 Mobile World Progress 2018 and who are you? Andrew Roberts from ST Microelectronics. So let's start the tour of our booth with the ST Experience song. Hello, I'm Frederic. I will show you our interactive table and we are still getting here on the wall some examples of customer products. So right now you selected this one and somehow it shows that your technology is inside the Note 8. Inside the Note 8 we have gyroscope, we have some pressure sensors, we have microcontrollers and then thanks to the interactive table you can know more about all the products we have here. So that was a hard drive just before? A hard drive so you are... Right over there from Seagate where we have power management solutions. So what do you do in the hard drive? We have power management solutions provided by ST and Seagate is one of our key customers for our hard drive. And using those elements we can go further into the ST portfolio here for the solution for the touchscreen. We can browse the portfolio and have more explanation on what we are doing at ST on plenty of use cases. What do you do in the shoe? In the shoes. Is it a step tracker? It's from Nike. We have some microcontrollers and some motion memes. It's a brand new shoes from Nike and we have components inside. Nice. And so there are some glasses that have something? Yeah the glasses are from Eitel from a French company. We can go to the glasses there. We have a microcontroller, we have motion memes and we have even some video running for this product and have more information. What do you do in the Switch? In Nintendo we have again plenty of microcontrollers and motion memes and we have also NFC. Our NFC solutions are made into this component. So actually ST is selling billions and billions of chips, right? Billions of components for microcontrollers, connectivity solutions, power management. We have also sensors. We have analog products. We have a very wide portfolio of products for microcontrollers. Awesome. Thanks a lot. So can we look around you both a little bit? Yeah sure. Let's go and have a look at an application for the smart home. This is an application that shows voice and edge computing. So Roberto, please. Hi. What are you showing here? Hi. Yes, we're showing a reference design of a smart home that can be controlled by voice or by gesture using sensors and actuators. So which sensors are the ST sensors in here? These are all ST sensors. We have sensors that can detect the gesture, time of flight sensor or sensor that can detect temperature and environmental parameters. You know, when I make a gesture I can open or close this window with a gesture. Nice. And the system is an IoT architecture where the control is local and on the cloud at the same time. So this gesture... What does it mean, the coin station? It has nothing to do with Bitcoin, no? No, it's called Bluecoin in this case. What's the Bitcoin? It's this board. It's an able board with four microphones that supports advanced audio processing with which I can call Alexa. Turn on the kitchen light. Just with this little like this? Just with the system. It has four microphones? It has four microphones. I can use one or two or four depending on how strong I want the interaction to be. Alexa, how old is Donald Trump? Donald Trump is 71 years old. I thought he was 73, are you sure? Okay, that's great. So with this system I can ask Alexa to do things for me. Alexa, turn off the kitchen light and the command is interpreted by the cloud and the light goes off. Nice. Alexa, turn on smart home demo. Sorry, I didn't find smart home. Alexa, I can't find my smart home. Let's try again. Alexa. Alexa. So sometimes Alexa has connectivity, but even when... It's very busy right here. It's difficult to connect to me. You know what, a nice feature of this system is that whenever there is no connectivity, I can always do gesture and get an instant response, because whatever happens in the cloud, I press also locally in this edge computing station implementing Amazon Greengrass in this specific case. Alexa. Alexa is sleeping. That's okay. If you give me a second, I can ask... So what is this board here that you're rebooting? We're rebooting a board implementing the AVS gateway that is the gateway allowing Alexa to interact with us. And this is based on an STM32 controller, so it's a low cost implementation, very simple, and it could be implemented in any existing system. How long is the battery life of this big battery? Depending on software optimization, this can last for several days. Alexa, how are you today? I'm doing great. Alexa. Do you know cats sometimes talk to me when there's nobody else around? So far, I've learned one word. Meow. Alexa, where's Ivanka Trump right now? At a White House governor's meeting, President Donald Trump said that he would have personally run into the Florida school during the shooting that occurred earlier this month, even if he were unarmed. So Alexa knows everything? Alexa knows everything, yes. And she's helping to run our smart home now? Alexa, open smart home demo. It's just a connection. It's a connection, yes, yes, yes. All right. Thank you very much. Thank you. Thanks. And you have a lot of other demos around here, right? Okay. Now let's go and have a look at how our mem speakers revolutionizing headphones sound. All right, so here it says actuate. Hello. So what are you showing over here? So today we are showing a demo done in collaboration with U Sound. As I've now said previously, we are manufacturing mems with SD technology, and U Sound is designing tiny mic speakers. So U Sound right here? Yes. And doing what? What is happening here? Micro speakers that are positioned inside the headphones. So there's many speakers in there? Yes, seven per ear. And what does it do? 360 degrees sound. And one of those things over here and there? What is going on with all that stuff? So you see that the sound is moving around his head. Nice. It's moving around my head. So you can hear it around you. I can feel it coming, hear it coming across, see? So what more technology is in the headphones? What