In episode 04 of the Let’s Connect! Podcast, Mikko Niemi of Silicon Labs joins us to talk about Predictive Maintenance, Human-Machine Interfaces, Asset Management, and all things IIoT.
Mikko Niemi is a Senior Product Marketing Manager at Silicon Labs, responsible for Industrial IoT solutions. Prior to joining Silicon Labs in 2019, Mikko worked at Landis+Gyr, a global market leader for smart energy and IoT solutions, for nine years. He held various leadership positions at Landis+Gyr in product management, technical sales and customer project management. Mikko is an inventor in two edge intelligence smart energy patent applications, and he has more than a decade of experience in delivering integrated end-to-end enterprise and IoT solutions. Mikko holds an M.Sc. degree in engineering from the Tampere University of Technology in Tampere, Finland, and an MBA degree from Texas A&M University Commerce in the United States of America.
Interested in connecting with Mikko? Reach out on Linkedin!
Silicon Labs is a leading provider of silicon, software and solutions for a smarter, more connected world. Its award-winning technologies are shaping the future of the Internet of Things, Internet infrastructure, industrial automation, predictive maintenance, consumer and automotive markets. The company’s world-class engineering team creates products focused on performance, energy savings, connectivity and simplicity. Key products include wireless SoCs, modules and MCUs used in various applications across different market verticals in both industrial and consumer applications.
Key Questions and Topics from this Episode:
0:00 Show Introduction
1:00 Mikko Niemi Introduction
3:40 IoT Healthcare and the Miracle on Ice
5:20 The Digital Transformation Revolution
9:15 Asset Tracking and Asset Monitoring
12:45 Supply Chain and Cold Chain
15:00 Predictive Analytics, Maintenance and Monitoring
19:00 Training the Machine Learning Model
22:20 Digital Twin and Edge Networking
27:30 Final Thoughts
- [Ken] This is The IoT For All Media Network. Hello, friends and IoT, welcome to let's connect the newest podcast and The IoT For All Media Network. I am Ken Briodagh, Editorial Director for IoT For All and your host. If you enjoy this episode, please remember to like, subscribe, rate, review and comment on all your favorite podcasts and platforms and to keep up with all the IoT insights you need, visit iotforall.com. Before we get into our episode, the IoT market will surpass $1 trillion in the next few years. Is your business ready to capitalize on this new and growing trend, use leverage is powerful IoT solutions development platform to efficiently create turnkey IoT products that you can white label and resell under your own brand, help your customers increase operational efficiency, improve customer experience or even unlock new revenue streams with IoT. To learn more, go to iotchangeseverything.com, that's iotchangeseverything.com, now let's connect. My guest today is Mikko Niemi, Senior Product Marketing Manager for Silicon Labs and we are gonna talk about all sorts of fun stuff within the industrial IoT and perhaps beyond, Mikko welcome to the show. - [Mikko] Thanks very much Ken. - [Ken] The pleasure is entirely mine, thank you for joining me. In case folks aren't familiar with Mikko or with Silicon Labs, can you give us a little bit of your background and sort of where you all fit into IoT? - [Mikko] Sure, so let's start with Silicon Labs briefly. So we provide Silicon products for the smarter more connected world, really focusing on low energy consumption, simplicity, connectivity. So we provide products for Bluetooth, Bluetooth mesh, ZigBee, Z-wave, Weiss and WiFi. The list goes on and on and on. And then we also have MCUs wireless modules and so forth for the IoT product lines. Myself, I've been kinda like, switching industries for a couple of times throughout my career. So I started my career really at the polar electro which was, or which is the pioneer for fitness trackers. So they invented a wireless heart rate monitors back in the 70s and they'd be in kinda like the market leader or the one of the top companies ever since. And from there, I went to a company called Landis and Gyr which for many doesn't sound like too much but it's actually the world leader in smart meters in the electricity space. So especially in the US if you go to your backyard or wherever your electricity meter is you might find their meter actually there. And there I witnessed kinda like a single industry that utility industries to shift from let's say from the manual reading collections to a fully automated IoT network solutions with integrated solution. So that's sort of like a, well it's a really good I would say platform to grow and understand how the IoT in the industrial space works. What are some of the requirements? And about a year and a half ago, I joined Silicon Labs leading industrial IoT segment at Silicon Labs and really I'm seeing the transformation now in other areas of the industrial space the industry four zero, obviously driving some of the landscape and everything. So I'm really excited about the opportunity to work for Silicon Labs and seeing create great growth in this space as well. - [Ken] So I know I promised that I would try to keep us on task, but first the digression. If you were working on some cool heart rate monitor stuff in the 70s and were you in Northern Europe at the time and more to the point, how involved were you with the Miracle on Ice in 1980? - [Mikko] Yeah so, I have... - [Ken] I think it help us beat the Russians, that's all I wanna know. - [Mikko] No, it's all good, so I am from Finland. So I'm from the country right next to the Russians, and