Ep. 40: Digitalisation could support the industry significantly Ankit Podcast

This transcription has been AI generated and therefore may have some inaccuracies.

Ep. 40: Digitalisation could support the industry significantly

Dr John Oyekan, lecturer in digital manufacturing in the Department of Automatic Control & Systems Engineering at the University of Sheffield, gives his insight on digitalisation within the textile & apparel industry.

Jessica Owen
Hello. My name is Jessica Owen, and I am the Deputy digital editor at WTIN, and this is the WTIN podcast. You music, as most events have been cancelled this year thanks to COVID 19 the World Textile Information Network has decided to host its first virtual trade show to keep the textile and apparel industry connected. The innovate Textile and Apparel virtual trade show will take place between the 15th and 30th of October, and the show will feature manufacturers of textile technology as diverse as man made fiber production through to garment assembly, alongside material producers for apparel, sportswear, personal protective equipment, through to smart fabrics and more, in order to get a taste of What has been happening, specifically in the world of digitalization today, I am joined by Dr John Oyekan, who is a lecturer in digital manufacturing in the department of automatic control and Systems Engineering at the University of Sheffield. In this special episode, John talks about technology such as AI and digital twins, how digitalization can help companies cope in a pandemic, the relationship between the need for both humans and robots and so much more. So let's get on with the show.

Well, hello and good afternoon, John. Thank you very much for agreeing to come on the show today. It's lovely to have you. So I have asked you here today so that we can talk about digitalization, specifically with regard to sort of the textile and apparel industry. But before we get on to that, just tell me and those who are listening a bit about yourself and what you do at the University of Sheffield. Good afternoon,

Dr John Oyekan
Jessica. Thank you very much for having me on this afternoon. My name is John, as Jessica said, from the University of Sheffield, and I am a lecturer in digital manufacturing. So basically, what that means is looking at various pillars in digital manufacturing and trying to apply to various sectors and various sectors, which includes aerospace, textile as well automotive and various other sectors, right?

Jessica Owen
Okay, and so, what did you do then at university? I think that you did undergraduate masters. I mean, what's your what's your background?

Dr John Oyekan
So I did my first degree at the University of Coventry. I studied electronics technology, and my my final year project was on robotic cells. So basically what that meant was that I developed a mini natural robotic platform that was able to move from one cell on a simulated factory floor to another cell on the factory floor, all autonomously. So after I finished my electronics technology bachelor's degree at University of Coventry, I worked in industry for just over a year before going back to the University of Essex to study embedded systems and robotics. And my my final year project then was developing an autonomous flying flying platform, and that was when quad copped up platforms were still breaking into the market. So I could say that I was one of the early adopters and early pioneers that developed autonomous behaviors for just flying platforms. So some of the autonomous behaviors I developed were autonomous landing, autonomous taking off, or to the most tracking of a target in all in real time. And then after that, I went on to do a EPSRC, which is the engineering physical sciences and Research Council sponsored PhD in swarm intelligence. So basically, what that I was looking, I mean, during my pleasure, I was looking into how you could use a swarm of robotic agents to track pollution in the environment. So if you think about a situation whereby there's. Hazardous substance in the environment, and you can't see that substance. How do you run away from that substance? Now, one way of solving it is by using robots, because robots can work in dangerous environments. They can do repetitive jobs continuously. They don't get bored. So one way of doing it was you send this flying robots into the environment. They find the pollution, and then once they find pollution, because they swarm. So swarm meaning multiple agents working together in coordination. So because they are working in a swarm, they can start tracking that pollution as it changes in profile, and that's where a problem that was unseen before now becomes seen, and then people can sort of keep away from dangerous areas so and then, after I finished my PhD, I did a Post. I went on a couple of post doctoral positions, including working at the one of the government funded catapults, and then I eventually ended up at the University of Sheffield,

Jessica Owen
right. Gosh, well, it sounds like you've done so much. I mean, what's what sparked that interest in getting into robotics and digital manufacturing then, well,

