EELS: AI-enabled snake robots and the search for life on Enceladus
01:02:04 Share

Could AI-enabled sneak robots help us investigate potential life on Enceladus? We'll learn more this week on Planetary Radio. I'm Sarah Al-Ahmad of the Planetary Society, with more of the human adventure across our solar system and beyond. Saturn's moon Enceladus is a compelling target in the search for life off of Earth.

Its subsurface ocean has all of the necessary conditions and ingredients for life, but determining if there is presently life on that moon is going to take some ingenuity and some out-of-the-box thinking. Today, Morgan Cable and Hiro Ono from NASA's Jet Propulsion Laboratory join us to discuss the Exobiology Extant Life Surveyor, or EALS, project.

EELS is a snake-like robot designed to navigate the treacherous icy terrain and vents of Enceladus, hoping to access and sample the ocean underneath to get us some answers. But first, it's time to celebrate this year's highlights in space exploration. Kate Howells, our public education specialist, will pop in to guide us through voting for the Planetary Society's Best of 2024 awards. Then Bruce Betts, our chief scientist, joins me for what's up and a new random space fact.

If you love Planetary Radio and want to stay informed about the latest space discoveries, make sure you hit that subscribe button on your favorite podcasting platform. By subscribing, you'll never miss an episode filled with new and awe-inspiring ways to know the cosmos and our place within it.

Every November, the Planetary Society recognizes all of the amazing things that have happened in the last year of space exploration with our Best of Awards. Thousands of people from around the world cast their votes for the best space images, missions, and achievements of the year. Everyone is welcome to join in, and this year's voting has officially begun. Here's Kate Howes, our public education specialist, to let you know how to participate.

Hey, Kate. Hi, Sarah. Thanks for joining me and for helping to put together this Best of 2024 Awards this year because it's always one of my favorite things. It's a great way to reflect on what happened in the last year.

Oh, yeah. It's also one of my favorite things to put together because I get to just kind of trawl through a bunch of old content, go through our image library, see what's been added in the last year and just be reminded of how stacked a year an exploration can be. And it's only November and we're just looking back to the beginning of January and so much has happened. So it's really a heartwarming reminder of how much we as humans do out there in the cosmos. Yeah.

So what are some of the things that people are going to be voting on this year? Because I can't even imagine trying to distill everything that happened in this year into a few questions. I know it's hard. But as always, we start off with the best space exploration image because everybody who loves space loves space imagery. It's one of the best, most accessible ways to appreciate space exploration. So we've got a handful of

Awesome pictures from throughout the solar system and from Earth, including an image from the International Space Station looking down at the moon's shadow over the Earth during the April total solar eclipse. We've got pictures from Mars. We've got pictures from the moon. It's really just a great collection of awesome.

awesome imagery that also reflect really cool moments in space exploration. We also then have a list that people can vote among of just straight up moments in planetary science. So launches, discoveries, advances in our knowledge so people can pick their favorite new advancement that happened. And we've got astrophotography. This year, we went to our members who often share their astrophotography work in our online member community.

And people just take the coolest pictures. So we've got photos from the eclipse and of the comet that was in the skies recently and of the aurora that have been so magnificent this year because of the solar maximum. And it's always just fun to get to highlight the stuff that our members do.

And then we've got favorite mission and, of course, favorite thing that the Planetary Society accomplished, because it's always great as well to look back on what this organization has managed to do over the last year.

It was a really cool moment for me going through, because I got to test this ahead of time. So I got to see the questions before all of you, but going through and actually kind of seeing the way that we've influenced a lot of these big space moments. And it's been a long time coming. It's not like we all did one really cool thing this year. This has been ongoing years of us advocating for things together to help make these missions possible. So it was kind of hard to figure out which one of those things I wanted to vote for.

Yeah, absolutely. I also did a bunch of test voting to make sure the page worked. And so I got to kind of pick like 10 favorites, you know, but I think if I actually were to go in and make my real choices, it would be a challenge because there are just so many highlights from such a great year. So how can people vote this year?

So if you go to planetary.org slash best of 2024, you will see our voting page and it's open to anybody in the world. You can send it to your friends and family. Anybody who likes space will be collecting votes through to the end of November. And then we will be releasing the results in early December. It's good to take a moment and reflect because honestly, we've done some cool stuff. We should be proud of ourselves.

Yeah, no matter what people are feeling right now, it is great to look back and be reminded that there is a whole lot of really impressive and awe-inspiring things that humanity accomplishes when we work together. Well, thanks so much, Kate. I can't wait to see the results. Thanks, Sarah.

The space community has achieved some really amazing things in the last year. One of the highlights for me was the launch of NASA's Europa Clipper mission, a spacecraft that is now on its way to explore one of the moons of Jupiter that has a subsurface ocean and potentially the right conditions for life.

But Europa isn't the only possibly habitable world in our solar system. Saturn's moon Enceladus is also a prime target. Like Europa, Enceladus has a thick outer shell of ice that guards a subsurface liquid water ocean. But on Enceladus, there are cracks in the ice that allow the water from the ocean to actually get out to the surface and spray out into space.

NASA's Cassini mission flew through those jets of water erupting out of that moon many times between the late 2000s and mid-2010s. It revealed that this subsurface ocean not only has all of the ingredients for life as we know it, but there's evidence of organic compounds and hydrothermal vents in the seafloor.

While there are no current NASA missions that are planned to return to Enceladus, a mission back to that moon is one of the priorities in the most recent Planetary Science Decadal Survey. It's a report that's prepared by the National Academy of Sciences, Engineering, and Medicine every 10 years. The report encompasses the most critical scientific questions facing the space community and a priority list of missions that we can build to answer them. But what kind of mission could help us unlock the mysteries of Enceladus?

