PDA

View Full Version : Robotics



ChunkyC
02-10-2007, 08:49 PM
YouTube video of a robot doing things I never thought I'd see in my lifetime ... at least not this soon.

Robo-skateboarder (http://www.youtube.com/watch?v=HBIW7As0nxM&NR)

Jamesaritchie
02-10-2007, 09:05 PM
About a year and a half ago, I heard a roboticist say the only thing stopping robots from physically doing nearly everything humans can do is an adequate, lightweight, long-lasting power supply. In other words, basically the same thing holding back electric cars, only more so.

ChunkyC
02-10-2007, 09:56 PM
They certainly seem to have worked out the balance issues for bipedal robots. Wowzers.

Will we see machines like Isaac Asimov's humaniform robots anytime soon? I'm thinking yes. Very soon.

Jamesaritchie
02-10-2007, 10:59 PM
They certainly seem to have worked out the balance issues for bipedal robots. Wowzers.

Will we see machines like Isaac Asimov's humaniform robots anytime soon? I'm thinking yes. Very soon.

It wouldn't surprise me a bit, but we still need that power supply. Then again, if some of the current research proves accurate, we seem to be getting very close to fusion power supplies that may well be safe enough, small enough, and more than powerful enough, to use in cars and robots. If this happens, all bets are off.

Development always lags behind research, and implementation always lags behind development, but I'd be willing to bet that in twenty years we'll have robots very close to Asimov's. Possibly even equal to his, though I wouldn't stake much money on this.

benbradley
02-11-2007, 12:10 AM
It wouldn't surprise me a bit, but we still need that power supply. Then again, if some of the current research proves accurate, we seem to be getting very close to fusion power supplies that may well be safe enough, small enough, and more than powerful enough, to use in cars and robots.
WHOA! Where did you hear that? I want references! Fusion has been the great boondoggle since the 50's or '60's - commercial fusion power generation has been "only ten years away" for the past 50 years. I've seen the video of Bussard's speech at Google, but even if that works it'll be for a traditional steam-driven power plant, not for something as small as a car.

Here's a couple of much more likely near-future power tecnologies, these for smaller devices, laptops and cell phones. First, fuel cells running a laptop available now (or rather, was announced over two years ago, dunno if it was actually available):
http://quote.bloomberg.com/apps/news?pid=10000080&sid=a5sZn3vGo1js&refer=asia
Then there are the mirco turbines running off butane (as in cigarette lighters) - these give more power for longer time than the best batteries of the same weight and volume. I expect personal robots would be run on something like this (or maybe a little larger):
http://www.technologyreview.com/Infotech/13865/
http://blog.scifi.com/tech/archives/2006/10/08/tiny_jet_engine.html

If this happens, all bets are off.
I'll be as surprised as anyone. You know what temperature it takes to do fusion? It's hot, I tell you, it's really hot!

kdnxdr
02-11-2007, 12:48 AM
I heard something about solar paint somewhere, wonder if that's some sort of possibility; maybe a solar fabric?

Check out Honda's website and go to the Asimo button. 'He' went on sale (really a rental program) to the general public in 05 or 06 in Japan. 'He' will be the first public model available, I would guess. I've already seen Honda's ads that they put in GoodHouskeeping to present Asimo as "one of the family". I think the ads were a bit before their time and haven't seen anymore.

I also think that the Rumba and derivatives are the first wave. Also, there are some dolls on the toy market that are precursors to a family robot.

The public has to be primed to really embrace the concept. I'm curious why I Robot was such a dud as a movie.

merper
02-11-2007, 12:59 AM
Which current research estimate is that? They're having containment issues with industrial size fusion plants as is, shrinking it down isn't going to make the issue much easier. The closest, barring any Eureka! sized discoveries, is ITER and IF that goes according to schedule it will only be finished in 2016:

http://www.iter.org/

You sure you're not thinking of micro-fuel cells?

Jamesaritchie
02-11-2007, 02:15 AM
Which current research estimate is that? They're having containment issues with industrial size fusion plants as is, shrinking it down isn't going to make the issue much easier. The closest, barring any Eureka! sized discoveries, is ITER and IF that goes according to schedule it will only be finished in 2016:

http://www.iter.org/

You sure you're not thinking of micro-fuel cells?

No, I'm thinking about the latest alternative fusion research by Dr. Robert Bussard. http://en.wikipedia.org/wiki/Robert_W._Bussard A video of his talk about the research can be found here: http://video.google.com/videoplay?docid=1996321846673788606

I've talked to a couple of physicists who say Bussard is going to be publicly blasted for this, largely because it may well take a lot of dollars away from many other scientists, but that he definitely seems to be onto to something very serious.

