New horizons in robotics

Assistive tech, preventing injury at work and biomimetics feeding into the design of future robotic systems – listen on to find out more about the robots of the future.

Digital Economy

Apple Podcasts EC Audiovisual Service Podcasts Google Podcasts Spotify

Transcript

CORDIScovery podcast- Episode #44 - New Horizons in robotics This is an AI transcription. 00:00:00:00 - 00:00:38:10 Abigail Acton This is CORDIScovery. Hello and welcome to this episode of CORDIScovery with me, Abigail Acton. Robotics is a swiftly changing field. Cheaper hardware is making research more accessible, and thanks to advances in AI, focus is now shifting from feats of physical dexterity achieved by expensive robots to building general purpose robot brains in the form of neural networks. 00:00:38:12 - 00:01:02:04 Abigail Acton A recent breakthrough from the Massachusetts Institute of Technology is pushing the boundaries of robotic intelligence by shifting away from traditional data training methods, MIT researchers have set their sights on mimicking the techniques behind large language models like GPT-4. This novel approach promises to revolutionize how robots learn, adapt, and interact with the world around them. So what are the latest developments in robotics? 00:01:02:04 - 00:01:26:04 Abigail Acton At home, at work, and in the fields? What can a novel assistive tech do to help people live independently? Can robots help prevent injury at work? And how can biomimetics inform the design of future robotic systems? Our three guests who have all received support from the EC’s science funding programs will talk us through their work to answer these and other intriguing questions. 00:01:26:06 - 00:01:40:24 Abigail Acton Mac MacLachlan is professor of psychology and social inclusion and co-director of the Assisting Living and Learning Institute at Maynooth University in Ireland. He's particularly interested in innovative technologies that are helping people remain independent for longer. Hello, Mac. 00:01:41:01 - 00:01:42:23 Mac MacLachlan Hello, Abigail. Lovely to join you. 00:01:43:00 - 00:01:59:19 Abigail Acton Lovely to have you. Arash Ajoudani is the director of Human-Robot Interfaces and Interaction Laboratory at the Italian Institute of Technology. He's interested in physical human-robot interaction, mobile manipulation, and assistive and tele-robotics. Hello, Arash. Welcome. 00:01:59:19 - 00:02:01:06 Arash Ajoudani Hello. It's good to be here. Thank you. 00:02:01:08 - 00:02:15:13 Abigail Acton Professor at the Dahlem Center for Machine Learning and Robotics at the Free University of Berlin, Tim Landgraf investigates the social behavior of guppies and bees in order to understand biological intelligence and refine artificial intelligence. Welcome, Tim. 00:02:15:15 - 00:02:17:05 Tim Landgraf Hi Abigail. Nice to be here. 00:02:17:07 - 00:02:36:03 Abigail Acton Great. Good to have you. Mac, I'm going to turn to you first. The SHAPES project aims to enable the large scale deployment of digital technologies for healthy and independent living. One tool that the project has explored is the Kompaï-3 Robotic Assistant. Can you tell us what challenges robotics could help us to meet Mac? 00:02:36:05 - 00:03:01:03 Mac MacLachlan Sure Abigail. Thank you. So I think one of the big challenges facing, all countries across Europe, are the demographic changes. So we have had very successful health systems within Europe, which means that we now have more people, living longer. But there's also a demographic shift so that there are fewer people being born. 00:03:01:05 - 00:03:32:11 Mac MacLachlan And so if you put those two things together, what you have is a greater level of need but a lower level of supply. So if we look, for instance, ahead to the year 2030, W.H.O., estimates that there will be a 4 million, shortage in the workforce. 58% of those would be, nurses. 28% will be other health and social care professions, and 15% will be, doctors. 00:03:32:13 - 00:03:42:02 Mac MacLachlan So the challenge is how are we going to provide personalized healthcare to people in their communities, ideally in their homes in the years ahead? 00:03:42:04 - 00:03:52:12 Abigail Acton Yeah. So obviously there's clearly an imperative. And you can feel almost that the clock is ticking. So can you tell me a little bit about the work that the SHAPES Project did to make robots a more feasible option? Please. 00:03:52:14 - 00:04:20:15 Mac MacLachlan SHAPES stands for Smart and Healthy Aging through empowering People by Engaging in Systems and empowering is the key word, in that acronym. Because we what we want to do is to give people technologies that allow them to live comfortably at home in their communities. But there's no single technology that can do that. Really. You need a whole array of different sorts of technologies. 