Subscribe
Share
Share
Embed
Welcome to In-Orbit, the fortnightly podcast exploring how technology from space is empowering a better world.
Dallas Campbell - 0:08
Hello and welcome to In Orbit, the podcast exploring how technology from space is empowering a better world. Brought to you by the Satellite Applications Catapult. I'm Dallas Campbell and in this series we shall be in conversation with some of the most inspiring minds in the country, exploring the ways that the UK is using space to make huge differences to our everyday lives, as well as gaining a better understanding of its role in shaping and sustaining our planet for the future. Now then, in today's episode, we are exploring the concept of digital twins, and I'm joined in the studio by Mark Hennen, Senior Earth Observation Consultant at the Satellite Applications Catapult. Also, we have Simon Evans, the Global Digital Energy Leader at the engineering firm Arup. And remotely all the way from Reading by Peter Bauer, the director of Destination Earth at the European Centre for Medium-Range Weather Forecasts. Digital twin, it's an interesting term. It's a virtual representation of a real-world physical counterpart. They use sophisticated data models to accurately simulate changes or problems using the virtual environment to test and monitor potential scenarios without ever altering the physical counterpart and this enables users to run various simulations and to plan for or prevent any issues before they become a problem. A digital twin is a powerful decision-making tool, but the system is of course reliant on data and can only be as good as the data that it's provided with. One potential source of information is Earth Observation data collected from space. Advancements in launch technology and the small satellite market have exponentially increased the amount of EO data that's available and this new wealth of data could play a huge role in the future of digital twins and their potential to manage and mitigate the risks of failure in all manners of human enterprise, as we shall see. Just before we start on talking about digital twins, actually it sounds quite matrix, digital twins. There is something quite sci-fi and matrix-y about it, but I'm quite interested in what it is you do and what your interests are. So I might start with Peter. Peter, just tell us a little bit about what you're up to at the moment and where you work.
Peter Bauer - 2:35
So I work for a European organisation that's called European Centre for Medium-Range Weather Forecasts, which is a long title, but in the end it's an operational weather prediction service that does global weather predictions and it happens to be the world leader. It's a European organisation that sits in the UK and we're doing this every day and obviously weather forecasts are very important for saving lives and property and the better we can do this, the more lives and more value we can save and we're right now running on behalf of the European Commission, a big program that's called Destination Earth, which is all about digital twins of the Earth system.
Dallas Campbell - 3:17
Well, we'll come on to how all that works and what that is, and digital twins in a moment. But whether, okay, so Mark, so Senior Earth Observation Consultant at the Satellite Applications catapult. You work here basically.
Mark Hennen - 3:31
It’s a long intro. Yes. Yeah. I work for the satellite applications catapults where my main role is basically to kind of champion Earth observation data to people that don't necessarily know how to use it or wouldn't typically use it in their everyday business and try to sort of promote growth in the UK space sector for my space, particularly from math observation.
Dallas Campbell - 3:52
So you're sort of selling it to the wider world. This idea.
Mark Hennen - 3:56
Essentially.
Dallas Campbell - 3:57
Okay, Earth observation. We did a podcast a bit about this before so I don't want to go too deeply into it of course, but Earth observation basically using satellites in orbits to monitor the earth in various different ways and human activity and all kinds of things.
Mark Hennen - 4:10
Yeah, yeah, so well, you know looking after the environments as well, but also seeing…
Dallas Campbell - 4:14
That’s a big one!
Mark Hennen - 4:16
Yeah, quite important. But also you're monitoring how humans and nature are kind of getting along and making sure you know to that sort of we keep we're alerted in a timely manner if anything's going wrong in the environment.
Dallas Campbell - 4:26
Great and Simon, Global Digital Leader at Arup who are a huge engineering from everywhere I look actually the Arabs one of those companies that you don't really hear much about and then everything you look at they're there.
Simon Evans - 4:38
Yeah you see the logo everywhere.
Dallas Campbell - 4:39
Yeah that's it. So just tell us a little bit about Arup and what you do for them.
Simon Evans - 4:43
So Global Digital Energy Leader that's the word you're missing out there.
Dallas Campbell - 4:45
Oh god, did I get it wrong?
Simon Evans - 4:46
That's all right.
Dallas Campbell - 4:46
Global Digital Energy Leader. Yeah shame on me sorry.
Simon Evans - 4:50
No problem at all. So as the kind of title suggests, I look after digital within our energy sectors at Arup and that means how we can use any type of internet-era technology to enhance and enable the work we do within the energy sector. But I've also got another role as well, where I look after digital twins for Arup, which of course is the reason I'm here today.
Dallas Campbell - 5:08
Is that like babysitting? I have to look after the digital twins.
Simon Evans - 5:11
A lot more in that, how we can see how this mega trend will really impact the work we do.
