How to 3D print fully formed robots

Shamini Bundell

Welcome back to the Nature Podcast. This week, the 3D printer that creates complex bio-inspired robots...

Benjamin Thompson

...and assessing the effectiveness of lifestyle interventions at preventing diabetes. I'm Benjamin Thompson...

Shamini Bundell

... and I'm Shamini Bundell.

Shamini Bundell

First up in the show, Nick Petrić Howe finds out how to 3D print a robot hand.

Nick Petrić Howe

If you look at your hand, from a mechanical perspective, it's quite complicated. It has bits of soft material, muscles and ligaments, combined with hard material, the bony skeleton that sits beneath. Together this allows us to do everything from banging with a hammer to threading a tiny needle. So, to make something like this, artificially, is quite a challenge. But a new paper in Nature has managed it using 3D printing. By overcoming a key limitation of inkjet 3D printing, a team has produced an artificial hand, with synthetic bones, ligaments and tendons capable of grasping different objects.

Robert Katzschmann

So, we could really show that we can get a functional, human hand inspired structure that we can pull on the tendons and then the structure moves and shows its range of motion.

Nick Petrić Howe

Now, you could make an artificial hand using conventional techniques, where you cast materials into moulds to form the desired shape, and then combine them with other materials. But these methods require tedious calibrations. And it's difficult to incorporate lots of materials with different properties, limiting what the final robot can do. Also, they lack the very fine details you would need to make something as complex as a synthetic human-hand. So, researchers have been interested in using 3D printing to do the job, as it can be incredibly precise and quick. But it has its own problems, such as the difficulty of combining different kinds of materials together, as Thomas Buchner, another of the paper's authors, explains.

Thomas Buchner

If we were trying to 3D print robots, we always had to use maybe different separate printers and then assemble the robot, or we had to do a little bit of 3D printing, and then cast around those 3D printed parts to get those different material properties that we were looking for to actually have a functional robot in the end.

Nick Petrić Howe

So, the team has been trying to streamline this process. Rather than multiple printed elements being combined together, they wanted to create a system that could print a complex robot in one go. To achieve this, they were working with a type of 3D printing known as inkjet printing, which sprays liquid from a nozzle — similar to home printers. The big advantage of inkjet printing is that it's faster than other kinds of 3D printing whilst still being precise, but there's a downside too. When something is inkjet printed, droplets of material can accumulate, so the surface becomes quite uneven. Typically, this is dealt with by using some kind of scraper to get rid of the extra material and get your would-be robot ready for the next layer of printing. But this extra scraping step can be a problem. For one thing, it limits what kinds of materials you can use — some materials can accumulate and kind of gunk up the scraper. So the team made a printer that could... keep an eye on the process.

Thomas Buchner

Here, we have a printer that kind of has eyes. So there is a laser scanner that scans the whole print bed then understands where there's too much and too little material and automatically corrects for that in the next layer that is deposited, so that it kind of fills up those unevenness in those lower spots.

Nick Petrić Howe

Instead of scraping away the excess, it doesn't make the excess in the first place. If one bit is too thick, the robotic eye can see that and make the next layer that little bit thinner, removing the need for a scraper at all. Which means a lot more materials can be used, which in turn means they can make all sorts of complicated bio-inspired robots. Here's Robert again.

Robert Katzschmann

We also build a little walking robot that has little arm with a gripper on top that can pick up boxes and other items. And also, we build a heart, sort of like a pump that is like a human heart with ventricles that we pressurise, and then we can make a fluid flow go through these ventricles.

Nick Petrić Howe

And all these robots were printed in one go, without needing to go back and add different materials for the different parts. They did, however, need to connect some electronics externally to do things like sense pressure in the hand, but nonetheless, the method detailed in this paper, impressed mechanical engineer and 3D printer aficionado, Yong Lin Kong.

Yong Lin Kong

I thought it was very impressive, and it's fascinating to bring a breakthrough into a technology that many have felt that is sort of in the mature state – it is already commercially available for decades. I think a lot of us might assume that, oh, this is the best this machine can do. But this paper show that just by incorporating machine-vision, one can bring material compatibilities that was otherwise not possible into this type of process and, and allows us to create not just multi-material construct, but in high-resolutions and higherthroughput approach.