Scientists and engineers are already trying to tackle this problem, and one of the possible technologies that we could use is EELS, or the Exobiology Extant Life Surveyor. Extant meaning "still in existence." Essentially, this robotic concept is designed to help us actually answer whether or not there is currently life in that ocean. EELS was a project designed and tested by a team at NASA's Jet Propulsion Laboratory.

It's a snake-like, self-propelled, autonomous robot that uses AI to learn and traverse its environment. I'll put images and video of eels on the webpage for this episode of Planetary Radio because it takes some work to visualize. It looks like a long snake or a worm with a segmented structure. Each section has actuation and propulsion mechanisms that allow it to move, grip, spelunk, and even propel itself in water.

The robot connects to a tether that it uses for power and communication, but much of its decision-making is autonomous. There's a pretty big communications delay between Earth and the Saturnian system. This robot can't wait for our input before making its next move, so it requires a high level of autonomy. EELS uses an onboard AI to learn on the job and explore potentially dangerous places that would be really difficult to traverse otherwise.

There are many applications for this type of technology, but EELS was primarily designed to explore Enceladus and descend into the cracks in its ice where it can gather data and samples for study. To tell us more, we're joined by Drs. Morgan Cable and Masahiro Ohno. Morgan Cable is a research scientist in the Laboratory Studies Group at NASA's Jet Propulsion Laboratory, or JPL. She is the science lead on EELS.

Her expertise is in organic and biomarker detection strategies. She previously worked on Cassini, but she's also part of the mission teams for Europa Clipper and the upcoming Dragonfly mission to Titan. Hiro Ono is a robotics systems engineer and the group supervisor for the Robotics Surface Mobility Group at JPL. His research focuses on developing innovative robotic systems for planetary exploration. He's the principal investigator of the EELS project and comes from a lengthy background of working on Mars rovers.

Morgan and Hiro recently joined me at the Planetary Society's headquarters in Pasadena, California.

Thanks for joining me at Planetary Society headquarters. I'm so glad to have you here. Thank you for having us. We are so excited to be here. And for both of you, not your first time on Planetary Radio, although your first time with me, you both spoke with my predecessor, Matt Kaplan. First time with you, first time in person. But it's so exciting. It is. The most recent time you were on the show, actually, it was basically making a case for why we should go back to Enceladus.

Why do you feel so passionately about the search for life? And specifically, why should we make Enceladus our next major target? Enceladus is just such a fascinating world. And I shouldn't say that there aren't other fascinating worlds out there. Every place that we...

go to explore in our solar system is unique and interesting in its own right. But for me, Enceladus, it holds such magic because it has this incredible liquid water ocean. It's got this tortured surface with these cracks in it. And it is the one place where we know that sample is coming from that ocean into space in a way that we can access it with technology today. And so that's one of the reasons why I think Enceladus is at the top of my list of

for exploration. From an astrobiology perspective, this is one of our first real opportunities to test that question of if you have the ingredients for life, if you have water, chemical building blocks, and energy, and you have those somewhere and you mix them together and you wait,

Does life form? And this is a place where we can go and test this. There are other worlds as well. Europa, Titan. There are plenty of other ocean worlds. We've already got a mission on the way to Europa. We've been to Titan with Cassini. And now we're going with Dragonfly. So it makes sense that Enceladus should be next on our list. Let's go.

I mean, adding all these things together, we can get a much better picture of the habitability of our solar system because I have been looking forward to Europa Clipper for ages. The Dragonfly mission to Titan is going to completely upend things, but we've been through the plumes coming out of Enceladus. We've seen what that data has shown us, and there's some pretty clear indications that this is

Just such a wonderful target for this kind of science. Cassini gave us so many incredible views of Saturn, its rings, and its moons. And I think the information that we learned about Enceladus during Cassini's 13 years in the Saturn system is one of the key pinnacles of achievement for that mission. But we have to remember that Cassini was not meant to be a life search object.

mission. It didn't have those tools. It had tools that were designed to understand the chemistry and, you know, the other properties of these worlds. But we didn't even know when we built and launched Cassini that there was liquid water out that far. That was one of the things that we were able to confirm with that incredible mission. And so I think any future spacecraft with a very carefully designed payload, tailored to look for signs of life,

would really just blow our minds with the kinds of discoveries that it might make. Right. And just adding the engineer's point of view, you know, I'm not a scientist, I'm not even an astrologist, but, you know, when I read Carl Sagan's Cosmos or Pilgrim's Dot, where he talked about, you know, finding life somewhere in the universe, what's the implication of it for humanity?

it gave me a chill, you know. So although I don't have, you know, deep scientific arguments about what's life, that kind of thing, but just the notion that the technologies that we create could contribute this, you know, civilization-scale science that will forever be recorded in history, that alone makes me so much excited.

But it is so valuable to have some really skilled engineers on this problem because the terrain of Enceladus is complicated. And if we want to find life, we're going to have to find ways to get down into those cracks and potentially sample things, check out the water down there. And that is a colossal challenge to tackle. It's an impossible challenge that we haven't been able to solve so far.

It's just impossible. We haven't been able to so far. In the future, nobody knows, right? It seems impossible until you do it. I mean, that's what space exploration is all about. Exactly. That's what's so exciting, to be a scientist working hand in hand with the incredible engineers at JPL. Life is not going to be...

just sitting out in the open on a planetary surface where it's exposed to the extreme conditions of space and UV radiation and things like that, chances are it's going to be in those hard-to-reach places, you know, under a rock, in a crevasse, down in one of these subsurface liquid water oceans. And by developing technology that can tackle those challenges, it really gives us access to these places that previously were so far out of reach.

