Robotic hand controlled by compressed air grasps the concept of delicacy
The Robotics and Mechanisms Laboratory (AKA the RoMeLa Project) at the College of Engineering at Virginia Tech has designed and built a prototype robotic hand that is controlled and operated by compressed air. Called RAPHaEL (Robotic Air Powered Hand with Elastic Ligaments), the robot can hold heavier, solid objects, as well as light or delicate ones such as a light bulb or an egg. The hand is powered by a compressor air tank at 60 psi and an accordion style tube actuator, with microcontroller commands operating and coordinating the movements of its fingers. It uses no other motors, and the strength of the grasp is controlled by a change in air pressure, making the hand quite dextrous. RAPHaEL -- which is part of a larger RoMeLa robot project named CHARLI -- has already won several awards, including grabbing first prize at the 2008-2009 Compressed Air and Gas Institute Innovation Award Contest. RoMeLa researchers envision CHARLI one day roaming the VT campus making friends with students and visitors. We look forward to that day, but until then, check out RAPHaEL holding some stuff after the break.



















came straight out of terminator movie
Gentle enough to hold a lightbulb, strong enough to crush an infant's skull.
a true terminator would be gentle enough to hold a woman, strong enough to crush a man's skull.
No sure I'm seeing the innovation here. The real problem with robotic hands is handling the requirement of delicacy (seen here) and strength (not seen here). There's a complex process we use to gauge how hard to grasp an object, we can catch an egg thrown 20 feet away (ok, some of us can), but we can also hold up our entire body (ok, again, some of us). We sense how hard to grasp something based on a mental catalog of data we keep on objects around us, empty soda can, tennis ball, baseball bat, etc. It's a complex mix of visual and physical cues. Until a robot incorporates both visual recognition and complex touch sensors (ie, texture, pressure, viscosity, etc) we're never going to have a human hand analogue capable of replacing many of the menial, but hard to mechanize, human tasks.
One small step for man, one less step for womankind.
8-
8--
8---
I couldn't disagree more, although an obvious step in the right direction, this thing looks like it would provide an aweful, chaffing hand-job.
I don't know about the rest of you, but I kept waiting for it to crush that Pepsi can.
too bad it was empty...
I was thinking the exact same thing.
It looked like they were trying to.
We weren't actually trying to crush the can in the video, it was only at about 60 psi in that shot, it's capable of 180, but the compressor used for that video is only capable of 100 psi. The next version of this hand should be able to crush cans at a lower pressure because the moment arm from the actuator to the hinge point is about two time longer.
Still looks creepy
Haven't we all seen robot hands that can pick up eggs and light-bulbs for years now?
[pun] including grabbing first prize at the 2008-2009 Compressed [/pun]
That fast without pressure sensors? Not often.
One of the guys who designed that is my RA (residential adviser) here at VT. That's awesome that their work is on Engadget.
I, for one, welcome our gentle, pneumatic overlords.
old hat:
http://robot.gmc.ulaval.ca/en/research/theme305.html
CHARLI BROW?
It's crazy because I used to design robots, and they could never be as gentle.
Impressive technology, clever name, RAPHaEL.
Handjob, anyone?
One step closer to sexbots..
dammit, is skynet coming that soon?
Anyone else find this rather un-amazing? I have seen many robotic hands that are capable of similar things as this one, yet they could move their fingers with much more accuracy. This one just seem to be able to clench or un-clench it's first, working rather like an air-powered litter picker, hardly dexterous tbh. I'll be interested once each joint in each finger can be moved separately and accurately, it doesn't need a constant supply of compressed air to operate (holds it's position) and is a lot stronger.
You've gotta be a lil' geeky or nintendo fanboy to understand this one:
WALLMASTER CAME TO LIFE! Everybody RUN!
http://www.youtube.com/watch?v=L5JHMpLIqO4
I liked the gripper they showed off in that video.
This is huge. What your seeing here will inevitable be somehow incorporated in to the future of an full AI robot. If or when we get to the point of robotics like were seeing in the movies this will have been the first steps a few of our grand children may look back on When their reading about the nano-microchip that was developed to do this task. They'll see how much effort and time it took to develop the stepping ground for that technology.
It flips you off every time it closes.
I find this a bit dated. 20+ years ago in college, we controlled the grip force of robotic hand using strain gauges, not air pressure. Strain gauges have a much higher frequency response and are far more accurate than air pressure, albeit they're probably more expensive.
Pretty crude video. Seems to not want to open all the fingers all of the time. The leaking air sound is anoying. The compressor starting up is an odd touch. Amatures. Interesting that we continue to want to reproduce the human hand of 4 fingers with an opposing thumb. Why not 8? Why not 3? Why a hand at all?
Now do it with Nitinol.
Looks like robots prefer Pepsi
Correction, Pepsi Perfect. The drink of the future!
I'm sure it's nicely engineered, but it's not the first air-driven hand.
The 1994 Utah/MIT Dextrous hand, built by SARCOS, had air-driven actuators that connected to the joints with tendons.
See http://www.cs.rochester.edu/~nelson/vision_lab/vision_lab.html
Why is the palm of the hand stiff?