00:04:20:15 - 00:04:51:11 Mac MacLachlan So some of these may be large data pools in the cloud, other aspects may be sensors. And some of these home and other aspects may be other forms of sort of data lakes or data mining. The key thing is that any piece of technology should be able to talk to all those other pieces of technology. So if you want to provide integrated services, integrated care, you'll want a robot that's able to communicate with the other information that's relevant to that person. 00:04:51:13 - 00:05:15:24 Mac MacLachlan So the SHAPES project sought to create, if you like, an ecology, that looked at the technical demands of providing a robot linking up to the other sorts of technologies available, but then also the social demands and indeed the governance issues and ethical issues around data protection, privacy protection and so on. 00:05:16:02 - 00:05:31:00 Abigail Acton Okay. Lovely. So can you tell us a little bit more about the Kompaï-3 robot, what it was able to do and whether you have any, feedback from users with regards to how they felt it helped them. It would be interesting to know, you know, the impact of this robot in people's lives. 00:05:31:02 - 00:05:54:06 Mac MacLachlan Sure. Yeah. Thank you. So we had, I think 56 different partners in this, project. And, one of them was, Kompaï robotics. We're very pleased to partner with them as they're one of the leaders in this field. And their Kompaï-3, robot, which is something that sort of stands almost shoulder length, if you like. 00:05:54:06 - 00:06:17:16 Mac MacLachlan It's got a wider base and then narrows and there's, a human like, face at the top of it. And really, what that does is a number of different things depending on the person's needs. So, for instance, it can be, a conversational companion, answering questions or providing information to reduce feelings of isolation. 00:06:17:18 - 00:06:26:23 Mac MacLachlan On the other hand, it could guide somebody who's working as part of a gait rehabilitation exercise. 00:06:27:03 - 00:06:33:13 Abigail Acton So when you say a gait rehabilitation, you mean like frail walking to try and make someone walk in a way that's more stable and robust? 00:06:33:15 - 00:07:03:22 Mac MacLachlan Yeah, exactly. And it can provide, both support and feedback to them about how they're progressing there. Similarly, some, people, as they age, may have cognitive problems. And, I mean, Kompaï-3 can provide sort of medication reminders, and monitoring. And it can look at the long term, adherence to, to medication. And then also, of course, to be aware that falls are a particular challenge for, for some people as they age. 00:07:03:24 - 00:07:33:24 Mac MacLachlan So another feature of Kompaï-3 is that it can sort of navigate somebody’s living environment. So clearly for some people, as they age, they may have sight difficulties or hearing difficulties or cognitive or indeed mobility difficulties. And so the Kompaï robot can help in different ways, for people to essentially retain their independence, and again, to live within their own community. 00:07:34:05 - 00:07:46:14 Abigail Acton Super. And did you have any feedback from users? Have you heard were there people are finding it easy to interact with the robot or do people feel reluctant to do so? What's the feedback on that. 00:07:46:17 - 00:08:18:04 Mac MacLachlan Across the project, but also, applying to the Kompaï-3, yeah, people were extremely enthusiastic about technologies. And that's a really good question, because we often think that older people are a bit less open to new ways of doing things and new technologies. I think one of the things we found in this project is when you co- design and co-develop technology with people, and they feel that they have much more of a say into it, and it's more relevant then. 00:08:18:04 - 00:08:44:09 Mac MacLachlan In fact, they're incredibly welcoming of these technologies. And through the project, for instance, as well as the gait rehabilitation that I mentioned earlier, other features that were added to the Kompaï-3 robot include facial recognition and, natural voice interaction with people as well. So they're key in that personalized, service, sense. 00:08:44:13 - 00:09:04:13 Abigail Acton Yeah. To make people feel more relaxed around them and more accepting of them. This is excellent. So then $60 million question when can we start to expect to see robots in care scenarios like this? My mother is elderly and at home. You know, maybe later in the future this would be quite nice for her to have. What's it looking like with regards to uptake in the market? 