And, you know, working side by side with scientists like Morgan is really fun because, you know, this gives the meaning to our work, right? We are not, you know, making random technologies for random things. We are, you know, doing this for a big reason. Yeah, we're definitely all wondering

one team. And it's so neat to see, you know, the engineers when people like Hiro come and it's not just about designing something, it's about working hand in hand to figure out, okay, what are the science questions we actually want to answer? And then thinking creatively about ways to address those challenges and get that robot where it needs to go or be able to extract a sample from a very specific place in a very specific place.

It's so fun to be part of a team where everyone's working towards that goal together and just really wants to achieve the dare mighty things together. That's what we do.

And then in this case, it's one of those things where if you actually find the answer to this question, and we should be clear, there is no mission currently planned to go to Enceladus. But if we want that mission to exist, we need to build the technologies to do it. And if we get these answers, if we realize that there is life in that ocean moon, and we don't know that yet, we don't know that yet, but how...

Amazing would that be. It would fundamentally change everything about space exploration, but also about the way we see ourselves in the broader universe. You're absolutely right. We call this civilization level science because you only get to make this discovery of life elsewhere other than Earth for the first time in human history. You only get to do that once. And we have the technology to be able to detect lifestyles.

life on another world right now. We just have to get it there. We have to go. We're not going to know until we go. And so it's an incredible opportunity to be able to take these technological advancements that we've been making over the past few decades and be able to apply them to these worlds where we think the likelihood, at least of habitable conditions being there is very high. And so these are the places that we need to go and do that search.

There's a few ways that we could tackle a challenge like Enceladus, which actually brings me back to the last time that you were on the show, Hiro. It was during one of the NIAC symposiums, NASA's Innovative Advanced Concept Symposiums. And people who listen to the show know that I love going to that conference for a few reasons. I've been lucky enough to host their webcast for the last two years, but more so because

these are the kinds of technologies that are really going to allow us to take our space exploration to the next step. And not all of them come to fruition, right? So you've been the lead on three of these projects. And the one that you were speaking about was the Enceladus Vent Explorer. Right. How

How did EELS come about and is it in any way connected to your previous work on that project? Let me explain what the NIAC project was. We broadly looked at the mission concept.

I believe, you know, we are the first, although very wrong, to seriously investigate the possibility, right, of going through that erupting vents of Enceladus, sending a robot down to the ocean. Now, on the NIAC projects, we are looking at the mission concept, right? You know, so it's abstract. How is the mission feasible? How the mission overall looks like?

but we didn't touch much about technologies. But the result outcome of NIAC is that it's likely possible to send the robot down there. But it's possible, being possible is different from we can do it now. So there's a wide opening technological gap. So how to fill that technical gap, that's the theme of EOS. And

On eagles, we particularly looked at the snake robot. In reality, you know, the robot does not have to be snake, can be some other forms, but

But for a few reasons, we decided to focus on snake and we developed the technologies that would one day enable the mission to the events of this. And so for those who don't live and breathe, EELS stands for Exobiology Extant Life Surveyor. And this is picture like an anaconda sized robot.

but one that is meant for good, not for, you know, killing and eating things in the Amazon. Eels is meant to be able to navigate both the features on the surface of an icy world, so dips and cracks and, you know, moving around and over obstacles through what we call unconsolidated media. So think sand, but made out of little ice pellets instead of silica. And then

get down into a vent, a crack in that ice shell where stuff is shooting out at your face at like two or three hundred kilograms a second. You need to push against the walls of that ice shaft and move down to get into the really exciting environment, that liquid water ocean. Now EELS was originally designed with Enceladus in mind which is why it's got the shape it has. But what we've learned

as we're developing this project is that there are tons of different applications in different worlds where an architecture like that or something else that's adaptable could be able to access all sorts of places. Think about lava tube caves on Mars or the moon. Think about just even here on Earth doing avalanche or earthquake rescue or reconnaissance, being able to send an adaptable robot to a place that's too dangerous for people. The possibilities are wide open.

Go to Google and search EELS, E-E-L-S, JPL, and you can see a bunch of movies. And I'll link these movies and images on the webpage for this episode of Planetary Radio. Because this kind of thing, it's faster to show the movie.

rather than speaking. It's hard to kind of wrap your brain around what this might look like until you're actually looking at it because there's all these different interesting segments and how they work together must have been one of the most challenging things. It's not just like you have a robotic arm that you have to worry about, not like Mars exploration isn't hard.

But trying to get this entire kind of bio-inspired mechanical device operating together, that must have been really rewarding. Yeah. So there are many reasons that it's challenging, right? But I would say the greatest challenge is the uncertainty, unknowns. You know, we've sent like 10 spacecrafts to Mars. We know pretty well, at least about the regions that we landed and explored.

On Enceladus, what's the best resolution we have on Enceladus? Like six meters or so? Six meters per pixel, about half a school bus. Only around really limited regions. And we don't know, for example, what's the surface topography, what's the geometry of the vent.

we still don't know how strong the jet is. So, you know, it's really, really hard to design something, right? Because we don't know what the environment is. Knowing that you can't fully anticipate that scenario, what situations did you account for?