Dr John Oyekan
what sparked the interest in robotics was ever since I've been younger, I've been quite fascinated about NASA, I'll say that. And the reason why I said NASA is because they send this rocket into space, but they also send agents to other planets. So they send robots, for example, sending robots to Mars, sending robots to the moon surface. So that got me thinking how I mean, that sort of sparked my interest. Now, if you see, from what I've said previously, there's quite a number of bio inspired techniques I've used in the past. I also had a very big interest in the sciences, so in biology, in chemistry and physics, but biology really picked really interested me the most, and that's how I've been using both robotics and bio inspired techniques in robotics. So digital manufacturing also got me interested in digital manufacturing, especially sensing autonomous agents and an AI was as a result of this background I have So robotics. I used my interest in robotics to support the digital I mean, to just support the robotics field, like digital manufacturing. I used my interest in bio inspiration to support my AI interest in digital manufacturing and sensing everything. Before you can control the robots, you need to be able to sense the environment as ends. My interest in sensing as well.

Jessica Owen
Okay, so, I mean, you just mentioned there you've got an interest in artificial intelligence, and, well, that was something I wanted to talk about today, because, well, if we start at the beginning, then digitalization can mean so many sort of different things. But within the textile industry, things such as artificial intelligence, digital twin technology, they're very interesting things. So I was just wondering if you could talk more about that and explain what they are exactly. So

Dr John Oyekan
artificial intelligence is a way by which you try to you develop an artificial agent that can do some of the things a human can do. Now I wouldn't say that you you'll be able to exactly replace the human but, but what artificial intelligence tries to do is to mimic some of the things that the human can do. So in the past, people have developed things like if else rules. So if the temperature is too high in an environment, you turn the knob down, which is typically what we humans do. If we're in a room and it becomes too hot, we go to the knob, we turn it down. So of course, we getting up and going to the knob. I don't want to say it's a bit of a pain, but some, some, I mean, we've gotten to a point whereby we have devices that can automatically do that work for for us. So in a nutshell, that is what artificial intelligence is. Now there are more sophisticated techniques whereby you look at data. And the reason why we look at data is because the world is very complex, and you we have to and to be able to understand the world in. One approach in artificial intelligence is to collect data about the phenomenon, various phenomenon in the world, and then develop mathematical models to explain that phenomenon. Now, by developing those mathematical models, looking at it from a simplistic view, is a function of input to output, and by using that function, we can now start using it to auto automate some processes that we want them to be automated. So that's in a nutshell. That's what artificial intelligence is. How do we develop how do we develop mathematical models that will give us a an understanding of a phenomenon in the world. And how do we use those mathematical models to automate processes now, from digital twins, digital representations of physical systems now? And the reason why we need to do that, especially in this increasingly connected world, the reason why we have to do that is because we have machines that could be on the other side of the world, or we have factories that could be on the other side of the world. So how do we create digital twins that will enable us to that will enable us to understand what those machines are doing, even though they are remote. So that's why we need digital representations of those machines now, one added advantage of digital representations of those machines is that we can now start carrying out scenario tests. So what if something breaks? What do we need to do? When will a component in that machine break? We can start using those digital twins to enable us to do those analysis and those predictions. So those so in a nutshell, that's what artificial intelligence and digital twins

Jessica Owen
are, right? Okay, that was a lot to get my head round. I'm not going to lie, but I think you have sort of summarized it quite well there. So looking at AI then first how, I mean, you've explained how it works and what it is essentially, and you've given a few examples. But what could this do specifically for the textile industry or manufacturing within that industry, for example? Okay,