00:09:04:18 - 00:09:33:08 Mac MacLachlan I think there are already, you know, robots, providing assistance. If you look, for instance, in Japan, the social acceptance of robots, I think is much greater. And so within Europe, you know, we have an interesting challenge. It's maybe less the technical challenge of what a robot can do and more of the sort of social challenge in terms of feeling that a relationship, if you like, with a robot, is an okay thing, that a robot is safe. 00:09:33:11 - 00:09:43:14 Mac MacLachlan A robot is something you can trust, and that overall, that a robot can be a meaningful contributor to, you know, enhancing your life and supporting your well-being. 00:09:43:18 - 00:09:51:04 Abigail Acton Excellent. Thank you so much for telling us about your work. That's fascinating. What do you think the assistive robots will be able to do in 20 years time, for example? 00:09:51:06 - 00:10:16:15 Mac MacLachlan Yeah. So that's a very nice question. I mean, we can distinguish between assistive robots that are helping people with particular tasks that need to be done and then companion robots, which really make people feel, well, that they have a companion, that there are others around, whether they are other people or other things that care for them and, and about them. 00:10:16:17 - 00:10:42:12 Mac MacLachlan And I think that that distinction of crossing the mental barrier and thinking that a robot cannot just only care for me, but it can also care about me, is going to be very important for those companion robots. So I think assistive robots will do more in terms of more ways of assisting people, but I suspect that they will also take on more features of companion robots. 00:10:42:18 - 00:10:49:14 Abigail Acton Excellent. Thank you very much. I'm reminded of that robotic seal that came out of Japan for people to help people with dementia. And how. 00:10:49:14 - 00:10:50:20 Mac MacLachlan Paro, yeah. 00:10:50:22 - 00:10:51:18 Abigail Acton Paro, thank you. Yeah that’s right. 00:10:51:18 - 00:10:53:01 Mac MacLachlan Yeah yeah. 00:10:53:01 - 00:11:16:14 Abigail Acton And how massively people loved it and accepted it and were happy to have it. Lovely. Thank you so much. Super. Thank you. Arash I'm going to turn to you. The SOPHIA Project set out to develop a new generation of socially cooperative human robotic systems, which can enter into direct socio physical contact with humans to perceive, understand and react to their distress or needs. 00:11:16:18 - 00:11:37:05 Abigail Acton Now, here we're shifting from assistive tech to look more at industry and the use of robots in that scenario. Your project has two main aspects. Can you tell us about the element to do with preventing injuries such as back strain, something that hangs over millions of workers and which has implications for long term health and financial consequences? What did SOPHIA develop for that? 00:11:37:07 - 00:12:12:01 Arash Ajoudani So thanks for the question, Abigail. SOPHIA wanted to address work related health issues. So the alarming factor was in Europe, we have about €240 billion lost every year in lost productivity and problems associated with the, you know, all the insurances and everything that workers, the absenteeism and all this is related mostly to, work related muscle disorders that the workers of today have to do repetitive or heavy tasks in manufacturing, in warehouse and in logistics. 00:12:12:03 - 00:12:38:09 Arash Ajoudani That is an alarming factor which is not really paid attention to. I think the biggest technological gap for this was today's, monitoring systems that are looking into how people are performing their jobs in terms of ergonomics are quite outdated. They are mostly very, let's say, posture based, that they are not really taking into account how external loadings and how different dynamic factors in the environment can actually influence the well-being of people. 00:12:38:09 - 00:13:03:01 Arash Ajoudani And in general, the economics of people performing their jobs. So our first really important goal was to can be revolutionized this? Can we provide some, new methodologies that, in real time, they can tell us how a worker is actually undergoing some fatigue in short or too long term, and how these physical interactions can actually, change the, the well-being of the person. 00:13:03:01 - 00:13:06:24 Arash Ajoudani So this was the first important let's say milestone for us to achieve. 00:13:07:02 - 00:13:19:01 Abigail Acton Okay. Thank you. So can you tell us a little bit about the wear-robots, where does somebody put them on their body, and how does that send data back to some system that can make a change? 