00:00:21:14 - 00:00:28:03
Hi, everyone, and welcome back to Tech Will Save Us podcast. I'm your host Dr. Tom Helliwell, as Callum is having the day off.
00:00:28:05 - 00:00:32:19
And I'm joined today by Gianmarco from the AMRC, introduce yourself.
00:00:32:19 - 00:00:55:05
Hi, Tom. I'm Gianmarco, tech lead at the AMRC in Sheffield. I'm a tech lead in novel robotics and sort of my area of interest is any sort of non-conventional robotics. Specifically, I try to look at what research happens in university and especially all of the research in robotics and how can we apply that to industry.
00:00:55:07 - 00:01:08:10
So I'll try to look at a low tier research that is again very academic based and I try and see how I can push the research towards the higher tier spectrum like 6 to 7, and I give that to industry.
00:01:08:10 - 00:01:16:08
Cross of the chasm, as we say. And joined with Gianmarco, is Mark Gray from Universal Robots. Thanks for coming down.
00:01:16:08 - 00:01:16:18
00:01:16:18 - 00:01:37:12
Thanks for having me. So my name's Mark Gray. I'm the Universal Robots UK and Ireland manager. I've just set up our UK office in Sheffield, which is our trading hub and we've got a demo suite there. I've been with the Universal Robots seven years, which makes me a dinosaur in Universal Robots terms, and I'm here today to tell you a little bit more about what Cobots can do. Fantastic.
00:01:37:12 - 00:01:59:01
So today we really want to talk about robotics and the area of robotics, especially around manufacturing, these applications industry and where do you think we've been as a as a human race, what's our relationship been with robots and AI up to press, what what's your sense of where we are today and where we should go next?
00:01:59:03 - 00:02:17:08
Where do you want to start, Mark? So, yeah, my my take on it is that robots have always been seen as quite negative I think, in the media and in the press where people relate them to job losses. But actually technology's always enabled how people are going to do their work and be more efficient. And I think that is changing.
00:02:17:08 - 00:02:43:22
There's a sort of global movement to look at automation generally, and robots are now being seen as a good thing. Potentially, like our robots, we make collaborative robots so they can share the workspace with people and that's opening up new opportunities for people really. Yeah, no, I, I agree with what Mark just said, there's always been sort of a mistrust in robots because the common sort of of public says robots as they're going to take jobs away from us.
00:02:43:24 - 00:03:16:10
But that’s never the case especially with collaboration robots, but robots in general, there there to sort of make lives of employees better. So we're trying to remove or like robotocists is in general trying to remove the physical labor from from from humans. We're trying to make make sure that humans are more comfortable, a little bit in their in daily jobs. And we don't want we don't want people to think that robots are going to take over the job because, that’s, that's never going to happen.
00:03:16:10 - 00:03:25:18
I think for me as well, I think there's robots in software in a way. Yeah. When I think about software, some of the work that Razor does, and the AMRC and everyone else really
00:03:25:23 - 00:03:56:18
Is that in a way you've seen robotics in software, in automated processes. I think when it comes to robots, classically we think about things that are in physical tasks, but it seems like the software world is almost automating quicker than the physical world. I think we've seen that happen in recently, things like robotic process automation. So when you do something, you order something online, suddenly it will trigger a chain reaction of events and you know, it basically means less people are required fundamentally.
00:03:56:20 - 00:04:16:08
Yeah, whereas robots, I think that the arms that Universal Robots are now doing, the cobots. I think the thing that people think about are you know, your car factories where there’s a structured environment and they're just doing body in white, you know, grabbing the big panels, welding, spot welding, assembling the parts together.
00:04:16:08 - 00:04:29:07
I think one of the amazing things with what Universal Robots are doing is the fact that you know, you’ve been using environments that you would never imagine robots being. Places where you'd expect people to have to do it.
00:04:29:07 - 00:04:51:00
Well, sanding was one, wasn't it? Yeah, we've seen them used in really quite small companies and they take over tasks that traditionally you would have thought that a robot wouldn't do and it might be sanding garden furniture or assembly of furniture, that sort of thing, all very labor intensive but quite dull tasks and difficult to recruit people to do those tasks nowadays and also difficult to retain people.
00:04:51:02 - 00:05:12:13
So if you can take that strain away by using a robot to do that, it means that you can deploy the people that you've got. So we're seeing more and more of those kind of smaller companies adopting robotics, you know, rather than the big sort of major users of automation in the past, like you say, have been big car production lines, it's all the sort of smaller manufacturers in different fields that are starting to adopt this technology.
00:05:12:15 - 00:05:28:15
And that's what's really exciting because if we do that domestically, it means that we're sort of competitive on the world stage as other companies that are automated in other countries that are automation as well. So we have an advantage over other companies in sort of Turkey or China in the fact that we're closer to the rest of Europe.
00:05:28:20 - 00:05:46:23
So if you make things for less money, we can export them without the extra logistics cost that China has to get products halfway around the world. So that's a big change for us. I suppose the labor costs as well in the West, Western Europe actually supports return on investment. Exactly. Which is the reason why you should look at it.
00:05:46:23 - 00:06:08:08
Yeah, I do, this concept of automating tasks I think is really powerful because I think there's a, there's a sense within manufacturing that you almost need to build a brand new factory and it’s got to be completely automated. Yes. Whereas you know, as you say, with Universal Robots thinking, what tasks can I actually automate? And actually it just fits quite nicely within an existing, you know, manufacturing system factory.