So we have a whole host of data from the Cassini mission. And we've used that, scientists have used that to generate models of what they think these cracks in Enceladus' ice shell look like, what the dimensions are. We have a few different models that fit the current data. And we can't, right now, we aren't able to decide between them. One of those models says that it's a wide open crack.

about five meters wide, and it just goes straight from the surface to the ocean. And in that case, it's a really easy problem, right? You could just chuck a probe down or something like that and call it done. The problem is that on the other side, we have models that say, no, instead, these jets, these vents and Enceladus are more like volcanic eruptions here on Earth, where you could have the ice converting to a choke point that could be on the order of about the width of your hand if you spread your fingers out.

about that size. And so we had to design a robot that could handle either of those extreme cases or anything in between. Let me generalize a little bit. There are two types of unknowns, right? Known unknowns and unknown unknowns. Known unknowns means you know what you don't know. For example, you don't know what the strength of the jet is.

You don't know what the width of the event is, but you know what you don't know. And this is, I won't say easy, but relatively easier to handle because you can design your robot to be robust over the range of possibilities. What's really difficult for space missions are unknown unknowns, right? Something that you don't know what you don't know.

And one example, probably, you know, I spent probably half of my time of my JPL career on Mars, both on research and flight. One surprise that happened on the Curiosity mission

around Seoul 400, was that suddenly we saw a lot of punctures on the wheel, right? And I know how, you know, we panicked around that time. That's totally something we, you know, didn't anticipate. And when you go to a planetary environment that you didn't know, those unknown unknowns do exist, right?

And to be clear, we've adapted and overcome. Curiosity is still making incredible discoveries on Mars today. But yeah, Hero is absolutely right. Thinking about what that option space could be, the things that keep us up at night, and doing our very best to engineer something that is so adaptable, so robust, that in a way it can handle those unknown unknowns, even though you have no idea what they are. And then now we are talking about

You know, going down the vent of this icy moon that you never have visited, the level of unknown unknowns are enormous. So we have to expect something, you know, much more spectacular unknown unknowns, right? So the biggest theme that we've been working on these days, and actually this is a theme of EOS, how to make your robot...

extremely more adaptable to crazy unknowns. We label it Space Exploration, Robot Exploration 3.0, rather than, you know, it's just a cool label. But on Mars, we had a luxury to send like, you know, 10 or 15 spacecraft over many decades. That's fine. It's only like six or seven months to get there. So we could do that. We learned incrementally to make it happen.

For instance, it could take like 10 years to get there. Can we send like 10 spacecrafts and learn? I would love to. That would be great. If only. If only we can wait for a few centuries, right? No, no. Faster. Exactly. So we somehow got to figure out to make the robot enormously more robust and adaptive to unknown environments.

That's a big theme that EOS is having. Part of the solution is that, and why snake question also, EOS is very different from Walgars in that it has

many ways of moving around, right? War war can only drive. Eagles can, you know, crawl, you know, move like trains, many other ways. So if something unexpected happens, we can change the mode and figure out the way to interact with the environment.

I think that's a big point of wheels. Well, I mean, the number of times that we've accidentally gotten our wheels destroyed on Mars or even got stuck in some situations, had to roll backwards. Or failed drilling. Imagine if we could have just rolled sideways with opportunity instead of, you know, it could have changed a lot of things. But there's two things going on here. One is the actual...

physical design of this object, which we can adapt for each world that we go to. But there's also the second part, which is the kind of AI brain behind this technology, because there's A, getting there, but also the time delay in communication with something like this makes it... It's not like...

It can hang out in the crevasse waiting for orders for hours on end, right? So we need to give it this element of autonomy, which is what makes this so fascinating. Yeah. What have you done here to enable this device to create its own plans? All right, that's my favorite topic. So stop me. I might keep talking for three hours, but let's go. You know, there are lots of remarkable progress in AI in the past 10 years, right? It's a mixed bag of hopes, right?

and hypes, weird things, and just talking. But it's real that AI made amazing progress like chatGPD or all the generative models. What's called, there's someone's paradox. What is that? More of the extra, you have to Google it.

Something easy for humans tends to be hard for AI. Something difficult for humans tends to be easy for AI. For example, if you do a metal mass of multiplying three, two, four digit numbers, it's enormously difficult for us. But it's just one line code for computers. Playing chess, that's

very hard for us, not easy thing, but computer-beated human beings in 1997, I believe. Yeah, I remember that moment when I was younger and when Deep Blue and Kasparov were fighting each other. That was a moment in history. Exactly. And there was a huge hype that, hey, computers are taking over. But the reality is that, funny thing is that there was a man sitting in front of Kasparov

who reads the computer screen and moves the piece, right? Why is that? Because visually recognizing the board, pick up the piece and move it. That's still a very difficult task for AI or computer, right? So that's what I mean. You know, those things that every three-year-old can do is still hard for computers, AI's robots. And it turned out that it is those people

Physical interaction with the world, which is so easy for humans. You know, as a Japanese, I can easily pick up tofu with chopsticks, right? It's easy for us. Basic skills for Japanese. Turned out to be very difficult for robots. And it's exactly those capabilities, right, that we need to make the robot-like eels work in a physical world on Mars or on Enceladus.

And on that regard, AI or autonomy or whatever you call it is far behind humans. There are lots of room for improvements. That's what we are working on. We'll be right back with the rest of my interview with Hiro Ono and Morgan Cable after this short break.

Hi, y'all. LeVar Burton here. Through my roles on Star Trek and Reading Rainbow, I have seen generations of curious minds inspired by the strange new worlds explored in books and on television. I know how important it is to encourage that curiosity in a young explorer's life. That's why I'm excited to share with you a new program from my friends at the Planetary Society.

It's called the Planetary Academy and anyone can join. Designed for ages 5 through 9 by Bill Nye and the curriculum experts at the Planetary Society, the Planetary Academy is a special membership subscription for kids and families who love space. Members get quarterly mail packages that take them on learning adventures through the many worlds of our solar system and beyond.