00:13:19:05 - 00:13:52:14 Arash Ajoudani Absolutely. So, the wearable robots, we call them wearbots, we're supposed to get all the necessary data to be able to build these, let's say digital models and to tell us how the dynamic states, in human articulation and human joins are changing. These wearable robots will tell us about the posture of the person, about let's say acceleration and possible, let's say some muscular activities of people, doing that and why we try to use wearable robots 00:13:52:14 - 00:14:16:07 Arash Ajoudani Was that due to the privacy issues we try to minimize the use of cameras, and also in cluttered spaces the use of cameras usually is not really easy to detect postures of people, especially, for instance, if you're imagining a person, in between, let's say, pallets or inside, very narrow aisle of an airplane, which is trying to mount things. 00:14:16:11 - 00:14:28:08 Arash Ajoudani So the wearable systems not only preserve privacy, but also can give you meaningful data, but it comes also with some issues, of course, in terms of, you know, sanitization and others. 00:14:28:10 - 00:14:37:06 Abigail Acton Lovely. Thank you so much. So what, it straps on around the back or it's something that's something that straps on around the arms so I mean, where does it actually go on the body? 00:14:37:08 - 00:14:56:00 Arash Ajoudani So it goes on important joints and important, let's say key points of the body that we want to measure. Let's say some quantities that we want to elaborate and give some indication of the load. And these, wearable systems are not only for measurements but also can provide feedback that we have. 00:14:56:00 - 00:15:13:07 Arash Ajoudani Some vibrate tactile modality that tells a worker that, for instance, your shoulder is being overloaded. If you continue this way of working, in such a way, you're posed to some issues in long term. And this feedback will be reduced if the worker can change the posture. 00:15:13:07 - 00:15:30:13 Abigail Acton That's fantastic. So it's like a haptic feedback the buzzes you if you're doing something not so great and the buzz gets less and less and less as you correct that positional, whatever it is. I think this is wonderful. But it's not the only thing SOPHIA looked at. You were also looking at the notion of collaborative robots. Can you tell us a little bit about Cobots, please? 00:15:30:17 - 00:15:56:13 Arash Ajoudani Exactly. So Cobots are basically collaborative robots that create synergies between humans and they can perform the jobs. But what SOPHIA did differently was Cobots were not only trying to focus on the task, Cobots were also trying to improve human state in terms of ergonomics and in terms of physical and cognitive state. So we brought these models, digital human models into the optimization and into the control frameworks of these collaborative robots. 00:15:56:19 - 00:16:10:14 Arash Ajoudani So these robots could reconfigure, could change the way they were doing, so that this new way of working with humans not only contributes to speeding up or making better productivity better, but at the same time, it can improve the ergonomics of people working with it. 00:16:10:16 - 00:16:14:04 Abigail Acton Okay, so I'm not quite sure I understand that. Could you give me a concrete example? 00:16:14:04 - 00:16:35:03 Arash Ajoudani Yeah. So imagine a very simple example. A robot is passing you a two kilogram drill. Right. Or any, any tool that you're supposed to work with it. So there are infinite number of places in the space that the robot can handed to you. But there's only one which is the most optimum in terms of workspace that you prefer to get this. 00:16:35:03 - 00:17:01:00 Arash Ajoudani And also it has the minimum impact on your body in terms of ergonomics because each two kilogram or each gram that you add as an external load, it has some consequences in your joints, right? So if you multiply it with eight hours per day and to 300 days per year, this will have really bad consequences. So the robots in this easy example will pass you the object with minimum impact on your health and maximum impact on your productivity. 00:17:01:00 - 00:17:01:14 Abigail Acton That's lovely. 00:17:01:15 - 00:17:04:05 Arash Ajoudani Yeah, this can scale to many, many dynamic and challenging tasks. 00:17:04:09 - 00:17:21:11 Abigail Acton That's great. And your underlining of the notion of repetitive it is really important here because absolutely it's something that builds and builds and builds. So yeah, so the robot would hand me the drill to my hand in a way that was the best way for me to be receiving it rather than elsewhere or whatever. Super. 00:17:21:12 - 00:17:28:12 Abigail Acton So it's like fine tuning the interaction. Excellent. Thank you. And how is this working in practice? Have you had any pilot studies? 