He believes he's solved most of the problems, and his test results are certainly impressive. If he's right, or as some are saying, merely on the right road, neither containment nor size is going to be a problem. Neither is radiation.

ChunkyC
02-11-2007, 02:19 AM
The public has to be primed to really embrace the concept. I'm curious why I Robot was such a dud as a movie.
I thought it was okay as a typical Hollywood sci-fi action film, but it had almost nothing in common with what Asimov wrote, other than a mention of the three laws of robotics and a few other bits and pieces here and there.

A much better robot film (IMHO) based on something Asimov wrote was The Bicentennial Man, starring Robin Williams as Andrew the android. I thought their portrayal of a household robot was much closer to what we might see in the not-too-distant future. And the script was a much better reflection of what Asimov wrote than was I, Robot.

But like y'all say, it all comes down to a suitable power supply.

merper
02-11-2007, 05:39 AM
Intriguing about Bussard. I don't know if he's right, but he's certainly no scam artist. Where's he been looking for funding? I'm surprised Richard Branson hasn't picked up on it. After all it's only $200 million in funding.

I hope he's set up a good system to protect his work, as he is getting on in years, and I'm sure there are certain people out there who might feel inclined to help him along.

benbradley
02-11-2007, 06:30 AM
Intriguing about Bussard. I don't know if he's right, but he's certainly no scam artist. Where's he been looking for funding?

Did you see the whole video? He was speaking at Google because Google has lots of money. As the slashdot article on this was titled, "Should Google go nuclear?"

merper
02-11-2007, 06:48 AM
I mean other than google.

Jamesaritchie
02-11-2007, 08:26 PM
Intriguing about Bussard. I don't know if he's right, but he's certainly no scam artist. Where's he been looking for funding? I'm surprised Richard Branson hasn't picked up on it. After all it's only $200 million in funding.

I hope he's set up a good system to protect his work, as he is getting on in years, and I'm sure there are certain people out there who might feel inclined to help him along.

The military was funding him, and there's now some private money, but I'm told he owns some patents, and they're taking things slow.

I've also been told he's going to have trouble getting money from traditional sources, primarily because of flak and threats from the "fusion the old way" crowd.

But he certainly isn't a scam artist, and neither are those who have been working with him. Every last one of his colleagues have impeccable credentials. His results have been very impressive, and I suspect things will heat up in the next couple of years.

Even if he's made some mistakes, it still looks like he's onto something real.

merper
02-12-2007, 12:47 AM
He says he can show proof-of-concept with just $1-2 Million. I don't know why he doesn't just ask for that first. It's the other $198 Mill that's scaring away investors.

Dawno
02-12-2007, 02:31 AM
*snipped for brevity*I also think that the Rumba and derivatives are the first wave.

The Roomba (and Scooba and Dirt Dog) are the consumer products of iRobot Corp. (http://www.irobot.com/) who are doing a great deal of development for the military (http://www.irobot.com/sp.cfm?pageid=109) as well.

These guys are ones to watch - they've got all that military R&D $$, marketing oomph, as well as a large nerd (http://news.com.com/Roomba+takes+Frogger+to+the+asphalt+jungle/2100-1043_3-6049922.html) fan (http://www.techeblog.com/index.php/tech-gadget/video-irobot-create-the-build-a-roomba) base (http://www.myroombud.com/) (count me in, I even wrote a poem and blog about my Roomba).

Thump
02-13-2007, 01:25 AM
What I wonder is how far the robot industry is going to go. There's plenty of money to be made but will we see porn-bots before janitor-bots?
My bet goes to the porn-bots :D There would be such a huge demand for them IMO. Seriously.

Love Bicentenial Man btw.

I think that unless they start selling afordable robots very soon there could be some trouble in the future. We're all pretty open minded and intelligent people here but the masses have mostly thought of robots as metal creatures that kill their creators for quite a while now. I predict that an android will be a scary thing unless they slowly get us used to the idea of something looking a bit like a human but without a soul.

IMO though... I want one that looks like Data ;) "fully functional and anatomically correct" *rowr*

greglondon
02-13-2007, 05:41 AM
by the time the the sun's light hits the earth, its putting out something like one horsepower per square yard.

Or something.

Cant remember the exact number.

It's a fairly diffuse power source, better suited for long term accumulation, not surges of unlimited power that might be used by, say, a super-jumping morphing robot. with fricken laser beams.

Mac H.
02-13-2007, 07:22 AM
His results have been very impressive, and I suspect things will heat up in the next couple of years.

