EDGE AI POD
Discover the cutting-edge world of energy-efficient machine learning, edge AI, hardware accelerators, software algorithms, and real-world use cases with this podcast feed from all things in the world's largest EDGE AI community.
These are shows like EDGE AI Talks, EDGE AI Blueprints as well as EDGE AI FOUNDATION event talks on a range of research, product and business topics.
Join us to stay informed and inspired!
EDGE AI POD
How Weeteq is Revolutionizing Microcontroller Technology for Industrial Automation
What happens when you push AI to the absolute edge of networks? Meet Weeteq, a Scottish startup revolutionizing system optimization by bringing intelligence directly into microcontroller control loops.
In this fascinating conversation with Martin from Weeteq, we explore the concept of "Ultra Edge" - technology that operates beyond what we typically consider Edge AI. Unlike conventional approaches that process data at the device level, WeTech's innovation works within the microcontroller itself, detecting and responding to error signals in less than one-tenth of a millisecond. This isn't just incremental improvement; it's challenging fundamental assumptions about what's possible at the lowest levels of our technology stack.
The implications are profound across multiple industries. For industrial automation, WeTech's technology optimizes motor performance without requiring expensive hardware replacements. In electric vehicles, it extends range by reducing energy consumption and minimizing stress on traction motors. Even safety-critical systems in autonomous vehicles benefit from their comprehensive monitoring capabilities. All this comes from a lightweight implementation that works with existing infrastructure - no ripping out and replacing required.
Martin also shares insights into Scotland's vibrant technology ecosystem, tracing his journey from Honeywell to startup founder and highlighting how universities like Strathclyde are fueling innovation in everything from robotics to renewable energy. It's a reminder that groundbreaking technological advances aren't limited to Silicon Valley or other well-known tech hubs.
Ready to discover how ultra-edge computing might transform your industry? Listen now and glimpse the future of real-time system optimization.
Learn more about the EDGE AI FOUNDATION - edgeaifoundation.org
Well, awesome, it's good to be here, martin. Thanks for joining us here on our Partner Edge AI session and, as I mentioned before we started rolling. You know, this is really about understanding the diversity of partners that we have in the foundation from big tech to startups and it's great to have you here from WeTech to kind of talk a little about what you're up to and what's happening out there for you in the Edge AI space. And so thanks for joining us. You're very welcome. It's good to be here, and you were on our Career Edge live stream recently, which was pretty cool. So that was like we do this thing where and I think there's another one coming up in May where we do kind of a live stream for how to build a career in the Edge AI world, and you were representing the kind of the how do you get into the startup biz? Yeah, pretty cool.
Speaker 2:No, it was good it was. I think that was my first live broadcast during the career edge piece. It's been a week of firsts. I had my first panel with Arm yesterday down in London.
Speaker 1:Silicon Catalyst event down there. So here here we are, my second live broadcast. So it sounds like, from the clues that you're giving, that you're in the uk and somewhere in the uk, is that?
Speaker 2:yeah, so we're based in scotland, uh, we're in glasgow, uh, the office is in glasgow. So, um, yeah, we've, uh, we've been going for two or three years. So, as I said, we're only a startup, but I think the fact that we're getting a lot of interest from big industry names like ARM and starting to get revenue as well this year proves that we're moving in the right direction, pete. So, yeah, we're a startup with big ambitions.
Speaker 1:So give us a background on WeTech. What's the pitch, what's the elevator pitch on WeTech?
Speaker 2:So, in a nutshell, we do power and control system optimization and it's taken us a long time to get to the point where we kind of, you know, just consolidate what we do.
Speaker 2:But this is what we do, but we do it in a completely unique way. It's a completely unique way because we're talking about a shift from system optimization being the domain of cloud, local and even edge compute to one way we're doing system optimization at the microcontroller level. So we call our underlying technology Ultra Edge, as it's really, as we see it, the next evolution of Edge AI. Edge AI is really current and really doing some amazing things. Our technology is slightly in a different shape in terms of we're at the microcontroller, within the controller, so that's where all of that action happens, and then we make use of that data for all applications of the chain, whether that's Edge AI, edge Cloud, you know wherever really. So we're really at that kind of like furthest extreme of the network.