00:17:28:14 - 00:17:55:01 Arash Ajoudani Absolutely. So we had three pilot studies. We also brought our technologies to Volkswagen. They were producing electric cars. We had a pilot study in the Netherlands. It's a SME. They were producing gears of 4 to 20kg. We had a pilot study in Slovenia. They were producing machinery tools. So all these, they use different tools and let's say technology developed by SOPHIA in terms of cobots, in terms of wearable systems. 00:17:55:01 - 00:18:34:09 Arash Ajoudani We also developed some exoskeletons, wearable exoskeletons, single joint exoskeletons, all these through some really, deep impact analysis. We understood that we are able with these technologies to not only maintain and improve the productivity, but we were able to improve, significantly, the ergonomics and well-being of people performing the same jobs. Obviously, if something can be automated, should be automated, but until we can automate everything, we need to go towards synergistic, workstations where humans have all the dexterity and all the skills and robots can start learning and getting this transfer of knowledge from human. 00:18:34:11 - 00:18:44:01 Arash Ajoudani Until then, robots can definitely improve productivity and ergonomics. And all our use cases showed significant benefits of using SOPHIA technologies in multiple aspects. 00:18:44:01 - 00:18:46:18 Abigail Acton Fantastic. That's excellent. You must be very proud of that. Absolutely. 00:18:46:18 - 00:18:47:11 Arash Ajoudani I am. 00:18:47:13 - 00:18:58:10 Abigail Acton Yeah. I bet. Excellent. Thank you. And it's obviously a big team behind you. As you said you were part of a consortium. Lovely. Does anyone have any comments or observations to make to Arash? Yes. Mac? 00:18:58:12 - 00:19:27:02 Mac MacLachlan Yeah, I really interested to hear in your work there, Arash and as you were describing it to some listeners, it may seem like we're doing different things and that I'm working with people with functional difficulties, whereas is you're sort of optimizing performance. But in another way, I think robots is helping us, remove that divide because an old context is actually about optimizing performance. 00:19:27:04 - 00:19:40:22 Mac MacLachlan From, from my perspective, that's really important. So that, people who may have functional difficulties can think about getting better at things than they've ever been as opposed to just making up for a deficit. 00:19:40:24 - 00:19:42:04 Arash Ajoudani Absolutely. 00:19:42:06 - 00:19:42:13 Mac MacLachlan Yeah. 00:19:42:14 - 00:19:48:14 Abigail Acton Excellent. No, it's a good observation there. Yeah. It does really directly overlap in that way. Thanks very much for that observation. 00:19:48:16 - 00:19:49:10 Arash Ajoudani If I can comment on that. 00:19:49:11 - 00:19:50:02 Abigail Acton Please. 00:19:50:04 - 00:20:10:24 Arash Ajoudani This is a this is an excellent observation. Of course, when you built these digital models, of course this can scale to healthy people to monitor. If something can happen something bad can happen to your limbs. But you can also scale it to people that come back to work after an injury. So you can make sure these robots adapt in a way that they're using the minimum 00:20:11:01 - 00:20:32:08 Arash Ajoudani The joints that have been injured or elderly or people with, you know, physical difficulties so you can optimize the way the people want or best should work. And this, as you mentioned, really nicely, everything comes down to an optimization. And for that we need models. And, and that's why our projects I think are contributing very well to this. 00:20:32:10 - 00:20:52:15 Abigail Acton Absolutely. No. No. Really great. And from such different angles. And I'm going to turn to Tim, who's coming from a completely different angle, because we've talked about people and how they can be assisted by robots in various ways. But I'm also interested in the use of robots in agriculture and biomimetics, in other words, the way we can imitate, what's out there already thanks to nature. 00:20:52:20 - 00:21:15:18 Abigail Acton So, Tim, by studying swarm intelligence, HIVEOPOLIS, your project, has developed the concept of a robotic dancing bee as part of a futuristic honey bee hive, helping pollinating bees to perform well in very challenging environments. So can you tell us a little bit more about biomimetics? What are biomimetic robots and how can they be useful? Tim. 00:21:15:20 - 00:21:45:03 Tim Landgraf Yeah. So biomimetic robots that, you know, the name already gives it away, robots that mimic, biological, templates like animals, plants, even. And why they do that? There's a ton of different applications. Right. So you can have like a legged robot that walks like a dog. And, you know, excels in, in rough terrain for a variety of purposes. 