I have to say that, back in the 70s, I was a small baby so I was not inventing heart rate monitors back then but the company was so, but to be fairly honest with you, I think the team that back then has done, pretty groundbreaking innovation by, figuring out how you can transfer the heart rate from those belts to your wrist unit. And then they even had these reading devices where you place basically the wrist unit and then it was like a serial connection I just saw pictures of it and some very old versions of the devices. It was pretty cool to see like how things have evolved from there to the modern watches, where you have WiFi and sometimes even say little connectivity. So it's pretty awesome. - [Ken] That's really, really cool. But to the point and I wanna start really really high level before we dig into some of the technologies that we're gonna talk about. I have been, sort of bandying about the idea of a new term for IoT and more specifically I think industrial IoT, where a lot of people have been talking about digital transformation lately. And I think that there's an argument to be made that, digital transformation, that there is a digital transformation revolution coming, that is going to include IoT technology and industrial IoT implementations, but will also include lots of other, technological innovations also that involve being interconnected, that it will involve AI and machine learning things, that will involve connectivity and sometimes probably 5G and other times you NB-IoT or all ZigBee, Z-Wave all of the various things where appropriate lots and lots of LoRaWAN probably. But I sort of think that were on the, maybe not near the cusp but we're approaching, needing a more umbrella term to talk about this sort of ecosphere. And I wanna get your opinion, feel free to tell me I'm wrong but I'd love to hear your thoughts about the idea that we're currently working in IoT industrial IoT industry 4.0, but I think that we're actually a subset of this larger digital transformation revolution that is really becoming transformative of everything that everybody is working on. So I'm curious about your opinions on all of that. - [Mikko] Yeah, it's a wide topic so thanks for bringing it up Ken. So, I see that I kinda like share your vision of the bigger digital transformation the whole I think the digital is the key word in all that we do. IoT, the connectivity is just an enabler of many of those things, but it all boils down to the transforming many of the business processes that are underlining the transformation. So supply chain is a great example like how that is changing from just transporting stuff from one place to another and then the other one, opens up the doors of the truck and sees like what's inside and some things get lost or whatnot and all that. And at the same time, reading how the manufacturers, how they want to approach kinda like the batch of one, like, you order a device or a product which is not even get produced and it gets mass customized for your needs and you can basically produce just a single product from the production line instead of pushing through like millions of the same product out. So I think there is like all kinds of unique things that are going on and how these manufacturers of these different types of products have to change or are changing their processes and businesses to be more, let's say customer oriented and more friendly in that way. So I think that's an interesting time what we're seeing right now and to some extent, maybe the whole pandemic has actually accelerated that development because of needs for remote monitoring and those kinds of things that may have been neglected in the past, because why should we do it? We have the people on the shop floor but nowadays may not be the case. So it's all great, it's really exciting times in the industrial crisis. - [Ken] I think you're right and I couldn't agree more about the last 10 months or a year really forcing acceleration, although I'm not entirely sure that isn't my sort of innate optimism looking for a silver lining to the cloud of 2020, but we'll see I think, and I think that you'll turn out to be right that this whole experience sort of forced everyone to work closer to innovation. But let's move into some of the specifics, that we wanted to touch on and I wanna start with asset monitoring. That is a talk about a broad topic, there's just a million different well assets to pay attention to and if you wanna get granular, I mean, you can say that that goes straight down to software code and like network monitoring, that's all data is asset, I think. So I think there's an argument to be made that this is goes beyond the physical and into the digital, like everything else does. So let's narrow our focus a little bit. What do you mean when you say asset monitoring? What is sort of the crux of that for you? - [Mikko] Yeah, so I would maybe narrow it down to few topics. One really interesting thing is like in the logistics space, like if you think that the United Nations, Food and Agriculture Organization, they've estimated that 30% of the food is going to waste in the supply chain. That's horrible that happens because, there are people suffering from hunger and everything in the world. So, that kinda like the whole connected logistics I think, it's a really excellent application for asset monitoring. You're looking at the refrigerated containers and trucks and everything, making sure that once you are moving the goods from one place to another they stay in the right condition. Also in the logistics space detecting, if someone tries to tamper with the containers and stuff having sensors there and it's going to be an interplay