Dr John Oyekan
so in the textile industry, and one of the things I've observed about the textile industry is that, and not only, not only this specific textile industry. I mean, if you look at look as well as aerospace industry, there's a similar trend there whereby there's a lot of legacy machines being being used. Now because of those legacy machines, it's quite difficult to to get data from from those legacy machines. So one way by which artificial intelligence could be used is through the placing of, I don't know whether our listeners know anything about IoT sensors. So I Otis sensors in a large I mean telling I mean I Otis sensors means Internet of Things. So basically, we could put these sensors on legacy machines, and that enables us to start generating data. Now from that data, we can start developing models, mathematical models, as I've explained previously in our conversation, we can start developing those models, and use those models to start enabling us to see when a particular machine is going to break down, or we could also use that data to start predicting when, how the quality of a fabric that's been produced, how is going to what is going to be at the end of the manufacturing line. So that's one of the ways by which AI could be used on a manufacturing line. Now in a bigger context, so I've spoken about just on the manufacturing line. Now if you take an holistic view of the entire system, we can start thinking about, okay, how we can use AI to also help the supply chain. The reason why I say that is that currently there is, there is a bit of reactive element in the textile industry, okay, we see celebrity wearing a particular fashion, or a particular type of of design, or we wait till a particular event or cause before the manufacturing the manufacturers start producing a particular type of fabric. But what if we could predict in advance what's going to happen? It will make the manufacturers prepared to produce a particular fabric so that they're able to meet demand just in time. And that's where I also see AI becoming very, very important in a textile industry, yeah. So

Jessica Owen
I mean what you're saying there is, really is just helping to make companies and manufacture manufacturers just more efficient, exactly.

Dr John Oyekan
Yeah. So making them. More efficient, making the supply chains more connected and as well as more efficient, because once you I mean adding to what I said previously, once you start seeing what, once you start getting a prediction of what a trend could be, what the demand trend could be, that could also help you to start asking or start looking for raw materials so that you are better prepared to deal with this demand, and that translates to be more competitive, on, on, on the global scale, okay?

Jessica Owen
And going back to digital twins, then can that again, is that sort of just helping the industry to be more efficient. Or what else could it help with? Okay, so

Dr John Oyekan
yes, it will make the to make the industry more efficient. But also, one of the things that that I'm currently seeing increasingly being used is how you could actually use digital twins of fabrics. For example, I use that to see whether it fits a customer or not without them actually wearing it. So for example, currently, if we want to buy a particular a particular cloth or a particular shirt or a particular dress, we go into the store, we try it on, and then we try a number of designs on before we say, Okay, this is one I like. But what if we could have like, a digital twin of that material, and people just stand in front of a sensor or in front of a camera, and you have that and that design you see superimposed upon upon them, so they can actually try that virtually before they before they actually buy it. And this could be done at home. And just think about how much that could save us having to go down, I mean, from an Amazon perspective, or from a regard. I mean, is it boohoo perspective, instead of you buying a particular cloth and then returning it again, getting another one in and then returning it again, there's so much that that kind of technology could save. I mean, could could save us in terms of transportation costs and in terms of time wasted, in terms of impact on the environment itself, because customers can try things on virtually at home and then decide whether they actually like it or not before actually placing an order. So that's one of the exact exciting things that I see coming up more in the textile industry, and certainly that also could actually feed down back to the manufacturing lines, because when people start trying on these things, they could actually go on their computer, do a bit of modification here and there, and they could start having bespoke fabrics, bespoke designs, dress designs or shirt designs that fit them, that makes them them. Because currently we are moving towards a situation whereby we're going for mass manufacturing to more customized manufacturing for people. So this is where else I see our digital twins or fabrics could become quite very, very important in the textile industry.