00:21:45:04 - 00:22:22:00 Tim Landgraf Or you could have, biomimicry because, you want to integrate a robot into an animal society, an animal group that's, that's our case. And, why would you want to do that? Well, I mean, you could think of agriculturally used animals so you can control or use those robots inside those groups to gather information on elsewhere, you know, other important aspects, how the group is doing. 00:22:22:02 - 00:22:34:14 Abigail Acton Okay. So tell us a little bit more about HIVEOPOLIS. So you developed the concept of a robotic dancing bee. What work did you do to model and replicate the performance of the bee dance? Perhaps. Maybe start by telling us what the bee dance actually is. 00:22:34:16 - 00:23:02:05 Tim Landgraf Yeah. So, Bees? Fly out, explore the environment, for food and other resources. And if they have found something, they come back. And if that something is really valuable, they do something that we call the waggle dance. That is very specific behavior that other bees observe. They follow that dance basically. And also you know, body to body contact with the dancer. 00:23:02:05 - 00:23:28:10 Tim Landgraf And then they kind of learn where that forager had found the food or whatever resource, and they fly out. So it's basically communicating location, direction and distance of that food source. And that's like from a basic fundamental science. But it's really interesting because those animals are super small grain the size of a pinhead. So how does it even work? 00:23:28:10 - 00:23:49:23 Tim Landgraf Right. So it's just like a symbolic kind of communication. So, so that's there's a lot of interest in understanding how that works. And does it actually work. And if we then know how it works we can obviously also use it. So that was the idea at HIVEOPOLIS, integrate technology into a beehive. And there's many, many things. 00:23:49:23 - 00:24:21:10 Tim Landgraf And the bee robot, the dancing bee was just one part. But then use this technology to the one hand learn a lot of you know about the health and the status of the of the hive, but also use the robotic components to basically feedback information to, steer them to, a healthier colony state. So tell them where they can find food, or stop them from going somewhere where the food is maybe treated with pesticides. 00:24:21:12 - 00:24:36:13 Abigail Acton Right. And so in the project, did you how far did you get with actually creating the notion of a dancing bee? Was it that you did the modeling and worked out, you know, the programing behind it, or did you get as far as actually developing a robotic dancing bee? How did this come? 00:24:36:15 - 00:25:00:20 Tim Landgraf We built so many different prototypes. I think, I lost count. So, we basically went into the project with already a working prototype. So in my PhD thesis a long time ago, I built a robot that could perform the waggle dance. But it was huge and clumsy and had to have a lot of manual control to basically do what it was promising to do. 00:25:00:22 - 00:25:19:05 Tim Landgraf And so our goal was to minimize everything, put it in a standard beekeeper hive so that it actually could work without having to open it up, you know, just to have the bees, because we wanted to help them. Right. So that's the whole system was to basically give back to the bees because we hurt them by, you know, using them like we do. 00:25:19:07 - 00:25:45:17 Tim Landgraf We make them unhealthy and give them parasites and all these things. And so the goal was to give them back. And so, our work with software development, modeling the dance behavior, but also minimizing all of the different components such that, you know, we can actually use them without opening it up. I think, like outside robotics, you know, especially with what Mac is doing but also with what Arash is doing it's so different from lab robotics. 00:25:45:17 - 00:26:00:18 Tim Landgraf And, you know, if there's so you know, if there's animal stuff, you know, wax and propolis and all these things, things stick together and they don't move anymore. And then there's like a whole bunch of different problems that you never think of in the lab. 00:26:00:24 - 00:26:17:11 Abigail Acton Now, that's a very interesting observation. Absolutely. So in practice it's yeah, there are different contexts. Of course. I was just wondering what the results were. Did you did you manage to direct the bees to certain places? And I noticed also there was something to do with weather forecasts and I mean, was there, you know, how was the information all cross shared? 