Dallas Campbell - 5:16
So you're all in the same sort of field, but slightly, you know, we've got weather, we've got energy and we've got promotion, if you like. So let's nail down if we can this term digital twin that you've all you've all mentioned. I kind of can assume I know what it means, but who wants to have a crack at defining it for those who maybe not a… Peter’s shaking his head. He looks…
Simon Evans - 5:39
It means many things to many people.
Dallas Campbell - 5:41
Does it mean many things? Okay, well, okay. You volunteered then. So give us a definition.
Simon Evans - 5:46
My favourite frame in this when I talk about it is, for a lot of people, it is a bit like a unicorn. It could be this huge concept that could do so much. But in its simplest definition, it's about having a digital version of something physical and that digital version, it can be a really detailed model of it, or it could be something really simple, like a 2D map or system, is connected, two ways. So you have data flowing from a physical space to a digital space, and then you have a connection somehow from that digital space to the physical space.
Dallas Campbell - 6:17
How do you mean a connection?
Simon Evans - 6:19
Connection, a way of making an intervention into your physical world based off what you decide to do, the insights that you derive.
Dallas Campbell - 6:25
So you get information from your digital twin and then you make decisions based on that information?
Simon Evans - 6:31
That's it, in essence, yeah. Somehow you get data from a physical world, a physical environment, make decisions from that data when it's in a digital space, whatever that digital space might look like, and then make those interventions. Which in itself sounds quite complex.
Dallas Campbell - 6:44
It does sound like my brain's going to...
Simon Evans - 6:45
The simplest way of saying it is about connecting cyber-physical things together. Blending our cyber-physical world.
Dallas Campbell - 6:50
All I can think of is the matrix though. All I can think of is that kind of model, that other sort of separate earth. Peter, can I ask you, are you happy with that definition? Because as a weather person, you've always used models when it comes to predicting the weather. So what's the difference between a normal weather prediction model that you might've used and a digital twin? Are we just being, is it just a fancy name?
Peter Bauer - 7:14
No, no, no, it's exactly the right question actually, because we're struggling explaining digital twins to the level where people actually understand what the difference of a digital twin or the added value of a digital twin is compared to what we presently do with our weather systems. So, the way weather prediction works is, we use a lot of satellite data, about a hundred million observations per day, we have a big simulation model and we combine the two every day, to kind of reflect what's going on out there in every detail we possibly can simulate and observe and then we use that knowledge to predict the future and that's tomorrow's weather forecast. But that's only one way, you know, that's exactly what was described before. So here you observe and you simulate and this information goes only one way from the real system into the digital system. So it becomes a digital twin if you actually, with that knowledge, drive certain decisions in your real system like you would say, you know, I have my digital replica in the computer because I'm predicting serious floods in the Netherlands tomorrow, I changed, if I could, I changed my land surface prediction or the dikes near the surface to then see what happens and I can play through whether actually for tomorrow's system that would be enough.
Dallas Campbell - 8:33
So it's not just having a load of data to compare the two, you can actually change parameters, you can actually fiddle with things.
Peter Bauer - 8:40
Absolutely, you can play through and optimise it and I think that's a good example, but for me what works almost better if you would take the human body as an example, and you would say, I have a digital simulation of you, and I observe you at the same time with heart rate and everything you can observe, constant x-rays and ultrasounds, and that gives me the best possible digital replica of you and then I want to understand actually whether a disease is coming up or if something happens to you and you're not feeling well, what's actually going on and what kind of medication I could give you or what kind of surgery I could apply to you to make you feel better. I can play through this in the most realistic context and then try it out and that becomes extremely important for us when it comes to whether extremes or climate change. We want to know what's going on in 30 years and how to prepare our society in the best possible way and the digital twins are the way to do this.
Dallas Campbell - 9:38
I'm just trying to still get my head around, is it different to a computer model then? I suppose.
Peter Bauer - 9:43
No, no.
Dallas Campbell - 9:44
Is it the same thing?
Peter Bauer - 9:46
So that's your matrix analogy. At the heart sits a computer model that is supported by observations. As we heard before, you know, you can play this in different levels of complexity. It can be a very simple thing that's only based on observations. But ultimately for the Earth system, because you have to consider vegetation is important, the ice in the Antarctic is important, the urban heat effects over Redding are important, so all of this is important. So you need to have a very complex simulation, in the end, to realistically represent everything that matters.
Dallas Campbell - 10:21
Got it. And then you can simulate various scenarios and see how things pan out.
Mark Hennen - 10:25
I think a very good real-world example of Digital Twin is your satellite navigation. If you're going to drive somewhere, you plug into your satellite where you want to go, and the digital version of your car on your journey will be created for you and it will use all kinds of different information based on road traffic, the sort of fastest route there, and predict when you're going to arrive. But as you continue your journey, you'll continually monitor where you are on your journey, how how fast you're going, what are the traffic patterns around you and continually update how you, when you're going to arrive and you need the best instructions of how to get there.
