Berkeley Talks transcript: Learning from nature to design better robots

Listen to Berkeley Talks episode #148: Learning from nature to design better robots.

[Music: “Silver Lanyard” by Blue Dot Sessions ]

Intro: This is Berkeley Talks , a Berkeley News podcast from the Office of Communications and Public Affairs that features lectures and conversations at UC Berkeley. You can subscribe on Spotify, Apple Podcasts or wherever you listen. New episodes come out every other Friday.

[Music fades]

Matthew Shears: Good morning, OLLI at Berkeley, and welcome to our Friday special event series. Today, we’re very excited to be hosting professor Robert Full. Robert is a Howard Hughes Medical Institute professor in integrated biology. He received all of his degrees from SUNY Buffalo before joining the Berkeley faculty in 1986. He directs the Poly-PEDAL Laboratory, studying animal motion science to inspire the design of novel materials and robots.

He is Founder of Berkeley Center for Interdisciplinary Bio-inspiration in Education and Research, serves on the National Academies of Sciences Board for Life Sciences, and is an elected fellow of the American Academy of Arts and Sciences. Robert received Berkeley’s Distinguished Teaching Award, is a National Academy of Sciences Mentor in the Life Sciences and leads education programs whose goal is to expand the STEM workforce by fostering inclusive excellence. His talk today is called “Bio-Inspired Design: Compressed Cockroaches, Gliding Geckos and Smart Squirrels.” Take it away, Robert.

Robert Full: Thank you, Matt, and welcome everybody. Thanks so much for coming. Here are some of the creatures that I’ll talk about. I’ll start off by explaining my motivation. That’s me when I was young, and we used to come from western New York and drive all the way to the beaches of Florida during spring break.

And I really liked weird animals, like these crabs and cockroaches. I also had a lot of interest in BioMotion, because that’s what I really wanted to do. I wanted to play Major League Baseball and I tried out for the Minor League Pittsburgh Pirates. And well, I’m here and not there, but I did, when I came to Berkeley, got to, using the same techniques I’ll show you, analyze the motion of a pitcher from the Oakland Athletics, which was pretty fun.

As mentioned, I got all my degrees at SUNY Buffalo, partly because I was a first generation college student. My parents didn’t finish high school and I didn’t really know what college was. But SUNY Buffalo was great, and I had a wonderful mentor. I went and did a postdoc at the University of Chicago, a little bit further west, a little bit less snow, a little more windy. And this is my 35th year at Cal, and I feel incredibly fortunate to have that journey.

I’m in the Department of Integrated Biology. What is that? Well, we integrate the level of organization of organisms from genes to ecosystems. We look at a diversity of organisms, and then we also look at them through time, through evolution. And we’ve gone on to integrate beyond the biological disciplines with engineering, physics, chemistry, mathematics and more.

What this allowed us to do, is to really look at this integration to show, really, how it showcases nature as a library of designers. So, this is the area I’ll talk about… some call it biomimetics or bio-design. It’s just basically learning from nature, and here are some of the classic examples, like Velcro and flippers that led to wind turbines, and Lotus leaf, the wonderful paint, and I’ll show you our gecko example. This area, biomimetics, is becoming the leading paradigm for the development of new technologies that’s leading to significant scientific, societal, and economic impact. If you look at the publication rate, it’s pretty extraordinary, it’s doubling every two to three years, whereas most areas in science it’s about a dozen.

There was a journal founded in 2007, I got to be that founding, called Bioinspiration & Biomimetics , out of London in the Institute of Physics. And I’ve been really fortunate to be its editor in chief since 2013, and it has wonderful examples of this area, and I get to see everything from around the world, all the different discoveries and how they’ve been translated. And the area in particular of most rapid growth, is bio-inspired robotics.

And so, that’s what I’m going to show you some examples of. There’s really enough, like a zoo of amazing robots, and every day there are new ones that are being discovered. And so, to this end, we created the center called the Center for Interdisciplinary Biological Inspiration in Education and Research, to look at this. I’ll show you the research, but I will talk about the education at the end.

And more recently, American Association of Advancement of Science, AAAS, founded another journal in 2016 called Science Robotics , because robotics was so game changing. And Berkeley got the inaugural cover showing here from my colleague in Electrical Engineering Science, Ron Fearing, for his hopping robot. I’ll show you a video of that at the end. And so, I’m very fortunate to be on its science advisory board, to participate in the publishing of the wonderful discoveries. So, what I want to do this morning, is to basically show you the areas of the field through the discoveries we made and the lessons we’ve learned from nature.

So, I’ll go through these four. The first lesson we’ve learned is that, animals are really good at managing energy effectively. So, this is the area I’ll talk about as bio-inspired field robotics. And so, this is what we do in my lab. This is a cockroach running on a treadmill. And you probably should ask, “Why are you doing this?” This is another cockroach. This is an eight-legged scorpion. Forty-four legged centipede running on a treadmill. And an eight-legged sideways running ghost crab, the ones I used to chase on the beach. Fastest terrestrial invertebrate. And we want to understand their motion. In fact, the mantra of our lab is “diversity enables discovery.” Looking at these diverse creatures gives us hints about novel principles.