00:06:08:08 - 00:06:34:17
Whatever you want to call it. Exactly. There are a lot of tasks that previously just wouldn't have been able to be automated that we can and we talk about a low impact installation where you're not having to change your entire factory floor because that's sort of be it for anybody in manufacturing. You must see this in research where, you know, you're introducing robots into a process, they don't want to change the entire manufacturing culture, they just want to automate certain parts of it to try and increase the productivity.
00:06:34:19 - 00:06:38:10
So that's why we see a big change in. Yeah, and to add onto that, you
00:06:38:10 - 00:06:57:08
mentioned, there's smaller SMEs starting to implement robots in a sense. But we've had projects with even big, big companies where they have to use, for example, for aerospace they have to inspect blades, the turbine blades, and they will spend hours and hours spending on that and are paid, people doing that are paid very well.
00:06:57:10 - 00:07:38:23
But after six months, they get bored with such a tedious job. Imagine doing that for 8 hours and we've used robots to, to sort of replace them, but it’s not replace the humans. People don't want to do that job. So that's why we need robotics. And another thing, another reason why we need to implement robots is that as we mentioned before, there's there's a massive gap in, in sort of industry and in the manufacturing industry specifically. I was running a sort of workshop in Denmark of the ERF and I realised that there's I think by 2030 there's going to be a gap of 80 million engineers that sort of missing or will be missing
00:07:39:01 - 00:07:53:08
from the job market. We're talking about 2030 and 80% of the work sort of force of 2030 is already working. So we can't replace that by the 20% can’t replace the 80 million people that are sort of lacking so robots is the answer to
00:07:53:08 - 00:07:59:23
that. Yeah, And they also, this is the answer, as you mentioned, like sort of third world countries where there's very cheap labor.
00:08:00:00 - 00:08:18:24
Obviously, we don't want that to happen here. So robots is the answer. If you want to be competitive with those countries where people are paid $1 a day. Well, obviously we can't do that, but we can do robots that automate those tasks that people don't want to do anyway. Yeah, Mark what's the market penetration? Because I think it's something we've talked about before, isn't it?
00:08:19:00 - 00:08:48:11
Yeah. The UK is lagging behind really sort, you know, adopting automation, robots, cobots. We are traditionally behind the G7 countries. Interestingly I found two graphs about two weeks ago and side by side the UK was the lowest out of the G7 for productivity, but out of the G7 robot density we were the least as well. So there is a correlation by having more automation in robotics, these actually means that you've got more productivity. It’s about having the more efficient machines.
00:08:48:13 - 00:09:10:07
Throughout all, from the industrial revolution, we've seen there’s, as we've increased technology, we've managed to produce more, at less cost with less waste and that's what we want to work towards, it’s everybody wants to work towards that. We're talking, it's interesting you mentioned about the lack of people that are going to be in the economy by 2030. We're all going to be driving electric vehicles by then so that manufacturing will have to be done locally.
00:09:10:07 - 00:09:23:16
And all those high tech roles are going to have to be sought. We can't we know we can't fill those roles now. So automation is going to have to play a part of fulfilling some of those roles in that sector. It’s driven it, it’s just an inescapable fact really.
00:09:23:16 - 00:09:34:07
Yeah. I think that the sort of systems you're talking about. Visual inspection type stuff where actually the robot, the robot concept and the software concept are really starting to come together.
00:09:34:12 - 00:09:58:11
Yeah. Outside the physical systems type stuff that we're familiar with, with you know, the automatic and systems control engineering department in Sheffield you know this this idea of so visual systems where robot arms, vision system, software and whatever the newest software architectures are and cloud all that sort of stuff are really brought together to do something totally different.
00:09:58:12 - 00:10:19:16
Yeah. And it's also, I think it's a standardisation, isn’t it, of it. Yeah. Every single time it does the same inspection. Yeah, exactly. And there's no human error either. So there’s a compelling case for it as well. Yeah exactly. You're not only reducing sort of the amount of unpleasant jobs that people don't want to do, but you’re also making sure that that job is done repeatedly because that's what robots are good at. They are very repeatable.
00:10:19:16 - 00:10:48:04
If they do one tasks once, they're going to do pretty much the same thing forever. So yeah. Yeah, I think Mark mentioned “automate what you hate”, that was your, sort of saying. Yeah. So what sort of things are you seeing? What are people hating and automating? So if you've got any sort of material removal, whether it's sanding or deburring or, or grinding and grinding is interesting because the vibration white finger statistics show that it's actually you can't use a grinder for 8 hours a day.
00:10:48:04 - 00:11:06:04
There are only a certain amount of time limit that you’ve got and the regulations are going to change in the future where it’s going to become more stringent. So people are going to have to look at another method of doing this or employ more people to do it. But one of the anecdotes that we get told more and more when we speak to people who are doing these kind of tasks, they just don't want to do them.
00:11:06:09 - 00:11:24:18
You know, you can't put someone on a job for more than two weeks before somebody else has to have a go at it because they just leave. Yeah, I mean, I've seen the opposite. I've seen the tier one aerospace supplier. You’ve got eight guys leaning over and, part of an airplane where I won’t give any more away, but they’re stood there sanding away with aluminium.