00:26:17:16 - 00:26:44:21 Tim Landgraf Yeah. In HIVEOPOLIS it's not all about the bee robot. It's basically, an array of different systems from maps that basically collect different sources of from different sources of information, like where was pesticide being brought on the fields and, you know, where are other colonies that we probably should be kind of coordinating with. And so all this information is collected in these maps. 00:26:44:21 - 00:27:06:21 Tim Landgraf And then we have different components like a honey harvesting module. And like the bands, for a module that could also stop dances and, and food dances. So see where they are going. And then everything is integrated. Humans stakeholders are also, sort to, you know, participate in this. Right. So give us information. 00:27:06:21 - 00:27:08:07 Abigail Acton You mean farmers, for example. 00:27:08:07 - 00:27:11:00 Tim Landgraf Farmers. But also we pulled in hipster beekeepers. 00:27:11:00 - 00:27:11:16 Abigail Acton So beekeeper. 00:27:11:19 - 00:27:15:10 Tim Landgraf So those who have one, 3 or 4 hives on the roof. 00:27:15:12 - 00:27:16:23 Abigail Acton Like urban beekeepers. 00:27:16:23 - 00:27:42:06 Tim Landgraf Urban beekeepers. Yeah, right. Everyone might benefit from the system. And so, yeah, the robot works more or less, you know, mechanically, is it as efficient and as useful in the actual world? No, there's a lot of research to be done. But yeah, I think this project was defined in a very crazy kind of perspective. 00:27:42:06 - 00:27:45:08 Tim Landgraf And it was clear from the beginning that it takes more than 5 years. 00:27:45:08 - 00:28:00:19 Abigail Acton Oh yeah. Sure. But I mean, they all do. It's always just laying down the groundwork for the next stage. It's like building a brick wall, you know, it's a layer of bricks. And then you have to put the next layer on top. Absolutely fantastic. Thanks for explaining that so well. Excellent. Well, I have one final small question for you, Tim. 00:28:00:21 - 00:28:10:14 Abigail Acton If you could look down the line in ten years into the future, for example, how do you see biomimetic robots evolving? If we do some blue sky thinking, where could we go with this? 00:28:10:16 - 00:28:35:23 Tim Landgraf Well, I mean, to be honest, I think, the natural world is like a technology that is super far advanced. So we should probably be biomimetic most of the times. And we are not yet, because we can't, sensors are not as good. And actuators are not as, you know, good. And battery life is slow and materials are bad and robots fall apart after five years and not 50. 00:28:36:00 - 00:28:38:01 Abigail Acton Oh, they get sticky with honey. 00:28:38:03 - 00:28:53:16 Tim Landgraf Yeah, yeah. And so, what I see is, is like an integration much deeper and you can't see those sensors anymore and more control over our environment to keep it healthy, to detect diseases early and all these things. Right? 00:28:53:19 - 00:29:08:22 Abigail Acton Yeah. Excellent. Okay. Thank you very much. Well, that's a positive hope for the future anyway. That's great. Well, listen, thanks to all three of you for sharing your time with us today. That was really super. And I particularly like the way that we looked at these various elements of robotics. So thanks very much and all the very best with your researches. 00:29:09:03 - 00:29:15:13 ALL Thank you so much for having me. You very much for joining us. Real pleasure. Thank you. I think we should all have a waggle. 00:29:15:13 - 00:29:25:10 Abigail Acton Yeah, we should have a Waggle dance yeah to celebrate. We should. Pointing people over to CORDISCOVERY online. Absolutely. Thank you very much, guys. Thanks for your time. Goodbye. 00:29:25:11 - 00:29:28:10 ALL Okay. Bye. Excellent. Bye. Bye bye. 00:29:28:12 - 00:29:50:17 Abigail Acton This week. If you've enjoyed this podcast, follow us on Spotify and Apple Podcasts or whichever service you use. And check out the podcast homepage at the CORDIS website. Subscribe to make sure the hottest research and EUs funded science isn't passing you by. And if you're enjoying listening, why not spread the word? We've talked about what baby baboons can tell us about language development. 00:29:50:19 - 00:30:12:14 Abigail Acton Listen to the episode of mind reading to find out more. In our last 43 episodes, they'll be something there to tweak your curiosity. The CORDIS website has articles and interviews that explore the results of research being conducted in a very broad range of domains and subjects, from proteins to protons. There's something there for you. Maybe you're involved in a project or would like to apply for funding. 00:30:12:19 - 00:30:32:08 Abigail Acton Take a look at what others are doing in your domain. So come and check out the research that's revealing what makes our world tick. We're always happy to hear from you. Drop us a line [e-mail]. Until next time.