Dallas Campbell - 10:59
I think that drives me crazy about that. So I drove from London to Oxford today and I left really early. I left to, you know, I can't remember what it was like six, half past six and it said, Oh, it's only going to take an hour and take an hour and of course I drove for an hour and it was still going to be an hour after an hour. So it never gets, I want it to predict better than it predicts of at the moment.
Mark Hennen - 11:15
Get a better sat-nav
Dallas Campbell - 11:17
Is that what I need to do?
Simon Evans - 11:20
Well, I think that's the interesting thing. Reflecting on both those kind of definitions of it, what makes a digital twin different from a model or simulation is probably the difference between, let's say, something like Google Maps and your sat-nav. In that there has to be some type of feedback loop. If you're just taking data from a physical world and kind of modeling it, that's just a model. That's just a simulation. You have to have a two-way feedback, physical to digital, digital to physical. And that loop is what defines a twin over, as I say, a model or simulation.
Mark Hennen - 11:46
One of the powers of a digital twin, as well, is connecting that live data into the digital twin. So you can actually…
Simon Evans - 11:52
Does it have to be live, that's the question.
Dallas Campbell - 11:53
Yeah, I was gonna ask you about the live, 'cause I suppose that's the, when I think of, you know, a computer model, for example, I'm not necessarily thinking of live data that's continuously…
Mark Hennen - 12:03
Well, a model will be based on some physical aspects that have been parametrized into the model, but the digital twin model, you can personalize that model to exactly to that thing that you're looking at. So if you're developing a model on how a potential, house might react to certain situations, it will be generic to all houses, for example. But if you make one for your house, which has a certain, you know, aspects or has certain features inside, then that model will be much more accurate because you actually use real time data to continually kind of train your model to be reacting to your house.
Simon Evans - 12:35
So I think it's an interesting thing here about real-time versus right-time. Yes. And I say that and in Peter's example of the weather system, having real-time data, instant data, millisecond data is sometimes applicable. But in a lot of cases about having right-time, the right time data based on decisions and the questions you're trying to answer. So if you're modelling a city, you might not necessarily need to know pedestrian movements down to the nearest second because that's irrelevant. But you wouldn't need to know water flow, for example, if you're doing flood monitoring. So having the appropriate amount of data in maybe right-time to answer the questions you're trying to answer.
Dallas Campbell - 13:09
Wow. Who was it? I think it was there was a mathematician George Box, I think it was he said all models are wrong, but some are useful. This was his thing. I understand what he means by that. I mean, a model is an approximation of a physical system and yet you're never going to get it right. But it tells you stuff. But I'm sensing from what you're saying about digital twins, this is taking that but making it much more by the fact that you have live data and it's real-time and…
Simon Evans - 13:35
Right time.
Dallas Campbell - 13:35
Right time.
Mark Hennen - 13:36
I mean it depends on what you're trying to do for example if you're twinning a glacier which moves very slowly you don't need data instantly you can get it monthly.
Dallas Campbell - 13:45
That's interesting but something like so with Peter for example I mean the thing with weather systems, it's so I mean weather systems are so ridiculously complex I don't I mean is this a digital twin in inverted commas that you use and work with is it does it give you sort of perfectly accurate predictions?
Peter Bauer - 14:02
No, well to the limit of how accurately we know processes in the atmosphere because you know, we don't know everything or can't cast everything in an equation or model. So our models are imperfect, our observations are imperfect. So there are uncertainties which limits our skills and weather prediction and climate prediction, you know, and of course, with digital twins, not only through this two-way functionality but also by using and investing in much better observations and much better ways to combine this with very sophisticated models, we try to increase the skill and actually make that digital twin interactivity work. But I wanted to say one more thing about the distinction and life and the right time and things, you know, I'd say maybe one distinction could be whether you're operating something or whether you're designing something, you know, operation has a lot to do with real-time and life because you're operating something and you want to operate that smoothly without failures and let's say energy efficiently if it's a factory or something. For this, you need live data and there's other things where you're designing something and your digital twin can actually help you make the right design decisions and if we think about a sustainable society in the future, green deal, these kinds of things, how do we restructure our energy systems and renewable energies and water management in the future. You know, their digital twins also have a role in designing the system right, given the present knowledge and how we can predict future states and all that.
Dallas Campbell - 15:35
Mark, can you just give us a few, I mean, we talked about weather, obviously, we talked a little bit about energy, just give us a few real-world examples that people who are listening can maybe be a bit surprised about that we, that digital twin technology, is that the right word?
Simon Evans - 15:49
Never technology.
Dallas Campbell - 15:52
Oh Simon, he's quite finicky Simon if you say something.
Simon Evans - 15:55
There's a lot of misnomers around it and I think that that can be really challenging because as kind of as Peter and Mark was saying the concept has become quite clouded in platitudes you can basically use the word digital twin and it makes it sound glamorous yeah and kind of avoid the point of what you're trying to do which is answer some particular use cases I guess I'd say…
Dallas Campbell - 16:14
But why technology? Why is that the wrong word?