00:11:24:20 - 00:11:49:12
Some of them have got masks on, some don’t. You think this is crazy for for such a modern, incredible product they’re creating, the manufacturing process was just almost really, really, really far behind. I think because we've got a labor gap in the UK and we've got an aging population of workers that have got skills. We've started to change how we look at people and we've really valued people. Like human skills are
00:11:49:12 - 00:12:05:18
the biggest asset that any manufacturing business can have. But also you want to look after their wellbeing because you want to retain people. So if you want to take away those horrible tasks they don't want to do, that's, well we should automate them really. And that's, I think robots and technology is a good way of sort of fixing that problem really.
00:12:05:20 - 00:12:36:08
Yeah. Another trend I'm actually seeing is, talking about terrible jobs that no one wants to do, is sorting waste or picking up and recycling waste. It’s something terrible. You don't want to do it. Again third world countries have people doing it, like spending loads, with no safety on and not paid much. I can see them, in fact, I’ve been sort of in contact with some companies, local companies who want to do it here.
00:12:36:13 - 00:12:59:24
Most of them want to do it safely and they want to do it properly so people don't get don't get injured and they're looking at using robots. And by using robots, you also remove the sort of necessity for hazmat suits because you don't know in some companies at the moment their employees will have to use, will have to wear for eight hours a full hazmat suit because you might be touching something that is contaminant.
00:13:00:01 - 00:13:21:07
So if you use robots instead, you don't have to use it. And same thing when it comes to nuclear and nuclear inspection. There's a lot of robots they use at the moment because you don't want humans around that. And then robots are much, much safer around them instead. And yeah, I think it’s getting those guys who maybe would be doing that work doing something which is classed as high value.
00:13:21:10 - 00:13:39:18
Yes, higher value activities, things that, you know, humans are really good at and robots are not very good. And even all this AI that we are now getting excited about can't do. Yeah. You mentioned problem solving, things like that, things that you know they just struggle with. Yeah, that's why you can't really get rid of humans really.
00:13:39:18 - 00:14:04:00
Because humans are always going to be behind the scenes controlling the robot and training the robot, making sure the robot does the right thing because sort of problem skills, the problem solving skills that we've got ,robots probably will never be able to do. And you mentioned software. Software is going very fast now and hardware is really slow, is trying to catch up. When it comes to robots, they are very hard systems.
00:14:04:00 - 00:14:24:19
You’ve got the electrical, you need to know about electrical, in engineering fields, you’ve got a bit of control, bit of mechanical, bit of software. And then if you talk about like ChatGPT or whatever, yes, you can integrate that with the software of the robot. But then how do, will the robot be able to understand your command? But then how do you translate that into a motion of the robot?
00:14:24:20 - 00:14:45:19
Because doing that specific thing, like picking up an object. To a robot is quite a complex because it can pick up an object 1 million different ways. So how do you choose which one It's so it's very hard to sort of cash out the software. I think we've seen that with Universal Robots and some of the business around robots that really made the barrier to entry.
00:14:45:19 - 00:15:09:11
Yeah. A much lower I mean Universal, you guys have got some amazing software. Yeah. Where you can literally train it by guiding it through a path, can’t you? You’ve got the ecosystem around the end effectors and. Yeah, it makes it just easier to deploy. I mean that's the whole point of our offering really is to have a pathway to make it as simple as possible for companies to buy, you know, piece of equipment, put accessories with it that are tried and tested.
00:15:09:13 - 00:15:31:03
That'll work and they can get results from quickly because that's one of the things that's really important to small companies is that they don't want to put something in and spend two years developing it you know, it needs to be within a matter of weeks and then they're saving money and, you know, they solved that problem. So by making it easy and simple for people, that's the that's the sort of mainstay of what our business philosophy is all about.
00:15:31:05 - 00:15:56:14
And I have to say, as a sort of user, I really like what Universal does because I'm also sort of working on collaborative workspace in general. And we look at the sort of the psychology of robotics and what the attitude towards robots are from humans and sort of research shows that sort of trust and understanding of technology go hand-in-hand.
00:15:56:16 - 00:16:19:18
So the easier you make something that is novel, the more people are going to accept it and you guys at Universal, you probably have got one of the most sort of intuitive sort of robotic system to use and even the interfaces yeah, they're very simple to use, which we use all the time. So you're working on collaborative workspaces?
00:16:19:18 - 00:16:42:00
So this concept of humans literally working side by side with cobots, okay. Which is really interesting, because that's, that's one of the areas where you can increase productivity massively by blending tasks between, task balancing between a human and a robot working together. That's where the the traditional sort of thought process is, it's either a person or a robot.
00:16:42:05 - 00:17:09:23
But having both together means that you can produce so much more and let the robot take over the things that it's good at and let the human carry on doing what they're really good at. And that's that's where we see the main growth in robotics over the next 5 to 10 years. Yeah. Sorry, I going to say the economic benefits almost immediate because if you just think about if you have conventional industrial robots that need to be like caged in between within buyers but if you sort of merge the work spaces, you can have the robot and the human working in the same space.
