keep wasting time on these stupid little robots? There's very little actually learnt from making an autonomous robot. So it's not for science. There's very few products that benefit from this research. So it's not for commercialization. It's basically a hobby and, as this hobbyist has shown, best left to the hobbyists. It always surprises me how little scientists actually work together. By now, all these autonomous robot researchers should have put together a simulation package to do their research in. The mechanical engineers, who just can't help getting their hands dirty, can take the designs that have been tested in simulation and ensure they work in hardware.. thus giving the simulation as much credence as hardware, and yet allowing the reproducability necessary for doing actual science.
How we know is more important than what we know.
I just can't imagine what people that take that claim as a challenge will call the next four microbots. Well, at least it keeps a simple naming scheme. :wq)
I'd like to see Emacs, the microbot, though... It makes toast, does taxes and raises your children, but unfortunately it has the size and weight of a phone book, draws a kilowatt of power and the wheels don't quite reach the ground.
(Disclaimer
In a fair world, refrigerators would make electricity.
Yeah, but Macs are better!
I would really like to see this tagged "indie". puuuuuhleeese
Under the influence of Post-Cyberpunk Gonzo Journalism
I wish that thing was bigger so I could turn it into something useful like a vacuum cleaner.
warning: The above content may test positive for sarcasm and/or could be a failed attempt at humor and so should be taken with a pound of salt.
Someone made a robot half the size of what Sandia made did six years ago. So this person had six years worth of COTS component improvements and six years to refine designs.
I'm impressed. Truly. In high school I built a computer out of TTL chips. Wow, I sure one-upped Eniac.
Sure the product is cool, but this is hardly a one-uppance.
Think about it... a bright person allowed to:
- concentrate on a project all day long
- without PHB shaking deadlines in front of him
- without being burned out on meetings all day
- without the distraction of phone calls, personalities or politics
imagine what *you* could get done. I'm drawing a parallel to the busy workplace - not in any way do I mean to detract from this guys accomplishment. By all means, he's done something remarkable.
boycott slashdot February 10th - 17th check out: altSlashdot.org
Kids really enjoy problem solving for things that move. This creates a great learning environment.
Even plain old bump-and-turn robots have some very interesting control problems, like getting trapped. THis really helps people extend their problem solving skills.
I also work in real-world robotics (big multi-ton mothers).Sure we use simulations for developing control ideas, but those are pretty limited. You can test out various theories, but simulation only takes you so far. You need the real thing to get the dynamcs correct. For some real fun you want to see a huge robot go out of control.
Some of the most interesting research in robotics is being done at the hobbiest level. Lejos http://lejos.sourceforge.net/ has some very interesting abstractions and models for defining and controlling behaviour. Then there's also http://www.seattlerobotics.org/
Engineering is the art of compromise.
I shrunk the Roomba
Then it'll be used to clean the sewage drains
Look at the history: The Wright Bros. built their airplane in a bicycle shop. HP was started in a garage. What about Apple? Most inventions of merit have come from backyard hobby/hackers/dreamers. They all had talent whether mechanical or electrical and ran with thier ideas/dreams.
Professional Politicians are not the solution, they ARE the problem.
I can't believe they still make Pabst!
My dear Frodo. Hobbyists really are amazing creatures
Jesus Saves
Here's to hoping that they DO make one that has it's own power source.
You learn from your mistakes, and there's no substitute for "the real thing." You can crash the robot in a sim over and over, and it's no big deal. Real destruction with sparks and fire teaches you life-lessons ... I had a summer job (many years ago) with a rather large gub'ment contractor. We were working on a robotic arm tasked with de-palletizing ammunition - ammunition of the 5-inch diameter variety. The second day at the site, the programming team had gotten the arm all twisted up after several hours of fussing with the control software. The design lead saw the "home" switch on the control panel, and decided that starting from a clean setup would streamline things.
...
... which was thankfully de-energized.
*** CRASH ! ***
Nobody had programmed any obstructions within the cell, and some Genius had put the servo drive rack within the robots hemisphere of motion. The shortest path to the home location went directly through the servo drive rack. And when I say "directly through," I mean "ripped the rack in half." Literally. Big multi-ton mother, indeed.
So there's a bunch of down time while equipment is replaced, and we're back on-site after about 3 weeks. To my surprise, the servo control rack is still within the robot's operating envelope, but the obstructions have been properly programmed. There's even a short demo where they try to move the arm into the obstruction, but the machine refuses (rather politely, I might add.) Several days of progress are made before the Brain Trust is at it again. One of the programmers decides it will be "cool" if the robot commits suicide - he'd been reading an Asmiov book if I recall correctly. So they program the BMTM arm to reach over and press the main power switch on the servo control rack. It refuses. So they place a piece of 3/4" black pipe in the end effector to create the necessary tool offset. Attempt number two goes [click]
*** CRASH ! ***
They had shut the main power breaker off, resulting in a rather ungraceful de-energizing event. Apparently the servo drives can "lurch" if power is abruptly removed. The arm stuffed the black pipe and the end effector halfway through the servo control rack
Not surprisingly, we were not invited back for a third attempt to program the arm.
Cool project. It reminds me of another hobbyist that beat Epson (http://www.epson.co.jp/e/newsroom/news_2003_11_18 _2.htm ) to the smallest helicopter http://pixelito.reference.be/
... SKUNKWORKS! :)
Really, this is why these kinds of settings work.
Paul B.
It's been a long time, so I doubt the project is still online, but in the '90s one of the MIT kids built a dozen or so robots that were a couple of cubic inches... like tiny bulldozers.
He called them "ants"... all in all it turned out very well and he derived many interested behavior patterns out of only a few sensors and actuators.
Nothing in the world is more dangerous than sincere ignorance and conscientious stupidity.
This was the first thing I thought of as well..... good job so far and all, but really. ~
Why is the "smallest" consistently an issue when it comes to electronics and motorized equipment? They're being made smaller all the time. This will continue to be non-news headlines for some time to come.
I'm not impressed at all. They're all using microcontrolers and ROMS and such. A researcher at Los Alamos developed stepper-motor driven insect-like robots using 12, 14, or 20 transistors that'd "learn" to walk, some with different speeds/strides, with no preprogramming and within a few attempts steps. They even developed what amount to social behavior when operated in groups. The more agile ones that could run would run over the smaller, slower ones. In groups, the latter gathered together and backed themselves into a circle, which prevented the larger, faster ones from running over them. The "beheavior" emerged from some very simple conditions, and stretches the definition of "behavior", much like the light-sensing toy cars exhibited in an old SciAm article. In both cases there's no real learning because there's no collecting of information to be used other than immediate feedback through hardwired circuits. But when you saw a table full of these "bugs" circle the wagons to protect themselves against the attacking "lobsters", it was hard to not think of it in those terms.
"You want to see real artifical intelligence? Make it warm and soft and fuzzy and cuddly." -- Karl Pribram, who understood the fault lies with us, Dear Brutish peoploids, not with our toy cars; it's what we "think" they're "doing".
"I may be synthetic, but I'm not stupid." -- Bishop 341-B
Dangit, you beat me to it. It is definitely the severed, shrunken head of the Pinnochio robot Johnny-5. Mod parent up please. http://en.wikipedia.org/wiki/Johnny_5.
the beginning of a real terminator....
of the bump-and-turn cars I played with as a kid.
What makes this a robot instead of a child's toy?
Clear, Dark Skies
At least the hobbyist has good taste in beer! Mmmmmmm Peeber!