Jessica Owen
I mean, that's quite amazing, really, if you could do that. And like you said, I think I've also seen recently that there is this trend to sort of become more customized in the things we in the things we want to wear. And if you were to take old fashioned, I'm going to call it old fashioned machinery and processes. It's there's a lot of manpower that's probably involved with that. It just doesn't make financial logistical sense. But if you can input implement these types of technologies, and it's, it's just making it so much easier. And like you said, as well, it helps the environment, which is such a huge topic at the moment, yeah,

Dr John Oyekan
and I think tied into that will be you raise a quite important point there, because currently, a lot of this, a lot of fabric, or a lot of dress designs are manual, but we're seeing a trend whereby automation of robotics could actually support that, that could actually support that as well going forward. And I know there's been some some some sort of reluctancy about trying to do automation or trying to use robotics in manufacturing lines. But one thing, one one aspect to look at it is that humans, we can be totally replaced, basically because there's some things that robots cannot do. If you remember earlier, back in my conversation, in our conversation, I said that we're trying to use machine learning to build a model of a phenomenon in the environment, but the world. Is very, very complex. We can't build an entire model of the entire world. Now think about it. A human being understands the world, and regardless of what unexpected scenario throughout a human being, the human being will look at that scenario and find a way out that a robot can't do that, even in a manufacturing environment that we could say is very constrained. There's only so much a robot can can do. So what we see in some other sectors is that we see that whether we like it or not, the human element is very, very crucial, is very, very important, even in even in automation. And we see that actually humans could be brought in as a way of program of, I would say, program training a robot to get better at various tasks on the manufacturing flow, and then that human becoming a supervisor of of that robot. So we have this human robot collaboration, relationship happening. The robot doing things that are very repetitive, very boring, things that could cause repetitive strain injuries to the human. We see the robot taking up that role, and we see the human providing that dexterity, that creativity, that solution to unexpected events on the manufacturing floor, we see the human coming, working collaboratively, collaboratively with the robot, to solve that kind of, those kind of problems and those kind of situations that arise so So, following on from, From from all that, I see that as time goes on, there will be a need for manufacturing systems to be, to be to be flexible, to be flexible and to be able to adapt to various product changes that might come as a result of maybe a change in trend, a change a change in Fashion, a change in the weather, or a change, as we said, customized, customized designs, the manufacturing systems we need to be flexible enough to be able to do that. And this is something that certainly various researchers and various R and D centers are looking into

Jessica Owen
that's very interesting, and it, it's, it's funny that you mentioned about the idea of replacing humans, because that was something I wanted to talk to you about, because I think some people might be worried that if they were to implement all this new fancy technology, they'd have to get rid of their workforce. And you don't really want people to lose their jobs, but by the sound of it, really these machines are just going to replace the boring, you know, low level jobs. And actually, these companies might have the ability then to train these people up into sort of better, more intelligent, I'll say, positions.

Dr John Oyekan
Yes, certainly, and that's the goal of of human robots collaboration research. So the goal is to not replace a human, because, as I said, There's no way we can replace, we can totally replace a human, but we have a situation whereby we can get the robot to do the boring task. We can get the robot to do the repetitive tasks that could lead to injury. I don't know whether our listeners are quite aware of this, but when you do a task repetitively over a period of time, especially when it involves a lot of heavy forces, a lot of lifting, lifting, it causes work related strain injuries and and that's cost the government, the UK Government, or even UK PLC, should I say, a lot of money per year, because a lot of people have to take sick days off in order to to get better, from, from, from these injuries. And some of them could actually be life changing in itself. So what we are looking at there is not to replace the human, but actually to give the human a better quality of life by giving him a robot companion that can help him out in those tasks that could actually cause him injury in the long term.

Jessica Owen
So looking at the current crisis, then with with COVID 19, how can these technologies that we've mentioned today, such as AI and digital twins, help to support the industry there?

