Hank: When engineers are designing new technology, they often look to nature for help because evolution has had a lot of time to make some really great designs. A hundred years ago, we barely knew how to make an airplane fly, but insects have been doing it for hundreds of millions of years. So researchers figure, we can probably learn a thing or two from all that experience. If nature has already figured out how to do it, why do all that work over again?

There's a whole field of research focusing on creating technology based on biology. it's called biomimicry and it's particularly useful in robotics, because robots are just machines designed to accomplish a task, but often those tasks have already been done by a living thing like a human or an animal, like rescuing a hostage or flying through the air. So in one of the fastest growing areas of robotic science, engineers are actively studying nature to see how it could teach us to build better robots.

(0:48 SciShow intro plays)

If I say "flying robot" you might think of like the Iron Giant or like, the cute little Eve thing from Wall-E, but in real life some robots are designed to fly like birds and bugs. And there's a lot more to flying than just flapping wings, which is why a group of Swiss researchers modeled the AirBurr robot after insects. its designers were trying to solve one of the biggest problems robots face when they're exploring unknown territory: how do they get around without crashing into things?

(1:20) To avoid collisions, robots usually have to be able to tell where they're going, and they'll often build a map as they go. But mapping systems tend to be complicated, and fragile, and expensive, so if a robot does crash which happens kind of a lot, it breaks and it's kind of a big deal. But, getting inspiration from insects turned out to be a promising but unlikely solution because insects aren't really that good at flying.

(1:40) If you've ever looked at a light hanging outside on a warm summer evening, you might've noticed that bugs tend to swarm around it, like bounce off and then pick themselves up and do it all over again. Part of the reason they do that is because insects tend to have very limited senses. Since they can't see too well, they have to navigate using whatever cues they can pick up and big, bright lights would usually be a sign of a clear path.

But since they can't rely on their senses to navigate properly, they need fail-safes. So most insects don't get knocked out of the air by just bumping into something, and they aren't to bothered by crashing or falling, they just get back up again. Well that's what the AirBurr does. It's housed inside of a big, flexible frame that is designed to bump into walls and stuff and survive. If it does fall out of the air, four legs extend to get it upright so it can start flying again. In this way, it doesn't need complex systems to get around, it just bumps around but eventually makes its way, like those bugs on your front porch.

(2:30) Eventually robots like the AirBurr might help with robot search-and-rescue missions, flying through unknown debris-filled places and bumping into things as they go. Not exactly the Iron Giant, but the hope is it'll get the job done.

Another talent that's useful for navigating strange places, and you find it a lot in nature, is jumping. And one of the most scientifically mind-boggling jumpers in nature is the water strider, a type of insect that you might've seen leaping across the surfaces of ponds and streams.

A group of South Korean researchers decided to build a robot that could do the same thing, but first they had to study the physics of how water striders move. The insects take advantage of water's surface tension, a force that binds together molecules on the top of water. That force provides some extra resistance, so a water strider's four longer legs can lay gently across the water's surface without breaking through the layer and sinking.

(3:15) When the engineers filmed water striders jumping with a high-speed camera, they found that even thought the insect's jumps look sudden, it actually starts with a long preparation process. Instead of just pushing down against the water to get to force they need for the jump, the way we would push off from the ground, water striders drag their legs inward, toward one another. Then, the water acts like a stretchable rubber sheet. As the strider's legs move closer together, they push down on the surface making tiny indentations that bend it a little. And then, the force of the water trying to bounce back to being flat helps propel them into the air.

The water strider robots do the same thing, with a little electronic clasp that quickly draws in their legs and they work. The little one centimeter high bots can jump 14 centimeters into the air.

Jumping has also been important in developing a whole different branch of robotics: soft robotics. Now, most robots are made out of hard metal or plastic which are useful because it gives them structural support, but since there's very little give, they also break more easily. And jumping, if you think about it, is a lot like throwing your robot into the air and letting it fall to the floor, just over and over again. And if you did that with your cell phone I'm pretty sure you'd void the warranty. So soft robots get around that by being... squishier.

(4:22) A group of researchers from Harvard and the University of California in San Diego created a jumping soft robot that takes the cushioning concept one step further. Even soft robots have to have some hard components, like the electronics that make them work, but the problem is, the forces that interact between the softer and harder parts can break them. But here, too, nature seems to have found a solution.