Even if he's made some mistakes, it still looks like he's onto something real.Sure, but to keep things in perspective he talks about an effective design being about 10 metres in diameter - not including the control system, etc. (The 1 metre diameter design won't be able to produce power - just demonstrate that the technique can generate a decent energy well) Fascinating subject, though. Hope his design does work.

Mind you, we make enough robots now. We just manufacture them biologically instead of mechanically.

Mac

kdnxdr
02-13-2007, 07:28 AM
In the Age of Spiritual Machines, Ray Kurzweil predicts that by 2010, or so, that robotics will be so commonly used as body part replacements that we will have bionic droids, and that the legal system will be fighting out who is human and who is not human.

benbradley
02-13-2007, 09:22 AM
In the Age of Spiritual Machines, Ray Kurzweil predicts that by 2010, or so, that robotics will be so commonly used as body part replacements that we will have bionic droids, and that the legal system will be fighting out who is human and who is not human.
I've got that book, it's in my to-read-queue. I forget when it was published, but it's 2007 right now, and I don't see this happening in another three years or so. I'm optimistic about science and technology (at least as far as how much more we'll have, not neccesarily about who will benefit from its use), but that seems a little too optimistic to me.

Dawno
02-13-2007, 09:58 AM
One thing we should all do, regardless, is have a copy of this (http://www.amazon.com/How-Survive-Robot-Uprising-Defending/dp/1582345929/sr=8-1/qid=1171346256/ref=sr_1_1/002-2917669-4516019?ie=UTF8&s=books), just in case.

Zoombie
02-13-2007, 10:23 AM
maybe a solar fabric?


I really have to steal that idea.



I'm curious why I Robot was such a dud as a movie.

I can rant about that for hours, but I don't need to. Maddox at “The Best Page in the Universe” has already done it for me. I'd link it, but he's kinda profane.


This movie at the top of the page makes me suspicious. There are too many cuts. If they really want to convince me, they should just film it continue. Just so I can tell they're not doing some of the 'intermediary' steps for the robot. Maybe I'm too paranoid...nah.

Anthony Ravenscroft
02-13-2007, 10:25 AM
My own intro to how far "intelligent machines" have come in recent years seems very mundane on the surface. A music-supply catalogue showed up, wherein they offered a "squeal stopper" for sound systems. This thing could track four audio channels, determine when unwanted feedback was starting, & filter down that chunk of audio spectrum in a few microseconds.

On sale for $89.95.

Holy crud -- this hunk of hardware thinks faster than most of my relatives!

Weirdly, there are some areas where AI still hasn't come up to speed. Some 20 years ago I was working on precog algorithms -- pattern recognition -- & the problem remains as tricky as ever. There's big military bucks to be made, but green recruits (for all their problems) are still far better in "shoot/don't shoot" situations than today's microprocessors -- whence I should note that your cellphone probably has more raw computing power than the typical ten-ton computer of 1975.

FWIW, there's some hints (like the original "Ginger" that led to the Segway) that Sterling-type heat converters will be the way to go for self-operating robots in the near future. Such machines should be able to scavenge from whatever sources are available outlets, solar, thermal, wind, even chemical.

Mac H.
02-13-2007, 11:15 AM
A very impressive example of how good we are getting at making 'intelligent' machines is the competition to make a robotic vehicle to drive itself.

The first competition in 2004 - not a single vehicle made it to the finish line!
The second competition in 2005 - five finished.

That's a huge leap in ability in only a year.

For the new one in 2007 - the robot cars will have to obey all traffic laws and cope with other vehicles on the road.

So I predict that by 2010 they'll have a vehicle that can drive itself fine. They won't be available to the public, but they will exist.


FWIW, there's some hints ... that Sterling-type heat converters will be the way to go for self-operating robots in the near future. Such machines should be able to scavenge from whatever sources are available outlets, solar, thermal, wind, even chemical.Remember, though that there is no such thing as a Sterling heat converter. Only a Sterling heat DIFFERENCE converter. Thus, you need one part of the machine in heat and the other in cold to extract energy from the difference.