Simon Evans - 16:17
So because what that alludes to is you have to buy something, let's say, and install it onto your computer, and that's your digital twin. But if we think actually a digital twin is about connecting a physical and digital world, it's not a product, it's not a technology, it's actually really a methodology and an approach, and it's that approach of connecting physical digital worlds, rather than, I'm gonna buy a piece of software from X-vendor, install it, and bam, there's my digital twin.
Dallas Campbell - 16:41
Got it.
Mark Hennen - 16:42
I think a key point of that is that there are new use cases for digital twins all the time. So like some of the kind of more traditional use of digital twins, which is mostly probably in the built environment in infrastructure, digital twins are starting to go beyond those now. We're going to different things. You know, Peter's already mentioned healthcare and people producing digital twins of human beings for medical care. But you know, we got to things like manufacturing and sort of looking at process lines and stuff like that. We've also got the kind of things of public good. So I know that Fujitsu, for example, developed a digital twin for a tsunami warning. Where you take real-time data, people using an application, telling the twin where they are, a tsunami alert happens, and it can tell those people where to go, the best routes for escape. We've got other things like, we mentioned Google Maps, but everyone's carrying a smartphone these days, you know, and that can create data which can inform you on things like purchases or healthcare, things to do in your personal routines. So there are digital twins being created all the time, and from the catapult, We're trying to… we’re really interested to see how space can get involved with that and how we can actually start using observation data to really help those twins.
Dallas Campbell - 17:51
Yeah. Well, maybe, maybe you could just expand on that. So where does Earth Observation and Artificial Intelligence fit into digital twin methodology?
Mark Hennen - 17:59
It's very, very new. I mean, we would kind of just come in. I mean, Peter and the guys at ECMWF and Destination Earth, one of the first kind of real big users of EO data in a digital twin kind of environment. The thing to remember with EO data is mostly it's at scale. You can start…
Dallas Campbell - 18:14
How do you use it at scale?
Simon Evans - 18:15
Massive.
Mark Hennen - 18:16
So a lot of digital twin stuff is used from point scale data, so you get sensors on a piece of machinery which will feed the information to the twin about what that machine's doing. But with satellites, we can observe landscapes in quite good detail and very rapidly and it's about how we use that data and what's that most useful for. To date, most of it's for environmental stewardship, so keeping an eye on the environment, making sure that the digital twin is aware of what's going on, provide timely alerts about certain things, factors that are happening, improve our modelling ability, and be able to make informed decisions when it comes to our environments, whether that's policy or developments and stuff like that.
Dallas Campbell - 18:58
Okay. I mean, all these different areas of human life, I suppose, as well as earth monitoring and weather and weather systems, earth systems. Do you guys sort of talk to each other? Is it kind of like the idea is you'll kind of create a sort of perfect replica of the earth in digital form down to the sub-millimetre and then, you know, whenever a problem arises, we go off, we go to our digital twin and sort of poke it a bit and see what happens.
Peter Bauer - 19:26
Yes and no. I mean, it comes back to the question, do you have to talk to everybody? Is everything relevant for anything else or everything else, you know? So I think in our context, we certainly want to go down to city scale. We want to use earth observation data for everything that relates to weather and land surfaces and I think there's a lot of overlap there. We haven't tapped into mobile phones quite yet, but that will come, car sensors as well.
Dallas Campbell - 19:50
What do you mean tapping into mobile phones? Why would you do that?
Peter Bauer - 19:54
Because they map as much your internet traffic as they could map surface pressure or temperature or moisture or pictures you take that contain clouds wherever you are, you know. So in theory all of this is an observation that we can use in the future of digital twins. So you basically carry along with you in your car, if you have a clever car, and your mobile phone and observatory, you know, in all kinds of ways and that can help us to support the information we're getting.
Dallas Campbell - 20:25
How important do you think digital twins will be in things like helping us figure out solutions to climate change? I mean, is this going to be a really big thing?
Simon Evans - 20:36
In the future, every asset process or system in the world will have a digital twin and that's the way it's going and what that means is, again, it's going to have some type of digital version of itself. And that might sound like a bit of a big scary movement. But what it's really talking about is how you can connect digital twins together. So everything we've talked about today...
Peter Bauer - 20:55
Where they need to be connected.
Simon Evans - 20:56
Exactly where they need to be connected. So sometimes you have schools of thought of people thinking we're going to make it one giant digital twin of the world, a planetary digital twin, if you will. But that's a ludicrous notion because from a computational basis it’s impossible.
Dallas Campbell - 21:10
Yeah, but it would look awesome.
Simon Evans - 21:12
Well, but what is more about is the federated effect.