Now, in terms of movement, you might think, “Okay, well, if you get more legs, are you then closer to the perfect way to move a wheel?” And many people have hypothesized this as one possibility. That’s not what we see at all. What we find is that, running with legs is like a bouncing ball. You actually take the energy of motion, you store it in a spring, and then you release it back. And you say, “Well, how could that be for these different creatures?” But here’s what we discovered. We discovered 2-, 4-, 6-, and 8-legged animals all bounce along like a Pogo stick. We call it a mass spring system. So, your one leg works like two legs of a trotting dog working as one, or three legs of an insect if you use tripods, or four legs of a crab, all bouncing along like this kangaroo. And they all have about the same relative stiffness when we got this.

We wondered then, since robots couldn’t really venture outside, if we took the principles we discovered here from the creatures and made a robot with a leg that was similar in terms of the stiffness of a spring, could it run outside? And here’s the first robot that could maneuver effectively outside using legs. We called it RHex for Robot Hexapod, it can go two or three meters per second.

And here’s the commercial version of RHex, it’s owned by Boston Dynamics. You may have seen that they also make humanoid robots and dog robots that do routines like dancing and things. But this one was tested to show the advantage of legs over 40 different surfaces, and in this difficult terrain, it’s far superior to anything with tractors and wheels. So, we really understood the advantages of legs in rough terrain. We also took it to even rougher terrain in Capitol Hill, to show the benefits of bio-inspired design and the wonderful capability of this area of being inspired from nature. So, there’s RHex on the steps.

Now you say, “Well, why is my taxpayer money going to this? What can this really accomplish?” Here’s just one scenario to show you what we’ve been working on for years. Here’s Rhex, and you may remember there was a subway explosion years ago in Spain, a series of bombs, and they didn’t know what to do, the first responders. Is it chemical? Is it biological? Is it nuclear? The idea is to send the robots down and to begin to detect what’s there quickly. And now, RHex also has a little arm on it. It’s trying to save people, so it doesn’t have to pay, it just goes right in and analyzes what’s there, or can take a sample and then quickly return it.

In addition to the ground robots, there’s now a coordination with drones that can also fly in and have a different perspective. So, RHex now uses a LADAR or laser range finder to find out where the stairs are, and then it actually changes its gate from that tripod gate to a wave gate in order to maneuver back up the stairs and return the sample. So, you can imagine a bunch of these coming in, in whatever entry point is available, to quickly determine what the risk is, what the hazard is.

Now, in our study of the cockroaches, we found something quite unusual. Which is, when the American cockroach goes really fast, it’s bipedal, it runs on its back two legs. And the pattern of bouncing is still, whether it’s six legs or two legs, just like what you see when you run. And so maybe, actually, cockroaches were the first biped. It’s possible. And of course, to demonstrate this general principle, if you work with amazing engineers like we do, they can actually make a robot do this too, even when they’re six-legged. And this is a bouncing robot using a [inaudible].

In fact, the principles we’ve extracted are so general, that they can be used to set the stiffness of prosthetics, to allow people to regain function. And so, the principles are so general, that they are part of the influence of many fields, and we’re incredibly excited about the potential application of these principles. In addition, when I published that paper, that cockroaches can run on two legs, I had a call from the company called Pixar, and they said, “Hey, we’re making a movie about bugs, and you have bugs running on two legs. Could you help us make the characters?”

And so, that’s what we did. We spent a lot of time making all these characters and helping them, which was pretty interesting, pretty exciting. So, here’s one of the Bipedal bugs. And we learned a lot about how we use motion to develop characters, not copy their motion, but develop their personalities. And so, that was a wonderful collaboration. If you remember anything from my talk, this is what you should remember, which is: You never know where curiosity-based research will lead. And hopefully, just in that first example, I’ve shown you that you could not anticipate those kinds of connections.

All right, the second lesson I’ll talk about, is the fact that animals and the environment are really inseparable. Even though we study things in the lab, we know that we need to look at the environment. So, I’ll talk about attachment, and show an example of the incredible advancement in bio-inspired materials. In fact, there’s so many things being developed, here’s just a list of the different capabilities of the bio-inspired materials and the creatures that are inspiring them. And this list is just, every day, new discoveries come out about the wonderful capabilities for biomaterials. And that’s in part because, natural technology are now finally moving to a place where nature can be a wonderful teacher, and that’s happening at light speed.

So, attachment. What was our inspiration? Not this guy, but this one. So, this is a gecko. It’s a gecko, gecko. And, as you know, geckos are really good at sticking to things. So, here’s one from a TV show from France, where they’re climbing up a plexiglass sheet and slipped, and it’s holding on by one toe, and doesn’t seem to be too disturbed. So, it can hang on by one toe, it just puts the other one down and reconnects.

