Slashdot Mirror
Quadruped CHEETAH Robot To Outrun Any Human
cylonlover writes "Robots are faster than humans at a lot of things, but up until now running hasn't been one of them. That is set to change with robotics company Boston Dynamics recently awarded a contract by DARPA to design and build a quadraped CHEETAH robot that is faster than any human. The contract also includes the creation of an agile, bipedal humanoid robot. It's hard to say which one might ultimately be creepier."
Will it transform into a music cassette?
Can it transform into a micro tape cassette?
I have a strange suspicion this DARPA robot isn't going to have Asimov's laws integrated into it...
I think he is the perfect guy to reveal and control this new robot.
Outsmart, outrun, outlast.
Belief is the currency of delusion.
What? You have no faith. These are geeks designing this. Surely, one of them wished they had a Daggit of their own.
Bearded Dragon
Semantics...
I, for one, welcome our CHEETAH overlords...
The wheel called and said "Uhm, so fucking what? I was outrunning humans before there was language"
I love how this is creepy, yet put wheels on it and it's normal. Legs? Creepy. Wheels? No big deal.
Tiger Blooded Bi-Winning Machine
These aren't the giant robot camels I ordered!
A robot moving that fast will display a sense of instability and correction mechanisms that could force a change to the current robot design philosophy. That change is the incorporation of "skin". Skin places a huge role in bipeds and quadraped of that performance. Unfortunately, none of the research nor robot firms are considering skin, it's all about novel, complex mechanical designs due to longevity needs.
BD's has a good base design, but to achieve that DARPA objective is beyond their current system such that they'll end up with a complex mechanical structure that will be too big and heavy, and likely slow--assuming they go with the current thinking and not incorporate skin.
Why are we comparing a robot cheetah's speed to a human's speed?
Is it faster than a real cheetah?
I vote for QWOP-ASKL.
Will it outrun a human on the open savannah or through an urban city? I know the TFA mentioned tight turns and immediate stop & go, but what about in a building, over a fence, through the neighbor's back yard, up the stairs, from one roof to the next? I'd really like to watch something like this outrun an urban freerunner.
That 2000 film with Val Kilmer, where that robotic cat like thing AMY went all mustang and started murdering everyone. Efficiently.
It's called CHEETAH now, but when you refuse to burn down your house and destroy your contraband information, this technology will be much more useful in the mechanical hounds.
Figure it probably weighs >300lbs traveling at a high rate of speed. Unfamiliar with the pre-programming of it but what happens when it hits something or someone?
Is there an incentive on the contract that gives a bonus for creating 5 cheetahs that combine into one giant robot?
Pretty sure DARPA has that prize sown up. The robots will look like cheerleaders by comparison.
Glad to know my tax $$$ are being put to good use.
After due consideration, Boston Dynamics decided to name the new robotic creature the Rapid, Advanced, Vigilant Autonomous Guard Entity or RAVAGE for short.
Motorcycles, Robots, Space Gossip and More!
While I can acknowledge cool engineering when I see it, we discovered millennia ago that the wheel is better when designing mobile tools. Why do robotics researchers keep going back to animal shapes? This might be faster than a human, but that doesn't mean much. I guarantee it isn't faster, more efficient, or more practical than an AI-controlled motorcycle or trike.
Even Rosie on the Jetsons had wheels. Isn't that what we are supposed to be aiming for?
Same company created the four-legged robotic pack mule. I've seen it live. No running CHEETAH robot can ever be as creepy as that thing.
http://www.youtube.com/watch?v=pZYbp1dKRZA
Although the exact designs are of course kept under wraps, a likely model might look like this.
I am officially gone from
Once the human problem is solved, Skynet can start work on dressing the cheetah-bots in people clothes and making them do other adorable things.
Ask me about my sig!
Boston Dynamics is also building the LS3, which is the militarized version of BigDog. Stronger, faster, more range, but not much bigger. That's a tough engineering and mechanical problem.
