Stanford Engineers Develop Gecko-Inspired Robotic Hand Grip
Stanford University has made great progress over the years with its gecko-inspired robotic hand. In May 2021, a version of the "gecko grip" even went to the International Space Station to test its ability to perform tasks such as collecting debris and securing satellites.
A huge number of robotic hands and grippers share a common enemy, what scientists call the "heirloom tomato problem." You may have seen a robotic gripper deftly pick up an egg or handle a basketball well, but unlike human hands, a single gripper is unlikely to do both, and the key challenge remains hidden in the middle.
“You will see robot hands make a powerful grip and a precise grip, and then assume they can do everything in between. We wanted to figure out how to create manipulators that are both dexterous and strong”
The result of this goal is "farmHand," a robotic hand developed by Ruotolo and Dane Brouwer (a graduate student at the Biomimetics and Dextrous Manipulation Lab at Stanford, also known as the "Farm"), and described in detail in a paper published in Science Robotics. In their testing, the researchers showed that farmHand can handle a wide range of objects, including raw eggs, bunches of grapes, plates, jugs of liquid, basketballs, and even an angle grinder.
FarmHand benefits from two types of biological inspiration. While its multi-jointed fingers resemble a human hand, albeit with four fingers, the fingers are coated with a gecko-like adhesive. This sticky but not tacky material is based on the structure of gecko toe pads and was developed over the past decade by the Biomimetics and Dextrous Manipulation Lab led by Mark Cutkosky, a professor in the School of Engineering who is also the senior author of this study.
Using gecko adhesive on a multi-fingered anthropomorphic gripper was the first challenge that required special attention to the tendons controlling farmHand's fingers and the design of the fingertip pads beneath the adhesive.
Like gecko toes, gecko adhesive provides strong grip through microscopic flaps. When fully in contact with a surface, these flaps create van der Waals forces, a weak intermolecular force arising from subtle differences in the position of electrons on the outer side of molecules. As a result, the adhesive can grip strongly, yet it requires only a small force to release. Another bonus: it doesn't feel sticky to the touch and leaves no residue.
“Early applications of gecko adhesive involved robot climbing, human climbing, or gripping very large and smooth objects in space. But we always thought about using it for more practical applications. The problem is that gecko adhesive is actually very finicky”
The point is that gecko adhesive must attach to a surface in a specific way to activate van der Waals forces. This is fairly easy to control when it is applied smoothly to a flat surface, but much harder when several sticky parts must contact an object at different angles to grip, as with farmHand.
The farmHand fingertip pads, located below the adhesive areas, help solve this problem. They are made of a folding ribbed structure that bends with little force. Regardless of location or contact angle, the ribs continuously bend to deliver equal forces across the sticky pads and prevent any of them from slipping prematurely.
“If you shift these ribs, the bending results in the same force regardless of where you start. This is a simple physical behavior that could be applied even outside robotics, for example as a shoe tread or off-road tire tread”
The tendons of the hand are also critical because they allow compression to be stretched excessively. While many robotic hands and grippers squeeze objects in a "C" shape, as if lifting something with fingertips, farmHand presses a pad with its fingertip against another pad. This gives the adhesive a large surface area to work with.
Getting the right design was especially difficult because existing computer simulations make it hard to predict real performance with soft objects, another factor in the robotic gripper problem. But the researchers greatly benefited from being able to 3D print and test many rigid and soft plastic components in relatively quick cycles. They go so far as to suggest that their success might have been impossible, or at least much slower, just five years ago.
Further improvements to farmHand may come in the form of feedback features that help users understand how it grips and how to grip best when using the hand. The researchers are also considering commercial applications of their work.