Each package includes images and factoids, hands-on activities, experiments and games, and special surprises. A lifelong passion for space, science, and discovery starts when we're young. Give the gift of the cosmos to the explorer in your life. In this case, we could try to teach one of these objects or one of these machines to do this kind of navigation, knowing what we know here on Earth, but...

It's far more valuable to just have it understand its local environment and make decisions off of that. Right. So how much of this is something that you've pre-programmed into this based on your experiences with testing? And how much of it is going to be just learning on the job? So when we specifically talking about what we have done on EELS, it's mostly designed by human engineers.

We also experimented machine learning techniques, for example, using reinforcement learning, which is just a difficult way of saying that learn from experiences.

to discover new gate buttons. We experimentally hooked up LLMs to EOS, which is cool, by the way. MELANIE WARRICK: That's really cool. GARY ILLYES: Done by our engineer named Rob Royce. He's awesome. Mostly still-- so first thing forward, I got to keep talking three hours.

definitely are interested in using the latest machine learning approaches. I think in short term, the reasonable solution is the combination of machine learning model-free, black box approaches, which is good at seeing humans' intuitions being adaptive to the environment.

combined with the model-based classical approaches. Why? Because we can understand and we can validate what's safe, you know, how it performs. So things like EELS or near-term spacecraft would probably be driven by, not by, you know, gigantic general AI, it would be a mixture of classical control, model-based planning, and interesting AI algorithms.

Yeah, I'm thinking of this in the context of what we saw with the Ingenuity helicopter on Mars, right? Oh, yeah. That was a scenario where, you know, humans made most of the plans for where it was going to go. But in the short term, actually flying around, actually grappling with what's going on on Mars, we gave it the autonomy to kind of do its thing. And those two things in concert was what made that very successful. Yeah.

Certainly, from the science perspective, it's always helpful to have science in the loop, what we call it, for some prioritization and decision making, in particular, say, where you would like to sample. But certainly, I could envision a future mission concept that could land on an ocean world like Enceladus. We may give it instructions of

Oh, if you see, you know, discolored ice that might be indicative of a briny deposit of salts or organics, that would be one class of material that we would like to sample. Or as you're going down into the vent of Enceladus, maybe collect a sample every 10 meters or something like that so we can have a transect. There are certain things that we could certainly do.

a priori know that that's something that the robot would want to do. Given that the one-way light time between Earth and Enceladus is somewhere between, I think, 68 to 86 minutes, depending on where Saturn and Earth happen to be, it's much more efficient use of the robot's time to be able to conduct a lot of these things autonomously. But certainly, we on the science side and as humans are going to want to be a part of that story and a part of that experience as

And so having scientists in the loop at some point is definitely still in the plans. I am, just from a science perspective, really excited about the different ways that we can work with these kinds of technologies and these new robotic architectures. And so I think as things like eels need to adapt to their environment, we need to adapt as scientists to working with these new tools and be able to use them to get the most science out of every single mission that we send to these other worlds.

Well, there are so many things that we just can't do ourselves, right? There are so many wonderful places we'd love to explore that even if we went there in person, we wouldn't be able to go down those crevasses because it wouldn't be safe for humans. Yeah. So this is a complex situation. But there's two kinds of data we're going to need to collect. Well, many types of data. But there's the data that the system for navigation needs.

And then there's the actual science data that we're going to be collecting. What kind of basic sensors do we need to put on a device like this in order for it to understand how to navigate its terrain? Well, and Hero can certainly add to this, but a lot of those sensors can actually be used for science too. So as eels or robots like that are visualizing their environment, they use LIDAR, which is basically like echolocation, but with lasers. So it's a way you bounce laser beams off of

the structures around you and can form a 3D map of your environment. So we use that, we use stereoscopic vision as well. So if you were standing or if you were a snake slithering on the surface of Enceladus, you would get that 3D map, being able to use that to interpret your environment. You can also use that for understanding the geology or the rheology. So the ice properties on the surface of Enceladus and that can give you some idea of how certain structures forms or how they're evolving and changing over time.

So those are some of the primary things, but there are also other sensors that we've talked about integrating. For example, having a heat map, so some sort of thermal imager would be useful for identifying hot spots. These would be places that are warmer at the surface. Usually that's indicative of heating from underneath where potentially plume activity is happening or close to the plume activity. What other sensors, Hero, do you think...

Sensor-wise, I think that's complete. I would just add some engineering insight, particularly on mobility with some heat recover. So...

On Mars rover, the minimum thing you need is the geometry of the terrain. You have to figure out where the rocks are. When you walk around in the room, you have to figure out where the walls are. So that's geometry. And the Mars rovers use cameras, the stereo cameras, to figure that out. Every time it takes pairs of pictures,

and process them to reproduce 3D geometry. On the ground robots, we often use lighters, as you said, or we can use radars. There are many other sensors that we can use. There are a few more information that I think that can help in the future.

One is the semantic understanding, right? You know, which means that even though the geometry is the same, for example, the flat ground, walking on a flat, hard floor, right?

And walking on a flat, sandy floor, sandy ground is very different, right? And you have to walk in a very different way. If you walk on flat ice, right, in Pasadena Ice Wing, which where, you know, we did the first experiment. We did, yes. It's very different. Was it the one nearby at the convention center? Yes. Exactly, yeah. Right down the street. At like five in the morning, yeah. That, yeah.

We tested from like 10:00 PM to 5:00 AM. MELANIE WARRICK: That's right. That's right. MARK MANDEL: Oh my gosh. Yeah. HANNU RAJU: Anyways, so you've got to understand that kind of information, which is lacking from the current Mars war.

