Robots that "think for themselves"

Fly's eyes: Can the nerves, eyes, and brain function of a fly be modeled within a computer circuit?
Courtesy NeilsPhotography Engineers are trying to design machines that can "think for themselves" when on surveillance or search and rescue missions. Somehow the machines has to look at its environment and decide what to do.
Have you ever tried to catch a fly? They are pretty good at seeing your hand and knowing just how to escape your grasp.

If we can figure out how a fly can do it ...

Can we figure out how a fly is able see, and find food, and escape from our fly swatters? With today's super microscopes, I am sure that we can visualize and model every nerve connection, muscle fiber, and eye facet.

Computational biology

David O’Carroll, a computational neuroscientist who studies insect vision at Australia’s University of Adelaide has been studying the optical flight circuits of flies, measuring their cell-by-cell activity. In a paper published in Public Library of Science Computational Biology, O’Carroll and fellow University of Adelaide biologist Russell Brinkworth describe an

algorithm composed of a series of five equations through which data from cameras can be run. Each equation represents tricks used by fly circuits to handle changing levels of brightness, contrast and motion, and their parameters constantly shift in response to input.

The "fly brain" circuits are small and use only a fraction of a milliwatt

“It’s amazing work,” said Sean Humbert, who builds miniaturized, autonomous flying robots,

“For traditional navigational sensing, you need lots of payload to do the computation. But the payload on these robots is very small — a gram, a couple of Tic Tacs. You’re not going to stuff dual-core processors into a couple Tic Tacs.

Learn more - mathematical modeling of insect biology

Secret Math of Fly Eyes Could Overhaul Robot Vision Wired Science
Robust Models for Optic Flow Coding in Natural Scenes Inspired by Insect Biology Computational Biology

We have constructed a full model for motion processing in the insect visual pathway incorporating known or suspected elements in as much detail as possible. We have found that it is only once all elements are present that the system performs robustly, with reduction or removal of elements dramatically limiting performance. The implementation of this new algorithm could provide a very useful and robust velocity estimator for artificial navigation systems.

Tags : Russell Brinkworth, robotics, nanotechnology, nano, math, David O’Carroll, DARPA, computational biology, artificial intelligence, AI

read more... | add a comment | share

This is not BEAR: It's how I think of BEAR.
Courtesy JGordonHave we never talked about the uncanny valley on Science Buzz? I searched for the term, and got nothing. (Although… I’m beginning to suspect that my computer doesn’t accept voice commands. “Computer, display LOLcats,” gets me nothing, and I know that there are LOLcats out there.)

So… the uncanny valley. It has to do with robots, and human-simulation thingys. It’s like… like… well, here’s an example:

Think about factory assembly line robots—big arms, repetitive movements… it doesn’t do much for you, does it? They’re just boring ol’ machines.

Now think about R2D2, Star Wars’ trashcan robot. Beep beep, whistle! Cute, huh? He rolls around, and does sassy things we can’t understand, and we know he’s a robot, and he’s pretty likeable.

Now think about Johny 5 from Short Circuit. He can talk, he’s got a face, and expressive eye-flaps. And we still kind of like him, despite the attitude. (Great, you can read fast. Clean my kitchen before I have you recycled, robot.)

Now think abut C3PO, Star Wars’ deeply uncomfortable, shuffle-gaited robot. He’s pretty much human shaped, he speaks human (with an accent too…), and he’s clearly grappling with some of the same personal identity issues we real humans deal with. And… he’s just a little bit creepy, isn’t he? He’s like us, but not like us… How do we deal with this goldbot?*

And then there’s the “Simroid,” the Japanese robotic monstrosity used for dentist training. See the Simroid:

Clearly Lady Simroid has discovered what it means to be human, and she is, appropriately, horrified. And it doesn’t help that her existence is limited to sitting in a chair and having dental students see what hurts.

But, see, robots like the Simroid, in their appearance and limited behavior, are quite like humans. And it’s weird! They make us uncomfortable. So like us, but they’re absolutely missing the piece that makes a person a person. Brrrr

And then, moving on, we have healthy, living humans. Or maybe Blade Runner replicants. And they aren’t so weird any more. We’re back up to something we’re comfortable with.

It’s the Simroid point on this scale where the familiarity/comfort level takes a huge dive. That’s the uncanny valley.

(Another way to think about it might be cartoons. Stick figures. Disney’s Aladdin. Toy Story. The Polar Express movie adaptation. Pirates of the Caribbean. Which of these are you least likely to see on a poster in a kid’s bedroom? Well, maybe stick figures, but do you see what I’m getting at?)

See the dip in the graph?: That's the uncanny valley. It's full of zombies and simroids.
Courtesy Smurrayinchester
There are different theories as to why objects in the uncanny valley creep us out so much. The remind us of dead things. (Like zombies!) They are similar enough to us that, on a biological level, we perceive them as a threat (because a genetically similar creature is more likely to pass diseases to us, I guess), and so we feel revulsion towards them. Or they’re no longer like robots, but when we judge them on the human scale, they come up disturbingly lacking. Basically, they’re weird.