Nick Petrić Howe

Yong Lin would like to see more materials being used in this approach in the future. For example, they could attempt to incorporate electronic components during the printing process, further speeding up the path from idea to robot. The team are already looking at this as they progress this research, and Robert is optimistic that this technique, even beyond robots, could help make medical implants... and even jazz up your shoes.

Robert Katzschmann

You could think of medical implants, you could think of much more intricate shoes, for example, intelligent shoes, or you could also use this in terms of prototyping things for tissue engineering, and provide structures in this direction. So the technology itself will only grow by now people developing more and more materials that can be printed on this contact-free printer. And the more you bring different materials in it, suddenly the systems or just the designs that are being printed, they become relevant for all kinds of fields, that's at least what I imagined will happen.

Shamini Bundell

That was Robert Katzschmann from ETH Zurich, in Switzerland. You also heard from Thomas Buchner, also from ETH Zurich, and Yong Lin Kong, from the University of Utah, in the US. For more on this paper, check out the show notes and we'll also put up a link to a video we've made about it showing that the robots in action.

Benjamin Thompson

Later in the show, testing the effectiveness of diabetes lifestyle interventions. Right now, though, it's time for the Research Highlights with Dan Fox.

<Music>

Dan Fox

Astronomers have teamed up with citizen scientists to identify an object that blurs the line between asteroid and comet. Quasi-Hildas are asteroids located beyond our Solar System’s asteroid belt, but within Jupiter's orbit. Around 300 have been spotted, but only a few are known to be active, displaying comet-like behaviour like having a tail made of dust or gas. As part of the Active Asteroids project, researchers worked with volunteers from the public to identify an active Quasi-Hilda by examining archive images. A thorough image search and follow up observations then revealed that this asteroid was active when was closest to the Sun in March 2016 and April this year. The researchers conclude that these periods of activity probably occurred as the Sun's heat converted solid forms of compounds like water and CO₂ into gas. They say that this finding could improve our understanding of the distribution of these chemicals in the solar system. Read that research in full in the Astrophysical Journal Letters.

<Music>

Dan Fox

Sometimes, you don't need to understand the joke to find it funny, and sometimes no amount of explanation can make a gag work for you. Now neuroscientists know that understanding a joke and finding it funny are distinct processes in the human brain. Researchers studied the brain activity of 26 people as they consumed humorous material. The participants first listened to 40 jokes and 14 neutral sentences that were played in a random order. After hearing each clip, participants had to decide whether it was a joke and rate how funny it was. Next, they watched an episode of the sitcom Seinfeld and completed a questionnaire to assess whether they enjoyed the comedy. Brain activity during both experiments showed that understanding a joke involves activity in two brain areas, the dorsal striatum, which has a role in memory and cognition, and, the ventral striatum, which responds to reward, but actually enjoying the joke only involved the ventral striatum. Both areas are rich in the chemical messenger dopamine, and the team suggests that dopamine signalling could play an important role in humour processing, opening avenues for further investigation. What's the deal with that research? Find out in the Journal of Neuroscience.

Benjamin Thompson

Type 2 diabetes is a leading cause of morbidity and mortality. Estimates suggest that type 2 diabetes affects hundreds of millions of people around the world, significantly burdening health systems. But behaviour change programmes – also known as lifestyle interventions – could potentially play a large role in preventing people from developing type 2 diabetes. And this week in Nature, there is a paper assessing just how effective this kind of intervention might be. Joining me to talk about the work is George Caputa, a senior editor for the journal Nature who handles a lot of papers involving metabolism and physiology. George, thank you so much for being on the show.

George Caputa

No, it's really great to be here, Ben.

Benjamin Thompson

So, we've got this new paper then focusing on lifestyle behaviours and clinical trials have shown that behavioural interventions can make a difference in preventing folk from going from being pre-diabetic, where someone has elevated blood sugar, to full type 2 diabetes.

George Caputa

Yeah, this is not the first epidemiological study or even just a scientific study in general, showing that behavioural interventions can help individuals lose weight or even prevent the development of full type 2 diabetes. I think that something that we have gotten to, especially at this point in time, is that we do we have access to huge amounts of data. And so we can start to make potentially more definitive conclusions. And these are important, obviously, because a lot of these studies are the basis for public policy, public health, governments, they decide about what to fund depending on these types of studies.