So what I did before was to train the machine learning, relatively simple machine convolutional neural net, to tell whether the terrain is rock or sand or... That would certainly help because you never know if the surface of Enceladus is made of fluffy snow or hard ice, if it is sintered. We don't know.

And sintered means that those little BBs, basically, those bits of water droplets that freeze as they come out in the plume, a lot of those, the majority of them, fall back onto Enceladus' surface. And they're not fluffy snowflakes. They're most likely little ice BBs. And what Hero's saying is that sometimes if they get hot enough or in certain conditions, they can form little necks and connect. That's called sintering. And so the surface of Enceladus could be like creme brulee,

It could be hard all the way down or it could just be this fluidized media where none of those grains are actually connected to each other permanently. And so we have to think about how we would navigate across any of those. Thank you, Morgan. This is the benefit of sitting next to a scientist. I do my best. So then in looking further in the future, I mean, from the experiences of EOS, I think what robots need in the future to go down the vent

It's a tactile sensing. Last year, we brought EGGS robot to a glacier in Canada called Athabasca Glacier. That was a success. We made a robot go down the natural crevasse over a few meters. I think that was the first in the world. It was successful. But that was a lot of teamwork on

Honestly, I was not a part of it. I was just talking. It was, you know, great engineers who made it possible. Wuhan Ducker, Mike Payton, you know, I have to name everyone probably. But yeah, so we used a few sensors, like five or six sensors to sense how much force you are receiving on the joint.

But compared to humans, do you know how many sensors we have on the palm? It's got to be... The palm of your hand? So many. I'm guessing it's probably like millions or something. I'm going to be on my own. Not millions? How many is it? I don't know. About 15,000 or so. Oh, wow. I learned something. Yeah. Not bad. But an enormous number of sensors are there compared to just a few.

And in fact, there was or there is a completely blind work climber who topped El Capitan, right? Only using the sense of touch.

And we often forget how much we depend on the tactile sensing from our foot, right? In fact, if you completely numb your foot, you cannot walk. So I believe in the future, we need a robot that has dense tactile sensor array that processes the information with neural net to feedback. And we don't have that technologies yet.

Yeah, you'd have to put some kind of like sensing skin on top of the robots. Yeah. But I think Kira touched on something that is really critical as we test and explore these new technologies is testing them in realistic environments. And for us, that is a glacial environment. It's not a perfect analog to these ocean worlds, but it's pretty close.

And so we were fortunate enough to have a field campaign in September of last year of 2023 at Athabasca Glacier, which is near Jasper or in Jasper National Park in the middle of Canada. It was an incredible trip and we learned so much by being able to test the

prototypes of the EELS robot as well as separate pieces. So we had a team that was testing different screw dimensions and compositions, how those interacted with ice, because ice that you find in the fields in these glaciers, sometimes it has, you know, different contaminants, it has different structures, because it's been exposed to the elements in the sun over time. And so we

Those are things that we can't necessarily recreate in the laboratory. And so having those testing campaigns where we do these realistic operations and see how the robot really behaves are incredibly, incredibly important. I've seen some of the videos of this thing. This was such a moment. And trying to get up and down a vertical, slippery face like that is a huge undertaking for a human, let alone for a non-human creature, this robot. Yeah.

How is it designed to actually allow it to stick to the ice and not slip straight down into this thing? So the principle is simple. You know, kids often climb up or down between two walls, right, by pushing the walls against each other. That's basically it. You push the walls against each other to create sufficient friction to support yourself.

It's easy to say, hard to implement. What you need is what's called force feedback control. You have to sense how much force you are giving to the ice by sensing the reaction force. And you've got to control it. Controlling it is also not easy because it's not a direct control. You have to apply the torque on the joints.

that translated into the force at hand, right? Which is, sorry for jargons, a highly nonlinear problem with enormous uncertainty. Plus this vehicle moves around, right? So the geometry of the wall changes continuously, which means that you have to actively adapt, change the joint torques

Because you cannot push too hard, not to the wrong direction. You have to keep pushing, right? Again, this is another example which is easy for humans, not easy for robots, right? And EELS also has the ability, so each module of the snake you can picture in your mind is a cylinder. It sort of looks like a two liter bottle of soda with the ends cut off.

that size cylinder. Each of those modules has these fin blades that are angled. It sort of looks like an inverted screw, or no, it would be a screw blade. And those are heated, can be heated. And so that's one way that we can get purchase

on rather slippery walls. As you heat those up a little bit, it melts some of the ice. Here on Earth, it melts it. On a place like Enceladus, it would likely sublime that water ice away, and then it gives you a groove that now you can use for being able to hold yourself up or down, depending on if it's plume forces pushing you up or gravity pulling you down. Heating might not be a necessity.

That's one of the things we learned, yes, from the field campaign. So in the particular case of our Athabasca descent, we used a robot, the interim version of a robot, which doesn't have sufficient torque to let the screw blade go into the ice. So during the engineering iterations, like a few months before that test, we figured out that, okay, so let's add the heater to compensate the lack of torque.

Now, we could remove that and we can, you know, give a stronger torque, but that also uses energy. So there's a trade-off, right? Whether you choose to have stronger actuator to let mechanically go into the ice or melt in. And yes, you know, there are lots of interesting trade-offs and challenges that we haven't figured out yet.