There is a very pertinent tale in Stanislaw Lem's collection "The Mask".
Welcome back, Mike!
You coming back to apply your time wisely again in criticizing every story posted on Slashdot?
I mean, is continually criticizing a site that you've been calling "stagnated" for a long time now a good use of your valuable time? Do you think the editors, being (as you say) morons, are likely to listen to your sage and useful feedback?
Umm, yeah, good luck with that. When you get bored, you might want to visit Niagara Falls. I hear if you yell REALLY LOUD at the Falls that they are too loud, they'll stop. It might take you a while, though. Don't give up. Keep yelling. Bring a sandwich in case you get hungry.
A robot dog would be more impressive because cheetah's have small heads and brains and because cheetahs are so fragile. We have livestock guardian and herding dogs based - think wolf. They can hit 60 mph (clocked running behind a car for one half mile) and they can run long distances at slower speeds. Yes, that's not the 85 mph that a cheetah has been clocked at but cheetah's only do short sprints. Our dogs also have coordinated pack behavior, something else cheetahs aren't known for (although it does happen). You don't want pure speed but a more general attack and defense animal - the wolf.
My secret test word was "autonomy" which is rather appropriate.
you know what else can outrun a human? A CAR.
Saying your "phone ran out of batteries" is like saying your "car ran out of gas tanks".
Ready to form Voltron!!
robots can go much much faster using the simple wheel.
"They set a Slamhound on Turner's trail in New Delhi, slotted it to his pheromones and the color of his hair. It caught up with him on a street called Chandni Chauk and came scrambling for his rented BMW through a forest of bare brown legs and pedicab tires. Its core was a kilogram of recrystallized hexogene and flaked TNT.
He didn't see it coming. The last he saw of India was the pink stucco façade of a place called the Khush-Oil Hotel. "
So a contract was awarded to build a robot; so what? It is very easy to write a contract but very difficult to build to the spec.
Talk to me when you have something to show. Till then it's just words on paper.
Why is it that in all these stories, it just strikes me that they always seem to be thinking, "what kind of robot devices would be useful to Skynet when it takes over?"
Despite binging and googling with all my might, I was unsuccessful at finding what the acronym CHEETAH stands for. (Actually, I'm still working on the HAM radio so often mentioned here...)
These proof of concept bots are universally a waste of money. PID loops and Computer Vision don't need a $20,000 anthropomorphic test-bed to be custom made every time that someone makes a pot with higher resolution or ups the camera by a gigapixel.
Inverse Kinematics is obnoxious to program for but it's hardly the bottle neck here.
Processing speed/$ and energy density are the only things holding up robotics. Not a lack of mechanical engineers willing to waste grant money reinventing the wheel with faster electronics and .0002 tolerance square holes because they don't have the talent to design something which has a high performance/cost of production ratio.
Unfortunately for the state of robotics, it's saturated with computer scientists with very few mechanics. A motorcycle engine and a power steering pump are a more worthwhile power-plant than tethering the fucking robot to the ceiling(and less expensive).
You can solve the problem of batteries on a small scale by throwing money at it and rotating the Li-Polys every five minutes. That doesn't scale well.
It's affordable for a $200 robot to chew through $8 worth of batteries/5 minutes. It's not affordable to maintain a $200,000 robot which chews through $8000 worth of batteries/5 minutes.
Big Dog is a rock-star and Boston Dynamics should get this contract if anyone does, because at-least with them they'll accomplish some research in the process, but the only way this "Cheetah" is going to outrun a car is if it's a hydraulic powered carbon fiber motorcycle with legs.
They ought to create a robotics center in an only nuke containment dome and let robotics nerds play with RTGs like NASA gets to(but with no shielding).
If the hydraulic valves can't keep up with the PID, then their robot should be made of lighter materials or they should get more creative with the structure. A penny of inherent mechanical stability is worth a pound of ARM7s and accelerometers.