Dr John Oyekan
Okay, so in terms of of AI, one of the things, one of the challenges that COVID 19 has brought about, mostly is about how we could do social distancing. That's one of the challenges is brought about. Another thing he has brought about is how companies are to change their manufacturing lines rapidly to deal with. Crisis, because there was a time when there was there was shortage of PPE and and our and then various manufacturing lines, or various companies are to think about how to adapt their manufacturing lines to get to get into PPE manufacturer. Now one of the ways that digital twins, for example, could help in that situation, will be hypothesis testing. Now, what I mean by hypothesis testing is that we have a manufacturing line, but we don't want to change the manufacturing line physically right now, what we want to do is to first do a series of experiments in a virtual environment to see what a manufacturing line can be reconfigured to do. So instead of you trying to tweak the actually physical manufacturing line, which will cost a lot of money, you do your tweakings in a digital environment. You get it you get it right, and when you actually deploy it in the physical environment, is done right the first time. So the reason why I brought that up is because in the textile certainly in a lot of some of the PPE equipment that were eventually deployed or manufactured, manufactured came from a textile industry. So things like those masks, masks, protective gowns and so on and so forth. So digital twins, where I see digital twins becoming quite important in the COVID 19 response, is how we could actually use digital representations of physical systems to do to conduct rapid hypothetical testing of manufacturing lines to see what we could actually redeploy them to do. Yeah,

Jessica Owen
well, I mean, it makes so much sense what you've just said, if you could do that, it really does sound very clever. So some people that I've spoken to over the last few months, since the pandemic, have said that they think that beforehand, I think people were quite skeptical to try out these new things, but the pandemic has sort of pushed them to do it because they can help, can help them in this time when you're having to do things a bit differently. So I mean, would you agree with that? Yes,

Dr John Oyekan
I agree with that. And one of the reasons why is because I think majority of the people that switched, or majority of us switched to remote, working with sets, we switched to online, working, working from home and where I see, and a lot of people as a result of that. I mean, there's so many things that we thought that requires, for example, face to face meetings. We could actually do them online. There were some, and that sort of saved a lot of transportation costs so people weren't traveling going from one meeting to another, you could just meet virtually online. Now looking ahead as well, one thing I also see is that, and this links back to what we were discussing earlier in our conversation, that you could actually use, you could start using digital twins to monitor remote manufacturing plants. So you have a manufacturing plant that is still running, but because you have this digital twin at all, you could actually understand what that physical manufacturing plant is doing in a remote, I mean, from a remote location. Reason is digital twin is digital representation. And you could also actually control that factory from a distance as well. And that's and this kind of technologies will start becoming, or became even more important as a result of of the COVID 19 situation. It I think the COVID 19 situation enabled us to start thinking that, okay, what could I mean? It unleashed some creativity in us in the sense that, okay, how do we continue production while still being remote? I think that's when we think that the COVID 19 are sort of highlighted, right,

Jessica Owen
interesting. So I mean, as I said, There people have previously been quite skeptical about these things. What advice would you give to people or companies, businesses who have no idea where to start using this technology? Okay, so

Dr John Oyekan
there are two advices. Well I can give the first one is to start to, first, have a business case. First think about, okay, what do you want to do? And first of all, ask yourself, what is I mean, don't just apply industry for for the sake of it, first of all, try to understand what processes that you think that might require it. Is it to increase efficiency? Is it to enable you to be able to get data, to get insights and to give you a competitive advantage? First, I. Understand what you want to do, and then pick a very small, a small section of that big problem, and then try, try applying some and try applying some techniques to eat so and that way build, build a business case and confidence to actually go bigger. Another advice I will give is, I mean, the government has set up high value manufacturing catapults in the UK, and some of them, we have seven high value manufacturing catapults in the UK, and the University of Sheffield also has one called the Advanced Research manufacturing center. So what I advise our listeners to do is to get in touch with one of these centers, and because they've been set up to actually do translational research, so research coming from an academic base and going to industry. So they are like a bridge in between academia and industry. So I'll also say that you get in touch with them and and and see how you can develop some pilot studies to suit your your need. Also academics as well universities as well, have departments that are quite good at doing applied research, and that as well, is another advice I will give our listeners to look for universities which have departments that have applied research, and they could also help you to come up with some pilot pilot studies, some knowledge exchange programs, and use some knowledge transfer partnerships that could enable you to start small, try some pilot studies, and then once you start gaining some benefits from those pilot studies, you can now expand it to the entire organization.