Benjamin Thompson

And in this case, then the researchers behind the paper and looked at a data set gathered from health data from folk who referred to different interventions, but primarily England's NHS Diabetes Prevention Pprogramme. What do we know about that?

George Caputa

Yeah, so the NHS Diabetes Prevention Programme is a behavioural change programme with weight loss, diet, physical activity goals, and this consists of 13 group sessions over the course of nine months.

Benjamin Thompson

And overall, what was the big question that the researchers wanted to answer in this paper?

George Caputa

Yeah, so the question is, if an individual is diagnosed with pre-diabetes, if they are provided with a behavioural change programme, can we prevent them developing full diabetes? Can we kind of stop the progression.

Benjamin Thompson

And so in this work, then they take the anonymized patient records from about 2 million people to analyse the effectiveness of behavioural change programmes. And they've done a lot of statistics, in particular, something called a regression discontinuity design. So broadly, what have they done to show the effectiveness of behavioural programmes?

George Caputa

Obviously, in an epidemiological study, you have to be looking apples to apples, because you have to make sure that the groups that you're looking at are being intervened in the same way that they all qualify the same, because then you start to get behavioural and environmental factors that can confound your analysis. And regression discontinuity design, very generally, it is a way to within this population be able to make potentially statistically significant claims about the changes that they see, and that they are very much influenced specifically by the intervention of this programme. And not because the individuals that they're looking at are different groups or influenced in different ways. And one of the reasons that we considered this paper is because it is using state-of-the-art statistical analyses. So this is something that was very appreciated by the reviewers that what was being done here was using some very new and modern techniques.

Benjamin Thompson

And what's the headlines of what they saw then in terms of folk taking part in these programmes?

George Caputa

So, I think to say that they saw a very dramatic results is not the case. But what they did see is that there was statistically significant changes in the measures that they were using to qualify these individuals for pre-diabetes. They're looking at HbA1c which is a blood marker for how high your blood sugar has been basically over the past month, and so it's a good indication of whether your body over the past month has been able to properly maintain its blood glucose, which is a measure of how pre-diabetic you are. And what they saw was a decrease in the levels of HbA1c in the individuals who participate in this programme, they also saw a decrease in body weight, and they also saw a slight decrease in blood pressure and these associated metabolic measures. And I think what the researchers were able to definitively say is that this intervention had an effect. Obviously, in a situation like this, for some individuals, it probably was a big effect. And for some individuals, it probably was quite small. That's why when you get the average, it tends to be more modest.

Benjamin Thompson

And as someone who is an editor who looks at science papers then, what is this paper done maybe differently to things you've seen before?

George Caputa

What I tend to see as an editor is a lot of proposed pharmacological treatments. So like new drugs, new targets, new potential pills that people can take. And so to see a study that's really looking at another type of intervention, that can be just as powerful as putting someone on another drug, or even combined can be very powerful. That is something as an editor, that's really important, because this is a space that we don't talk about enough in terms of public health, in terms of these types of behavioural interventions. I think that as a society, we're kind of very focused, and we associate the most effectiveness with a drug.

Benjamin Thompson

I'm sure, though, that it's not the case that this is kind of everything sort of wrapped up with a neat little bow, there are those who doubt the effectiveness of behavioural intervention programmes. And I'm sure there are questions that remain to be asked when other people get their eyes on this paper.

George Caputa

Yeah, I mean, you know, I think there are people who will look at this paper and say, we're not seeing the same dramatic changes in blood glucose levels in this HbA1c that we would see if we were to immediately put these people on, we would consider frontline, first choice, diabetes drugs. The truth is, the individuals that they're looking at in this study aren't diabetic, they're pre diabetic. And usually, in many countries, the clinical intervention here is that your doctor says you have pre-diabetes, you need to change your behaviour, maybe gives them access to some information, but that's probably it. You know, this is like a full programme. I think the critics of the paper would say that this change is not so big. But when you're looking at the numbers of people who have diabetes, if you can reduce maybe 2% of those cases, that is a huge amount of savings for a public health programme, and for people's lives.

