Snakes have got to be some of the most creatively mobile animals ever evolved. They can move fast. They can move stealthily. They’re good climbers. They’re good swimmers. They can squeeze into very small holes. Some of them can even fly, a little bit. And all of this despite looking like a lizard that’s missing 100 percent of the limbs that it’s supposed to have.
Roboticists have been working on snake robots for a long time, primarily with a focus on versatile mobility in constrained spaces. With that in mind, we’ve seen a variety of limbless robots that can mimic snake “gaits” fairly well. But it’s not just the lack of limbs that makes snakes so special—it’s also their scales. In a new article in Science Robotics this week, researchers from Harvard show how mimicking snake scales with kirigami-inspired deformable materials enabled them to make a limbless soft robot that can crawl by simply inflating and deflating itself over and over.
A snake’s scales are all pointed the same direction, providing a substantial amount of favorable friction that makes it easier for the snake to move forwards than it is to move backwards. This makes moving backwards occasionally inconvenient, but it also means that the snake is able to achieve forward motion by generating a wave along its belly that first pulls its scales forwards, and then pushes them backwards. If snake’s scales had a symmetrical amount of friction, it would just move forward a little bit and backwards a little bit over and over. But since the scales are effectively slippery when they move one way and sticky when they move the other, the snake is able to move forward as long as it can get some grip on the surface. This is also how bristlebots work, incidentally.
The Harvard researchers from Katia Bertoldi’s group leveraged these “anisotropic frictional properties” of snake scales to turn the repetitive pulsing motion of an inflatable soft robot into forward motion, in much the same way that snakes can crawl forward on their bellies without using their trademark side-to-side slithering motion. In order to make scaly skin, the researchers manufactured a variety of different stretchable plastic sheets, each laser engraved with a unique pattern of flat scales. The pattern was structured such that when the robot inflated itself and the sheets stretched, the flat scales would deform and pop up away from the robot’s body, gripping the ground and turning that inflation and expansion into forward motion. It’s simple, it’s cheap, and it’s effective.
The researchers tried a bunch of different scale patterns to find the one that worked the best, including linear scales, circular scales, triangular scales, and trapezoidal scales:
You can compare these designs to the scales on an actual snake—my black pine snake, Satin:
It turns out that the most efficient design (for this particular robot, anyway) is the trapezoidal one, but apparently not because the trapezoidal shapes generated more friction—rather, the trapezoids allowed the skin to stretch more, which meant that the robot would elongate more when it inflated, yielding a longer “stride.” As long as the scale design can effectively anchor themselves in place against the ground while the robot pushes itself forward, moving quickly becomes about stretch instead.
The researchers also noticed that their morphable pop-up scales came with a downside, which was that the amount of friction that the scales could generate decreased significantly when the robot was in its deflated state since the scales all popped back down again. This can be solved somewhat by just never completely deflating the robot, but the researchers are also experimenting with a different technique. They found that over-stretching the triangle scale design strained the plastic skin to the point that it permanently deformed a little bit, meaning that the scales remain popped up even when the robot deflates.
We found that this plastically deformed skin affected the response of the system in two ways. On the one hand, it increased the elongation that the crawler experienced at the beginning of the inflation process. On the other hand, its permanent directional texture resulted in highly anisotropic frictional properties through the entire actuation process. Hence, the efficiency in locomotion for this crawler was optimal when the supplied volume [of air] was cyclically varied between 0 and 12 ml, resulting in ~22% improvement in comparison to the best performance of the corresponding crawler with a purely elastic kirigami skin.
While this research is focused on the skin and scales, we very much appreciate that Harvard took the time to make it work outside on an untethered robot. They point out that their method is easy, affordable, and, yes, very scalable. It works for miniature robots and much larger ones (with thicker and stiffer sheets), possibly even as larger as planetary rovers. It’s also exciting to consider non-mobility applications, like applying this kind of skin to things like inflatable robot fingers to adjust their grip on demand.