Jessica Owen
Well, that's really interesting. I didn't actually know that the UK had these programs in place. And I'm assuming, if the UK has it, then probably many other countries across the world also have these sort of schemes there for people. Yes,

Dr John Oyekan
Germany as as a scheme run by different Hoover Institute. So it's similar to the high value manufacturing catapults we have in the UK. And I'm sure as well that places like the US, China and Japan also have their own schemes in place that work towards bridging fundamental research with

Jessica Owen
industry, right? Okay, um, so moving forward then, I mean, there's so many things that you've highlighted already that are very cool, very new, just they could do so much for people. But, I mean, are there any other interesting developments that you've heard of recently that you've been really impressed by and you think, wow, that's that's pretty innovative.

Dr John Oyekan
Yeah, I think, as we I think one of the one of the things that I found quite interesting in textile is the integration with wearables as time goes on. Because textiles, everybody wears clothes and wears dresses, but what we are seeing increasingly, is outdoors fabric could actually become more intelligent and actually start collecting data right from the source itself, to improve the designs of clothes, or even to make better to inform the design of better fabrics. So, and that's one of the cool things I see coming that's slowly coming up in the textile industry. I mean, currently it's not the smart textiles are not widespread yet, but as time goes on, as more and more research, as critical mass in that area increases, it's most likely that we'll start seeing some smart textiles on our on our doorsteps, very soon.

Jessica Owen
No, I think, yeah, I've spoken to so many people about smart textiles, and they all say the same thing, that they reckon that. I mean, all the research is happening now, and there are a few companies out there doing things with like medical smart textiles and smart yoga pants, even. But yeah, I think it's probably going to be another five, maybe even 10, years before people, before all of us, have at least one piece of that sort of technology in our wardrobes. Yeah. So finally, John, I was just wanting to ask you so WTI n is hosting a virtual trade show for the first time this year, and obviously you having such an interest in digitalization. I mean, what do you think to that? Do you think it's an interesting move, considering the pandemic and other shows having to be cancelled?

Dr John Oyekan
Yeah, I think it's a very good move, because a lot of I mean, for. Example, a lot of conferences this year, and I'm talking about academic conferences, even even the flagship conference, if I've have to, have had to move online. And I think that the pandemic, the pandemic has enabled us, has raised the awareness of how we need to think about how to do things differently, but also efficiently and also still have the same impact as we would have if we had a face to face conference. And I think that, I think that's what I'm increasingly seeing, that okay, a lot of conferences, or a lot of events, even though, I mean, are currently moving online, and majority of them are still getting that impact that they will have had if there was a face to face meeting, basically,

Jessica Owen
yeah, I think that's exactly it. Isn't it? We've just all had to, we've been forced to do things differently. And actually, I think some of it's working out quite well. And speaking to some someone else recently, and they said it's probably not going to replace the traditional events and trade shows. But I mean, maybe we'll have a hybrid system in future. Who knows? Yeah,

Dr John Oyekan
and I think certainly that's, that's one of the, one of the ways to go. I mean, for example, I mean, if one thing, one of the things as well, the pandemic has raised is that it has made us understand how, I mean, from a climate change perspective, from an environmental impact perspective, we've seen that, okay, some things could still go on and still be effective, even when we couldn't see face to face. So why I see, as you said, there is a situation where we're going to have in a hybrid system whereby, yeah, there'll be an there'll be a face to face meeting, but also there will also be that online things that we can do from an online presence or an online perspective,

Jessica Owen
yes, definitely. Well, I mean, John, that's all we have time for really today, but I mean, thank you so much for coming on and enlightening me and hopefully the listeners about digitization and all this stuff that it can do. I mean, it's been great to have the opportunity to speak with you. Thank you. 

Dr John Oyekan
Yeah, thank you very much for having me. Jessica.