Robots, from the challenge of physical dexterity to neural networks

Robotics is a swiftly changing field: Cheaper hardware is making research more accessible, and thanks to advances in AI the focus is now shifting from feats of physical dexterity achieved by expensive robots, to building ‘general-purpose robot brains’ in the form of neural networks.

A recent breakthrough from the Massachusetts Institute of Technology is pushing the boundaries of robotic intelligence. By shifting away from traditional data training methods, MIT’s researchers have set their sights on mimicking the techniques behind large language models such as GPT-4. This novel approach promises to revolutionise how robots learn, adapt and interact with the world around them.

So, what are the latest developments in robotics at home, at work and in the field? What can novel assistive tech do to help people live independently? Can robots help prevent injury at work? And how can biomimetics inform the design of future robotic systems?

Our three guests, who have all received support from the EC’s science funding programmes, will talk us through their work to answer these and other intriguing questions:

Mac MacLachlan, is professor of Psychology and Social Inclusion, and co-director of the Assisting Living & Learning Institute at Maynooth University in Ireland. In the SHAPES project he explored the role of digital technologies and robotics.

Arash Ajoudani is director of the Human-Robot Interfaces and Interaction laboratory at the Italian Institute of Technology. He is interested in physical human-robot interaction, mobile manipulation, along with assistive and tele-robotics. Ajoudani coordinated the SOPHIA project.

Professor at the Dahlem Center for Machine Learning and Robotics at the Free University of Berlin, Tim Landgraf investigates the social behaviour of guppies and bees in order to understand biological intelligence and refine artificial intelligence. He worked on the development of a robotic bee in his project, HIVEOPOLIS.

Happy to hear from you!

If you have any feedback, we’re always happy to hear from you! Send us any comments, questions or suggestions to: [e-mail].

Keywords

SHAPES, SOPHIA, HIVEOPOLIS, robots, human-robot interaction, assistive tech, GPT-4, digital technologies