Speaker 1:And so what kind of what are your products or deliverables, or how would you describe those?
Speaker 2:Yeah, so really we're very much a microcontroller level technology. So that's um from an ip perspective, that's creating our own microcontroller, our own companion ic. But really the target applications for ourselves so far has been motor drive applications, power inverter applications, so very much focused on industrial automation, automotive From a power inverter perspective, getting into data centers and group utilities kind of space. You can imagine that even with the amount of investment and development that there's been in, you know, edge, ai, but just really in terms of industrial automation, there's still inherent inefficiencies in power electronics because, you know, technology has not really been there to solve all the problems. So you know, we've got these inherent inefficiencies that we just assume are inevitable and UltraEdge is here to tell everybody that it's not inevitable, because if we can do clever things like not only understanding the data, the operational data that exists within the control loops of microcontrollers.
Speaker 2:What if we could change the output of those control loops and do some really clever things, um, such as using ai, a ultra, um, you know far reaches of the network to take an input and change the output of the controller in real time, you know? And then what can we do with that?
Speaker 2:so it's all about, um, getting rid of error signals that we don't need you know really, getting really smart on energy efficiency, on productivity and also not putting stress on motors, and that's really kind of what we're about. It's like the typical three things, you know saving energy, increasing productivity and minimizing downtime and really that's kind of a focus for Edge AI applications. But we're really at the microcontroller control loop level. So just taking it to the next stage, so are you fabbing?
Speaker 1:basically, you're designing chips to be fabbed and you're also licensing IP, or do you do both?
Speaker 2:Yeah, I mean, we're a very, very small team, so the the motivation to um, you know doing any kind of fab work ourselves is pretty low, but um, we, uh, we're not writing off that, you know, in the future.
Speaker 2:We've got some pretty good partnerships now in place. That would be um good partners if we did go down that approach. But really, what we're about in the short term is number one, validating the technology. So we have a number of pilot projects which we're completing at the moment and we're looking for more. And then it's really about how does the customer want to use our technology? That's very much an IP play and deploying our you know, our algorithm, our innovation and our models in our existing chips. And absolutely fine, if they want to kind of keep the status quo and have that level of resiliency, then we can apply our technology to supplementary chips and eventually our own chips. So whether that's complete IP outsource of the microcontroller to somebody else or we start, as you say, fabricating our own microcontrollers sometime in the future, that's something that we'll definitely look at.
Speaker 1:Yeah, so do you have kind of your own ML models to kind of optimize power and things like that? Is that kind of how it works.
Speaker 2:Yeah, we take inputs from the microcontroller control loop things like torque and speed and current, et cetera and we apply various ML models created with TensorFlow Lite and various other different applications to really work on those models. And we talk about this idea of our ability to detect and respond to error signals in less than one tenth of a millisecond, which in our opinion, is real time. We talk a lot about edge being real time, but we're still observing what's going on when we're in the control loop. There's no more real time than that. So if we can do this and we can start to respond to these error signals in less than one tenth of a millisecond, this is really playing real time. And what we're able to do is be able to apply these models in a really really lightweight way to be able to still achieve that detect and respond timescales as in one tenth of a millisecond, plus applying a model and applying our algorithm using systems or processes like TensorFlow Lite in a really compact and light such way.
Speaker 1:Cool. How did you find Edge AI Foundation or you know? Did you guys get involved, like how did that come about?
Speaker 2:Well, I guess we found TinyML first.
Speaker 2:You know, as in you know, we've been looking, we've been closely monitoring, I mean, what's going on with TinyML for a number of years now. I guess when we started off, we were looking at TinyML as being a really good collection of organizations that are doing really exciting things that we wanted to be involved in. As we've been growing the products over the last sort of two years, I think we've reached a level of maturity that it was the right time for us to sort of reach out and make contact. And we were involved in a webinar I think it was not not christmas gone, but the christmas before, in fact where, uh, we shared a bit of our technology.
