(calm bright music) - We're able to introduce a pretty robust roadmap. We're working on the next generation, adding vision capabilities, adding more performance, adding more interfaces, and this is something we are looking into for the next couple of years. (calm bright music) - Hi, folks. I'm Fabian and you are listening to the "Podcast4Engineers," the podcast you just have to listen to if you're interested in what's going on in the semiconductor market.
Today, we have a very special guest, Omar, Omar Cruz, and he will enlighten us with regards to the developments we see in microcontrollers. Omar, why don't you introduce yourself, please? - Thank you, Fabian. I'm Omar Cruz, I'm the product manager for PSoC Edge inside Infineon, a line of microcontrollers within the PSoC portfolio. Happy to be here. - Well, great to have you here, Omar. So, not everybody might know what PSoC is.
Maybe you can give us some idea of what this microcontroller is all about. - Definitely, definitely. We are a pretty successful brand of microcontrollers. It's quite known, we've been here for more than 20 years.
We are characterized for low power microcontroller in the consumer and IoT space, and we have a pretty compelling roadmap now that covers AI, artificial intelligence, we are focusing also on security, we're focusing on integrated connectivity, and, you know, we are excited that we keep coming up with new and new stuff. - So am I, and very good to have you here to discuss all these. Well, if you're looking into the industry, there's obviously plenty of microcontrollers.
Where would you say the PSoC is different or where are the differentiators? - Yep, definitely. From the inception, we've been always focusing on the reduction of bill of material costs, and the way we do this is through integration, right? You take a look at the traditional microcontroller. It's gonna have its processor, it's gonna have its memory, it's gonna have its, you know, buses and interfaces. But what we do is we have more. We have programmability.
We have customization inside our microcontrollers. We have analog modules that, you know, are easy to interface between an analog sensor and what happens in a microcontroller. We have our very known CAPSENSE™ technology for capacitive touch. So that's something that, you know, people know about. We always focus on ultra-low power capabilities, battery-operated applications. We have also integrated connectivity, whether we're talking about BLE or this, you know, mix of BLE plus Wi-Fi combos.
And all of that, that's a lot of differentiation. We have a best-in-class HMI and this is something that we're known for and will continue to be known for. - Well, where do you put all these advantages? What kind of applications does a PSoC™ go into then? - Yep. So we have been quite famous in the smart home and consumer IoT. You know, you look into the home appliances, the odds are very high that if it has a touch interface, most likely it's gonna have a PSoC microcontroller in it.
So if you think about smart home appliances, we are there, HMI differentiation. If you think about also the medical devices, we have ultra-low power capabilities. Something that is also really important is the reliability of our microcontrollers. - Obviously, yes, yeah. - Exactly. So we have that set with PSoC family. We also have the capability to support automotive infotainment. Similarly to the home appliances, right?
If in an automotive you have touch, most likely it's gonna have a PSoC in it. So, you know, across the space, we're well wide known and we are happy to, you know, have this presence in the market. - Obviously, yes. Okay, so we've talking about the present. Let's have a look into the future. We see the world is evolving. Everybody's talking about IoT, well, basically also artificial intelligence. So where's the next step for microcontrollers like PSoC?
- Yeah. And let me tell you a little bit about the inception of PSoC. From the inception and from the beginning, we started as a peripheral microcontroller, right? If you added or the customer needed some sort of capacitive touch sense capability, we will integrate cap touch. We will have compute capabilities as well with PSoC 4. If they needed more compute power, we introduced the PSoC 6 with even more connectivity, right? And now our customers, they need more.
We are seeing a trend that they need more performance and more computation capabilities at the edge of the devices. - Okay. - We are seeing the status quo, what we call, that most of the applications today, when we're talking about artificial intelligence, they're made at the cloud. - Yep. - They're cloud-based. You know, you have your end device, it sends the information, it sends the data to the cloud. All the performing and all the processing is gonna be done on the cloud and then taken back.
Now we're seeing the trend that everything is gonna be edge AI, people who are talking about AI and machine learning. So this is something that we're looking at. Of course, this will bring benefits in terms of cloud latency, right? Like you have a natural latency when you talk to your device. It goes to the cloud, it goes back. You have latency, right? So we are seeing benefits on that, having at the edge, you know, being done, performed at the edge.
We have security concerns also addressed with this. You always have that data coming back and forth, right? The expectation is to be doing all this processing at the edge, you don't have this transportation of data happening. You know, you might have sensitive data, you know, going back and forth through the edge, so we're gonna avoid that. We are gonna improve energy efficiency with this approach. And we are also planning to have the computation capabilities at the edge.
