A robotic gripper uses a balloon filled with granular material, like coffee grounds, that can be switched between a loose and a rigid jammed state using an external vacuum or drive. Pressed against an object, the grains flow to conform to its shape, then jamming locks that shape in place so the gripper can lift items a standard mechanical hand could not, including irregular objects like a spring or an LED light. The hosts note this design, from around 2010, sidesteps the assumption that a gripper must copy the human hand and instead exploits the jamming transition directly.
- 01
The hosts contrast this approach with decades of robotics research aimed at replicating human hand dexterity, noting the granular gripper achieved general-purpose grasping without that complexity.
407 words · auto-generated from the episode video
1:53:35>> Okay. Okay. So, here we're looking at a video of a robot that's built from a human hand. It can't pick up those random items, right? But instead, they've got a sphere with a bunch of coffee grounds that they can control whether it jams or not. The key idea here >> is that because you've got a sphere with all of these little tiny grains >> and you can you can tune the transition between whether it's jamming or not >> using some kind of external drive. >> You can now fit that sphere. You can fit it to any arbitrary shape and you can pick up anything you want. Here it's
1:54:17picking up an LED light putting it. Brilliant. >> Yeah. Here it's picking up a spring. This is notoriously hard to do and you like we used to think that you know the solution to creating like random hands is like emulating the human hands but mate we've had millions of years of evolution >> right it's incredibly difficult to do this I mean now with the advent of AI and modern robotics sure but this is 2010 this is 16 years ago >> they had figured out >> it's also clever how they got there because it's not intuitive to me that >> yeah a ball a ball of coffee grounds cuz that's what this is ball of coffee grounds, >> right? >> It's like a hacky sack. >> And it's it's a surface that basically can arbitrarily based on whatever the
1:54:59object is that it comes in contact with create the exact construction in order to be able to lift it. >> Yeah. Yeah. And there's there's other examples. And here those are the objects. This is a standard hand back in the day. It couldn't lift those up. But this a single thing is just going to go and lift it up. >> Brilliant. >> Isn't that cool? >> That is very >> And this is directly from that jamming transition. Yeah, that's very >> we had understood using >> using using this mathematics. >> That is very cool because again it's it wasn't intuitive to me that that's where you go with it. >> Mhm. >> But obvious it makes again in retrospect it's like oh >> yeah exactly. So the next paper that we're going to talk about is 2005. This is in physical review letters. Um and
HistotripsySpaceSoft Matter Physics