All you have to do is look at fighter planes to see the cost of performance in moving in the opposite direction. It's not reasonable to develop this shit at the cutting edge.
Then again, the MEMS gyro in your wiimote can probably be traced back to the economies of scale and CHEETAH-type research from the 1990s.
All the same, the philosophy of throwing money at an area you want to see advance only works if you throw the money at the right area. In this case, you have a bunch of mechanical engineers and computer scientists drinking at the trough of diminishing returns.
What we are seeing is public relations barbie dress up for the same boring battery technology.
They'd get a better ROI by putting that money in material science/superconductors/battery research and enjoying the benefits of SMES.
http://en.wikipedia.org/wiki/Superconducting_magnetic_energy_storage
Magnesium Diboride looks very promising, and liquid hydrogen cooling kills two birds with 1 stone. Put an RTG on board and watch out.
But don't let me spoil the fun, I'm sure Boston Dynamics will make a better use of the money than anyone else. Bonus points if it can climb stairs.
Those three little rules are awful vague. Seems like you'd almost have to be omnipotent to be able to full weigh them. And we all know the kinds of problems that can lead to. I mean, a little ol' lady with Alzheimer's wandering around in a construction site is a lot different than a teen skateboarder shredding in the park.
the preceding comment is my own and in no way reflects the opinion of the Joint Chiefs of Staff
My first thought: AMEE from Red Planet http://thecia.com.au/reviews/r/images/red-planet-1024.jpg
Why is this article tagged Idle? Surely this is most important News For Nerds!
Yet again, the Slashdot geniuses rush in to shit all over something they a) don't understand and b) couldn't achieve if they tried.
Oh, how I love this site.
Hey, geniuses: shitting on something isn't valid criticism, and you're not equipped anyway. You're just being d-bags.
Of course it will outrun any human, it's Cheetahing.
Would make a cool cruise missile...
Comment removed based on user account deletion
..why the spec for the robot called for an auxiliary chainsaw mount.
The roboticists will have a real milestone when they make one that can outrun a real Cheetah, and maneuverable enough to catch a Thompson's Gazelle. When it catches the gazelle, it has to do it as quickly as a cheetah. No fair just wearing it down by having more stamina. (Okay, building a robot to chase down gazelles might be cruelty to animals, something I'm against. My point is to put the achievement in perspective. Building any robot on legs that can outrun a human is an achievement, I admit. I'm just kind of struck by the idea of them using a cheetah form to do it. I would've read the article but the link is slow.)
In theory, theory and practice are the same; in practice they're different. (Yogi Berra & A. Einstein)
so, something that has not happened yet, is news in such a definitive fashion ...
i would like to remind you that there had been endless projects that were 'to' a lot of things, and never realized, or canceled, or half-completed, or failed.
Read radical news here
Time to invest in some bolo lessons
Have you fscked your local propeller head today?
As part of Mr. Lee's good neighbor policy, all Rat Things are programmed never to break the sound barrier in a populated area. But Fido's in too much of a hurry to worry about the good neighbor policy. Jack the sound barrier. Bring the noise.
My exception safety is -fno-exceptions.
What is a quad-RAPED cheetah, and why would I want to see that?
Who cares if a cheetah, robotic or otherwise can run faster than a man. The real question is whether it can run faster than a real cheetah.
Why? Why would you award a contract for such a thing? Is there a practical application for this thing?
These are not the droids we have been waiting for....
You mean AMEE, and you ruined the joke I was going to make, you insensitive clod!
The Admin and the Engineer
I have a strange suspicion this DARPA robot isn't going to have Asimov's laws integrated into it...
Terminator, Transformers, ... these are the wrong movies. Try Red Planet and the AMEE robot. Given the following snippet I bet you can guess what happens next. Hint: Asimov would not approve.
..."
"The landing craft is damaged entering the Martian atmosphere, veers off course, and crash-lands far from their landing zone near the habitat. In the process, they lose track of "AMEE", a military combat robot re-purposed to serve as their "Mars surface navigator"
http://en.wikipedia.org/wiki/Red_Planet_(film)
and the sad thing it's been years since i read fark regularly....