So, when you’re building your humanoid, you have to decide early on where you’re going to shoot for on the uncanny valley scale. If you aim too high, you may end up dooming your creation to the same hate we have for ventriloquists’ dummies. (In my opinion, you should probably set your expectations somewhere around R2D2, unless you’re making a baby. And even then…)

Enter the military-funded “Battlefield Extraction-Assist Robot,” or BEAR. BEAR was designed to be able to rescue wounded people in combat areas, and to do heavy, potentially dangerous tasks. It’s basically some big treads and a torso with arms, and each new version is a little stronger, and more nimble and damage resistant. And the newest versions have bizarre teddy bear heads, apparently because that’s the sort of thing that’s reassuring to an injured soldier.

So where does this fall on the uncanny scale? We like teddy bears. But teddy bears are usually soft and fuzzy, not six-foot-tall human-torsoed robots, able to dead lift 500 pounds. Also, their dark lifeless eyes aren’t usually set in hard, urban camo faces. For me, at least, a face like that seems to promise physical dismemberment with utter, robotic detachment (pun intended, I guess?).

Am I alone? Am I relating too much (but not enough) to the BEAR? How do y’all feel? Anything else in the uncanny valley that you feel deserves a shout out for its creepiness? Let’s have it.

*I’m aware that R2D2 and C3PO are supposed to be spelled out phonetically. I won’t be doing that. Ever.

Tags : uncanny valley, robots, bear, robotics, Battlefield Extraction-Assist Robot, military

read more... | add a comment | share

The scariest of robots: And how do I know there's a monkey brain inside? Look how angry it is.
Courtesy litmuseOh, you’re probably the same way—how often do you find yourself thinking, “I wish monkeys were more terrifying. Sure, they’re all fanged little were-men, with hand-feet and clever brains, but there must be some way that they could be worse.”

Pretty often, huh?

And, when you watch the news, don’t you constantly find yourself musing, “Hmm. The future is looking a little too bright.”

Well, don’t worry, Buzzketeers. The future promises to be just as dark and bewildering as ever, and horrifying cyber-apes are part of it.

“Now, JGordon, it can’t be that bad.”

Hey! Don’t sound so disappointed; it is that bad. Skeptical? Check it out for yourself—Sciencemen and Techladies have trained two macaque monkeys to control huge robotic arms…using their monkey brains!

Macaques have shown their evil little faces on Science Buzz before (murderous enthusiasm and enthusiastic murder), and I don’t think a refresher on robots is at all necessary—because there’s no escaping them.

Robotic limbs are becoming kind of a big deal these days, but even the most advanced of them rely on nerves remaining in a partial limb, or another part of the body entirely; which muscles to activate for a certain function must be relearned, or an operation like gripping with a robotic hand can be linked to a movement like shrugging the shoulders. It’s tricky to do, and it pushes the brain’s flexibility, especially considering that the only feedback the limb gives might be a hot or poking sensation at the connection point (this in place of a real limb’s feedback, like the pressure, friction, or warmth one might feel through their hands or feet).

Wiring a prosthetic (or any robotic device) directly into the brain—as was the case with these monkeys and their robot arms—overcomes some of the problems with existing prosthetic technology, while adding some new challenges.

With electrodes implanted right into the brain, relearning limb function can come much more quickly and naturally (awful little monkeys can do it, after all). A little too quickly, actually—a monkey at Duke University was similarly wired up this winter to make a robot in Japan walk, and the robotic body actually received the signals to walk before the monkey’s actual body did. Limbs wired the same way could be too fast or powerful for the brain to initially cope with. You might, say, run into a wall before your brain has time to create another route for your robo-legs; the speed of the limb action would be faster than the speed of thought.

However, if the prosthetics operated with a “closed neural loop,” that is to say if they could be made to provide natural feedback to the brain (like heat, pressure, strain, etc), scientists think that the brain could adapt much more quickly, and could even learn whole new pathways of motion. So a person wired up in the right way might be able to control a plane, or a nanosized robot directly with their mind. And it wouldn’t be something where you would think about walking forward and the plane would fly forward—you would learn the plane’s movements of flying, feel the flying, and control it as if you were the plane. That sort of things is still a long way off, and unless new technology is invented to sense and input to the brain in another way, it would require having a bunch of electrodes stuck through your skull and into your neurons.

This, of course, is all scientific blah be de blah, and if distracts from the real issue behind the story: cyborg monkeys. Do you know what the monkeys were actually taught to do with their metal limbs? Feed themselves. How horrible. Why not just teach them how to operate guns with their minds, or remove human brains through our nasal passageways?

In time, that too will come to pass. Look forward to it.

Tags : cyborgs, transplant, robots, robotics, prosthetics, health, monkeys, the future

read more... | add a comment | share

Watch this video of BigDog, the robot pack animal

BigDog has been improved. It now carries 340 lbs. on ice, through snow, over obstacles, and can even jump over things.

Tags : BigDog, robotics, Boston Dynamics, DARPA

read more... | add a comment | share