Peter Bauer - 21:14
It's not too far out though. I agree with the statement before, you know, it's probably not going to be a single digital twin, but in the end it doesn't matter. I guess for the, you know, in the end it's probably a collection and a hierarchy of digital twins where what we do in Destination Earth is your global picture that goes down… that covers everything down to let's say 100 metre scales or something. But then you have specialized digital twins underneath you know one is for the city of London or one is for an energy farm in the North Sea and the third is about agriculture in Spain or something and then you have micro twins on a specific factory or a house or something sitting underneath there, you know, but they had they share common elements in terms of methodologies and this is why I liked what was said before about it's not a technology, it's a methodology. But it relies, of course, on technology to work, you know, and depending on how big or complex your system is, it actually indeed requires enormous computing and enormous amounts of data to work, particularly if it's done on the fly, you know, in real-time.
Simon Evans - 22:19
Absolutely. And I think the interesting part in that is that every single one of those twins is going to be owned by a different person in some capacity and they share the relevant information in a secure way with each other to answer system wide challenges. So for example in the UK and coming back to my work in energy now there's a lot of work happening in the UK around how you can make an energy system digital twin or the virtual energy system as it's called where you can allow every asset within the energy system to talk to each other like the Internet. like the internet is how computers talk, to allow you to answer system-wide challenges. Now we're talking about Earth observation here. If you can connect Earth observation and actually every sector together, you could, for example, ask questions about climate change and using all the array of different information and modelling you have across all the different sectors and systems, come up with various scenarios and solutions. So to your original question, what is the role that digital twins as a methodology or an approach have in climate change or in the future, they're gonna be intrinsic. We're gonna see them more and more in everything we do.
Dallas Campbell - 23:23
And is it gonna speed up the transition, particularly the energy sector, away from traditional energy through to new types of energy?
Simon Evans - 23:30
If we look at digital twins again being connecting physical digital things, we know, and there's literature that says this, that in order for us to decarbonise, we need to connect systems together and connecting systems requires some type of exchange of data and it doesn't matter what sector you are. So you can draw a line there saying, if we really want to achieve these net zero targets, we need to really lean into digitalisation and arguably digital twins at some capacity. But proportionately, because of course, you can latch onto that and go, great, I'm gonna spend all this money answering use cases that aren't relevant rather than answering really relevant use cases that will help you on your journey.
Mark Hennen - 24:05
I think an important thing to remember with climate change is that we are sort of having to make decisions in increasingly pressured environments, whether it's fiscal pressures or just time running out of time to be able to provide all the sustainable infrastructure that the globe is Going to need going forward with increased pressures from climate change but also from population growth economic pressures and these kind of things and digital twins allow us to kind of plan and predict in a kind of digital space So we can kind of get it wrong there until we find the most optimal solution and implement that in real life.
Dallas Campbell - 24:37
So it's a bit a test bed almost a way of practising advice.
Peter Bauer - 24:39
Yes.
Dallas Campbell - 24:40
Could one of you maybe just give me, I mean we've talked a lot about various areas and aspects, as a working example of when someone has done this and used a digital twin in order to do something or to try something or to test something.
Simon Evans - 24:54
Yeah, so there's a number of examples. One classic one that people resonate with is how you could use a digital twin to improve building performance or its environment for the occupants.
Dallas Campbell - 25:03
In terms of what? Performance for what?
Simon Evans - 25:06
So let's say you have a building, your office block, let's say, is going to have lots of sensors in it, as they all do, that are collecting information about the CO2 levels, the oxygen levels, the temperature, humidity, etc. and you can take all that information, you can use that to understand and model what happens if you made this change, for example, let's say we pump more oxygen to this place, or we reconfigure the building this way, how would that provide a better outcome or worse outcome for the occupants? So you can basically inform your decision making through scenario modelling. And the key back to the digital twin is that you have to have that connection back into the physical. If you just do a model, well, we've been doing that for a long time.
Dallas Campbell - 25:43
It's just SimCity.
Simon Evans - 25:44
It's just SimCity, but it's SimCity that then updates the real world again. It's like connecting SimCity to your real city.
Dallas Campbell - 25:51
That's, there we go. That's, that's it. SimCity, but plugged into the real world.
Simon Evans - 25:54
And if it's not plugged in, it's not a digital twin.
Dallas Campbell - 25:56
Got it.
Simon Evans - 25:57
It has to have that feedback loop.
Dallas Campbell - 25:58
Nice. Okay. I want to know about… because I know the ESA, European Space Agency, they've got their digital twin of Earth project. Are any of you guys involved in that?
Simon Bauer - 26:10
Yeah, we have been involved marginally because the program was started at the same time as we started the thinking around Destination Earth and I would say the distinction between the two is that, for instance, an ESA-only program that's subscribed to and funded by ESA member states as opposed to the European Commission, as is the case for Destination Earth and it's mostly Earth observation focused. So I would actually claim that the ESA program doesn't really look so much into this two-way connection that we just spoke about and that we consider an important distinction between shadowing the real world or observing what's going on, monitoring change and things, as opposed to this two-way direction with the modelling part at the centre. So I think it's a more complementary program.
Dallas Campbell - 27:04
I'm just sort of wondering what the aims of it were, like why they started it, like what specifically were the aims of the ESA project that's different to what other people are doing?
