The Sandia Hand mimics human fine motor movements. If a finger pops off, the hand can use its remaining fingers to pick it up and resocket it. The fingers (attached to the hand with magnets) fall off easily because they're designed to do so, in case the operator smashes the hand against another object. This prevents the hand from breaking. The fingers' modular design also means operators can mix and match fingers with different shapes and functions or attach other tools, such as flashlights, screwdrivers, or cameras. Operators control the hand with a glove.
Sandia National Laboratories has created a low-cost, highly dexterous robotic hand to aid soldiers in disarming improvised explosive devices. (Source: Randy Montoya/Sandia National Labs)
Developed by a team led by principal investigator Curt Salisbury, the Sandia Hand has a tough outer skin covering a gel-like layer that mimics human flesh. This helps the hand grab and manipulate objects the way a human hand would.
Hands are the most difficult part of a robotic system to construct, because of the need to design high dexterity while keeping costs down, Salisbury said in a press release. Sandia puts the cost of most robotic hands at about $10,000 for "each independently actuated degree of freedom" -- much too high for distributing to troops. The Sandia Hand, which has 12 degrees of freedom, costs about $800 per degree in low production volumes, Salisbury said.
To drive costs down that far, Sandia developed the hardware in partnership with researchers at Stanford University. It also worked with the consulting firm LUNAR to select motors, research skin-simulating materials, perform cost-of-goods and cost reduction analyses, and research design-for-manufacturing considerations. The Sandia Hand was funded by DARPA's Autonomous Robot Manipulation Program.
The hand's dexterity may also help investigators track an IED's makers after the hand has disarmed it. Instead of simply blowing up the device, the hand could perform delicate disarming operations that preserve evidence. Researchers did not say what kind of evidence could be used to detect IED makers, but chemical evidence is at least one kind that comes to mind.
Definitely looks like we're heading into some serious improvements in terms of the dexterity and flexibility of robotic hand movements. All good for those tasks that require precision and fluidity of movement. I'm stuck on the discussion about the "fingers" breaking, however. As these robots are built and marketed to be more human-like, those human-like descriptions become interchangeable and in cases like this, is can be jarring!
Ann, you talk about cost of most robotic hands being $10K and this one being $800. I wonder, what is the difference? Are those hands fully autonomous, or is it something else? Don't get me wrong, this is a very interesting and seemingly useful development. It is always interesting to know what was done differently to get this much cost advantage.
Lou, it looks to me like the key means of driving down costs were making an effort to do so. Since these are aimed at the military, costs had to be low, and there was a concerted effort to reach that goal. As we state in the article, this was the effort in partnership with the consulting firm LUNAR to "select motors, research skin-simulating materials, perform cost-of-goods and cost reduction analyses, and research design-for-manufacturing considerations." What I'm curious about is the opposite: if this can be done by a concerted effort, why the heck have they been so expensive before? My guess is that earlier robotic hands were developed either for fundamental research, or for very high-end low-volume uses.
From what I can see in the video, weight is an issue for this hand. The receiver of the princess phone was picked up but not the base. Even the case, which seemed heavy, was almost dragged by the handle. The ability to carry heavier weight may be a key difference for price.
Magnetic attachments for the fingers are a great example of design thinking. I wonder how they'll solve the problem of the fingers flipping back (very noticeable in the video).
Nadine, I agree about the weight issue, especially since the fingers are attached magnetically. The fingers' ability to fall off instead of break means--that they can fall off. OTOH, how much weight they will be dealing with when performing delicate IED disarming tasks may be a moot point. Their design seems to be aimed more at delicately picking up small objects, like the key, than at dealing with weight.
I wouldn't get too caught up in the magnetic attachment. That really seems to be a separate feature for specific applications where 1) force limitation is vital and 2) dropping a finger isn't going to cause other issues.
In addition to the other suggestions posted, I could see a few generations down the pike focusing on scalability. Same concept, for example, for microsurgery except 1/10 the size and the operator is using a zoomed-in video feedback.
Just wish the video would have shown more of the glove control mentioned in the article. There was just a brief glance at the end. It seems that there is an added layer of complexity for the user, though, since there are only 3 fingers.
Ann, Great article! I see variety of applications for which the Sandia Hand can be used in addition to bomb detonation. One application I see this hand being quite useful is in working with hazardous chemicals. A scientist will not need the protective gear when handling hazardous chemicals, the Sandia Hand could be used. The scientist can work remotely and manipulate the chemicals using the gloves to control the handling process. Also, the International Space station can make outside repairs with these flexible hands as well.
mrdon, I think those are interesting apps you mention. Of course, the materials would have to be designed specifically for use in those specialized environments: with hazardous chemicals and for the vacuum of space. Interestingly, the GM Robo-Glove my colleague Chuck Murray wrote about earlier this year http://www.designnews.com/document.asp?doc_id=240915 was derived from NASA's humanoid space robot Robonaut 2, which is up in the space station and one of whose tasks will be conducting EVA (extra-vehicular activity) dangerous operations.
This is a pretty neat development. What a clever approach to a robotic hand design. Interesting that this came out of Darpa - it seems that all the financing for robotic development is coming out of the government these days.
Holy cow, this is impressive! From the detachable fingers, to the amazing dexterity, this is a great development. I was impressed right up to the key pick-up, after that I was amazed.
The key pickup does hightlight one aspect that will require some more effort - feedback. Trying to pick up something without having the force feedback creates a sense of separation and isolation. Getting pressure feedback from the fingers to the operator's glove interface is the next big challenge.
Article states: "Traditionally $10,000 for each independently actuated degree of freedom, The Sandia Hand, costs about $800 per degree and has 12 degrees of freedom". Guessing that means each of four fingers is two degrees (one per joint), plus wrist rotate, plus wrist bend, plus elbow and plus shoulder equals twelve. (My guess). This entire robotic arm at $9600, is affordable for such a sophisticated piece of technology. Compare that to the prior (traditional) estimate at $120,000 for what we are seeing and the benefits become quite clear. TJ took the words right out of my mouth: first I was impressed, then I was amazed. It dropped a AA battery into a flashlight. THAT's dexterity sufficient to diffuse a bomb.