And I think that's also one of the important things about doing these developments early. So the most recent planetary science and astrobiology decadal survey, they recommended that a mission to Enceladus should follow Uranus orbiter and probe. But importantly, one chapter of that entire large report includes recommendations for technology developments. And I think that this highlights how...

how advancements in energy storage and things like that can really help us leapfrog forward in terms of the science that we can get out of it because it allows us to do these simple things that you might not necessarily think about, the importance of torque, the importance of having enough energy to heat to be able to grab a hold of something, those kinds of things. And so by developing all of these technologies in tandem, it really gives us opportunities

further down the road to say, okay, not only have we tested a concepts, but if we're able to implement these other synergistic technologies all together, now we can do even more. We can really explore this environment in the way that it deserves.

Yeah, we're in a bit of a pinch with a lot of these worlds. I mean, we've only been by Uranus and Neptune once. You know, if we want to go back to these places, it sure would be a shame if we realized, say, that there was some really cool terrain on Triton that we really want to go check out. There definitely is. There definitely is. There are these places that if we could not only just

put something into orbit if we could or fly by them, but also have this option to have this multifaceted thing that we can kind of iterate on for each world and have with this existing technology the ability to just drop them on locations as we go. That could be very powerful. And then that added...

benefit of having the AI not only be able to take care of itself, at least while we're not paying attention, but also maybe help us figure out what is most important out of the data that we're collecting. Because again, sending data back from that far is also a huge challenge. That's true. Yeah. These worlds that we're discovering are so diverse and fascinating. And each time we send a spacecraft somewhere new, we learn more than we could have possibly imagined. So just think about if we had dedicated missions for each of these

worlds around Jupiter, Saturn, Uranus, Neptune, and beyond. And as well, the asteroid belt, you know, Ceres, there's so many exciting worlds to explore and each one is deserving of its own mission, I think. And I just wish we could send them all right now.

Right. May I go a bit personal? Yeah. You mentioned the Voyager's Neptune encounter in 1989. That was actually the event that brought me to space. I was six years old that summer.

And my father is a space geek, so he watched everything about Voyager. He was so excited. So I sat next to him and watched TV back in Tokyo, Japan. And I was so fascinated.

you know, beautiful blue planet, the volcano, the crew of volcanoes on Triton, the gigantic storms on Neptune. And that was so fascinating. So I thought, you know, I want to build a spacecraft like Voyager when I grow up. And I learned that Voyager was built by this place called JPL, which I've never heard of at that time. So, okay, you know, in the future, oh, I wish I could be there.

And that happened somehow, like 20-some years later. And here I am. Who knows? One of these days, maybe this technology that you've both worked on could go off to these worlds and show the next generation something that inspires them to then break those boundaries. Definitely want to go back to Neptune in my lifetime.

That would be amazing too. That's my favorite planet, honestly. And one of the coolest things I've discovered just working on the EELS project is that the snake has already grown legs in a certain way that some of the technologies that we developed, particularly some of the autonomy technologies, as well as a few of the others on the engineering side, we're already seeing those being incorporated into other mission concept developments and other engineering and robotics projects that we're working on at JPL.

And so I think that's important to remember that not everything that you work on, that instantiation, that moment in time that you've created may not itself be.

explore, say, the events of Enceladus one day, but pieces of it are going to live on and emerge and be able to help us answer some of these cool questions in ways that we can't even predict right now. And I think that's really cool too. But I definitely hope to see an EELS-like robot on another world in my lifetime. I really think that would be incredible.

And, you know, honestly, you never know how soon or how far, right? In my childhood, when you look at the book, right, you know, by 2020, humans are walking on Mars, you know, living in space, which didn't happen, unfortunately. But I sometimes think in this way. Sometimes it's easy to predict what's going to happen in the future, but hard to predict when.

So I think, you know, we're going to build a city on the moon. Humans are going to be on Mars. That's one part. I'm sure it's going to happen. I think a mission to the subsurface ocean of icy moons will happen. Why? Because there's this fundamental question that we want to answer, right? This is probably one of the oldest questions of humankind. What I don't know is how soon it's going to happen.

Right? If NASA gives us like 5 billion bucks tomorrow, then we can probably do it in three years. But, you know, it's not happened, of course. It could be, you know, many decades, many centuries. But, you know, no matter how long it takes, it will happen. And I believe, you know, I think it's fair to say that we...

made a great leap in technologies toward that direction. So I never know if that spacecraft that actually goes to the ocean will look exactly like eagles or not. Probably not. But, you know, our contributions will be there.

NASA's mission has always been driven by the science. That's one of the reasons why I think we work at JPL is because science is the driver. It's the questions. And how do we answer those questions? So as long as there is NASA, as long as we are still guided by these questions,

the science and looking up at the night sky and wondering, as long as that is still our guiding light, our candle in the dark, I think, yes, we will be sending spacecraft to these ocean worlds and exploring their watery depths and hopefully being able to answer this question. Just got to dare enough mighty things one step at a time. Exactly. Or one slither at a time. One slither at a time. Yeah.

Well, thanks so much for joining me and for working on this technology, whether or not we get to see it on Enceladus someday, or maybe someday scouting out locations on the moon when we build our city, right? It would be wonderful to see this technology put to use and just to see all the spinoffs that come from it. So I really appreciate both of you working on this, but also for being here with me today to talk about it. Yeah, thank you for having us. Thanks so much. It's been such a pleasure. And now it's time for What's Up with our chief scientist, Dr. Bruce Betts. Hey, Bruce. Hey there, Sarah.

I love that people are trying to come up with innovative ways to explore these places. Because not only do I want to know what's going on down in there, but real talk, even if we had humans straight up on the surface of Enceladus, who wants to volunteer as tribute to go down that crack?