00:17:10:00 - 00:17:45:22
You don't need space, you don't need as much space, as big as a factory, you can just reduce space. And that's already a massive increase in sort of, of reducing costs. Yeah, I think the first thing I sort of think about is kitchens. Having spent a bit of time as a chef. I think the idea of having a robot arm cutting and maybe peeling vegetables or something quite, quite compelling, but surely there's some safety challenges around that and I know that Cobots have a lot of safety systems in place, but if a robot was holding a knife, you know, is the, did you mention some research going on in that area in order to sort
00:17:45:22 - 00:18:13:08
of manage that type of situation potentially? Yeah, I think I would say we are quite far from having a robot in the kitchen yet. What a shame. Mostly because as you say, as soon as you give a robot sort of a knife in its hand, if it was controlled that robot can just cut someone because, obviously Mark can tell us, especially collaborative robots have got sort of four sensors at each joint so they can sense a collision
00:18:13:08 - 00:18:44:14
at any point, they can just stop. But if you've got a sharp knife, and you're touching someone's eye, you don't need that much force, you can still harm someone. So the research that is sort of happening now is not it's not just focused on the robot workspace, but it is also looks around humans and look human tracking technology or even looking at the human if is getting close to the robot, if it feel like keeping his distance, but shows that if the humans keep some distance, maybe there's something going wrong so the work can be slowed down.
00:18:44:16 - 00:19:11:21
And there's also obviously lots of AI just to understand facial emotions, so people’s response to the robot. And so again, tracking of the person and that's going to again it’s a work in progress. But people are looking at incorporating that sort of AI algorithm to understand the sort of human behaviors into the robot controller so we can have a more comprehensive or more comprehensive understanding of the environment, which includes the human as well.
00:19:11:21 - 00:19:37:07
So yeah, I understand, yeah, because I think Universal Robots has a mechanism for when it touches some, it has some sort of mechanical. Yeah, we've got force monitoring throughout all the joints. So if the robot collides with something, it detects that and then it puts itself into a stop. But if you've got something that potentially could injure somebody on the end, like a rotating tool or a screwdriver or something, you can protect a person from an area scanner.
00:19:37:09 - 00:20:07:16
So if they go into the zone, you can make the robot stop. We call that speed and separation. And it's interesting what you were saying about that dynamic, being able to monitor a human's behavior and then adapt the robot's program to that, because that is the future. That is where you've got a complete collaborative cell and the flexibility of that, of course, means that you can make much, you know, many, many different types of products in that one type of cell and having that dynamic link to the human it’ll actually like they can interpret what the human's going to do next so it can set up the next part of the process.
00:20:07:16 - 00:20:28:15
And so it's a really interesting research area I think that. Yeah, there's also some research that's looking at sort of monitoring the human behavior and see if the human sort of, if it changes this behavior, that means that maybe the person is tired. So the robots should adapt to the tiredness. So making sure that it can go this same sort of, it can synchronise with the human.
00:20:28:15 - 00:20:57:14
They can just slow down so the human doesn't get tired. Also that because we're talking robots are here not to replace jobs, but to make jobs easier. So there's some research is looking into monitoring the person and for example, in attending exercise when there’s a robot and a human sharing a workspace, when they pass objects to each other, there's lots of research looking at the sort of most comfortable way for the human to pick up the object and give it to the robot and then the robot sort of optimisers looking at the height or the posture.
00:20:57:19 - 00:21:14:06
And then it can just tend the object in a much better way. So the human posture is better, you don't have to sort of bend or like because in the long term, if you do that for 8 hours, that's very detrimental to your health. So is it doing some sort of real time control type stuff there? Yeah. Rather than just running a program.
00:21:14:11 - 00:21:44:03
Yeah. So real time control, which is sort of connected to also human tracking and with some some, at the moment they're doing with sort of IMU sensors and also some biosensors where you can just try to sort of heart monitor of the person or even like the posture if that person is sort of bending a bit too much, because, you know, even when you work on a monitor, you're supposed to look at the monitor at a certain height because for 8 hours a day, every day, you might start sort of causing pains to your back or to your spine.
00:21:44:05 - 00:22:04:23
And we're trying to, people in research are looking at the same exact thing but with robots. How can we make sure that this collaboration is as efficient and also healthy for the human as well. That's important because you can have different people working in that same cell. So you could have a, you know, lady who's five foot six who's, you know, weighs about seven or eight stone and a guy who's six foot four, weighs twenty stone.
00:22:04:23 - 00:22:25:24
How their posture and how they are going to interact with the cell is completely different. You know, the ergonomics will need to change dynamically. So that research does that take place primarily within universities now, what is is it you know, is the industry taking the lead on that sort of thing? I think as well, the other question I was going to ask is and is part of that is, where does the robot operating system fit in with all this sort of stuff?
00:22:26:00 - 00:22:57:03
Because that's been going on for a long time and I know it's been developed by Universal around the world for a while. But you know, where does that fit in with this, this whole ecosystem we're seeing now? Yeah. So from what I've seen and so I'm looking at roadmapping some collaborative workspaces for, for the AMRC in general, also for the UK industry, unfortunately it's mostly academia that's not, industry in a sense as they're looking at sort of this this aspect.
00:22:57:03 - 00:23:23:02
Well, they’re just putting their funding in and they're not getting involved as much. The trend is changing. More industry are understanding that that's where we need to go, but it's mostly academia. And that's another thing that another trend that you see with ROS, Robotic Operating System, which I'm a big fan of, but again it’s not really implemented with with industry.
00:23:23:04 - 00:24:02:03
And just for the people like watching, ROS is this medium where, that’s been developed to sort of unify and make a universal how you control robots because many robots brand, many robot brands have got the wrong proprietary language and they use different ecosystem to control them. So then when you're trying to build a cell, especially people like us integrators where you've got different systems working together, everyone uses its own proprietary language, the integration takes ages and ROS was born just to sort of tackle that, to make sure that every sensor, actuator, every robot shares the same language and it can talk together.