Speaker 2:But again, I kind of look back on those uh presentations and, uh, you know, often go through the old pitch decks from even from three months ago and I think, yeah, back a year ago it's, yeah, it's a different message to the one that we have now. So I think now's the right time to get really fully involved in AJI foundation. Obviously, we're this side of the pond so it's not possible for us to attend. You know everything that you guys are doing, but things like the Milan event coming up is perfect for us and you know it feels like the right time to get involved.
Speaker 1:Yeah, there's a. I mean, you know, one of the cool things I like about the foundation is we're a very worldwide community. Like we're building community in Taiwan and Asia, europe especially a very hot area, especially with the Milan event coming up and we just had the one in Austin, you know, then we're actually going to be planning on, in 2026, getting into India. Actually, we'll be in India in December of this year. It's kind of a huge community out there as well. So, yeah, it's pretty cool to see the diversity of folks. Like I always talk about the stack, you know, from metal to cloud, when you want to do edge AI, there's lots of it's a, it's a team sport, right, and so you know you guys are in there in the slice. They're pretty close to the metal, um, doing your thing, which is you know then, and there's all this, you know apps and models and devices and things on top of that, all the way to the cloud. So it's pretty cool that you guys are involved like that. How long have you been involved with WeTech?
Speaker 2:Yeah, so WeTech's been in existence really for kind of two to three years. Before that. The previous incarnation was called 5G3i, which was a different technology, different model. It was more connectivity solutions. But yeah, my background really is I was at Honeywell for quite some time 13 years with Big Red H and left Honeywell to do something a little bit more closer to home. Plenty well, to do something a little bit more closer to home. Um, so looking for some kind of like startup position, uh, here in scotland, because I always had kind of european, you know, roles doing a lot of travel and um, there didn't seem to be so much so I ended up working for dialogue semiconductor for a while, which was good, but that was during lockdown and then I got an opportunity to work for a startup accelerator in Glasgow called Stack, which is
Speaker 2:the Smart Things Accelerator Center and one of the first companies through the door was 5G3i. So my now co-founder Tanner is ex-Dialogue Semiconductor himself, has had startups himself and really the technology really resonated with me. So, long story short, having supported him in a mentorship role for quite some time, I took the leap to really take a step into founder territory myself, exciting what a big step it was.
Speaker 1:Yeah, that's cool, as I understand Scotland. I mean there's quite a bit of tech happening in Scotland. There's a lot of space tech, there's a lot of ocean tech, robotic tech. You mentioned connectivity. I mean it's a pretty, pretty, pretty hotspot these days for a lot of startups, right.
Speaker 2:Yeah, I mean you know, even going back, I mean everyone talks about, you know this rich technology that Scotland has developed. You know the TV, you know the telephone, you know all this kind of crazy stuff, the, you know the Watt steam engine and all this kind of great stuff. So there's such a rich heritage in Scotland, in Scotland in general, in innovation. And then you throw into the mix the universities that we have here, which are extraordinary. You know Edinburgh, glasgow, st Andrew, um, you know st andrews. They all get the, they all get the kind of the, the spotlight, but, um, here in glasgow we have strathclyde university that are doing great things as well, um, and so around that you have this great um, startup or technology kind of ecosystem which, unless you kind of stop and have a look around, you don't necessarily see.
Speaker 2:I was always in tech when I was at Honeywell, but I didn't really know what was on my doorstep. So getting into it and taking a step in and seeing what's here is phenomenal. As you say, a ton of renewables, robotics. We've got the robatorium in the Heriot-Watt University down the road, so it's a really good place to be. Usual kind of challenges around attracting funding and trying to get folks from maybe even from London, you know up to Scotland to take a look at what we're doing. But yeah, it's very much a vibrant sort of startup. Yeah, that's cool.
Speaker 1:I know there's a lot of hydropower stuff there going on too.
Speaker 2:Yeah, yeah, a lot of natural resources from Scotland.
Speaker 1:Like, the geographic positioning of Scotland gives it some unique advantages for like hydropower, for data centers and things like too, which is pretty cool yeah, and as you said this, this space industry is huge.
