The ability to grasp and hold objects is one of those human skills that robots have never quite got the hang of.
Most attempts to tackle the problem have centred around giving robots human-like fingers. That seems sensible, given that it’s the solution discovered by the multibillion-year optimisation process we call evolution.
And yet, we clearly haven’t yet uncovered all of nature’s tricks. Robotic fingers generally require a centralised control system to handle data streams from touch and visual sensors and which then calculates how to carry out a grasping action.
Humans clearly have a way of short-cutting or simplifying this process. When was the last time you reached for a cup of coffee and had to think about how far apart your fingers should be or how hard you should grip?
Perhaps there’s another way robots could do it, suggest Eric Brown at the University of Chicago and a few buddies. These guys have developed a robotic gripper capable of grabbing objects of almost any shape with an appropriate level of force but without any system of tactile or visual feedback.
The new device is deceptively simple. It consists of an airtight rubber bag filled with grains, in this case, of ground coffee, connected to a pump. The idea is to dump the bag on top of the object to be gripped so that the grains form loosely around it.
As the air is pumped out of the bag, it contracts with two consequences. It first tightens the contact between the bag and the object and then forces the grains to jam, so that they harden into shape. “Applying a vacuum enables the gripper to gain remarkable rigidity while almost completely retaining its shape around the target,” they say.
The grip then comes from three factors. First, there is the friction between the bag and the surface of the object, there can also be a suction effect if the bag forms an airtight seal around part of the object. Finally, the interlocking shape of the object and the gripper simply prevents it from falling. As a result, it is hugely adaptable. “Only a fraction of an object’s surface has to be gripped to hold it securely,” point out Brown and pals.
And it certainly seems to work on a huge variety of objects. Brown and co used their gripper to pick up everything from a ball to a spring and to do it with the object in any orientation and without any prior knowledge of the shape. “A granular system can move with ease from gripping steel springs to raw eggs, and it can pick up and place multiple objects without changing their relative orientation,” they say.
The device has its limitations, of course. It failed to pick up a flat disc or a ball of cotton, for example. Neither is it able to manipulate objects while holding them.
But for many applications, none of this will matter. What will count is its utility, adaptability and above all its cost relative to systems that require both feedback and computationally-hungry centralised control.
Ref: arxiv.org/abs/1009.4444: Universal Robotic Gripper based on the Jamming of Granular Material
Capitalizing on machine learning with collaborative, structured enterprise tooling teams
Machine learning advances require an evolution of processes, tooling, and operations.
The Download: how to fight pandemics, and a top scientist turned-advisor
Plus: Humane's Ai Pin has been unveiled
The race to destroy PFAS, the forever chemicals
Scientists are showing these damaging compounds can be beat.
How scientists are being squeezed to take sides in the conflict between Israel and Palestine
Tensions over the war are flaring on social media—with real-life ramifications.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.