00:24:02:05 - 00:24:37:18
The problem with that is ROS is very, again, academic because it's open source. And with open source is sort of, its greatest strength because it's open to everyone in research, that people, researchers can just reapply to other tasks so you don't have to start from scratch. You can just implement whatever people have done already. But then the open source is also the biggest weakness as well, because people like aerospace companies, they can't really translate because you don't know the... Security is the big issue.
00:24:37:20 - 00:24:58:11
Yeah...You know laboratories, high tech manufacturing, secure manufacturing they want to have something that is secure and has traceability in it as well. So you (? Universal?) do this Robots have spent a lot of time and effort on their software because it's just a barrier to entry, it’s so much lower. Yes. Get a robot in and train it in the way you want to.
00:24:58:11 - 00:25:26:09
We want it. Yeah, we wanted to create something and what's interesting about where we started from, we actually before we were a robot company, the three guys who started the company actually did a research project to ask people why they wouldn't use industrial robots and the answers they gave formed the nucleus of what a cobot was. They decided to make the architectures have force monitoring, so it didn't have guards around it, but they had, they wrote a software package called PolyScope
00:25:26:11 - 00:25:49:09
to make it simple for non robot programmers to use it so they could have released it just with ROS and the uptake might not be as quick. But by making it really simple for people to use because we didn't have anything previously, we weren't an industrial robot company that's adapted a product, it’s being created from scratch. So that's one thing for development of products is when you actually go to the market before you've made a product, say, look, what do you need?
00:25:49:09 - 00:26:09:19
And they say this and we come back to the package to say, look, now we've made this, it is easier to use. And that's the thing. The ROS, we have a ROS driver package that we have for ROS and ROS 2, which is really useful for researchers. And it’s almost like a, well essentially like a plumbing system, messages going between all the different parts of the robot and different systems.
00:26:09:21 - 00:26:29:15
But we wanted to make something that was out of the box ready that anybody could simply use, whether you're in a machine shop or whether you're a welder or whether you're in a food and packaging company. They could all use that software. The software we created straight out of the box and we keep adding bits to it, you know, all the templates and wizards to try and make it easier.
00:26:29:20 - 00:26:43:16
It's a bit like the old Paperclip used to say on the Microsoft Word, it’s the thing that gives you hints and tips and things you can’t do wrong, but it actually just makes it easier for you to use that piece of software rather than, you know, just giving you a manual and say, there you go.
00:26:43:18 - 00:27:09:02
That's the way that we try and make it as simple as we can. Well I’d say it's a bit better than that... really annoying. What do you think the future will be of robots? Will we see them more frequently out there in the wild and in manufacturing system, do you think there will be a platform, a standard that emerges as being the one that everyone has or do you think it will stay sort of quite mixed?
00:27:09:04 - 00:27:15:00
I think it's quite fractured, isn't it? At the moment There's there's lots of different manufacturers in different fields of robotics.
00:27:15:00 - 00:27:30:07
I put a poll on LinkedIn about a week ago and put sort of four different examples of areas where people were interested in robotics being used and by far the winner over anybody else was agriculture, a very labor intensive sort of industry.
00:27:30:07 - 00:27:55:02
And again, we're facing that where we can't get labor to come into the country to help us. And and agriculture is the backbone of any economy. You've got to feed yourself domestically. So that's one of the areas that people are really interested in developing robotic systems for. I think there are others that are quite labor intensive like construction and logistics and, you know, just general sort of like other sectors that we still rely on human labor for.
00:27:55:14 - 00:28:16:00
But agriculture was one that was really, really strong in that and I think that's going to be a big area for investment over the next ten years, I think, where people are going to need to sort of develop robot systems. It's interesting, because I think agricultures, it's like robots in the wild and quite literally in that case and then with logistics it’s again similar.
00:28:16:02 - 00:28:40:11
If you’re going down the street, I know that in America they've had a few other sort delivery robots kicking around and also, you can have automated cars and things like that. But then you look at an Amazon Warehouse and they have robots, but it's a very self-contained environment. Yes. I think with manufacturing we are seeing a lot of challenges with interoperability, getting data across from CNC machines, robot arms, conveyor belts, whatever it is.
00:28:40:11 - 00:29:06:04
Yes. It's been a big challenge. I think there's definitely a lot of demand for something that maybe grease the wheels of that data and interconnectivity and control that we don't seem to have right now, so. I think so. It's an interesting area. And again, that's, I think what ROS is sort of the answer to, because that's why ROS was born, was born for. It was just meant to be an interoperable sort of system that can communicate with everything.
00:29:06:04 - 00:29:40:24
And again, sensors, actuators, robots. And I think there's sort of a sense that industry is getting more interested in it and in ROS more specifically and maybe we're going to take a big take on sort of not necessarily away from open source, but more rigorous open source that's more structured because I think just at the end of last year, Open Robotics, which is a company that was looking after sort of the ROS maintenance, has been purchased by or has been acquired by Intrinsic Robotics.
00:29:41:01 - 00:30:08:17
And Intrinsics is sort of a sub-division of Alphabet, which is also the company that owns Google. So you can see that people, people like, who work there in Google, they can see a future in growth because they can see, as I said, there is this need of a common language for for system, because as the complexity increases, we need to make at least a basic simpler because the stuff to solve is actually so complex in itself.
00:30:08:17 - 00:30:29:22
So if you just spend years just making sure everything communicates with each other, then we haven't solved the task yet. We're not going anywhere. Yeah, I can see. Yeah, I think as well, I think a lot of people see robotics as, you know, they think of the humanoid robots. I was going to ask you about this as well, who do you think is leading in terms of robotic excellence in the world now?