Speaker 2:So you go up to the far north of scotland. There's not a huge amount um in terms of built areas around there, so they're building space ports and things like that yeah absolutely awesome.
Speaker 1:So what else is happening? Like you know, you mentioned that the industries that you're selling into so you know, automotive is a big one. I mean, what are your hot verticals these days for for?
Speaker 2:your stuff. So definitely industrial automation, so anywhere where we have kind of nonlinear control and this is the whole kind of premise behind what we're doing is, you know, pid control and things like this. There's been technologies that have been used for 100 years and I guess it's one of those. It ain't broke. Why fix it? But now we've got good reasons to fix it, which is sure that we have uh, you know, advances in uh ai, uh models etc. That allows to do things differently, but um, so, uh, linear control has worked really well. But um, in in the modern world that we're in this ton of applications now which are nonlinear, and so if you think about things like even in a factory where you have a conveyor with a different load we were speaking to a tire company this afternoon, that is- mixed in rubber and all this kind of stuff.
Speaker 2:So, you know, nonlinear loads, we have this ability to, you know, use existing equipment. So it's a very light touch, non-disruptive technology which just enhances things. Rather than you know, it's a complete rebuild of what's going on. So in industrial automation in factories, that's big. In automotive it's also big because we it's easy to just think about the traction motor in an electric vehicle, for example and a real good example of a non-linear control system.
Speaker 2:So every time you put your foot on the accelerator and depending on the terrain that you're on, you know you have a response and what we were able to do is extend range in those vehicles because we're, you know, decreasing energy consumption but also not putting stress of these kind of heat current overshoots within electric motors and traction motors. So that's the kind of obvious ones. But then even in safety systems, so certainly in electric vehicles and when we're going through to autonomous vehicles where you have motor driven safety systems like steering and braking, now we're not advocating doing any kind of takeover of the controller but with the technology at the microcontroller level, giving 100% coverage of stress conditions, certain coverage of stress conditions. This very much plays into a SLD which is the IOS ISO 26262 requirements that you need for all of these steer by wire, break by wire type systems, which is going to facilitate autonomous driving. So tons of applications. Our challenge is Trying to be laser focused at the moment of tracking one or two of these.
Speaker 1:Yeah.
Speaker 1:And then, building on from that, Well, it makes sense to go after a lot of these what we call brownfield deployments. I was just talking to someone about that. It's like in the real world out there, people have invested CapEx and a lot of equipment and process and you know they want to add, you know, intelligence, they want to add AI into your process to optimize it, to save money, save power, like you said. Um, it all adds up to a lot of cost savings and efficiency, and efficiency is the new currency. Uh, so it sounds like you guys are in the right place yeah, hopefully, and they don't.
Speaker 2:They don't want to rip it all up and start again.
Speaker 2:I mean that's the thing you know ours you can start with a single motor. You can start with a pilot project, have it work and then really quite quickly put it into your R&D line and then quite quickly try it out on an active working line. It's more about having the ability to be able to get into the systems, to be able to deploy. That's the thing that, that that we find, um, uh, I guess the challenge and that's the the kind of pilot projects that we're looking to do is for those, um, those innovators really, that you know, get into the pcb level, uh, you know, uh, tinkering and uh, uh, being able to experiment. That's really where we're at.
Speaker 1:Yeah, that sounds fantastic. Wow, it's cool. I feel like I need to make a trip up to Scotland at some point too, because there's so much going on.
Speaker 2:You'll be more than welcome and we'll definitely show you the sides, Pete.
Speaker 1:Yeah, I'm not a golfer, but I heard the golfing is pretty good up there too.
Speaker 2:Well, you can always spend some time in the 19th hole, as we say, which is ah yes, there you go. Yeah, but after.
Speaker 1:After the round well, martin, thanks for the background on we tech. Uh, appreciate kind of digging in a little bit for our listeners here to learn more about what you're doing and uh, yeah, I just appreciate your engagement in the foundation and and what you guys have been contributing, so appreciate that I appreciate it.
Speaker 2:Pete, Thanks very much for the invite.
Speaker 1:All right, sounds good Thanks.