Mac

Lhun
02-13-2007, 04:24 PM
Sure, but to keep things in perspective he talks about an effective design being about 10 metres in diameter - not including the control system, etc. (The 1 metre diameter design won't be able to produce power - just demonstrate that the technique can generate a decent energy well) Fascinating subject, though. Hope his design does work.I was going to comment on this but you were faster. This problem won't be solved soon either, since it's a geometric problem. The better the volume to surface ratio, the better the energy needed for containment / energy output ratio.
I think our recent advances in nanotechnology will make capacitors the energy storage devices of the future. If you can create particles that are only a couple of atoms in diameter, that'll give you a real boost in surface area compared to the old tinfoil&paper capacitors.
Weirdly, there are some areas where AI still hasn't come up to speed. Some 20 years ago I was working on precog algorithms -- pattern recognition -- & the problem remains as tricky as ever. There's big military bucks to be made, but green recruits (for all their problems) are still far better in "shoot/don't shoot" situations than today's microprocessors -- whence I should note that your cellphone probably has more raw computing power than the typical ten-ton computer of 1975.
Well our brains evolved over several hundreds of thousands of years into optimized pattern-recognition CPUs (by now we're actually overshooting the mark. Everytime you see a face in a cloud that's your pattern-recognition programming producing false positives ;)), while the programmers only had a couple of decades to play with computers. Give'em time.
I predict that an android will be a scary thing unless they slowly get us used to the idea of something looking a bit like a human but without a soul.I don't think we'll see humanoid robots anytime soon. Not even after robots become more common. "Form follows function" is pretty important, even more so when it's a complicated and expensive machine. And a humanoid robot is pretty versatile but still besides the point.
Sure, he could mow lawn as well as vacuum the floor and wash the dishes, but he still needs tools for all that. It's far more efficient to just make a robotic lawnmower and a robotic vacuum and skip the android.

MattW
02-14-2007, 02:12 AM
Robotics reached it's pinnacle in 1983 with Herbie Hancock's "Rockit."

Pthom
02-14-2007, 04:38 AM
... maybe a solar fabric?

I really have to steal that idea.
Larry Niven already did in Destiny's Road (http://www.infinityplus.co.uk/nonfiction/destiny.htm) and Legacy of Herorot (http://en.wikipedia.org/wiki/The_Legacy_of_Heorot): He talks about a fabric that, when in direct sunlight, generates electricity. He called it "Begley Cloth (http://www.jerrypournelle.com/archives2/archives2mail/mail392.html#Begley)."

PeeDee
02-15-2007, 03:21 AM
I think that "humanoid" robots, which is what generally comes to mind with the word robot is a while off, not because it's impractical or impossible (although the power supply issue is a big problem) but because it's not necessary. We may look at humanoid robots and go "That's so cool," but until we have an actual physical use for them, robots will continue to look like Roomba, or the mechanical arms that build cars.

Now, when robot A.I. is being developed, how advanced will it get? If you look at computer-controlled-player AI in video games, there's a lot of possibilities for human emulative behavior, even if it's just calling pre-programmed responses.

WILL robots, when they are finally constructed as humanoid AI-controlled robots have something built in such as Asimov's Three Laws?

...

The one big use for building humanoid robots is that they could serve multiple purposes. A Roomba cannot build a car, a car-building-arm cannot vacuum your floor. on the other hand, a humanoid robot of sufficient design could vacuum your carpets, carry your groceries, fix your car, move your furniture, and clean up messes. So that'll be the use for humanoid robots.

...

Another question: When robots take on humanoid forms, will they be accepted by people, or will they meet with the distrust that robots so frequently receive in classic sci-fi novels?

One startling and puzzling discovery that has video game developers bamboozled right now is that a video game developer released an in-game graphics video of stunning video quality...and it met with horrible reaction from the audience, who may have loved the graphics, but were instantly turned off by the woman who speaks to the camera. The problem was that she was so human, people were alarmed by her eyes (which still, without understand how, seemed "dead") and the way that she moved.

In other words, she was so humanlike that people disliked the video without actually knowing why.

merper
02-15-2007, 07:32 AM
PeeDee -

This is why:

http://en.wikipedia.org/wiki/Uncanny_Valley

It's a good sign. :)

I wonder if there will even be a need for real actors 20 years from now.

ChunkyC
02-15-2007, 08:23 PM
I wonder if there will even be a need for real actors 20 years from now.
Judging by some of the movies that top the box office, I think there's a need for real actors right now. ;)

PeeDee
02-15-2007, 09:26 PM
There was talk about the danger of Synthespians (which is what Samuel L. Jackson called 'em) and how they would take over Hollywood. I guess I would lament the loss of, for example, Johnny Depp or Hugh Laurie.

On the other hand, if a robot takes the place of Jessica Alba, Hollywood loses no great acting ability... :)

merper
02-15-2007, 09:55 PM
I think those guys will still be fine. They'll still have the huge developed fan bases that support them. This is of course, assuming that a studio can't just take their image and make a computerized version of them.

American Idol Season 25 may be completely synthetic though. As opposed to the 95% synthetic rating it stands at now.

