Podcast: Robotic Pioneers on the Moon's Surface

In this episode, we discuss a joint endeavor led by ETH Zurich to leverage robotic teams to explore the lunar surface for precious materials. Our favorite bit about this implementation of robotics? It is a role based team with redundancy built-in as team member backups  in case of a failure during the mission.

EPISODE NOTES

(0:50) - Robot Team on Lunar Exploration Tour

Transcript

Folks, it's been over 18,500 days since we've sent a man to the moon to explore and understand what's going on there. In this episode, we're talking all about a team from ETH Zurich and a bunch of other universities in Switzerland that's developing a pack of robot dogs that can go to the moon and do the exploration for us. So, we better understand what's out there on the surface and if it's more worthwhile for us to send more missions to the moon in the future. I think it's really interesting, so let's buckle up, put on your space helmet cause we're launching into this one, folks.

I'm Daniel, and I'm Farbod. And this is the NextByte Podcast. Every week, we explore interesting and impactful tech and engineering content from Wevolver.com and deliver it to you in bite sized episodes that are easy to understand, regardless of your background. 

Daniel: What's up, everyone? Like we said, we're talking about this team from Europe. It's a group of a bunch of researchers from ETH Zurich, FZI Research Center for Information Technology in Karlsruhe, Universities of Basel, Bern and Zurich. It's a super team, folks. They've assembled the Avengers to create a pack of robot dogs that are training to explore the moon. It sounds way out there and it's because it really is his way out there up and out of space on the moon, but it's really exciting. And like super corny here, but I think it's one small step for a robot.

Farbod: Oh wow! You're going there.

Daniel: One giant leap for space exploration because we will potentially have the opportunity with these four-legged robots. I say dogs, they aren't meant to mimic dogs, but think of like Boston dynamic spot or the guard dogs in Black Mirror, that type of four-legged animal, but it's been designed by a team from Europe and they call it ANYmal because it can do anything. And there's four-legged robots, they're training them, these smart robots to be a team and work together on an exploration trip on the moon.

Farbod: Why? Why are they exploring the moon, Daniel?

Daniel: Well, there's a couple of reasons why obviously I think we as humans have this like burning desire to understand what's out there in space and the moon is the closest thing to us. It's large, it's vast, it's pretty much completely unexplored. Manned lunar missions are really expensive. They're risky for human lives. They have limited exploration capacity because you can only sustain life out there for so long. But even more so than that, we are interested in potentially harvesting and understanding what minerals there are on the moon, rare earth metals, rare moon metals, right? There could potentially be some really valuable resources on the moon.

Farbod: Absolutely.

Daniel: One for scientific value, right? For us to understand the geology of it all. And then in addition to that, we think that there are specifically on the moon, very, very valuable metals and minerals for us to use on Earth for things like electronics, etc.

Farbod: Yeah, it's actually kind of funny how little we know still about what's on the Moon. I think it was a couple years ago where we literally shot a rocket at the Moon because we thought there was like an ice cap somewhere and by blowing it up, we were like, oh, that actually is an ice cap. So, at some point there was water on there. Awesome. With the potential of rare Earth metals on there and our heavy usage of these rare earth metals and the affordability of going to space, you're starting to open up this entire new industry of space mining. And that's so exciting. But we first need to understand if there's even anything of value there. We need to understand what's there to begin with. And that's where it all started is the European Space Resource Innovation Center, they basically hosted this competition of like, who can come up with the best system to get this task done? Right? And that was the inspiration behind what this team has developed.

Daniel: Well, and I think if we think about where, how long ago it's been since we done, like as humankind have spent a lot of effort on doing exploratory missions on the moon. I know the Indian Space Agency just did another exploration mission, but outside of that, we haven't done many major explorations on the moon for decades.

Farbod: Yep.

Daniel: And if we think about how much robotics have improved over the last couple of decades, and we haven't yet had the opportunity to capitalize using these robots on a foreign surface, like the moon, we're not talking about like a robot that you can make with a Lego mindstorm kit with like a couple of tank treads here, we're talking about something that can run, can crawl, can jump, can traverse different types of terrains, is incredibly efficient from an energy perspective. It is also really, really durable in terms of dust and dirt and the elements, temperature wise, et cetera. So, we've got all these advancements in technology. This is why the European Space Agency and the European Space Resources Innovation Center, that's why they hosted this competition. And that's why this Avenger super team of these European universities is working together with this ANYmal team to try and create a pack of robotic dogs that can help leverage the most cutting-edge technology from robotics to help us understand what's out there on the moon.

Farbod: Yeah. And let's talk about this pack of dogs that won the competition, right? What makes them so special? Like you said, ANYmal is like this variation of a quadruped robot. We know MIT has their own variation of it. Boston Dynamics has a version of it. So, what was the secret sauce that made this team really win? And I think, based on my understanding of what went on, it was the coordination of the team that really made it special, right? Because they had three dogs, two specialists, one generalist. So, you had one specialist whose entire job was mapping and geology, understanding the terrain and generally what it's you have big rocks here, small rocks here, whatever. Then you have another one whose entire purpose is spectral analysis, like we used to do in ChemLab. You know, you have the Ramon spectros, spectros, spectromity, what was it, the? Ramon spectros. Something spectra. Spectrometry, that sounds right. It's like the light hits the material and what's reflected back tells you the composition of what you're looking at.