So there's a lot of items that will allow us to address those trends and then looking to the next generation of microcontrollers. - Okay, so obviously understood, we can see a development where devices really get artificial intelligence in themselves, yeah? So you are introducing now PSoC Edge. How is this new device actually addressing these challenges? - Yep, yeah, that's a pretty good question. So we have different types of challenges, right?
We have the compute performance needed at the edge in order to address these challenges. So we have a new, also new type of operations. We have machine learning operations that are different from the general purpose type of operations that we have seen in the past, right? So now one of those drivers or one of those requirements that we see in the field have to do with how to address this type of machine learning operations.
So you're gonna require, of course, a general-purpose processor, which it's being done by most of the microcontrollers, but you also need the hardware acceleration for AI, machine learning type of operations. So that's also needed. The other aspect is like we are still talking about battery-operated applications, right? You still need low power. - Obviously, yes. - I mean, it's not like you're gonna be exchanging batteries every six hours, right? - Nope. - That's not optimal.
- That's what the consumer doesn't want to. He wants or she wants to have a device lasting at least four days. - Yeah, exactly, at the very least. The other challenge is security, right? How do you address that security? Of course, you have benefits by not transporting data back and forth between the device and the cloud. That's addressed. But at the same time, you need some sort of tamper protection. You need, you know, make sure that your information stays secure and private. - Yes.
- And that's something that we are also planning to address with PSoC Edge. We talked about integration, and that's the DNA of PSoC, right? A lot of integration. And you talk about the integration of different interfaces. You have now voice, you have a vision, you have sensors, different type of sensors, and all of that to be integrated in one microcontroller is key. On top of that, you need more memory, right?
The applications are becoming more complex, so you need more memory on your device as well, and this is also what we are planning to address with our PSoC Edge family. - So, Omar, so far we've talked a lot about the hardware side of this new microcontroller. Obviously the hardware is nothing without software. So where does the PSoC Edge support software topics? - Yeah, that's a great question.
And, you know, you can come up with that groundbreaking hardware, but if it's not addressing the software, it's not that useful, right? So something that we have done at PSoC Edge, I'm quite happy to have it, it's we are able to enable what we have in the hardware, right? Think about PSoC Edge. We not only have the general-purpose processor, we have also hardware acceleration being brought by Ethos-U55. So we have that enabled in our software.
We have a software, a ModusToolbox that allow us to have not only the drivers, not only the middleware, but also code examples to interact with our hardware acceleration, with our Ethos-U55. So we have that being addressed as well. The other aspect is that we also have a lot of documentation. We have support. We have also, we address the different type of IDEs that might be, you know, that our customers might rely on, and we have that support as well.
And then something that we also did, and I'm quite proud of that, it's just last year Infineon acquired Imagimob solutions. Imagimob has been working on software for machine learning for a while. We noticed that we thought that the combination of ModusToolbox with Imagimob would be really useful for the customers. Something that we have done with Imagimob of is now we are able to bring an end-to-end machine learning solution, from the importation of the data, right?
So customers may have some machine learning data. So import that data, we label that data, we pre-process it and come up with a training model on the Imagimob side. Now we work together with ModusToolbox and we are able to optimize that training model with our hardware, right?
With our Ethos-U55 for hardware acceleration, with our Helium DSP engine that we have at PSoC Edge, with our low-power domain that we also have a PSoC Edge, because we actually have come up with a low-power domain to address those battery power applications.
So we are able to optimize it and we are also able to deploy it in the field, to validate that optimization, put the code into our microcontroller, and have a really end-to-end machine learning solution that will allow the customer to expedite their time to market, to have a faster time to market, which is key, having a production-ready, machine learning solution that can rely on and then be ready to have their own device with machine learning capabilities out with a production-ready solution.
- Yeah, and I also understand that Imagimob has, well, what they call Ready Models. - Exactly. - So that you don't really have to be that programming nerd to write your code, but you just simply take them. - Exactly that, and that's one of the key aspects of it, right? You are able to rely on those production Ready Models that Imagimob has, optimize it with our hardware, and being able to, you know, just leverage all the knowledge that we have.
Talking about Imagimob software engineers working together with our hardware and software engineers in the Infineon side of things, although we are all one happy family. You know, being able to work together and bring the best of our two different worlds, hardware and software. - Yeah, so working hand in hand, that's what it is. We've talked about this new device. You've introduced us to use cases to the, let's say old PSoC. Where are the use cases for this new device,
for PSoC Edge, then? - Yes. We're always gonna see improved use cases, right? And you can think about the old cases now being optimized with hardware machine learning, right? Think about a thermostat, right? Thermostat, you need a microcontroller. You know, what we are seeing is that trend in the industry to add machine learning capabilities. We have what we call the explicit interaction.