00:30:29:22 - 00:30:48:18
When you think about humanoid robots. I know Boston Dynamics is the one that always comes up in my mind, but I know Tesla is investing money into it now. Are there any others that come to mind? Yeah, I would say definitely. Pushing the envelope sort of thing? Yeah. So Tesla is definitely not leading and there's there's a few I can see
00:30:48:18 - 00:31:03:12
actually in the past few months, I've seen a couple have been popping up. I can't say. There are Chinese ones isn’t there? Yes, Chinese companies. Yeah. I don't necessarily remember the name, but I can see a lot of them coming up. To be honest I find them extremely cool. I don't understand why they need it.
00:31:03:12 - 00:31:18:09
Because having a robot arm, like if you are on top of a wheel and sort of HGV. It achieves. That’s all you need. It achieves a similar goal. You don't need that level of complexity to run sort of a humanoid robot. It’s ridiculous.
00:31:18:09 - 00:31:42:15
It's just.. just like walking, it’s completely complex. I mean, if you need to add weights on it and you’re spending loads of money to have so many degrees of freedom that aren’t necessary. If you just have some wheeled robot, they can just with an arm or a couple of arms, you've achieved what you need to do. So I don't really understand why we're going towards that because we don't need that. I guess they are aiming for this like completely generalised platform.
00:31:42:16 - 00:32:07:13
Yeah. But it's certainly a big ambition. But as we've, with AI, every time AI in general, even with my experience I realise that every time you try and make prototypes general and try and do everything, it’s just part of everything while robotics, you just, you just need to sort of simply decide on the task you want to do, and you can do that task amazingly.
00:32:07:15 - 00:32:24:00
I understand that some sort of flexibility is needed because tasks like the product, the object that’s coming in, you need to pick up might not come always in the same orientation, or might not, I mean, every single time is already quite complex on its own. So we just need to make sure that we're very good at task,
00:32:24:04 - 00:32:49:14
task-based projects rather than let's try to make everything, because we can't. Yeah. You touched on a good way of thinking about it so I guess, Mark, when you go to a manufacturing business, do you support them on identifying which tasks they should target? Yeah. How do you go about doing that? We actually did that yesterday. We were over in a company in Oldham and basically it's almost like doing an audit on their production line.
00:32:49:14 - 00:33:16:10
So you can you can rank them through one, through to five with easy right through to very difficult and difficult can involve, there might be multiple parts of all got to be fed and organised before they’re being assembled as opposed to a part that's just going into a press or having one screw inserted into it. So we always try and steer people towards using the the easy one first, because a culture of implementing a robot arm into a company is get one in first.
00:33:16:12 - 00:33:33:21
Let's show it’s successful. Let everybody see it and get comfortable to be around it. And then that sort of changes the culture in their output on things where it's taken over a task that perhaps they didn't want to do. And it's, you know, it's making more profit for the company, which makes them more sustainable. So we do help with that.
00:33:33:21 - 00:33:52:09
We go and do these audits on company sites to say this is easy, this is not. And then sometimes we find things that we've never done before. You know, they can be really quite complicated and we'll do an evaluation on that and see whether it's possible or not. That's the best way really, you know, you want to help everybody, but there are some tasks that are still, they look relatively simple,
00:33:52:09 - 00:34:11:20
humans are amazing. We were talking about the humanoid robot there. A human to be replaced by humanoid robot is really difficult to do because we've got our own built battery. We just feed ourselves and that provides energy. We've got wrists, fingers, joints, eyes and a brain that can make all these decision. And to automate all that would be incredibly complicated to compete with a person.
00:34:11:23 - 00:34:33:20
Yeah. So we are nowhere near... I suppose as well, with, with we've created this environment in our, in our vision, of ourselves so to some extent I suppose we've made that bed. Yeah. Because if environments were built purely for robot systems. Everything would be flat. Yeah. Yeah. Daleks can’t go upstairs can they? Everything would be flat. You could organise it better, that's for sure.
00:34:33:22 - 00:34:56:01
Yeah it’s just taken me back to the Amazon warehouses again where there's just no people. They’re basically shelves with robots with a shelf on top and they all just sort of bring the shelves to you and you grab it out and it disappears back into the crowd. Yes. And there's no humans in that area that is just purely sort of, I don’t know how to describe that, like a dynamic shelving system.
00:34:56:01 - 00:35:14:11
Yeah, like a hive or something? Yeah, yeah, yeah. It all comes down to cost per pick. So the more times that human has picked that product and put it down on something else, it's adding to the cost of the logistics of getting that product to the person. But the less people that touch it, means it's less expensive to produce and get out into the marketplace.
00:35:14:11 - 00:35:45:09
So automated warehouses and logistics is something that's that is going to increase. You know, almost everybody uses online shopping for one thing or another and you know for Amazon and for going to a place like Ocado for shopping all these automated systems behind them and that will increase because again it has to be has to reduce costs. So yeah, for the future that's going to see something we're going to see increase. I think as well with with that type of implementation it’s relatively easy to see the return on investment.
00:35:45:09 - 00:36:09:04
Isn’t it, because one frustrations that we have in Razor is that we are trying to sell data management things and I’m trying to say, Oh, you know, think about the continuous improvement abilities. If you've got the data in one place, you know what your company is doing, you can take action and become better, whereas if it’s automating something, whether it's software automation or it's automating a task, it makes it a lot easier to sort of make a business case to invest in it as well.