Daniel: Spectroscopy.

Farbod: Spectroscopy. Close enough. You know, ESL, I took ESL. Anyways, its entire role is understanding what something is composed of. Then you have the generalist, which can do both of those jobs, just not as well. So, I'm assuming there's some sort of neural network that's embedded in these systems. And you have a much more condensed version of it, of either task on this generalist. And the reason that the generalists even exist is for redundancy. Because, you know, space is kind of wild. It's kind of wacky. If one of them goes down, you still want the mission to complete.

Daniel: And what I will say is you mentioned there are probably some specific software tunings, like let's say inside the proverbial brain of these robot dogs that helps make the specialist a specialist or helps make the generalist a generalist. But there are also a lot of physical hardware considerations that went into designing each of these robots into their specialization. They had a modular design, basically meaning they have this base design, but they can remove certain parts and interchange them with different parts based on what that robot's role was going to be in the team, and then also what the overall mission requirements are. So, they're saying they've got one mission they planned for right now where you do mapping and geology and another one to spectral analysis, but maybe we can do another one where one robot is specialized to look for water. Another one's meant to specialize and understand the weather and another one's meant to collect rock samples. They can interchange the different hardware to make sure that each of these robots made on this interchangeable platform, right? They're all built on the same base technology, but you can specialize each of them to be able to be really, really excellent at one of these specific tasks. Or, like you're saying, have a set of generalists that maybe aren't as efficient at doing spectral analysis as the specialized spectral robot, but they've still got the capability to do that in a case where that redundancy is required.

Farbod: Yeah. And another obvious benefit of all this is in comparison to the standard Martian rovers, instead of having to rely on one rover that has to do all these different types of experiments sequentially, now you have a bunch these robot dogs that can just do the tasks in parallel. So, it's obviously more time efficient. One added plus there. So, what, but we already talked about it, they won this competition, but what's come after this one, right? Well, the European Space Resources Innovation Center gave these folks a one-year contract to keep continuing their work. They wanna see this solution of collaborative robots with redundancy and everything expanded and built upon to really make this a feasible solution for this type of exploration. And one thing that was pretty cool to me is, we've talked about this animal approach, this platform that can be adapted for different situations. This team has now actually said they would love to start adding roving and flying robots to this solution as well to really leverage their strengths for different types of exploration like you were mentioning, which is, you might be able to fly away far out and collect spatial data about water sources, et cetera, et cetera, whereas the rover might be able to do something else that's much better.

Daniel: Well, and it makes me think about how if we were sending a manned mission to the moon to do exploration, we would in all likelihood have very specific folks with very specific tasks to do certain things, right? You'd have a geologist there meant to help study the geology. You'd have a scientist there to help with spectral analysis. You'd have someone there who's meant to focus on communication. You'd have someone else who's meant to sit there and do like geoinformatics systems and combine this all together into data that we can send back to earth in the same way that we have very specialized human rules. We're trying to kind of mimic that collaboration, mimic that teamwork. But in this case with all a fully robotic team not a man mission there, on the moon. And we mentioned the moon. They think that this is a platform that could be interchangeable and used for other interplanetary missions, making exploration more sustainable, more thorough on Mars and on other planets, etc. But one of the things that I thought was really interesting is they've, at least in this iteration, had the hardware very specifically designed to be efficient on the moon. So, they use solar panels and battery storage that were optimized for the lunar environment. That means the intensity and the timing of the sunlight. The temperature range that's there on the moon, they very specifically designed this hardware to be successful on the moon to make sure that if and when we send these robotic dogs to Mars, we can ensure longer mission durations because of all the things that we do know about the moon from our previous missions, we've been able to collate that data and design hardware that's specifically tuned to perform at its peak in those conditions.

Farbod: Right. What do you think the future of this looks like? Are we gonna start mining every planet that's around us and then not start this in future?

Daniel: Well, I think that, I mean, I have a little bit of hesitation around that. I think that's definitely a potential future outcome. But one of the things that's most exciting to me here is just that the pure fact that we have not been able to truly explore and map and understand what's there on the lunar surface, at least beyond what we can see from a telescope and the small areas of the moon that we've covered with manned and unmanned missions. Something like this that's really energy efficient, that's purpose-built for exploration on the surface of the moon, could allow us to cover large lunar areas in detail. And that would enable us to increase the chances of a significant discovery in the future, whether that be ice and water, whether that be a certain mineral that we wanna go mine, or something else that we think is interesting and leads us to understand more about the origin of the moon or kind of the geological history of the moon and the earth, et cetera. All of those are very interesting potential outcomes, but they're all contingent on us getting a better understanding of what the surface looks like and what it's comprised of. That's something that this robotic team could help us achieve.