Simon Bauer - 27:14
I think it's trying to add value to the Earth observation data or extract value from Earth observation data…
Dallas Campbell - 27:21
Right, I see.
Peter Bauer - 27:22
Which is ESA's primary mission, how they launch and operate satellites and derive products from that and I think it's to up the game using such data for the purpose of system monitoring.
Dallas Campbell - 27:35
Yeah. Mark, as someone, I mean, your job is to sell this, I suppose. He's grimacing at me. Not sell it, not like, you know, sort of door to door. But I'm just sort of interested in what your day-to-day activities are.
Mark Hennen - 27:47
So I think at the Catapult it's a two-way street. So we are trying to promote EO data to those who maybe don't necessarily use it. But we're also doing research in-house and working within the sort of listening to the community and seeing what are the requirements of EO data and Digital Twins obviously is kind of an existing concept, but it's also, you know, it's evolving and could include EO data. But we need to understand what are the kind of use cases for EO data. It's important to remember with observation data at the moment, we're going through a kind of a new era in the kind of commercialisation of the space industry where, you know, launch, the space is much cheaper than it ever has been before. Even the development and construction of satellites is cheaper than ever. So the industry can be much more dynamic and reactive to the requirements of the users. So we want to keep our ear out and hear from you guys, the developers or the users, what data do you need? Because there's an opportunity here for us to start developing bespoke missions that could feed data directly into these digital twins for even smaller assets, for city scales. You know, there are new products being released all the time, you get satellites as small as 30 centimetres long being launched. Looking at kind of bespoke missions like looking at air pollution, for example, you can develop a constellation that looks specifically at air pollution, which would be…
Dallas Campbell - 29:03
So rather than having satellites that would do multiple things, you're getting much more specific things that we do because I suppose from all of your points of view, how important is this new satellite revolution, the fact that presumably, access to low-Earth orbit is getting cheaper and easier, satellites getting smaller and cheaper. Is it a real game changer for… Peter?
Peter Bauer - 29:27
Yeah I think it is. I think you need both though. I think you need specialists, like high definition high-spec laboratories like what we presently have, you know, ESA satellites and human satellites and NOAA and NASA satellites. You know these are hundreds of millions a pop you know plus launch plus operations and all that. So these are specialist observatories, spending years to develop. But at the same time, and I really agree, we need these complementary systems that are easier to deploy, easier to design and maintain, easier to replace. So if one of these big satellites falls out of the sky, it's hard to replace. If you have installations of smaller satellites, less complicated, much easier to have a reliable, continuous information source and then in the end also what you do with it. It's not just launching stuff and having it in space, but what you do with it in there, you know, away from the traditional, you know, public entities, launch satellites, collect the data, give it to everybody, you know, the diversification of use, that we will see for satellite data and digital twins in the end, you know, you could write a digital twin maybe tomorrow for something you care about and sell it to others, you know, having access to that huge digital twin technology that we're developing today. So there's unprecedented opportunities for commercial companies in the future, both in data provision and use.
Dallas Campbell - 30:48
Peter, you just said something I think is really interesting. It's what you do with it. That seems to be the thing. I often wonder that just the speed of transformation in terms of, well, methodologies and technologies, things getting smaller, cheaper, just AI becoming extraordinary. Is it kind of going too fast? Is the sort of politics and the business sector keeping up with how fast things are going.
Simon Evans - 31:10
It's a really interesting one because actually a common misconception we have around digital twins is they're a new idea. Well actually one of the earliest examples is from the space industry, the Apollo 13 missions, where they used what they called at the time digital mirroring to bring the astronauts home safe and since then other sectors like manufacturing and aerospace have been all over digital twins the last 30-40 years.
Dallas Campbell - 31:30
How do you mean for Apollo 13? What do they have?
Simon Evans - 31:36
Well, so they were using simulation and modelling to help bring the astronauts back.
Dallas Campbell - 31:39
Oh, I see! So having the guys in the room with bits of old pipe and tubes.
Simon Evans - 31:40
In essence, yeah, that's a very low-maturity digital twin and one of the earliest examples. Now why I say that and it's really interesting is the whole point of digital twins is, you know, they're not new, but what's made them really exciting these days, and as we've kind of discussed, is that you now have that greater connectivity between physical-digital, digital-physical, which is sped up because of things like storage, compute and connectivity, connectivity like 5G, for example, which means it's much farther than ever before. But it always has to be about the use case. Why are you doing it? Because you're not going to invest millions, billions of pounds if you have no real answer or no real possible use case and I think that's the part, for me, it's understanding that it's really a journey, not the destination when it comes to digital twins. You find your use case, you develop something that's valuable, and you keep on iterating and because of this kind of confusion, people thinking it's new and it's not new, you have a huge wave of hype being created traditionally.
Dallas Campbell - 32:35
Humans do. Humans love hype. We like to big things up.
Simon Evans - 32:38
As if it's like this new magical thing. It's like, well, actually we've been doing this for a long time. It's just now we can see more value from it because of storage, computer connectivity.