PeeDee
02-16-2007, 12:45 AM
I think those guys will still be fine. They'll still have the huge developed fan bases that support them. This is of course, assuming that a studio can't just take their image and make a computerized version of them.

Which they can, really. Witness things like Polar Express and the upcoming Beowulf.

On the other hand, with computer technology, I would rather have stylized movies like "Monster House," that perfect piece of Halloween movie making.

merper
02-16-2007, 01:17 AM
I mean legally. The technology for it is already pretty damn close.

PeeDee
02-16-2007, 04:09 AM
If it were needed, the technology would be there. Really, what they can do now could be applied to Hollywood. All they need is the driving demand.

PeeDee
02-16-2007, 04:10 AM
Robotics is of particular fascination to me right now. My serial that I'm doing is almost all robots, so it's been on my mind. Everything from "real world" robots, to Asimov's Robots (my personal preference) to the M5 "Five Alive!" style robots.

Jenan Mac
02-16-2007, 05:56 AM
Any of you familiar with US FIRST? It's a robotics competition for school kids (the website's usfirst.org). Some of the projects are pretty cool.

merper
02-16-2007, 07:17 AM
I tried to start a team in my high school senior year, but I had trouble getting support, and I got lazy. It is pretty cool though. It was founded(or at least promoted) by Dean Kammen(segway inventor),

benbradley
02-16-2007, 11:16 PM
Having been a regular attendee at the local hobby robot club, I've heard of First, and have almost been talked to being involved. There's also BEST Robotics which is a very similar competition.
http://www.usfirst.org
http://www.bestinc.org

My local hobby robot club:
http://botlanta.org

PeeDee
02-16-2007, 11:20 PM
I'd only heard of them in vague passing, so it's interesting to look at them closer.

I must admit that when I think of robots, my attention really only peaks with the news and information that can eventually relate to humanoid robots. For example, the robot head they built that had tracking eyes, and expressions that changed based on who it recognized coming into the room. I LOVE that.

I realize that more abstract-shaped robots are just as interesting and cool, but I suppose it's the twelve year old in me that gets giddy over one thing and not the other... :)

smoore
11-24-2008, 07:48 AM
I'm a little late in jumping into this thread, but you are all missing the one thing that robots need -- human-like vision. And I don't just mean video either.
Consider the following: if you shake your head back and forth while looking at a pencil, your brain can still determine it's not the pencil that's moving, but your head, yet to video the pencil moving back and forth would be indistinguishable. Now you might say that's more than vision. OK, maybe it is, but it's not just hardware, it's software. Another case is determining the speed (more or less) of an approaching car when you are driving -- yeah, I can write an algorithm to do it, but the one in your head is much better. Similar things can be said for peripheral vision experiments. And don't forget sound. I can pick out the violin in the string quartet as well as in the symphony orchestra, and I don't need an FFT to do it.
I guess you could argue that the software will eventually catch up. But maybe the power supply will come first.

lpetrich
02-09-2009, 01:48 AM
]
It wouldn't surprise me a bit, but we still need that power supply. Then again, if some of the current research proves accurate, we seem to be getting very close to fusion power supplies that may well be safe enough, small enough, and more than powerful enough, to use in cars and robots. If this happens, all bets are off.
I'll believe in that when I see it. Nuclear fusion is VERY difficult to run in a controlled manner, as opposed to being run in a bomb.


Development always lags behind research, and implementation always lags behind development, but I'd be willing to bet that in twenty years we'll have robots very close to Asimov's. Possibly even equal to his, though I wouldn't stake much money on this.
One would have to have human-scale AI, which is MUCH more difficult than many people thought in decades past. It's one of the great disappointments of my life. :(


Here's a couple of much more likely near-future power tecnologies, these for smaller devices, laptops and cell phones. First, fuel cells running a laptop available now (or rather, was announced over two years ago, dunno if it was actually available):
http://quote.bloomberg.com/apps/news?pid=10000080&sid=a5sZn3vGo1js&refer=asia
Then there are the mirco turbines running off butane (as in cigarette lighters) - these give more power for longer time than the best batteries of the same weight and volume. I expect personal robots would be run on something like this (or maybe a little larger):
http://www.technologyreview.com/Infotech/13865/
http://blog.scifi.com/tech/archives/2006/10/08/tiny_jet_engine.html
I'm not surprised. It's hard to get usable-energy densities from batteries that are anywhere close to what one can get from combustible fuels.

So fuel cells are a way to go, and miniature turbines may also be. Miniature piston engines? I checked out model-airplane engines (http://www.rc-airplane-world.com/model-airplane-engines.html), and some of them look rather high-maintenance.