between different technologies. So if you think that you would have like a container full of sailor connected sensors, how do you get the signal out from a container it's like a Faraday cage. So you're likely more having like a local network using some of the technologies like Bluetooth Low Energy or Bluetooth mesh or things like that. And then you have the Backhaul technologies like NB-IoT that you mentioned that will take the data back to the central office. Or some of our customers are even working kinda like the local network within the truck itself. So they will have like WiFi connectivity between the cargo space and the cockpit of the truck where you can see that, Hey the temperatures are rising so as there's something going on? So maybe sometimes the driver can actually find a place to fix the issue while he's driving it instead of like plainly driving the truck all the way to the destination to find out that Oh, I was driving like a rotten carrots now. So, that whole logistics space is super interesting, and I think Silicon Labs has really good product line to provide solution in that space. - [Ken] Well, I think you're right, I think that it is not just interesting, but critical. I'd like to think that the driver would pull over before he tried to fix the issue, but the nope that's exactly the response my joke deserved, that's fine. The really sort of salient point I think is one that we've realized recently the refrigeration point that you mentioned the cold supply chain is in desperate need of this kind of asset tracking and monitoring. And I think that they are starting to realize that now not just because there's a COVID-19 vaccine that requires extreme refrigeration for storage and transport, but that has very narrow tolerances but also because so much more stress has been put on the supply chain over the last nine or 10 months. I think that the monitoring and the paying attention of the state of those packages in transit is not just critical to the survival of the asset in the package, but also to customer service to satisfaction and business growth, to all aspects of the logistics line and company. - [Mikko] Yeah, absolutely. I think you're hitting the nail on the head with that common Ken so, I do agree that the cold chain is an area where you can gain a good payback for your investments? And in fact, in the food side of the cold chain monitoring there are increasing regulations. So there was a few years ago, this US regulation came into place, the food and food safety, something like that, Food Safety Act was coming in and the same kinda like goals for the European union region as well and it's likely then expanding to some other regions as well. So that's partly driving that one as well. One technical aspect to that conversation is then okay, so now we have all the data, how do we secure that? So cyber security becomes like a huge thing and a very important thing is to just to make sure that though the integrity of the reads can be kept because imagine that someone tempers the readings, they can fool the system so you want to make sure that the products that you deploy are highly secure and can face the challenges coming from the attacks. - [Ken] Yeah, I agree. And that actually transitions us beautifully into the next topic we're gonna talk about which is sort of the ultimate expression of execution on data, which is predictive analytics and predictive maintenance and monitoring, within not just within the industrial space, but sort of everywhere where you take your sensor data, your data collect your historical record and sort of train and run an algorithm to predict when there will be faults and avoid downtime, predict when there will be a need for repair or even just regular replacement of the batteries or whatever. So this is a concept that I think people are becoming very familiar with, but the actual Applications in and execution is still seems very rare. I can't off the top of my head think of a truly predictive maintenance system that I know of offhand, now that doesn't mean one doesn't exist it just means I don't know about it. But to one that isn't just programmed with a schedule of, they talk to the expert guy who works on the line and he says, "Yeah, we do this every three weeks." I want actual predictive machine intelligence to say it's actually you need to do this, two weeks in three days from now, even though you did three weeks last time. - [Mikko] Yeah, so that's an excellent segue to another excellent topic of industrial IoT, the predictive maintenance. And just for the listeners, you're basically trying to avoid unplanned outages because that's usually the most costly thing is the unplanned outage of any kind of industrial process. And why it is so costly is because it happens, all of a sudden and then you have to rush, like did you have the spare parts in stock? Did you have the people that can do the repairs in place? And all those kinds of things and things start to get crazy. So, but with predictive maintenance you're trying to predict that, Hey this thing is going to fail in the next three months or so. So, Hey I can place the order for the new bearing, I can go ahead and schedule the maintenance guy, I can get everything this lined up before the whole thing breaks down, plus I can plan ahead of time so when I have to maybe bring the whole process down for the maintenance I can plan to do something else as well, and get kinda like the biggest bang for the buck when you actually have to shut down the processes. So that's what you try to do. And growingly, people have tried to solve this problem by, Hey, I'll put these vibration sensors into my motors and pumps and everything and I'll send all