(4:42) The team realized that many living things contain gradual transitions between the softer parts of their bodies like our skin, and the more rigid parts, like our bones. All in between those softest and hardest parts, there are many parts that are firm but still flexible. So they decided to build a robot that worked the same way, using a 3D printer to create a plastic robot that was soft on the outside but more rigid in the middle with a gradient of stiffness between them.

They printed the robot using nine different types of plastic, each with a different amount of stiffness, and ended up with a tougher core for the electronics which turns into a rubbery shell as you get farther away from the center.

The soft outside is useful both for protection and for the jumping itself, which works by inflating the bot like a balloon. When it's time for the robot to jump, the butane and oxygen inside ignite and suddenly expand, forcing the soft plastic outward and propelling it into the air. There's enough power in that to send the bot 3/4 of a meter high and 15 cm forward and then land without breaking. When the team dropped it from 1.2 meters up, it still didn't break, so they tried it 34 more times just to be sure and it still worked! The researchers figured that robots can also combine soft and hard components, letting them be fast and powerful while staying safe.

But if squishy jumping robots aren't your thing, and I don't understand why they wouldn't be, well maybe you'd like to join one for a swim. You could jump into the pool with a humanoid robot swimmer, like the one with possibly the most descriptive name ever, the swumanoid. That's a swimming humanoid. Swumanoid.

A group of Japanese researchers developed the bot because humans are actually pretty good at swimming but it's hard to figure out exactly what we're doing right when we're swimming through the water. Attaching sensors to a swimmer's body could tell you how all the forces work, but they wouldn't necessarily stay put while the swimmer was swimming, so swumanoid is a kind of simulator.

The team used 3D scans of human swimmers then installed motors into a humanoid looking bot that would let it swim the same way. But getting the robot to make all the same movements was a challenge all by itself, since our joints aren't easy to replicate in a machine.

Swumanoid has elbows that move back and forth in two directions like ours and shoulders that move in four. That lets it do the front stroke, backstroke, and butterfly, but it can't quite manage the smooth circles of the breast stroke yet.

Still, the robot can swim more than half a meter per second, about a third of the world record speed in the 100 meter swim, and the researchers are hoping that by studying its motion, they can work out the mechanics of swimming techniques and maybe even improve them. So swumanoid is a kind of twist on biomimicry, it uses robots to study how life gets something done which can then help improve the way living things do their thing.

Now, flying, jumping, and swimming are all more general types of behavior, and it's useful to have robots that can do stuff like that, but some animals have very specific traits, like the ability to change color at will, and researchers are trying to make robots that can mimic that, too.

Like this little master of camouflage. Developed by a group of engineers at Harvard, it's supposed to mimic and octopus, but it only has four legs and it's a little differently. Octopuses, like other cephalopods, change colors by controlling the size of special pigment containing cells called chromatophores, making certain pigments more or less visible. They use their color change either to hide from predators or to scare them away or to communicate with each other.

Creating cells that act like chromatophores would be incredibly difficult and probably not worth it, so instead this robot is hooked up to a machine with different colors of dye and researchers control how much of each color is running through its body to either blend into its background or to stand out.

It can also change the dye's temperature, so it can be camouflaged in visible light but glow in the infrared part of the spectrum, or use dye that glows in the dark. And it moves around by pumping air through its body and moving its legs forward.

Blending in would be useful for doing recon without being noticed, but standing out might be more important in search-and-rescue operations, so say a survivor in the ocean would be able to see it easily. The human-controlled prototype is easy to build and only costs about $10, but right now the color change process is slow, taking about 30 seconds to pump in the dye that adjusts the robot's color.

Future versions might cut down on that, and by incorporating miniature pumps, it might also be able to move around independently without being attached to a central system. Some day, the whole system could be incorporated into a robot that automatically detects its background and decided how to change its color in response, almost like a real-life octopus.

And these are just some of the technologies that we are trying to improve by learning from nature's billions of years of evolutionary experience. It turns out that when you combine biology and robotics, you get some useful robots, and they're pretty cool, and there are plenty tricks left to learn from.

Thank you for watching this episode of SciShow which is brought to you by our patrons on Patreon. If you want to help support the show, you can go to Patreon.com/scishow and if you just want to keep getting smarter with us, you can go to youtube.com/scishow and subscribe.

(endscreen)