00:36:09:07 - 00:36:27:04
Exactly. You've got to measure something to know how the processes are first, haven’t you.. Yeah, exactly. And if you can show that in data sometimes where you know, some of it's paid for itself and after it's paid for itself is then earning money for the business, that's quite a compelling thing to look for further tasks that you can automate really. Is there anything else at the AMRC that
00:36:27:04 - 00:37:07:05
that’s going on in terms of research in robots and automation that you think is worth mentioning or discussing? Um, yeah, I mean we, we do touch on very broad, um, sort of, we try and look at the whole aspect of, of, of robotics. Um, one thing that I think has been mentioned before that we're looking at is reconfigurability, so having reconfigurable cells where you can just sort of swap the robot and almost like nothing has happened and you don't have to spend days or weeks re, sort of, rejigging and we, I mean, um, do an integration
00:37:07:05 - 00:37:32:14
There’s one that occurred to me just as you were speaking, which was the robotic sheening, that's been going on for quite some time now at the AMRC. Yeah. So yeah. The gist of it, correct me if I’m wrong, is it for the large geometry basically roughing base machining. Yes. So the, sort of the advantage of robots compared to like conventional CNC machines is that you've got a bigger work space, a much bigger workspace.
00:37:32:14 - 00:37:58:07
And so yeah, you can and also have got more flexibility um when it comes to sort of if you want to do different shapes of different sizes and you can just reprogram the robot, you don’t have to rebuild whole, because those CNC machines are quite specific now, very expensive as well. Robots can be much cheaper. Um, so when it comes to machining, you can, if you can have a robot doing it, you've got bigger environments they can work with and you can just reprogram it
00:37:58:07 - 00:38:31:08
and it’s much, much easier and much cheaper. But you've got a disadvantage of robots, especially with serial robots where they're not as stiff as CNC machines. So you get a bit of drop in accuracy because, as you said, robots are very repeatable, but they're not super accurate compared to say the CNC machines. So there’s always that trade off. One of, sort of the solutions that we found is, I don’t know if you've been to our site, where we've got massive robot Titan, which is enormous, I think it's got seven meters of walking volume.
00:38:31:10 - 00:38:51:23
Um, but then yeah, it's a massive robot so and it can, it's quite funny to watch it because it’s got a tiny spindle just to do like very precise machining. But then you've got this huge machine it's probably ten meters tall. Is there a challenge with the stiffness of it as well? Yeah, Yeah. Wen you’re cutting something that is really. Obviously it’s much, because you've got bigger mass
00:38:51:24 - 00:39:06:03
Obviously it's much more stiff than conventional robots, but it's still not as stiff as a CNC machine. So yeah. That's very interesting, I think it’s, because when I heard about it for roughing operations, it made a huge amount of sense to be
00:39:06:03 - 00:39:15:10
The AMRC is an absolute jewel for South Yorkshire I think, to be based here and that's one of the main reasons why we came to Sheffield, we wanted to be based here, our UK office.
00:39:15:12 - 00:39:44:17
Our head office is here and we see Sheffield sort of evolving as a, as a technical hub really. For you know, having innovations led by the AMRC and the universities here. That's why I wanted to be aligned here to sort of create a cluster because Sheffield's always been a maker city hasn’t it. It's been one of those places where innovation has started and there's no reason why we can't do that now, create the next sort of skill sets that people are going to be able to use and smart manufacturing that people can use can originate here.
00:39:44:19 - 00:40:10:18
Yeah, yeah. I think it’s obviously one of the birthplaces of the first industrial revolution, you know and it would be great to sort of take more control of the fourth industrial revolution and create an ecosystem really, I think. Where do you think we are going wrong? It's a classic question. I think, well, there are there are no fundamental reasons why we're going wrong, is that we we need to sort of change our perception of manufacturing really in the UK.
00:40:10:22 - 00:40:32:04
It adds real value to our economy and we what we export, we punch above our weight with the sort of GDP that we have. We export a lot. It brings a lot into our economy for that. But we need to sort of make it easy for people to gain those skills, to make it easy for companies to use technology to be competitive on the world stage.
00:40:32:09 - 00:40:53:13
We were in the past. We look at the Victorian era when we went through all the sort of industrial revolution, the things that we developed, you know, anything that I'm reading the book at the moment called Exactly by Simon Winchester, it’s the history of Precision. Precision Engineering started in the UK. Up until that point, people making bolts with all sorts of different types of threads and we standardised on it like we did with the time.
00:40:53:13 - 00:41:12:21
Greenwich Mean Time is a world standard that originated here. There's no reason why we can't get back to that sort of mindset and make engineering really interesting for people and be able to set the standard and domestically produced products that will expand what we export abroad. Yeah, no, I agree with that. I think education is the key. We need to train more people.
00:41:13:02 - 00:41:36:20
As we said throughout the podcast, there's a massive skill gap that we need to sort of address and the only way to do is if we just sort of tackle this by teaching new people, training new people and not necessarily with university. Certain jobs, you don't need a degree, you don't need a Ph.D., you just need to be trained and, you know, an apprentice to just do the work.
00:41:36:22 - 00:42:03:16
So, yeah, just like training and disseminating how much, um, sort of how interesting robotics is and how much you can achieve and how great a career it is. Because sometimes I can see that people are sort of poached from finance people with the skills that we can sort of, that can be used for robotics, engineering in general, they are just poached by like big company like finance companies because they're getting paid loads of money.