the data back to the cloud. So interestingly, it's a great idea because the cloud has to basically unlimited processing power and storage for the data. So, and you can train as sophisticated AI machine learning models and everything. So now, if you think that, okay that's only one application trying to compete of the airspace, if you will and there are other radio devices working at the same frequency band sending stuff that someone else wants for their processes. So now you have this aerospace competition and we've heard from some of our customers that they start to see that there could be some congestion at the network level because you have so many sensors trying to send the whole data back to the cloud. Well in predictive maintenance cases I would say 99% of the data is pretty much redundant. It's basically telling the cloud algorithm that everything is just fine. So you're basically overusing the network for just passing the role data, which leads to the growth of the embedded AI machine learning. So actually train the model and you bring it all the way to the end sensor node and then you train the end sensor node with the machine learning algorithm and the local conditions, and now it can actually start looking at the data that is being measured from the vibration or something else, and predicting from there that, Hey this actually provides the signature that is an anomaly I need to alarm, but for the most part it doesn't change or send anything out. And what we've seen and then some of the other folks in the industry have seen and witnessed is these can actually save your battery life. So it's pretty and intuitive that, Hey I'm doing more calculations at the edge, but if you think that the other alternative is okay, I turned my radio on and the radio is typically the most power consuming part of the whole sensor. - [Ken] Sure. - [Mikko] So you're basically reducing the data transmission which give you some, let's say capacity to run the algorithm at the edge and only send the anomalies out. So really that is an area that I think will grow dramatically in the next couple of years. When people start to understand that, Hey I can run a TensorFlow model in our 32-bit MCUs. You don't need like a GPU processor to do that. You can do it with this low power MCUs that are used in these sensors. So I think that's really interesting area of development that we're seeing and I'm believing it quite big time that it will pick up speed in the next few years. - [Ken] I think the low power consumption part of the equation is really sort of the silver bullet of the edge being effective. I think that we need this sort of almost nano cloud of edge devices and sensors and whatnot that are doing processing and not requiring frequent maintenance themselves because now you've created a whole another layer of need for maintenance so let's make sure that those guys are pretty much self-sufficient. The other aspects and since you brought up the edge and the importance of this sort of predictive processing happening there, the other piece of that equation that I'm always fascinated by and really excited by is the digital twin and how you can do modeling and experimentation and sort of learning outside the system based on those same historical models in the cloud using all that processing power, and then just send your outcomes back down to the edge, to change behavior and alter the actual working algorithm. That seems like just such a milestone in functionality. And I don't know how much you've worked with digital twin at Silicon Labs or elsewhere but I'd love to hear your thoughts about it. - [Mikko] Yeah, I think the digital twin is also a very interesting development, and I think it basically falls under the umbrella that we discussed early on which was the digitalization of the whole industry. And basically that is showing you details of the device as it, is aging and you can like you said change the settings and all that. I think that provides such interesting amounts of data to the engineers who are working with those processes and everything that it will likely be growing dramatically in the future. And what you need for that one is of course sensors and data because just providing runtime data and all those kind of things it's just not enough for creating like the full picture. So you're likely going to be integrating some of these Applications together like, digital twin, predictive maintenance. So, and all those kinds of things are likely somehow summing up to a bigger picture of the whole industrial processes and systems. - [Ken] Sure and I think that the larger and more interconnected your digital twin gets the more like the entire IoT system it looks, the more you can learn from it. And the more you can find those sort of unintended consequences that can be either ways to avert disaster, like downtime and other problems or product loss, or anything else or ways to discover new unexpected profit centers or opportunities or efficiencies. I think that kind of intelligence can go to the advantage of the company looking for it in a lot of ways. And there's not a lot of case against it except for scale and like ability to make it happen. - [Mikko] Yeah, no absolutely and I think the new business models, all kinds of asset service type of business models are definitely going to need the data. And like I think was it Rolls-Royce or whoever invented kinda like the airplane motor uptime kinda like they were selling just the uptime hours for the airlines you didn't pay anything for the motor or the engine. So I think the same to some extent can happen in the industrial space that you're buying like a production line as a service and then you're just