Benjamin Thompson

Nature's George Caputa there, to read the paper we chatted about by Lemp et al. and an associated News and Views article. Look out for links in the show notes.

Shamini Bundell

Finally, on the show, it's time for the Briefing Chat, where we discuss a couple of articles that have been featured in the Nature Briefing. So Ben, why don't you go first this time?

Benjamin Thompson

Yeah, I've got an article this week that I read about in Wired, and it's based on a paper in Science, and it's looking at a potential explanation for why an estimated 10 billion snow crabs disappeared from the Bering Sea off the coast of Alaska. Now, this is a climate-change related story, but maybe not in the way you might expect.

Shamini Bundell

Ooh, I haven't actually heard of snow crabs. Can you paint me up a visual description?

Benjamin Thompson

Yeah, so this species of snow crab we're talking about in this case are about maybe 10-15 centimetres across, look like a crab and they live on the sea floor. And they are an extensively monitored species and they're economically very important for fisheries and so forth–

Shamini Bundell

– oh –

Benjamin Thompson

– but populations, as I say, absolutely crashed following marine heat waves in 2018 and 2019, like 90% of the population, which had previously been quite healthy. And this corresponded with water temperatures rising by three degrees.

Shamini Bundell

So, these marine heat waves were particularly impactful on these crabs.

Benjamin Thompson

Well, absolutely. And it turns out, it's not just that it was too hot for the crabs, which is what I think you might expect hearing that –

Shamini Bundell

– that's what I would assume, yeah –

Benjamin Thompson

– well, actually, no. So, it turns out that these crabs actually might have starved. So the researchers pulled together a bunch of models and different strands of evidence. And what they've shown is that is likely what happened is, as the temperature went up, so did the crabs’ metabolism, okay, so they needed energy, and estimates suggest that the three degree rise then in this water temperature, potentially doubled the amount of calories that these crabs needed to consume. So there's obviously a lot of competition for food. And it turns out that in fact, before the heat waves, the population had been bigger than usual, right? So, a record high. So the amount of competition was astonishing for scarce resources, because of course, these crabs were just really, really hungry.

Shamini Bundell

Ah okay, so double bad luck there of the heatwave coming at peak population time in an accidentally not very adaptive response there.

Benjamin Thompson

Well, not just double bad luck, Shamini, there's multiple layers of bad luck –

Shamini Bundell

– ah right –

Benjamin Thompson

– so other problems of this article discusses suggests that warmer waters can alter the movement of nutrients up and down the water column, and that affects the food chain. So less food in general for the crabs to eat. And as we've discussed before, food chains are astonishingly complex webs of different things. And if you sort of lean on one part or nudge one bit in the wrong way, it can have some quite serious effects much further away, for example.

Shamini Bundell

And this is a specific example of like, they figured out why this snow crab population crashed in this specific point in time. But obviously, marine heat-waves linked to climate change are gonna get more common, this work must have quite sort of broad implications.

Benjamin Thompson

Well, absolutely. And in terms of the snow crab, I think researchers estimate that it could take years of normal, I guess, sea temperature levels before the populations come back. But as you say, it seems like that might not necessarily be the case. But there is lots to understand in general about this, because there are other negative effects as well of warming seas on different organisms, right? Like, for example, making eggs is very energy intensive. Okay. So it could be that for some species, they just don't have enough energy to breed. And changing temperatures of the sea could mean that populations move to different places to be at the right temperature they like to be at, and this could introduce animals to new predators, or it can introduce predators to new areas, of course, as well, or invasive species and what have you. But it's not necessarily all bad for all animals, some fish seem to do better in warmer waters. So it's a super complicated one.

Shamini Bundell

And you mentioned fishing as well. So I guess there's a more immediate impact on this local area of this snow crab population crash.

Benjamin Thompson

Yes, snow crabs are eaten by people and I think fishing seasons have been closed or curtailed recently, which is obviously having an economic impact on the fishers who are out there. So climate change is having multiple effects on multiple different areas once again.

Shamini Bundell

Yeah, I feel like we're gonna be seeing a lot more of these complex and kind of unpredictable at times effects of the different elements of climate change –

Shamini Bundell

– I don't think my story is gonna help, although it's to do with humans sort of escaping and living on Mars, which is not, not what we're going for.