Media that can be recorded and distributed can be recorded and distributed.
-kfg
OK... a company with a name like Boston Dynamics, working with DARPA building robotic quadriped runners, am I the only one who thinks Fringe? Are they into cybernetics, biotechnology and space/time manipulation as well?
I think you are beginning to realize that what you formerly considered an attitude of those growing old was in fact an attitude of those growing up. As we grow up our attitudes that are somewhat based on textbooks and storybooks are modified by real world experiences. Now this is often domain specific, its not an overall understanding.
For example a hobby of mine is SCUBA diving. In the "old days" divers used mechanical analog gauges indicating depth and tank pressure, and a watch and a plastic card with tables indicating a time limit for a given depth. When dive computers were introduced to replace the preceding the diving magazines were filled with articles, everyone was really interested and curious and in theory thought this was really cool. However on the dive boats the doctors, lawyers and business types had the computers while the electrical engineers and computer programmers were still using analog gauges, a watch and a plastic card. Was the later group technophobic, I think not. To borrow from your post I think I would describe the the later group as questioning the speed of change and the benefits, a more grown up perspective. The former group, whether having blind faith in technology or wanting the new shiny thing as a status symbol was acting more childish. I admit the preceding characterizations are overly general but from the on-the-boat conversations we had back in these days I think there is some fairness to these depictions.
I'm not impressed until it can run faster than a cheetah
to the Tiger... for the paws that refreshes
For justice, we must go to Don Corleone
This isn't pairing AI and an autonomous robot, so your question has no relevance. At this time I would say that yes, we should rush ahead with successive generations of robots.
Poor quadruped thing..
While I can acknowledge cool engineering when I see it, we discovered millennia ago that the wheel is better when designing mobile tools. Why do robotics researchers keep going back to animal shapes? This might be faster than a human, but that doesn't mean much. I guarantee it isn't faster, more efficient, or more practical than an AI-controlled motorcycle or trike.
There is a major flaw in your logic. You are not considering that the tools we make are constrained by our technical proficiency. Wheels are our best option because they are of a sufficiently low technology that we can make them. As for practical they are not, hence the need for use to build roads or lay tracks all over the planet. Roads and tracks being another sufficiently low technology. For true practicality you have to look at what nature developed to navigate natural terrain.
As long as it doesn't piss on the carpet.
And we have the carnivorous robot technology being developed by another lab.
See subject.
My God, it's Full of Source!
OUTSIDE_IP=$(dig +short my.ip @outsideip.net)
It's a neat benchmark of our progress in robotics but humans have never been fast runners. Our advantage is that we can maintain a decent pace for extended periods of time. Many animals which are much faster sprinters would completely fail to keep up with humans over large distances. So while I'm impressed, it's more of a neat fact than a major milestone.
...where two guys prance around in a cheetah costume..the big dog parody was brilliant. http://www.ebaumsworld.com/video/watch/80666772/
Lots of natural quadrupeds can outrun a human.
The question is can it outrun a real cheetah?
How about a real swallow?
Ray Bradbury saw this coming:
"It made a single last leap into the air coming down at Montag from a good three feet over his head, its spidered legs reaching, the procaine needle snapping out its single angry tooth. Montag caught it with a bloom of fire, a single wondrous blossom that curled in petals of yellow and blue and orange about the metal dog, clad it in a new covering as it slammed into Montag and threw him ban feet back against the bole of a tree, taking the flame gun with him."
--Ray Bradbury, Fahrenheit 451, p 146
A well trained human can outrun just about any other animal, for a sufficiently large distance. If this robot only has the short stamina of a real cheetah, it'll not be much use. Here is a good TED talk along these lines: http://www.ted.com/talks/christopher_mcdougall_are_we_born_to_run.html
No "quadRAPED"/"quadRUPED" jokes?
Nuff said.