paying like a monthly fee for it. But it's long ways ahead, but I think something like that could happen eventually in the future. - [Ken] I think that was GE actually, but they may make the engines that Rolls-Royce sales. I'm not sure, but I'm pretty sure GE was doing that uptime sales also. And my only concern with that is sort of the subscription fatigue problem that we haven't really hit yet but everybody's starting to talk about when I start talking about the, as a service model is folks are worried about, somebody in the some CFO somewhere going what do you mean we're paying 35 different monthly fees for all this and we don't own anything. And so, I don't have a solution for that problem, but I do see it coming and I'm thinking that the answer to that is going to be what IoT already does and is good at which is the sort of partnership economy and bundling a solution into one profit share solution model on the business development side. I think that's a strategy to address the subscription fatigue problem before we reach it. - [Mikko] Yeah, and I think as long as you can really solve a problem for your customer then the subscription fee is likely more successful model for you. Like in my past with my previous employer Landis and Gyr we had what we called the managed services customers. And basically the customer was just buying the meter reads. So the product that was sold to them was the meter reading file. And we had people who were taking care of the smart metering networks and everything along those lines. So that was really interesting way to do it because, if the customer which in this case was a utility didn't do it then they would have had to hire people who knows IT who knows how to run those systems and everything. So now you're actually solving a new problem for them that the digitalization is bringing. So as long as you keep that in mind, I think you have a chance to be successful, but if you try to nickel and dime more money out of your existing customer base just by turning something into a subscription-based you might not be successful at the end. - [Ken] Those very politically said, well done. Unfortunately, we're getting near the end of our time Mikko. And, I know we haven't even hardly scratched the surface of this topic. So maybe, I'll have to have you back on at some point later in the year and we'll talk about how things are going, but for now I wanna give you sort of the last word. And if there's something sort of that you really wanna make sure the listeners go home with, this is your opportunity to leave it with them. - [Mikko] Sure, so I think I would highly encourage, the listeners to visit our website sitelapse.com and start looking at kinda like we have we are going to be having a lot of good, there's already good content out there on our website but we will be definitely having more interesting stuff coming up this year, it's going to be a big year for us, and all that. So I really do think that the industrial IoT is now picking up the pace and it's going to outpace to some extent I think the smart home space, because there's just so many opportunities to develop this. And in addition to the asset monitoring there is the human machine interfaces that we didn't have time to discuss but it's also something that is doing a transformation, and we held a webinar last year on that one as well. So I really, I'm excited about this opportunity Ken to speak with you. I'm happy to come back and talk about some other topics later in the year that would be awesome. - [Ken] Yeah. - [Mikko] And anyone who kinda like wants to personally connect with me, you can find me on LinkedIn and send via connect request so I can definitely continue the conversation with you guys privately there, so. - [Ken] Yeah and we'll put those links into the show notes for you folks out there listening. And yeah I definitely wanna talk to you again about the human machine interface stuff because I personally welcome our Cyborgs, Overlords and I think everyone else does too. So Mikko Niemi a Senior Product Manager of Silicon Labs. Thank you so much for being my guest has been really a pleasure to have you. And thanks for talking with me today. - [Mikko] Thanks very much Ken. - [Ken] Thanks again to all of you listening out there. I hope you've enjoyed our discussion, and if you have please make sure you like this and subscribe so you don't miss out on any of our episodes, we post every week and I hope you'll leave us a rating review and comment and your favorite podcast platform. If you would like to be our guest please put down the link in the description. And we also have a great sister podcast network called the Island for all podcasts so make sure you check that out. - [Ryan] Hey, Ken let me jump in real quick and introduce your audience to another awesome show on The IoT For All Media Network. The show that started all the IoT for all podcasts where I bring on experts for around the world to showcase successful digital transformation across industries. We talk about Applications in IoT solutions available on the market and provide an opportunity for those companies to share a device to help the world better understand and adopt IoT. So if you're out there listening and haven't checked it out be sure to go check out the IoT For All podcast available everywhere. - [Ken] Thank you, Ryan, now get back to your show. And thank you all for joining us on this episode of let's connect. I've been Ken Briodagh Editorial Director at IoT for All, and your host. Our music is Sneaking On September by Otis McDonald. And this has been a production of the IoT For All Media Network. Take care of you. You are listening to The IoT For All Media Network.