Soft Robot Hand Can Sense Shapes and Snag an Egg With Ease

It's got silicone fingertips. 


The human hand is a machine-honed by millions of years of evolution, transforming from bony fish fin to a grasper that relies on the delicate interplay of mechanoreceptive neurons and fine motor coordination.

In contrast, the rigidity of components and less sensitive feedback loops hobble robot hands. But soft robotics — that is, using something that can squish in lieu of tougher materials — could offer a solution, as shown in a recent Massachusetts Institute of Technology demonstration.

An egg? No problem:

As MIT researchers will present at this week’s the International Conference on Intelligent Robots and Systems in Hamburg, Germany, the soft hand gripper has conquered all delicate ‘90s artifacts: It snatches Beanie Babies, compact discs, and white printer paper with ease.

Soft robots are more flexible than their hard-bodied compatriots, not only in their components, but in the way they approach tasks. To pick up an egg, a hard robot would first have to calculate how much force it could use or else risk crushing the object. The MIT Computer Science and Artificial Intelligence Laboratory developed its soft robot with sensors in the fingers. The robot fingers assess the shape of an object and compare that shape to items it’s already picked up.

MIT PhD student Robert Katzschmann likened the robot, in a press release, to a blindfolded human, who can still pick things up without seeing it.

Humanity draws ever closer to realizing its ultimate goal of robot masseuses.

Related Tags