So, we’ve looked at their feet, and what we found is that, they look like alien feet. These are real gecko feet, and they’re bizarre. They have bizarre toes. Why is that? We still don’t quite know why that is. That would be a nice follow up, the next Ph.D. project, but we’ve learned some about how they stick with these remarkable toes.

And so, here’s a quick video of what we found, that’s about 10 years of work in a few seconds. So, if you zoom in on a gecko foot, what you’ll see is, they have these leaf like structures called lamellae. And if you zoom in on a lamella, you see the secret. They have huge numbers, millions of hairs, and each of the hair have the worst case of split ends possible. A 100 to a 1,000 split ends, where they become nano size. And so, this is really a nano materials project. And so, of course, our question was, “How do they stick?”

And so, what I told an undergraduate at the time, Tonia, you just have to get the hair and measure the force that the hair can generate, despite the fact that you can hardly see the hair, it’s so small and you certainly can’t easily manipulate it. Well, she didn’t know I was joking, and we were working at the time with Stanford on a very fancy micro-electromechanical system, a MEMS, a sensor device for the smallest forces, which is quite a technical achievement. Tonia went, took the hair with her surgical hands, glued it to an insect pin, took a wire, and then pushed it on a wire and bent the wire, and measured how much force it required the bend to wire. And she got the same result we did, so she actually measured the force of a single deck of hair.

And I would say, throughout my career, that undergraduates are our secret weapon. And that’s because they don’t know what can’t be done. And that is incredibly important when you go for creativity. Yes, a lot of times they do things that aren’t productive, but they come up with things that we don’t see, because our mind is fixed.

And so, I’ve had just amazing undergraduates at Berkeley, and I feel so fortunate, I’ll show you some of them, but here’s Tonia. And so, what did we find in looking at this and making these force measurements? What we discovered was that, the gecko hairs don’t stick by Velcro, or by suction, or by glue, or capillary action. These 2 billion nano-sized split ends stick by intermolecular forces, called Van der Waals forces.

It’s just hard to believe, it’s just amazing. But what we did is, we went through and, of course, got a physicist who works on Van der Waals forces. We gave him the challenge of, “Give us an experiment that we could do that would convince you that, that’s really what’s happening.” And he did. He didn’t believe that it would work at all. It did, and of course, he wanted to be on the next paper, which he was. So, it was a wonderful, interdisciplinary collaboration.

Now, if you look at these, this is just one gecko. There’s a lot of geckos, and we are incredibly fortunate to have museums in our building, in the Valley Life Sciences Building on campus. Here’s one of them, The Museum of Vertebrate Zoology. And it literally is a library of design ideas. It’s not a public museum, but it has incredible research resources, where you have access to biodiversity and you can see how important it is.

Here’s my former graduate student and PD, who says, “Okay, I’m going to look and see all these different geckos and how different their hairs may be.” And that’s what she did for part of her thesis, and what she found, was this incredible diversity. So, the one I showed you is on the left, but all of them have different hairs of different structure, size, thickness, different numbers of split ends. Why? Honestly, we still don’t know. My guess is that it allows them to stick better to particular surfaces, but we don’t know. Tremendous diversity that we want to further explore.

So, what we end up doing then, is to take, actually, one of the simplest creatures. It wasn’t even a gecko, it’s a lizard, but it just had a hair without the extensive branching to study. And then, working with my colleague, Ron Fearing in engineering, he was able to make one of the first synthetic, self-cleaning, dry adhesives learning from the gecko, a nano array, that had a lot of the properties of what we see in the gecko.

Now, it takes a pretty long time to translate this into products, but one of the former postdocs who now directs the Max Planck Institute in Germany, started a company called NanoGripTech. And now, finally, it’s advancing in its association with another company to begin to make gecko-inspired adhesives for wearables, for different types of sports, as well as medical equipment. And so, here is the first human supported by a gecko inspired adhesive. That’s my former graduate student, now professor at Lewis & Clark, on the left, and that’s his firstborn child. This was done with the group from Stanford, incredibly exciting. And the reason we’re protecting her, is because the gecko has an interesting property. It sticks when you pull down on it, but it releases when you push up. And so, we didn’t want it to be bumped and have her ever fall.

And she’s going like, “Dad, what are you doing? Why am I doing this?” It was pretty funny, but we could show the incredible effectiveness of this kind of adhesive. And now, then you might ask, “Well, then how do they actually let go?” It’s also an important question. Here’s what they do, which is pretty weird. So, they peel their toes, in milliseconds, as they run up surfaces, and they just do this very strange thing. It works well, and we understand the mechanism by which this peeling occurs to allow them to detach. While working with our wonderful colleagues at Stanford, they designed the first robot named Stickybot, which is up on my wall, actually, that could climb up smooth surfaces. So, here’s Stickybot climbing up with these gecko inspired adhesives on its toes.

And they decided that, even though they thought it was sort of silly and ridiculous, that peeling thing wasn’t a bad way to pull off their toes. And so, this is a remarkable team, Mark Cutkosky and now Sangbae Kim was at MIT, amazing robotics engineers. It just shows how all of our projects, to really move things forward, end up being interdisciplinary teams, where each team in different disciplines benefits with their own field by the interaction, and the collective discoveries are beyond what any single team could do. So, no individual team could have done this, but the collection does, and that’s the way we do our science to make breakthroughs. And so, who would’ve imagined, 10 years later with Tonia, that we would be briefing the United States House of Representatives, a STEM Ed Caucus on Undergraduate Research and American Innovation.

Extraordinary, although it’s quite a challenge to be at Capitol. But again, I hope this shows you another example that, literally, you never know where curiosity-based research will lead. And so, in the geckos, Anne PD, that student I mentioned, she was trying to discover how they could run upside down. And it was hard, because they kept letting go. And she said, “When they let go, they always fall on their bellies. They go into this skydiving Superman posture.”

I said, “What? Okay, let’s look at this, let’s make them do that and see what’s going on.” And so, we put them underneath a surface, shook it a little bit, and here’s what we discovered. Now, that’s real time. If we slow it down, you can see what they do. They use their tail to do an air-righting response. They just swing their tail around, and then they’re in the skydiving posture. Now, this was unusual, because they were never have shown to glide in nature, they don’t have the gliding adaptations of other lizards. So, with my colleague, Robert Dudley, we used his wonderful vertical wind tunnel, so you’ve probably seen people practice skydiving in these, the air blows up, and we put a little tree trunk outside it, and then we wanted to see, could they glide?

And what we found was, absolutely. They could do, actually, beautiful equilibrium gliding in this wind tunnel. And it was amazing how they could be so stable. Here’s a picture of it. So, here’s the fact that they can glide, but you can also see their wonderful adhesive toes there, the lamellae, and here’s what they could do with their tails. So, it turns out they could steer with their tails. They swing it one way, they turn left, they swing it the other way, they turn right. An amazing capability with their tails. And so, of course, then we had to see if they glide in nature. And so, picking another graduate student, Ardian Jusufi, off he went to Singapore and Southeast Asia to look at these creatures and tried to see if he could see them actually gliding in nature.

So, here’s the first glide in nature that we captured. The animal’s at the end of the red line, and then, if you look on the right, you’ll see it hitting a tree. All right, let me show you what happens when they collide with the tree. It’s unbelievable, actually. They’re going six meters per second, and watch how they interact with the tree. They crash head first into the tree, and then they go head over heels backwards and they use their tail for stability. Because they’re small, they don’t get injured, and that’s the way they can adapt for landing. Ardian was able to build a physical model of this landing, a soft robot gecko, where we can manipulate the parts to test whether or not the tail was important and, really, how they attach to surfaces. And could this be a landing mechanism for vehicles that are gliding or flying?

Okay. So, that one is on bio-inspired materials. Let me talk about the next one, one of the hottest areas, are bio-inspired soft robots. And this is an area where it’s really important, because we’ve made a lot of great robots, and then you take them outside and they break. They’re really not usable. And so, one of the distinguishing features of animals, is that they’re so robust. They’re incredibly robust, and so the whole industry of designing devices is shifting from rigid structures to soft structures, like we see in the animals. In fact, it’s just an explosion again. And so, if you look at all of these different bio-inspired soft robots, and the number of publications in the new field of soft robotics, it’s just taken off. And in our journal, that’s one of the most frequent papers that we publish, is that capability of soft materials that you can use for so many different things.

And so, what did we wonder about? We wondered about the cockroaches. And you think of them as crunchy and things, but they’re actually good soft robots. Their skeleton provides support and protection, it anchors their muscles. They can locomote with it, they can sense with it. But if you look at the exoskeleton carefully, what you see is that it’s really just a bunch of skeletal plates and skeletal tubes with a compliant membrane attached to it. And so, again, collaborating with my colleague, Ron Fearing, he pioneered a new way to design a soft robot, but an origami approach called Smart Composite Microstructures. So, what he was able to do, is to take a design drawing, laser cut it, laminate it, and then fold it up to make a robot.

And so, here’s one of the first robots called DASH, for Dynamic Autonomous Sprawled Hexapod. And you could then design in robustness that would allow it to really function for a task in nature. And so, you have to be careful what you tell graduate students, because you said, “Oh, what happens if you throw it off the top of a building?” So, they did that with the computer science field, to hear their thesis hitting the ground. And we’re wondering, “Is it still working?” It’s still working. That’s pretty robust. Even more extraordinary than that, the group of graduate students realized that, “Hey, we should form a startup. This could be pretty neat for a design of an educational robot.”

And so, they did. They did a Kickstarter, developed a wonderful design, called it Dash Robotics. And amazingly, it worked really effectively. So effectively, that Mattel picked it up for a toy robot, they called it Kamigami. Here’s the actual robot. And I’ll tell you later, the cool thing is, in classes at Berkeley, the students get to build this origami robot from flat sheets, test it, and redesign with it. Okay, so that’s the notion of what you might do for designing with this incredible soft robotics approach.

Well, my graduate student, Kaushik Jayaram, wondered, “Can this explain how they seem to be able to go anywhere?” And so what we did, is we said, “Okay, let’s check that out. We’ll CT scan them, and we’ll have them go through the smallest crevice that we could.” They can go through a crevice that is the height of two stacked US pennies. They can compress their body by over 50%, which is extraordinary. How do they do this? And so, we took a look at this. So, here is the cockroach standing up on the left, its normal, free standing height. And we went through and looked at the smallest gap it could go through, which is three millimeters. And so, you’ll see that going through it in real time.

So, they can squish through almost anything, because they don’t have hard parts like a mouse might or something. They just squish, and they’re really good at going through these tiny spaces. How good are they going through them? Well, they can go through these tiny spaces, three millimeters or four millimeters, and my students wondered, my undergraduates particularly said, “We wonder if they could locomote through vertically compressed spaces.” So, it’s like two plates, one on top of the other, and go into that space and then run through it. And then I said, “Well, that’s silly. They’re squished in half, their feet can’t touch the ground, they can’t move. Now, that’s a waste of time to try to do that kind of experiment.”

And they said, “We already done it. And here’s the results.” Which is just amazing to me. So, the top one is undefined locomotion, so there’s no top to it. And it’s real time, which is so fast you can hardly see it, because they can run a meter and a half a second. And then, we slow it down 20 times. So, here’s the unconfined version, it’s just really fast. And you can see they’re moving along with it with a particular gate, this alternating tripod.

The next one is where it’s compressed a bit, still really fast, but now you can see how they’re altering their gate, and they’re using their legs differently. They’re using different parts of their legs. The bottom one is the smallest confined space that they could move in, four millimeters. So, I’ll first show it real time, and then I’ll slow it down 20 times. Okay, so this is real time. Now, it’s slowed down 20 times. So, yes, it is the case that in the smallest spaces in your walls, in your ceilings, in your floors, that they can zoom around that fast. And what we found is that, they’re not using their feet at all. They’re compressed and they’re using the spines on their leg to have sufficient friction in order to move.

So, then we wondered, “How robust are these?” Well, if you’re an engineer and you want to test that, you put them in a materials testing machine. So, we did that. And so, here’s the test of their robustness. No cockroach was harmed in this experiment. We tested them, actually, flying and running before and after this, and this did not do anything to them. So, we looked at their stress strain relationship at different rates. And what we discovered, remarkably, was that even at 800 times their body weight of squishing them, they weren’t injured. So, it’s just amazingly robust.

So, we said, “Okay, we have to build something like this to test, to see if it can also withstand this kind of compression.” So, my student looked at the architecture of the creature and built one of these origami robots. He called it CRAM, for Crawling Robot with Articulated Microstructures. So, here’s CRAM freestanding, and here it is compressed, and here’s a video of CRAM running uncompressed, and then compressed. In our interactions with first responders in FEMA, they told us, at first it looked much more like a cockroach originally, and they said, “You can’t do that. Imagine someone’s trapped and a cockroach is coming at them.”

And so, we made this shell that we think is a little bit easier to interact with. But it can do it, it can move, even though it’s compressed 50%, because of this wonderful nature of being a soft creature. And so, we now envision that bioinspired robots can have tremendous impact on conservation, monitoring, inspection, search and rescue, security. And we’re looking at this with professor Robin Murphy, who runs a wonderful center for Urban Search and Rescue, and people who’ve interacted with FEMA. So, we’re really excited about that capability of getting information that’s important really quickly.

Again, I hope that this is another example of the fact that you never know where curiosity-based research will lead. We just wondered how they could squish through stuff. And that’s amazing, actually, to then be able to take those principles and use them. Okay, the last one I’ll talk about, is our ongoing project, now our recent project, on the fact that animals can learn. And we call this area Cognitive Biomechanics, so we’re talking about bio-inspired learning, decision making, and even creativity and innovation. And so, you may have heard a while back, that DARPA had a robot contest called the DARPA Robotics Challenge.

It was challenging. It’s actually really hard to get robots to operate in unstructured environments. I know you see them doing amazing things, but they didn’t have three months to practice on the same situation, same scenario. They had to do novel environments. And it’s really hard, and yet, not so hard for creatures. And so, what’s the difference? That’s what we wanted to find out. So, this is a big challenge. So, we call this cognitive biomechanics, and the animal we selected is a squirrel. So, if you’ve ever been on Berkeley campus, there’s a lot of squirrels and people feed them all the time. And that was an advantage for us, because it turns out the squirrels don’t perform if you put them in small little cages.

So, we actually wanted to study them outside, and it turns out, we had one of the world’s experts, professor Lucia Jacobs, on squirrels from the psychology department, just incredible wealth of knowledge, that we worked together on this project. And so, why are we looking at squirrels? It’s because they can do this. It’s amazing, actually. How do they know what their capability of their body is? How do they know that the next branch is going to support them? Because if they make a mistake, the Hawk eats them. And so, how do they know this? It’s an incredible capability. And so, recently, Nate Hunt worked on this for his Ph.D., he’s now professor at University of Nebraska. And we published it in the journal, Science.

And so, here’s what we did. So, we took this magnetic board with perches on it, and we wheeled it outside to the eucalyptus forest to the side of our building. And so, here’s the setup. They usually take turns, but that one tried to cheat and cut in line. They take turns, and we give them peanuts. They like peanuts, and we train them a little bit to begin to get used to the structure, as you see here. But this is free ranging, this is outside, you can see a person, students walking in the background. And what did we do? Here’s one of the key experiments. And so, what we did is, we said, “Okay, we’re going to give them rods or branches of three different stiffnesses, put a peanut in the landing perch, and have them jump to get the peanut.

And the question was, “What would they do on these different rods of different stiffnesses?” Because if it was important for them to jump off the stiff part, then they have to jump a long way. Whereas they may want to run down to the more compliant part, so they only have to jump a short distance. And so, we hypothesized that there was a trade off between their gap size, how far they have to jump, and how compliant their surface, the branch was they’re jumping off of. We wanted to try to interrogate their brain about how they’re making these decisions. And what we found was, sure enough, animals selected a shorter gap distance when the branches were stiffer, and a longer gap distance when the branches were compliant. But what we assumed is that, “Oh, well, they’re probably equal in their decision making, in their trade off. They’re both equally important.”

This was completely wrong. And we could interrogate their decision making, and it showed with a simple model, that the animals cared six times more about jumping off a stable surface, than how far they had to jump. And so, that’s a beginning to try to understand their decision making. And then, we did another experiment. We did an experiment where we had a very stiff platform, we had them jump off of it, they landed perfectly every time from the stiff platform, not a difficult task. And then, we made the platform look exactly the same, except it was incredibly compliant. And so, here’s what happened for their first jump.

We did 652 trials and they never fell, but they just made it in this one. And it turns out that, all it took were five jumps until they were able to learn to land perfectly on the perch. So, a remarkable learning for a pretty difficult task. And so then, we were pretty excited and we said, “Okay, let’s do another experiment. Let’s do a transfer learning. Let’s move the perch back, let’s move it up, so it’s harder. And let’s see if the ones who did it before do better than the ones who didn’t.” So, we have the high speed cameras rolling, we’re collecting data, ready for the first trial. And here it is.

It parkoured off the wall, and added a control point. It was innovating right before. And of course, it landed and looked at us like, “Okay, we’re squirrels. We can get into bird feeders.” But it was remarkable, because that was not something that they did before. And so, we have lots of examples where they innovate like this, where we’re trying to do a particular experiment, and it’s more important that we find out how they get around it, to understand the cognitive part of it, and so we have lots of examples of that, which is pretty amazing and incredibly exciting.

And so, we have a wonderful interdisciplinary team working on this, of people that work on metamaterials, ones that are actually recording in rats, from the hippocampus, to try to understand what the brain can tell us, and we have roboticists. But we’re just in the middle stages of this development, and I can’t show you the agile robot yet. The most agile robot ever built is our goal, and we’re not there, but we have some hints about it. And so, let me show you the latest robots that we’re going to mesh together, that hopefully will also allow this kind of decision making and thinking.

So, here they are. So, the one on the upper left is the fact that we found spines are probably pretty useful. None of the robots we see have built the spines, and then there’s one that can maneuver and jump on the bottom. Both of those from our lead at the University of Pennsylvania. And then the upper right is a robot called Salto, from a grant we worked with Ron Fearing on at Berkeley, making just a monopod jumper. So, these are all the beginning robots to think about decision making, learning, and innovation. So, you can see the backbone bending one, you can see the one on the bottom doing jumping, and the monopod hopper just hops.

And so, they do power amplification for animals to that hopper. And so, we’re trying to put all these together to give it the agility that we see in the squirrels. And what partly we found was that, much of the capability of the control, is built into the morphology itself. So, you also have to have a brain and a control system, but so much of it is built in the morphology. We’re calling this computational morphology. And the more we work with the metamaterials people, the more we realize that you can build a lot of intelligence right into the materials, which is what we find in the creatures. So, it’s a very good lesson. So, another example of where you never know where curiosity-based research will lead. Stay tuned, we’ll see what we can develop for these in these squirrels.

So, let me finish by telling you how we try to share our approach in research, with these wonderful interdisciplinary teams, with students. How can we set up a program to do that? And that’s what HHMI has supported us on. So, I got this wonderful HHMI grant, and the program I called Eyes Toward Tomorrow. And so, it’s involve, imagine, invent and innovate. And what it is, is it really is the wonderful integration of discovery-based learning.

So, all our classes, we try to have students make original discoveries in the classes, through labs and field experiments. And it’s combined with design-based learning, which is what Obama really pushed that for a nation of makers, because the real power of the revolution is in its democratizing effect, because now almost anybody can innovate, and that’s what’s happening.

And we strongly believe that diversity is demanded for that innovation, that the real promise rests in uniquely diverse communities that are engaged in these interdisciplinary approaches to solve important societal problems. So, to this end, I created a course called Bio-Inspired Design, which, in fact, the labs are going on right now. As we speak, the students are doing a wonderful MakerSpace activity where they’re designing novel splints for finger injuries. And in the past, they’ve built a wonderful prosthetic device, I’ll show you that. But this is open to everybody, there’s no prerequisites.

Our aims are science literacy, that is we have the students use the primary research discoveries and extract the principles. We focus on giving them 21st century skills and teaming and community building, as well as a career path, a connection to the future. We have 180 students, 36 design teams, half freshmen and sophomores, half non-STEM students, the majority female, and about a quarter underrepresented and marginalized students, from over 40 different majors. So, we have incredible diverse interests across the campus, to allow people to get together and share their unique perspectives for discovery and innovation.

And we do it in the Jacobs Institute of Design Innovation. We have incredible space thanks to Paul Jacobs, a four story MakerSpace building for undergraduates, where they can do 3D printing and laser cutting and all these wonderful things. And there are interdisciplinary design related communities there. One of the things we did, is to have them initially 3D print a prosthetic hand for children. So, that looks like this. And they first printed a finger. And their assignment was to 3D print their finger, and then stand by the 3D printer and take a selfie. And they really liked this exercise.

And then, as a team, we had them assemble all the fingers onto the hand. And so, it was a great way to introduce them to the MakerSpace. And then, they have several design projects. They make, as I mentioned, the origami robot, they also make a gecko inspired adhesive and design things from it. Then they have a final project.

Here’s examples of a few of their final projects. One team developed a gecko-inspired tail balancing system for the elderly. Very cool. Or a spider-inspired sensing to restore touch in amputees. One on muscle atrophy-resistant casts based on seahorse skeletons. So, you want to keep your movement in ways that don’t harm you, but you want to restrict movements in other directions, is incredibly clever. And one on a voice recovery system for throat cancer patients, based on songbirds. They use a different part of their larynx to produce sound. Just amazing project.

So, they all do this showcase, they do a video and do a poster to show off their stuff. In the larger picture, that’s only one piece of what we’re creating, that’s the Bio-Inspired Design course, which is part of a student-centered creative action community, we call it, because that also includes DeCal classes. DeCal classes at Cal are where students teach the class; they’re student-led classes. And so, a couple of these that students teach, intro and take the big course, and also ones for furthering design.

And then, we also have coupled to registered student organizations, student-led organizations, it’s the Berkeley Bio-design Community, where a bunch of communities get together and they have wonderful activities, like design-a-thons, and bring in speakers and go visit different companies. We’re really fortunate, recently, to have Berkeley and a wonderful donor support a new program called the Berkeley Discovery Program, and here’s the arc that we want students, so when they enter in, they want them to follow over their career time at Cal. And we applied for one of these, and fortunately, we got one of these departmental innovation awards for integrated biology.

We called it Discovery for All: Empowering Inclusive Communities in Integrated Biology. So, stay tuned, we’re really excited to share the approach I showed you with the campus. And just to conclude, we really see that diverse communities hold the key to creativity, that each voice has a unique benefit, because we all benefit from inclusive excellence. Thank you, and go Bears.

Matthew Shears : Robert, this has been absolutely amazing. Thank you so much. I got to show my son about this parkour squirrel, he likes doing that all over the city. But we have time for a couple of questions. There’s some people here who are very concerned about the wellbeing of that poor cockroach.

Robert Full: Honestly, they were never hurt. And literally, we quantified it, that as they could run the same speed after they were squished, and they can fly if the temperature is hot, and nothing happened to them. Now, of course, when we enter our building, we see people that have stepped on cockroaches every day. And we could actually tell them how they could more effectively squish them, actually, but they weren’t harmed. And they’re just so robust, but now he can begin to make it. Really, the soft robotics area is just exploding. You’re going to see versions of this, and not just robots, but all devices having softer properties, because who wants a big metal thing interacting with people in your home, or with children, or your pets. That’s not going to happen. They’re going to have to be soft structures to make them usable.

Matthew Shears : No, that’s cool. Questions specifically about that cockroach, is about whether you were able to measure the effects of, say, trauma? If it’s able to demonstrate that it felt something, or that it even thought something?

Robert Full: I have no idea how to do that, or what that means, because they do this all the time. That’s what they do, they squish themselves, and so it’s an occurrence that’s part of their natural behaviors.

Matthew Shears : Well, that is cool. There’s a lot of people who are saying in the chat that this is their favorite talk of all. If you do have other questions for Robert, you can put them in there now and he can answer to them. There’s one at the very beginning here. Do you also study the movements of underwater creatures, like the octopus or squid?

Robert Full: Yeah. We don’t, but there’s a bunch of other people that do. And so, they do it in all the different environments. And there’s been a lot of progress on looking at swimming. Again, the big breakthroughs are with soft robots, ones that can bend very much like we see in nature. And yes, there has been incredible work in Europe. There’s been a whole octopus project. So, if you look this up, there are soft octopus robots that both run on the ground, but they jet propel too. And they’re amazing. And so, yeah, again, light speed development is amazing.

Literally, I get to see papers come in every day. It’s incredible. And they may demonstrate something about building an octopus robot, but really what it is, it’s a demonstration of a greater capability that can be used for many things. It’s just a really nice example of the capability of different materials and control systems and designs.

Matthew Shears : That’s cool. And there’s a couple of more practical questions here. There’s a question about whether high school and middle school teachers could observe this kind of teaching and curiosity building instead.

Robert Full: Sure.

Matthew Shears : Yeah.

Robert Full: Yeah, absolutely. So, we do tours of our center, that’s one way to do it. And also, for a plug for the Lawrence Hall of Science, we worked with the Lawrence Hall of Science, and they run, with us teaching it, a bio-inspired design camp for high school students. And so, that’s pretty fun. But for observing the class and things, we’d be glad to set something up, absolutely. So, our dissemination, you’ve been able to inseminate this class and program to four other universities, as well as, like I said, the Lawrence Hall of Science summer camp. But ultimately, we think it’s sitting in a sweet spot, where it can be used for a K through 12, no problem.

I did, years ago, teach a sixth grade class with this kind of approach. It totally worked. Obviously, the projects were simpler and we didn’t have the quite technology, but the interest is very high and it brings out the creativity. So, I would like to continue to think about how we can develop a K through 12 program.

Matthew Shears : And what about our contingent? There are some folks here who would be interested in joining a MakerSpace. Older learners, is there a way for them to join?

Robert Full: Sure. Yeah, we’ve talked about this a little bit, but now if there’s real interest, I’m on the Jacobs Design Institute faculty council, and we meet and discuss possible programs that we develop. And this is one of the things we talked about, but really didn’t have the interest that we needed. Now, we can follow up with you if there’s that interest, and they create programs all the time, so we created a Masters of Design program for two years that is amazing. It’s incredible in this amazing MakerSpace building. And so, there are tours there, I think there are tourists today of Jacobs Design Institute that you can take to just see what it’s like. But if there’s interest, that’s what they do, is we create programs to potentially allow you to do things.

And I just tell you that, that all the different programs in Jacobs and the classes are just so much fun. They’re so interesting. And it brings together such diverse people. I would say, in our design class, we have a lot of different people, but what you see, is that the people who think they’re the designers of a structure or something, especially the engineers, they’re not near the most creative people that come up with the ideas. I’ll take a history major any day for creativity. It’s just, you bring what your experience has been, we work on teaming really hard, teaming activities, so that we make sure, before there’s a consensus, that everybody has a voice, and everybody shares and they feel comfortable in doing that.

And then, when we do that, we have this thing called a collaborative plan, where everybody says what they can do, what they’re best at, and then they decide who does what when and things, and it works out really well. And so, it’d be interesting to develop a program with projects that you might be interested in. I think that would be great. Or even sit in on teams or something, or be a part of our bio-design community, that’s another possibility, our registered student organization. I think this is great. This is wonderful.

Susan Hoffman: Well, Robert, it’s wonderful to have you today and to know of your interest and to take up your suggestions. We have been working with the Lawrence Hall of Science on our National Science Foundation grant, which we’ll fill you in on, and we’ve worked with the Jacob Center for Design. We are very interested in bringing OLLI members together with undergraduates. We see a lot of potential on both sides, having a good exchange. We’ve done some co-designing work and we definitely see the future for that, and with you. So, thanks a lot for joining today.

Robert Full: Great, I’m incredibly excited about this, and I think that it could add a whole other dimension to the students’ education if we can set up something. That would be fantastic.

Susan Hoffman:: Okay.

Robert Full: Outstanding.

Susan Hoffman:: Okay. All right. Well, thanks a lot for today, I know we’ve just run out of time, but thank you again for all of this.

Robert Full: Thanks, everybody. Thanks so much for coming. Take care.

[Music: “Silver Lanyard” by Blue Dot Sessions ]

Outro: You’ve been listening to Berkeley Talks , a Berkeley News podcast from the Office of Communications and Public Affairs that features lectures and conversations at UC Berkeley. You can subscribe on Spotify, Apple Podcasts or wherever you listen. You can find all of our podcast episodes with transcripts and photos on Berkeley News at news.berkeley.edu/podcasts.