Dallas Campbell - 32:47
Yeah, exactly. Let's talk about the future. We've only got a few more minutes left, but I'm interested in the direction of travel. Peter from where you're sitting in Reading. What is the direction of travel? Where are we going to be in a decade, say, do you think, in terms of what we've been able to do within the digital twin world?
Peter Bauer - 33:05
So I would expect by then that we have a fairly accurate digital twin of the Earth system as a whole, at a global level, down to the scales that I mentioned earlier. So hundreds of metres, I would say, supported by very sophisticated simulation models that we don't have yet. and all the observations from all the satellites to mobile phones and things. So I think we're going to have that and we're going to have that hierarchy established as well, not everywhere all the time, but for important assets like, as I said, energy parks, like wind parks or wind farms, agricultural systems, flood protection systems. So everything that matters to society where presently all these structural investments go, I think in 10 years we will have a pretty good digital twin coverage of these aspects.
Dallas Campbell - 33:56
So basically it will enable us to make better decisions. We'll have a sophistication, we can make better decisions.
Simon Evans - 34:02
Faster, better decisions faster.
Dallas Campbell - 34:02
Better decisions faster, that's good. So in terms of things like climate change, because you're right here with this, using this, are you optimistic that we'll be able to use this to make things better to solve real big existential issues?
Simon Evans - 34:18
I am, but I think it depends on a couple of things and the main thing is industry, all industries collaborating. Now that might seem obvious…
Dallas Campbell - 34:28
But we’re terrible at collaborating.
Simon Evans - 34:30
But the thing is because digital twins mean many things to many people, everyone's creating their own ventures of it rather than let's say collaborating on the rules, then competing on the game. The results from it. I think it's for us as industry or sectors to come together and work out how we can collectively make digital twins work for us as humanity and help humanity flourish in the future and if we don't really look at that and we continue to be isolated and siloed, I think we won't really get there.
Dallas Campbell - 34:55
Mark, are you happy with the speed in which people are understanding that this, I keep wanting to say technology, I can't just stock myself, this methodology, or do we need a kind of an education program to…
Mark Hennen - 35:09
I don't think we necessarily do need an education, but I think people just need to make use of it, I think. I don't think we need to necessarily know exactly how a digital twin works. Much like my own internal work analogy about the SatNav, you don't need to know how that works. You just need to be able to make use of it. And I think if the industry starts to understand the kind of benefits from these things, then there'll be more opportunities to develop the right data sources and to get buy-in into the industry so that people will continually come back to look at Digital Twins. I mean, from an EO perspective, again, we kind of, we want to make sure that the capabilities are understood so that people don't get a kind of bad impression what we're actually going to create and so, I think that's the important thing, making sure that people can understand what it can do, but not necessarily how it does it. I think that's the important thing.
Simon Evans - 35:53
I think it's about bringing people along in that education as appropriate as possible and there are a number of communities out there, but there's one particularly in the UK, or centered in the UK, that's probably, I say, the world's foremost community, the Digital Twin Hub, which is the catapult, of course, no very well.
Dallas Campbell - 36:08
What's the Digital Twin Hub? Explain the Digital Twin Hub.
Simon Evans - 36:10
What is the Digital Twin Hub?
Dallas Campbell - 36:10
Wait, we've missed this.
Simon Evans - 36:11
The Digital Twin Hub is many things, but it manifests as a community of digital twin practitioners, enthusiasts, people interested in it, who then come together, exchange ideas and knowledge on the topic of digital twins.
Dallas Campbell - 36:23
Do they dress up like in Tron?
Simon Evans - 36:27
I'm sure, if you could, if you wanted to. But it's a really great, the satellite catapults are actively involved, as are many of the other catapults, as are Arup, incidentally and it's a great way of bringing together people, whatever their maturity in thinking and understanding, to help share the best practices, the ways of working, the kind of use cases to make us all move forward together.
Dallas Campbell - 36:47
That's really interesting. So the digital twin hub, things like that. What about, I mean, do you talk to politicians and decision makers? I mean, I look at the….
Peter Bauer - 36:54
We are, yeah, we are and I would like to re-emphasise a point that was made before, again, about collaboration. I think important for things like this to work in the end. So it's not compartmentalised, but there's a real like system of systems, if you wish, on the digital twin system of systems is that you agree on data standards, you agree on interfaces, you need certain reference systems that you demonstrate new developments with that can't be everybody but leading systems. You know, there needs to be a certain agreement for how to come together actually. Otherwise you will have, this will always be separate and there's 20 competing systems and they exchange nothing and cooperation doesn't happen. So it's very boring to talk about standards.
Dallas Campbell - 37:39
No, but it's really important. I mean, I think about low Earth orbit and what's going on there, and there are no real standards. I mean, countries do different things and have different systems and different ways of doing things and as we become more and more reliant on low Earth orbit for day-to-day living, I kind of worry about, crikey, there's no real standard yet.