00:42:03:18 - 00:42:35:00
But then at the end, they don't really enjoy the job because it's just crunching numbers, they're not working with robots and not seeing the future. I think is a bit of a disconnect as well with engineering, being able to communicate with people in finance to get funding. And I always said in the past that maybe people in accountancy qualifications should do one module of engineering and people in engineering should do one module in accountancy, and they can find some common language where they can discuss about return on investments, why we need to invest in this, what it will to bring to us because they don't really understand each other
00:42:35:05 - 00:42:52:08
I don’t think and once we saw break that cycle, then that's when it's going to be successful for people to get money from above and invest in the shop floor really. You know, I think it goes from the average business through to venture capitalism and all that sort of stuff, because the ecosystem isn't anywhere near as mature as it needs to be.
00:42:52:10 - 00:43:11:22
I think the other thing is we could do for end use of businesses. Someone selling a great product, you know, like, the only one that is coming to mind is Rolls-Royce but it’s because there's only two countries in the world that make, you know, engines of that type and we need more sort of those businesses, I think because we do have a lot of businesses in the UK supply chain.
00:43:11:24 - 00:43:25:12
But we haven't got so many that are selling the end product. Yes. I think we need more of those. And that's where the real value is add, is when you put something together and you export it all over the world. Yeah, that's where the real sort of like future lies, I think for our manufacturing
00:43:25:16 - 00:43:45:03
I think we should, one thing that we can lead on is like the sustainability aspect of manufacturing because the UK has always been well, in recent years I think we're striving for better sort of sustainable manufacturing and that's one thing that we can lead on and sort of teach the rest of
00:43:45:03 - 00:44:05:10
the world. Because there's a big argument that I hear often that some of, the UK only produces like 1% of the total CO2 emissions, so it's not even worth doing it because if the other countries aren’t doing it, we aren’t going to make a difference. But well it's not true because first of all, given examples in manufacturing,
00:44:05:12 - 00:44:29:11
And we can just lead the way and teach other people how to do it, but I was reading this article the other day and apparently, obviously one third of the sort of gas emissions are produced by China, But I think another one third is small, loads, loads of small countries like the UK that produce 1%. So if everyone has the same sort of attitude, oh it’s only 1% and it’s still one third of the emissions. Nothing will change.
00:44:29:11 - 00:45:04:17
So yeah, I think that's another thing that we should learn. And so when we start looking at robotics, we should think, we should think how we can use robotic system to sort of make it more sustainable. Manufacturing, for example, I've seen lots of research using robots for disassembly. So we've got part of any system, especially electronic system, like just like chucked as a whole, because the company use because we need to sort of rethink the way we create manufacturing, because at the moment manufacturers is not created for disassembly, we are just creating something to be just once and then chucked in the bin.
00:45:04:19 - 00:45:15:20
We need to first think about how we can create for disassembly, both how can we use robotics to sort of inspect a part, taking a part, and then disassemble it, and then reuse it, reuse what we can.
00:45:15:20 - 00:45:23:08
Yeah, you can imagine almost a standard that maybe supports the disassembly, like, like a circular economy type thing.
00:45:23:08 - 00:45:52:02
Yeah, exactly. And almost has like signposting for robotic systems. Disassemble it. Yeah. And then organise all of the constituent parts or whatever. Or BOM (Bills of Materials) of what material is in that sort of specific device and where everything goes. Yeah, yeah. That would be a really good idea actually. Is there anything else on sustainability where you can see it being applied? Um like robotics in general? Yeah robotics in manufacturing.
00:45:52:02 - 00:45:53:05
How could the robots be...
00:45:53:05 - 00:46:09:09
Well robots don't make as much scrap as humans do. They're more repeatable. So it's staggers me when we we, you know, when we go and see companies how some of them run at quite considerable scrap rates and just consider that as parts of their operations. But there are ways of automating, you know, some of the tasks out.
00:46:09:09 - 00:46:27:24
So making, you make much less scrap and now with energy and the cost it is, I mean it’s so terrible to get all the way to the final part of an assembly and then it turns into scrap. You've added value all the way through and time. But then automation and robotics is one way of being able to make products better for less.
00:46:28:01 - 00:46:38:14
And it's true about being like green, using sort of lean manufacturing principles to do away with waste. We are incredibly wasteful in manufacturing and we need to automate out of that really.
00:46:38:14 - 00:46:52:11
Yes, that’s a very good point. I think the standardisation of things and in automation, it does really help with quality issues because there is just less variation in the process. So you'd don’t get as much quality issues. Exactly.
00:46:52:13 - 00:47:17:24
I think there's definitely a bit of a misunderstanding almost about quality and how important it is. When you talk about big volumes of things as well, you produce millions and millions of things. You know, even one in a million starts to get problematic. And you could have a whole marketing issue, you know a brand issue if something really went wrong, you know, the company value of brand image could change very quickly if it's out there and fails.
00:47:18:01 - 00:47:37:11
So you aim for the best quality ever, it kind of solves the problem on the scrap anyway. Yeah. So all companies should steer towards that, shouldn’t they, to to try to adopt the best quality processes, the best manufacturing processes they can so that they don't make any waste. Right, well guys, thanks for coming down. It's been an absolute pleasure to have you here. Thank you. Same yeah, it’s been a pleasure.
00:47:37:13 - 00:47:40:21
Thank you guys for joining us today.