Robots can be trained to perform all kinds of tasks, but their rigid hands and fingers can make it hard for them to grasp and hold objects. Now, scientists at the Massachusetts Institute of Technology have created a robotic hand made out of silicone rubber that can lift and manipulate objects as delicate as an egg.

The robotic hands' three fingers are equipped with sensors that can gauge the size and shape of the object, and even identify it out of an array of other items. The grippers also proved to be able to pick up other objects, including a Rubik's cube and a Beanie Baby.

"Robots are often limited in what they can do because of how hard it is to interact with objects of different sizes and materials," said CSAIL Director Daniela Rus. "Grasping is an important step in being able to do useful tasks; with this work we set out to develop both the soft hands and the supporting control and planning systems that make dynamic grasping possible."

The work highlights the value of "soft robots" made from unconventional materials such as silicone and paper. These robots have a number of advantages over "hard robots," including the abilities to hold irregularly shaped objects, squeeze into tight spaces, and recover quickly from impacts.

"A robot with rigid hands will have much more trouble with tasks like picking up an object," Hosaid graduate student Bianca Homberg. "This is because it has to have a good model of the object and spend a lot of time thinking about precisely how it will perform the grasp."

One downside to soft robots is that they tend to have trouble accurately measuring where an object is or whether or not they have successfully lifted it. To overcome this obstacle, the researchers created "bend sensors" that "hone in" on an object and send back location data based on the curvature they measure. This allows the robot to pick up an unknown object and compare it to clusters of data points it has collected from past objects it has come in contact with.

"As a human, if you're blindfolded and you pick something up, you can feel it and still understand what it is," said PhD candidate Robert Katzschmann. "We want to develop a similar skill in robots - essentially, giving them 'sight' without them actually being able to see."

In order to control the robot's gripper, the researchers use a series of pistons that work to push pressurized bubbles through the silicone fingers, allowing them to expand and bend. In the future, the team hopes to improve the model by adding more sensors that allow the robot to grip a wider range of objects.

"If we want robots in human-centered environments, they need to be more adaptive and able to interact with objects whose shape and placement are not precisely known," Rus said. "Our dream is to develop a robot that, like a human, can approach an unknown object, big or small, determine its approximate shape and size, and figure out how to interface with it in one seamless motion."

The work will be presented at this month's International Conference on Intelligent Robots and Systems.

WATCH: