Best of: The future of underwater robotics
30:15 Share

Hi, everyone. It's Russ Altman here from the Future of Everything. We're starting our new Q&A segment on the podcast. At the end of an episode, I'll be answering a few questions that come in from viewers and listeners like you.

If you have a question, send it our way either in writing or as a voice memo, and it may be featured in an upcoming episode. Please introduce yourself, tell us where you're from, and give us your question. You can send the questions to thefutureofeverythingatstanford.edu. The future of everything, all one word, no spaces, no caps, no nothing, at stanford.edu.

S-T-A-N-F-O-R-D dot E-D-U. Thanks very much. Hey, everyone. This is your host, Russ Altman from the future of everything. You know, the field of robotics has a long history at Stanford Engineering, and Professor Usama Khatib has been a pioneering leader in that field, working on everything from human interactive robots to underwater exploration, pushing the boundaries of what robots can do.

Most recently, he's led the opening of a new robotics center at Stanford, and it's amazing. He's going to continue his work on Ocean 1K, a humanoid robot who now has a new home in the robotics center. Join us as we talk about his journey. That's Osama Khatib's, not the robot. His vision for the future of robotics and how his research is transforming the way humans interact with machines. I hope you enjoy the episode.

Before we jump into this episode, I'd like to ask you to rate and review the podcast. It'll help others figure out if they're interested in the future of everything. So building robots in general is tough, but when you're building robots to operate underwater at depth, even depths as deep as one kilometer or a thousand meters, then there are lots of significant challenges. You have to figure out how are you going to build a robot that doesn't get crushed by the pressure.

You have to make sure that the electronics work. You have to make sure that the vision system works and can look around. Lighting, electricity. If you're going to have a humanoid robot, you have to have hands that can be manipulated, grasp things, and also move them around.

It's not easy. Well, Usama Khatib is a professor of computer science and engineering at Stanford University. He's an expert at building robots, especially robots that work under harsh conditions like underwater. He's going to show us how he got a robot to manipulate and explore at 1,000 meters of depth.

Usama, you were with us back in 2017 and you told us about some very exciting work for robotics in the water. But I believe a lot has changed since 2017. So I would wonder, can you bring us up to speed about the latest exciting innovations that have happened over the last five years? Well, Russ, it is really a pleasure to be back with you.

In 2017, I was reporting on the mission we had on La Lune, a vessel of King Louis XIV that sunk off the coast of France in 1664.

And the plan was to reach there with a robot that has hands and head and capability of connecting to a human sitting on the boat through a haptic device. So we did this, we succeeded. However, that robot was limited to basically 200 meters.

And the truth is underwater exploration requires much, much more capabilities in terms of depth, in terms of abilities of the robot.

And following the success of that expedition, we decided to take on the challenge of going to 1,000 meters. So this would be over 3,000 feet. I don't know exactly the equivalence, but it is a big challenge, technological challenge in terms of space.

how to build a robot that can maintain the structure at that depth. Also, the challenge is

The material that we use to make the robot buoyant, that is, you don't want to use your thrusters to lift your robot. It is going to be almost 200 kilograms. And that means you need to be basically floating in the water. Right.

So the material that we needed wasn't there for the first version. That is, we were using some material that was conventional flotation material.

Well, we were looking for a solution. The problem is if you use the same material, you end up with a huge robot. Huge, yes. You need to have higher density. Higher densities means you need bigger volume. And, well, that wouldn't work.

So we discovered there is a new kind of material for flotation that is built around the idea of hollow microspheres. And that means now you have a much lighter material, but very strong. Yes. And we developed this solution. We had to redesign everything, in fact.

And it was quite amazing because our robot turned out to have almost the same shape, the same volume. We had to increase slightly the volume of the robot, but probably you saw images of the robot. It looks really like almost the size of us.

human diver. Yeah. So, so let me, let me ask a couple of things. First of all, about this new material that was able to withstand the pressure. Is it fundamentally metallic or plastic or what is the fundamental, these microspheres, what are they made out of?

So if you think about it, it is glass. Glass. Basically. Wow. Yes. It looks like glass. It feels like glass. And it is sort of a silicon oxide material.

material. It's a syntactic foam. Okay. And this would sustain a huge pressure resistance like 6,000 PSI and we can go high depth without any problem or challenge. Okay. And so the other things just for establishment, these are not tethered robots. These are free floating. Is that correct?

Yeah, the robot actually itself is free-floating and operating, but here is the biggest challenge we haven't addressed ourselves, and this is the battery. Ah. If you think about operation where you need to take this robot for like 1,000 meters, you need about seven hours, and our battery is like 45 minutes, so there was no way.

But because of that, in fact, the robot wasn't alone. I mean, every time we dive, we are not alone. We have robots that are ROVs that are following us to put light. It's very dark there. What does ROV stand for?

It is a remote operation vehicle, operated vehicle. Okay. And these are basically, I mean, there are a lot of underwater vehicles. And the difference with our robot, these vehicles are just mobility vehicles with cameras. They can navigate. Our robot is a robot with arms and hands and legs.

and cameras and stereo and everything. I mean, to give you an idea, this robot is sort of your avatar. Yes. You see through the eyes of the robot. So when you are looking at this monitor, I'm wearing the special glasses for stereo.

And I'm looking at the monitor and I see in 3D, I see through the eyes of the robot, my hands. Wow. So I'm extending the hands of the robot to operate and I can see the hands of the robot. And through the hands equipped with sensors, we are able to touch things.

So when the robot is going to touch anything in the water, you are going to feel it in your hands. And how are we feeling it? Well, we are feeling it through a device you're familiar with. This is a haptic device we use in surgery. Yes. Basically, similar to the haptic devices that we use in minimally invasive surgeries. So I call our robot POS.

except as performing minimally invasive exploration in the water. Yes. So, okay, let me go to that because there's some questions I have around that. I think we discussed this back in 2017, but you are very devoted to having these robots be anthropomorphic, to look like people. You talked about the hands, the head.

remind us why is this so important versus some arbitrary shape that might be more suited to some other exploration. So tell me why it's so important in your work to have this anthropomorphic robot.

Well, I can give you this image that I myself experienced. This was diving on the Crispy shipwreck. The Crispy shipwreck was a backboard, incredible, huge, big shipwreck.

I mean, the boat that was an Italian boat, that was sunk during the war, the Second World War. And it was, it is sitting, resting at 507 meters. Okay. So we're diving there and approaching the crispy, you see this magnificent boat floating

Sitting there with these colors, with life all around it. I mean, when it sunk, it lost a lot of lives. But basically, life came back in the form of marine biology, in the form of all these corals, special kind of corals, white corals. Anyway, it was just breathtaking approaching this boat.

And I'm looking at my hands. So I'm feeling like now I'm going to do the task. And the task was to place a marker so that we can do photometry, scan the boat and do the photometry. And I'm approaching this. I come to the rails. I see...

the marker with the position where I'm going to put it. I appreciate it. And when I touched it, it was just unbelievable. They told me, you were screaming. I mean...

It is like you are really immersed in this incredible thing where you are touching something at 500 meters and you are performing a task you're familiar with.

Through your hands, where you can see your hands and you are placing hands in the right position. Right. And you are feeling, I mean, this is the idea. The idea is you are really creating a sort of...

avatar of yourself, allowing you to perform the task in ways that you cannot do if you are building a block with an arm. And also, I mean, thinking about it, when we look at all the tasks, for instance, we

We developed in this expedition something really unique that also demonstrated the use of tools. So a camera placed on a perch that is about two meters away

We are taking it all the way inside enclosures where the robot cannot go. Right. And we are rotating the camera. So you need two hands. Yes. And with one hand, you're holding one side and with the other hand, you are rotating this camera. And we came back.

We came back with incredible images from inside the boat that the biologist was just like amazed. He said, our colleague, a biologist, was with us and she was saying that

These are iron-eating bacteria that we do not encounter in the Mediterranean. And here they are. They are in the form of tubes, long tubes, orange long tubes. And it was really, really a surprise for us. Everything you are filming, I was in another dive on...

an aircraft, the Baron aircraft, and we were able to film inside the cockpit from outside because we are able to maneuver and move this camera. So the use of the two hands requires your eyes

over the two hands, and now you see the upper body of a human shape. Well, it doesn't have legs. Right.

So it is a mermaid. So this really that I found that answer very convincing because basically millions of years of human evolution are being taken advantage of because you know how to coordinate your eyes, your hands, and you've taken full advantage of that control system and just transplanted it into the robot. Whereas if it was an unfamiliar geometry, you know,

you wouldn't be able to do any of this. So I really do. That makes perfect sense. So, so I wanted to ask, what are the big technical challenges from the robotics point of view in, in building this, this device? Um, because you've mentioned so many things you've mentioned the depth, you know, it has to be, uh, it has to be strong enough to withstand all the pressures. You mentioned the binocular vision, the haptics and the feeling, um,

the manipulation that you have. Which of these are you most proud of as a technologist? Well, one of the challenges that make this robot capable of doing things, and this is the arms and the hands. So we redesigned the hands and this was reasonable complexity. We did it in fact, in collaboration with my colleague Mark Kasky's group.

and with another group also in Pisa who has been working on hands and they adapted one of their hands for us. But the biggest challenge is how you take arms and place them deep in the water.

You can imagine, okay, I'm going to take these tubes, very strong tubes that will resist the pressure. But then you have a very heavy arm, you don't have the dexterity, you don't have the compliance, the safety that you want to have in your arms. So how can you make lightweight, gentle, compliant arms? Well,

I mean, the pressure is there. So the only way you can do it is by creating inside the arm the same pressure that you have outside the arm. Ah, to kind of equalize. Equalize. So how you do it? Well, you need to put a liquid inside

oil, we call it. It's not oil, it's a vegetable oil. It's very, I mean, it has nothing to do with oil, we call it oil. So it's oil-filled arms. So you take the pressure from outside through a compensator, you take tubing all the way through the arms and you place these tubes to fill the arms

and the hands all the time at the same pressure as you go down. Now this is great. Well, we worked it out before, but now we have to make it stronger to sustain the 1000 meter pressure. But then suddenly you remember, well, I have electronics. I have all these oscillators. I have all these capacitors.

Now you have to redesign your electronics to go to 1000 meters. Because they also need to withstand that same pressure. Yeah, they are swimming in the same oil at the same pressure. So this was actually a lot of work, a lot of concept that we developed.

And you do everything you imagine. You build the electronics, you build the pressure inside, you do all of that. And you test it in the pool at Stanford at one meter. And then you go, and then you go. And when we went to the 1,000 meter to Cannes, so we already went to 500 meters, the robot was doing fine. We wanted to go to

the 1000 meters and should see that our design works. So 500 meters, okay, 600 meters, 700 meters, 750. And as we were going there, it's amazing, everything, the robot didn't mind, the electronics was working, everything was going perfect.

And the only thing was we were hitting the ground and then the boat will move deeper. And then we kept going down, we kept going down and everything worked. And this is when you say, oh, wow, this is amazing. We did it. We were really right. This is the Future of Everything with Russ Altman. More next with Osama Khatib.

Welcome back to the Future of Everything. I'm Russ Altman and I will continue my conversation with Professor Usama Khatib about engineering robots that can work under pressure underwater.

So in the last segment, Osama told us about this remarkable technical achievement of getting a robot to manipulate and function at a thousand meters of depth. In the next segment, first I'll ask him how hard is it to get a human to control such a robot? And then we'll talk about the wide variety of applications that he's most excited about.

What kind of training does a human have to go through in order to run this robot? Is this highly, highly technical training and you're getting kind of experienced divers or is it much simpler than that? Well, the truth is both. So in one hand, high school student

can within maybe half an hour be able to operate this. And the thing, Russ, I wanted to tell you is we're going to do that. So we're going to make this robot available to schools, to universities remotely. So from your school, from your laboratory in a university, we will have the haptic device.

And this haptic device is connected to the robot somewhere in the world. And now you can explore and help with the exploration of underwater in the oceans. So this is going to take a little bit of time, logistic operations. We solved another problem in terms of the communication, in terms of the communication.

delay, time delay. We replace teleoperation by two independent loop. The robot is autonomous on its own and the haptic device interface is autonomous and they are communicating through a smart interface.

so that we do not have instability due to teleoperation delays. Right. Wait and see. That's amazing. Well, so that part of the answer sounds to me like half an hour for a high school student. That means it's pretty easy. But you said both. So tell me about the technical – where does it become tricky? Where it becomes tricky is the task. Okay.

And some of the tasks are really, really delicate tasks. So my colleague, Michel Lour, who is the archaeologist,

likes to approach archaeology in a specific way. He has his expertise. He touches first gently, he moves around the object, he doesn't go and like smash it. So there are, it is like the haptic device could be used easily, but then the skill of the needed for the task might require more.

And we still want a surgeon to operate on us in minimally invasive surgery. So we need the expert in every task if we are doing underwater archaeology or biology or whatever. That makes perfect sense. So...

But that's also, as we said before, that's taking advantage of all the hours and years of training of the human. And it's where the anthropomorphic shape of the robot, I'm sure, makes it easier for them to do that transition from in real life, so to speak, to the remote environment. Okay. I wanted to not – I didn't want to miss the chance to find out about the future. Not just future, but you've done many things with this robot already.

You've mentioned some of them already. You put sensors on the boat. It sounds like you looked at an underwater airplane. Tell me about where, and it sounds like there's an educational component with the high school students. What are the other applications that you're most excited about? Archaeology is a big one, and there are a lot of missions that we are discussing right now. We are discussing with the

National Science Foundation of Italy, a collaboration with Stanford University,

to explore sunken cities and other shipwrecks. And I have some upcoming meetings. We are talking with... So in the Mediterranean, the shipwrecks, the old shipwrecks disappear because of the temperature of the water. In the Baltic Sea, there are shipwrecks that are like...

very old shipwrecks from the...

many, many centuries back, and they are still in shape. Yes, they recently found that very famous ship that went down in the northern seas. The producer from Hollywood contacted me, and they would like us to go and explore some of those Baltic Sea shipwrecks. So this is one. The other one, and this is really a big one, is the environment.

And the biology, the marine biology, the marine biologists never, the scientists say, I never could have find those iron...

Eating bacteria without your robot. Right. So we are providing somehow a tool to the biologists, to the archaeologists. But also, once we are able to take this in a bigger number,

of robots, we will be able to have collaborative robot operating together with human underwater in placing sensors, in maintaining structures. I mean, the internet, all these cables requires a lot of maintenance.

All the structures we are building to bring electricity in the seas requires maintenance. There are a lot of cracks that will take place and now using tools we will be able to find these cracks and explore the structures.

And I have a lot of people coming from the oil and gas and other industries looking for solutions because we have to remember diving is very dangerous beyond some surface.

certain depth. And there are no other solutions than going with a robot that has the capability of the diver, the hands, the arms to perform those tasks. Okay. I have to ask a couple of quick questions because this is so exciting. Uh, uh,

First of all, what about search and rescue? We've seen some amazing stories about people stranded and divers, as you said, under very dangerous circumstances trying to make saves. Is there a future for search and rescue here? Yeah. I mean, the cave in Thailand. I had the question and I had a discussion with some colleagues at the university about

at AIT, I mean, at the Asian Institute of Technology. And in fact, we...

We have a tool. I mean, if a diver is able to help, well, our robot diver can do much of the work, not 100% of a human, but in many situations, yes. This is a tool that will bring together human intuition and capabilities collectively

collaborating with a robot to create the synergy and safety for everyone. Yes. The other questions I want to ask was, do you have, so people who might be aware of this might wonder, does the robot have to do decompression like humans do in order for the systems to re-equilibrate or is that less of an issue?

Not at all. There is no time. As soon as the robot is called to come up, the robot will come up. It goes straight up. It goes straight up. It goes straight up. And then I – so I don't want to at all – the one –

The reaching of one kilometer is an amazing milestone, and I don't want to be too greedy, but let me ask you, do you see a path for even deeper exploration or do you think we're hitting upon some fundamental limits?

Well, we are far from the actual limit. I think we will be able to go to 3000 meters without too much of difficulty. I mean, this is this is the beautiful thing that is happening today. When we think about the multidisciplinary nature of robotics.

material, new materials, new sensors, new methods of fabrication. And the material science is helping us.

And I think we will be able to go to 3000 meters. There is some application at 4000. I'm still discussing how we can do it. But I think beyond that, probably we will have some more challenges. I mean, technology challenges. But this is...

absolutely amazing experience to just imagine you are virtually there. You are touching places that are at so much depth

So many times I get a question, are you going to reach those famous boats that we dream about? And well, maybe they are beyond our reach yet at this time, but I think that is going to happen. There are some amazing locations

2,000 meters, 2,500, I think we should be able to do that. Thanks to Osama Khatib. That was the future of underwater robotics. You've been listening to The Future of Everything. I'm Russ Altman. You know, we have more than 250 back episodes in our archives, and you can listen to these conversations about the future of anything.

You can find me on a lot of social media like Blue Sky, Mastodon, Threads, at RB Altman or at Russ B. Altman. And you can also find me on LinkedIn, Russ Altman, where I announce all of the new episodes. And you can also follow Stanford Engineering at StanfordENG.

If you'd like to ask a question about this episode or a previous episode, please email us a written question or a voice memo question. We might feature it in a future episode. You can send it to thefutureofeverything at stanford.edu. All one word, the future of everything. No spaces, no underscores, no dashes. The future of everything at stanford.edu. Thanks again for tuning in. We hope you're enjoying the podcast.