You know, you would go to your thermostat and by voice say, "Hey, I want this temperature to be 22 degrees," and that's the explicit interaction. Or you would see, you know, you would rely on the touch capability and set everything to 22 degrees. We also have this now implicit interaction, right?
Now the device itself is gonna be ready to learn, to become really intelligent and say, "Oh, this person, you know, that I have identified with face identification, with face recognition, likes to have this temperature at this time of the day." And at night, perhaps, you know, you need something cooler. You know, you might wrap up in your blankets and you'll need a different setting for the night. You won't need to talk again to your thermostat, you know?
The thermostat will adapt, become more intelligent, become more intuitive, and add levels of intimacy as well. And that's something that is unique for machine learning. That's something that will optimize our way of living and you didn't even think about it. Also, another potential use case that we're seeing, smart speakers. The smart speakers today, and I have a five-year-old and a three-year-old, they try to talk to the smart speakers. They are Spanish native in terms of language.
They try to talk to them, and it's like, first of all, this latency, right? They talk to the smart speaker, "Hey, can you play this song?" So like three seconds, comes back, there is always not the one that they wanted, right? Now we're able to come up with natural language. Now we're able to optimize things, again, become more intelligent, have intelligence at the device and avoid that latency that we're seeing.
You know, being able to process those type of things at the real time with voice recognition, with natural language detection and recognition as well, and all natural, happening naturally. So this is something that on top of that, coming back to that implicit interaction, you know, we are able to have a profile setting, right? Now they're gonna know that my kid, by identifying his voice, he likes to play, I don't know, child songs, you know?
And all of that will add more levels of intimacy, intelligence, intuitiveness. On the other side of things, we have industrial applications, right? You have your factory automation line. They will be able to also learn that if your hand is misplaced and it can become dangerous, you can add safety capabilities on it. You know, you can add a lot of different scenarios to improve safety, to improve performance, and to improve efficiency at the edge.
And those are the type of use cases, along with the ones that we have been famous for for the last 20 years, that we'll play a really good role in. - Yeah, so PSoC Edge makes your life easier. - That's the intention, to make our lives easier. - Sounds fabulous, Omar. So, talking about microcontrollers, talking about PSoCs, where do you see the future of this? Is there any new development, new use cases, new opportunities, new products? - Yep. I mean, yes, yes, yes.
- Okay! (chuckles) - When you have the intelligence, it's almost become like a limitless opportunity, right? Like you have all these different capabilities to add intelligence at every way of life. You can think about if you stay at smart home. Your cooktop, your cooktop can become intelligent. It can identify the sizzling or the color of your meat or whatever your, you know? - Okay. - And then it can tell you recipes, it can tell you- - Turns a chef. - Almost, almost like that, right?
It adds so many levels of not only customization and intelligence. Think about your washing machine, being able to recognize soil levels, not needing a lot of soap or not needing this many cycles, you know? Knowing when to stop, become more energy efficient. All those different use cases are something that we're looking at.
I keep talking to customers every day and they introduce, you know, I was just talking to a garage opener, and they're interested in machine learning because you are adding new capabilities of safety as well. You may have your kid, you know, just running by, and it can damage the door and his head or her head. And we are adding these safety levels, we are adding this intelligence on those different use cases that it becomes, you know, so many different opportunities.
Now, with that said, we also have the expansion of our PSoC Edge. We have so far announced three different products for PSoC Edge. We're able to introduce a pretty robust roadmap. We're working on the next generation, adding vision capabilities, adding more performance, adding more interfaces, and this is something we're looking into for the next couple of years.
Of course, as I've said, software becomes really important, and software is also part of the investment that we are doing at PSoC Edge. So it's not only about the hardware; it's also about the software and the solutions we are coming up with by talking to customers, by understanding the trends in the industry, and this is something that we're looking into and really, really happy to be, you know, at the edge of things. - Well, Omar, thank you so much.
We can, well, at least I can, I can sense your excitement about this product. For the listeners out there, for the audience, where can they find more information about the PSoC Edge? - Yeah, please visit us at infineon.com/psocedge. We have documentation, we have product briefs, we have white papers.
So if you are as excited as I am to talk about artificial intelligence, machine learning, go to the website, infineon.com/psocedge, and, you know, we are happy to accommodate all the different questions as well. - Sounds great, Omar. Thank you very much for joining this episode. - Thank you, thank you for the invite - Okay. To the audience out there, thanks very much for your attention, thanks for joining. As we experience, the world is turning ever faster.
There will be more topics coming up which we try to put into different episodes, so please stay tuned. (calm bright music) (music fades)