Wow, even you say that? Okay, no one then, apparently. Yeah, no, even I would not spelunk down in there, but I do have a deep fear of getting stuck in a cave. Wow, we learned one more thing about your intriguing psychology. I know, right? Totally go to Mars, can't handle a cave.

But yeah, this is one potential use case of this technology and a good one. We need to find a way to get down into the cracks of Enceladus if we're going to do this kind of science. But there are a lot of places in our solar system that we are just currently unable to explore, even if we sent humans there. So what do you think are some of the good places that we could target with this potential technology? Yeah.

Oh, there are so many places that would be wonderful to go and hard to do and hard to ever get approved. But then, you know, there have been things that have approved and worked that I never believed would happen. So I'm not necessarily the best judge of that. But if you had these super cool things and you had a super cool snake eel, snake eel,

Then, you know, any more caves, crevasses. So the moon and Mars both have caves, lava tubes that one could explore that would be spiffy keen to get under the surface. You've got all these icy worlds that you'd love to dive down into with Europa and Enceladus and so many others. I mean, you could get crazy with Kuiper belt objects and

that are,

And you can always go to Venus and find weird things. I mean, it's like the Earth. If you wanted to explore the Earth robotically on the surface, there are all sorts of mountains that would be very challenging to do that. And Mars, all the terrestrial planets have the equivalent, and all of the icy bodies pretty much have some equivalent probably of the trying to explore glaciers with crevasses. Crevasse? Crevasse? Yeah.

Anyway, those things. So there's all sorts of things you could do with it. And plus, you could have a race on Mars. So, yeah, let's do it. I mean, seriously, though, there are so many places that we can't.

visit as much as we like and putting orbiters around all these worlds. That's so challenging. If we could just fly by Neptune and drop one of these eel robots onto Triton or something, that'd be so cool. The challenge you've always got, the realistic challenge is you tend not to start with the really hard, dangerous places because you have such higher risk of

I mean, it's dangerous enough to go to the surface of any of these bodies. And so when you go to Mars, you go to the lower areas so you have enough time to land. You go to places without rocks all over the place when you first land. And you get more advanced rovers and landing systems. They can do more. But you go to the moon with humans, you land in really flat places that you think are safe. And so...

That's the reality. But on the other hand, hey, we've got a quadcopter drone thing that's going off to Titan. And we had the crazy landing systems that JPL developed for Mars that worked, airbags and sky cranes. So who am I to say they can't do these things? They get crazy and figure them out. It's pretty cool.

Yeah, we'll just start with crevasses here on Earth. We'll move on to the moon as if that's not super complicated. And then hopefully maybe one day Enceladus. And then someday exoplanets. Far-flung future there, but how cool. Far-flung future. Band name. Band name. Well, all right. What is our random space fact this week? Hmm.

So I'm not sure what you discussed about Enceladus, but it's in your interview, but it's an intriguing thing. I'll give you a couple of facts just to be safe. One, it has the brightest surface overall in the solar system.

And we didn't know why until they discovered these plumes, bitten out ice, but it resurfaces with nice, bright, snowy ice. And that's what does it. The other thing is, it always strikes me about Enceladus that makes it so weird to me that you can do a subsurface ocean, is it's so small.

It's not that much bigger than what you need to make a round body. Its diameter, not to mention its volume or surface area, is like six times smaller than Europa. In fact, it would fit roughly across east-west the state of Pennsylvania. Or if you want to go from Philadelphia and Pennsylvania to Boston, it would be about that distance. And so it's a wee little pup when it comes to moons.

And yet it's got this amazing, interesting surface. So that's why we should start with Eels Exploring Pennsylvania. It would be cool to stand on the surface of one of the moons of Saturn and be able to actually see how reflective Enceladus seems in the sky. Because you could say it's the most reflective surface. But unless I have some kind of visual representation, it's hard for me to understand what that means. Yeah. Yeah.

This probably won't help, but overall, it's Bond albedo is like over 80%. So it reflects over 80% over the spectrum, the whole spectrum of the light coming in. That's compared to like Earth with its clouds and things. It's like 30% ish, 35. The moon we think of as really bright because it's so bright compared to the black around it, but it's fair.

It's measured in like a few percent in terms of reflection. It's on average very, very dark. And so if you put in, ooh, that's a good thing, put Enceladus where the moon is, that puppy would be bright, although it's much, much smaller. So think about that. Maybe a random space fact of the future. We'll just have to make sure that the first people to go to Enceladus have polarized glasses.

Nice. Boy, his glasses are cool. Yeah, they are. All right, everybody. Go out there, look up at the night sky and for no apparent reason, think about tree rings. They're kind of cool. Thank you. Good night. Dendrochronology. We've reached the end of this week's episode of Planetary Radio, but we'll be back next week with more space science and exploration.

If you love the show, you can get Planetary Radio t-shirts at planetary.org slash shop, along with lots of other cool spacey merchandise. Help others discover the passion, beauty, and joy of space science and exploration by leaving a review and a rating on platforms like Apple Podcasts and Spotify. Your feedback not only brightens our day, but helps other curious minds find their place in space through Planetary Radio.

You can also send us your space thoughts, questions, and poetry at our email at planetaryradio at planetary.org. Or if you're a Planetary Society member, leave a comment in the Planetary Radio space in our member community app. Planetary Radio is produced by the Planetary Society in Pasadena, California, and is made possible by our members from all over the world. You can join us as we support the search for life off of Earth at planetary.org slash join.

Mark Kilverta and Ray Paoletta are our associate producers. Andrew Lucas is our audio editor. Josh Doyle composed our theme, which is arranged and performed by Peter Schlosser. And until next week, Ad Astra.