Farbod: You're right, you're right. My main concern when I was first reading this was, I feel like you're gonna keep increasing the payload that has to make it on the lunar surface or like the surface of any planet that you wanna explore, but not only has it become more and more cost efficient to put stuff into space, but we're getting better at making sure that these missions are successful. And then on the payload side itself, computing has gotten so much better, we've gotten so much better at robotics that we can keep miniaturizing this, whether it's the hardware or the computational resources that we need to make the software run, which means that we can get more out of smaller robots. So, I don't know, this seems like a pretty feasible solution in general.

Daniel: Yeah. And especially in the world and in the time that we're living in, like you're saying with advanced AI algorithms that we even talked about it on our last episode on how we can kind of create a hybrid AI system where some of the AI computation is done at the edge, right? You know, we've got advanced AI algorithms that are possible to be streamlined and shrunk down onto something that fits inside a robotic dog or inside a robotic helicopter or something like that. One of the things though that this team mentioned, is at least in the current state, the team is still not fully autonomous. So, you've still got a lot of human input into understanding and basically what they do is they have a human working like a project manager, assigning tasks to each of the specialists, each of these specialist robots, monitoring the completion of those tasks and then providing a new set of tasks to help complete the rest of the mission. I think it'd be really, really interesting for them to create another like generative AI layer over all of that, that helps optimize a swarm of maybe hundreds of these robots to help explore the entire lunar surface in a matter of months I don't think that that's out of the realm of possibility. I think it's just something that this collaborative robotic team is just new enough that they haven't had the opportunity to develop it yet. I think that's something that could be really, really interesting for them to develop in the future. At least to where we've got the point where for something as benign as exploration, we can allow these robots to operate in a really efficient manner and help us understand as much of the lunar surface as possible or the Martian surface as possible and just like I said, any further action on either the moon or Mars, etc. is highly contingent on us having a really strong understanding of what the surface looks like and what it's comprised of.

Farbod: Perfect timing. You were talking about generative AI for training robots. Another throwback to our trip to MIT, where we were talking to a researcher who was leveraging generative AI to train robots. So yeah, folks listening, hint, hint. We got some content coming out from MIT. We made a little visit a couple of weeks ago, and yeah, this fits very nicely with that little piece of information.

Daniel: No, I'm with you, man. We keep seeing the world's smartest folks, the sharpest most brilliant folks working on very similar topics that butt up against one another and even have synergies with one another. I think it's really, really interesting to watch the technology world unfurl from where we are. You know, just two friends making a podcast. I think it's really interesting, but especially keep your eyes and ears peeled for our content from MIT because we had the awesome opportunity to speak directly with the team who are working on some of that and ask them some tough questions to help understand what it is they're working on, why they're working on it, and kind of what their secret sauce is and achieving and solving the big problems that they want to. So, keep your eyes and ears, eyes and ears peeled for that. And then also if anyone from this team, from ETH Zurich wants to jump on and have an interview with us, we would love the opportunity to meet you folks, dive in deeper with some questions from our community. And get a deeper understanding of what this team of robotic dogs is doing and how you're achieving the awesome work that you're doing.

Farbod: It would also give us just a good excuse to fly out.

Daniel: Yeah.

Farbod: A little The Next Byte trip.

Daniel: Yeah, there we go. Get ourselves out to Europe.

Farbod: Yeah, before we wrap up, please do us a favor with a nice little TLDR, what we just talked about.

Daniel: Yeah, so to wrap this up into the TLDR version of what we talked about today, we talked about how we're taking robotics that we've developed on the earth and sending them out to the moon. So, it's corny. I said it, but it's one small step for robot dogs and it's a giant leap for space exploration, because instead of sending a group of astronauts to the moon, which is really, really expensive, really, really risky and has a limited exploration capacity, this team from ETH Zurich is working to create a group of smart robots that we could send on our own exploration trip to the moon. These robots don't just take pictures. They're like detectives. So, they each have very specific roles. They crawl around investigating rocks, mapping out the land, sharing notes with one another, collaborating with one another, all while navigating the really tricky landscape and climate on the moon. This is obviously just been a theoretical development and they've done it in a testing environment here on Earth, but we're looking forward to seeing if and when the European Space Agency partners with this team to send these robotic dogs up onto the moon. We can learn a little bit more about the lunar surface, which will help us understand what's out there. How can we capitalize on it and whether it's worth sending more missions out to the moon in the future.

Farbod: Money, killed it!

Daniel: Thanks, my guy.

Farbod: All right folks. I think that is it for the episode. Yeah, thank you so much for listening as always catch the next one.

Daniel: Peace.

As always, you can find these and other interesting & impactful engineering articles on Wevolver.com.

To learn more about this show, please visit our shows page. By following the page, you will get automatic updates by email when a new show is published. Be sure to give us a follow and review on Apple podcasts, Spotify, and most of your favorite podcast platforms!

--

The Next Byte: We're two engineers on a mission to simplify complex science & technology, making it easy to understand. In each episode of our show, we dive into world-changing tech (such as AI, robotics, 3D printing, IoT, & much more), all while keeping it entertaining & engaging along the way.