Benjamin Thompson

There's the obvious segue, right. Okay, go ahead.

Shamini Bundell

It's a fun story. Now, we have talked before about robot chemists, and AI robot chemists that basically, you know, you send them away, and they can do a whole bunch of experiments. They can work out what experiments to do, they have a little robotic arm to do the chemistry with. This is a slightly different twist on this story. I was reading about this in Nature, and it's a Nature Synthesis paper. And these researchers in China, they've developed this particular robot chemist with the aim of basically helping us settle Mars, or at least for humans travelling and working on Mars, or potentially other planets.

Benjamin Thompson

And what can this AI chemist do then? Why do we need it on Mars, I suppose?

Shamini Bundell

Well, so it turns out that having things on Mars is very important. If we're working there, if humans are there, you know, they need particular resources and one in particular that is very useful is oxygen, not just for humans to breathe, but also rocket propulsion. If you want to get something off the surface of Mars via rocket, you actually need quite a lot of oxygen. And oxygen, like various other resources, you don't want to be having to take it there with you, you want to be able to get it from resources that you already have there. So, this particular robot chemist was set a task, which was basically looking at different ways where you might be able to get a set of chemical reactions that allow you to extract oxygen, using only stuff that's already on Mars, chemicals that it could literally get from the surface, if it was there.

Benjamin Thompson

And I'm imagining then on the surface, you've got what regolith is the name of Martian soil?

Shamini Bundell

– Ooh, well done –

Benjamin Thompson

– I think we've talked in the past, there are lots of different sorts of rocks, right. But I think that's what researchers are learning now.

Shamini Bundell

Yeah, so, if we were to go to Mars, one thing, yeah, we know we would have is whatever rock was at our feet. The other thing that we know Mars has is water. So ice under the surface, at the poles and that's pretty convenient. If we want to make oxygen, we could theoretically break down H₂O into oxygen. Now, the way that you would do that is you would need some sort of a catalyst, you know, some sort of chemical processes going on, that takes water and produces oxygen out the other end. And it's this catalyst that the AI robot has to basically invent from scratch. Now, if we were here on Earth, we could be like, here, great, catalyst go. But again, we don't want to have to be loading up our rockets with loads of catalysts and sending it over. These researchers wanted a robot that could go to a bit of rock on Mars, and see what's in that material and invent a catalyst and create a catalyst from what it's actually got there.

Benjamin Thompson

Ah, okay, I'm presuming that it's not testing it on Mars. As far as I'm aware, there isn't a robot AI chemist on the surface of Mars right now.

Shamini Bundell

No, they haven't sent it to Mars. This particular little robot may never go to Mars, but it is a proof of principle and in order to sort of demo it, they gave us some meteorites, either meteorites that are from Mars, or that are sort of equivalent to the Martian surface. That was the test basically, it was like okay, if you had this chunk of material, can you develop a new catalyst? And yeah, their robot AI developed a new chemical that acts as a catalyst that produces oxygen. It did a bunch of modelling simulations, it did a bunch of experiments, the kind of thing that a real, human chemist could do, but it would take a long time. Now, some people think this robot chemists going to Mars is actually very unlikely, because there's a NASA project that's working also on making oxygen on Mars, which uses a different method, which is basically producing oxygen from the Martian air, which is mostly carbon dioxide. And the lead investigator on that project says, you know, we can scale this up, we can produce loads of oxygen. So maybe this robot chemist will be set to other tasks, again, using local materials, whether it's Mars, whether it's the Moon, whether it's, you know, some other exciting planet in the far future.

Benjamin Thompson

Well, there's a story that combines a lot of things we've talked about a lot in the past, we got AI, we've got robots, we've got Mars, this is right in the centre of this Venn diagram and let's keep an eye on how it goes. But let's call it there for today's Briefing Chat, and listeners for more stories like these, and where you can sign up for the Nature Briefing to have more like them delivered directly to your inbox, check out the show notes for some links.

Shamini Bundell

And that's all for this week. As always, you can keep up with us on X we're @naturepodcast, or you can send us an email to podcast@nature.com. I'm Shamini Bundell...

Benjamin Thompson

...and I'm Benjamin Thompson. See you next time.