Simon Evans - 37:58
Yeah, and that was to my point earlier of collaborate on the rules, compete on the game. We need to collaborate on the rules, which are the standards, the common approaches, methodologies, without necessarily having our commercial interests in mind. But then when that game is created, we can then compete on it based off our own company's USPs. Because there's plenty of pie out there of digital twins to eat, if you will. But it's about having the common rule set, the taxonomies, the methodologies, to allow this integration at a fundamental data level that will allow all this to succeed. Because if that doesn't exist, we're just going to be creating isolated computers.
Dallas Campbell - 38:30
Just in a word, are you optimistic about the future? Working with this great thing, do you kind of look at what you're doing going, "Oh my god, this is awesome!" Or do you go, "Oh, crikey, this is giving me a headache."
Mark Hennen - 38:40
I think I look forward to it because the opportunities for innovation in this area. The more data that comes on, the more research we can do, the more interesting new technologies we can develop. I think that's really exciting, and I think that's definitely something I'm looking forward to.
Dallas Campbell - 38:58
Peter, I'm interested in your career, this has all come about pretty quickly, and I'm just, from where you're sitting, are you excited by it? Are you excited about the possibilities of the future?
Peter Bauer - 39:08
I'm very excited and I believe that we're in contrast to maybe 10 or 20 years ago, we're now in a situation where we actually have digital technologies that can really fulfil what digital twins require. And I think our society is ready as well, you know, in the sense of, I think everybody understood climate change and the impacts now. Everybody is affected by extremes, one way or another. We really understand the cost for our society now. It's hundreds of billions every year or trillions even worldwide, you know, the impacts of extremes, you know, stuff just have to look at the, at the documentation by the World Economic Forum, you know, or for like this. So we know what the cost is of doing nothing. You know, and investments in this technology is marginal, compared to what it will bring to our society and then this is the economic aspect. As I said before, you know, if we understand, we start to understand now the actual business potential in many of these areas. So apart from, you know, making our society more sustainable, there's actually money to make in that sector and I think this will, in the end, trigger the real change.
Dallas Campbell - 40:13
That's always the thing that triggers change, is if there's money at the end of it. Thank you so much. What a fascinating discussion, what a fascinating subject. It's been an absolute pleasure to talk to you.
Simon Evans - 40:23
Thanks for having us.
Dallas Campbell - 40:23
That's my pleasure.
Mark Hennen -40:23
Thank you.
Peter Bauer - 40:24
Thanks very much.
Dallas Campbell - 40:24
Cheers. Thank you.
Peter Bauer - 40:25
Have a safe trip back to London.
Dallas Campbell - 40:27
With my bad sat nav.
Peter Bauer - 40:30
You know, if you had a digital twin, you could regulate traffic in your favour, you see?
Dallas Campbell - 40:34
Really?
Simon Evans - 40:36
That’s a bias of the digital twin.
Peter Bauer - 40:37
You would monitor where you are, and you would see that the traffic is bad and you would reroute traffic in the real world based on your digital expectation.
Simon Evans - 40:44
I liked your analogy of a sat nav. I think that's really good. Google Maps is great at this as well, because it's bringing in different sources. But you kind of see with some of these systems, you’ll pick a route, let's say to a destination like today, you're coming from London. But then the interesting thing, if you had a traffic jam, it reroutes you onto another route. But at the moment, as far as I can tell, it reroutes everybody onto the same new route.
Dallas Campbell - 41:04
Correct.
Simon Evans - 41:04
So all you've done is create a new traffic jam elsewhere. So there's not that kind of overarching intelligence.
Mark Hennen - 41:09
But they are getting better though. I think like, yeah, like sometimes 10 years ago would be terrible, right? But now they're getting much better because there's more and more data that's like training the algorithms that are getting smarter. The AI is improving because you've got more data to build.
Simon Evans - 41:18
But it has to be two way though, because let's say all four of us here drove here today and We'd all drive down the road, the same road for argument's sake, and there's a traffic jam ahead. I guarantee all of our sat-nows would reroute us onto exactly the same road, because none of them would talk to you and say, "I've put Mark here, I've put Peter here, I'm going to put Simon here." So you ended up causing the problem elsewhere, rather than that two-way communication of, "Right, we have 100 people trying to move, let's send 20 down here, 20 down here, and 20 carry on where they are." and that kind of connectivity depends on the data exchange.
Dallas Campbell - 41:50
I can see how that would work. Anyway, listen, we shall leave it there. It's been a pleasure to talk to you, so thank you very much indeed.
Simon Evans - 41:56
Well, thank you.
Mark Hennen - 41:56
Thank you.
Peter Bauer - 41:57
Thanks very much for the opportunity.
Dallas Campbell - 41:58
That's it for this episode. Thank you very much for your company. To hear future episodes of In Orbit, be sure to subscribe on your favourite podcast app. And to find out more about how space is empowering industries between episodes, why not visit the Catapult website or you can join them on Twitter, LinkedIn or Facebook. See you next time.