is up here and there? Up there, there is the amplifier. The amplifier? Yeah, at the moment it's a discrete components, but they are going to embed it in a smaller form factor. All right. So what is a Petra technology? Petra technology is the one that is used to manufacture the speakers. So there is a piezo material that is moving a membrane that emits a sound. Is this going to revolutionize the headphones? Yes, exactly. You have a better sound quality with a smaller form factor and... A more immersive experience? Yes, exactly. A more directional sound. Well, cool. Awesome. I definitely want to try that later. Awesome. Thank you. Thank you. Thanks. And there's a lot of other things. You have wireless charging, for example. There's a lot of other things. Touch display. There's more over there. Just a second. Here it says something about the artificial intelligence. So... No problem. So how do you integrate AI in the ST? There's all these new networks. We can jump in there just after. We can't don't? Yeah. Just saying. Let me look over here. Over here there's something to do with the flight sense. Right here. There's some sense and touch display. All right. So can you talk about this? Yeah. So do you have a sensor in water? Yeah. This is a waterproof pressure sensor. That's metallic one. Yeah? So not needed to coat it. But it's water resistant by default. Waterproof? Yeah, waterproof. By default? Yes. The whole board is? Did you coat the board but not the sensor? Exactly. Exactly like that. So it's waterproof? What pressure? What pressure? It can resist up to 10 bar. So you can swim with it without breaking it. So it's to show when you watch how deep you are or what? Yes, also. Or it can work also in air. Okay, so you use it as a standard pressure sensor. But you can use it for waterproof devices as well. All right. So this is also something to do with the touch display. This touch display? Yeah. Yes, it's meant for industrial application. So you can use a glove and touch? Yes, you can use a glove and use the touch in a harsh environment. All right. Cool, thank you. What are you showing there? What are we showing here? This is like how NFC technology is revolutionizing. Maybe you can tell us about it. So what are you showing right here? So here we are showing NFC smart tag. So it's a tag with different sensors that can control the environmental data. For example, during logistics. Yeah, can you hold it right here? So what is on here? So on here there is one STM32 microcontroller. A big one there? A big one there. There is three environmental sensors like 3D accelerometer, 3D magnetometer, humidity, temperature sensor, and also nanopressure sensor. And around is NFC? Around is the antenna. And there is NFC chip. So it's a dynamic tag, ST25DV. This is a dynamic tag which is used to communicate and to exchange data between the sensors, in fact, and the smartphone. All right. You want me to show them? Yeah. So what is this? So this is emulating some environmental data that are changing during transportation, for example. So here it's emulating the cold temperatures. So first I need to set up my tag, saying that I will check temperature, humidity, pressure, vibration every one second. So my system will sample every second, it will take some measurement. So now it's measuring. Then I will emulate the shock. Nice. That's the shock. Then I will emulate hot temperatures. So now it's emulating like a transportation of goods. Then at the end, when my goods arrive, I can control what depends during the transportation. So I read my own data, and then I can print all the data. So we can see there's a lot of shocks, and temperature is variation. So then I can decide if something happened during transportation. It's used for cold chain monitoring and so on. Food, transportation. Yes, and cold chain, transportation, and containers. Ice cream. Ice cream. I don't want the ice cream to melt. It's not a good thing. Yes, so thanks to NFC. Malanaz, transportation, everything. Yes, exactly. Maybe the pet in the airplane to see if you need anything else. Yes, so you can put a reason. All right. So it's like thanks. Thanks. Thanks. So let's have a look at artificial intelligence R&D. So maybe you can tell us a bit more about ST and artificial intelligence projects. So our approach regarding artificial intelligence is to propose to customers the solution according to their needs. So we declared in three different segments. The first one is based on the STM42 and only on the STM42 when we integrate. What is this CubeMX AI? CubeMX AI is development tools we use for STM42 applications. And we add to CubeMX and then neural network layers. And the users have access immediately to those layers to implement an AI solution. That's an Omkortex M4 based microcontroller. But is AI possible on this? Exactly. It's not an extra CC. It's only the STM42. On the second solution, what we are proposing is free. So on this solution, it's still on an FPGA. So it's not yet at a silicon level. So it's emanating a future chipset? Yeah. It's all R&D projects. So first we'll have a future microcontroller with AI. With AI, and if you want to reach a good level of performance. So basically with this application, you can manage with a few classes, a few 10. With this application, with this IP, you can manage with hundreds of different classes. And with the last solution, you can go to thousands. Let's go over here. So here's the time right here. This is called the Deep Learning ST Orlando V1. So it's a deep convolutional neural network SOC. So it's a SOC solution. Is it a new SOC? It's a new SOC. It's right here? Yeah. This SOC, what is the SOC in here? Well, the SOC is this... Is this a microcontroller? It's a heterogeneous system based on ST architecture. That includes hardware accelerator as well as ST DSPs. Is it no ARM or something else? There is an M4 ARM that is used for control of the system. And what is it used for neural network? What is it designed for that part? Okay. The networks that we can deploy are a variety of different topologies and training sets. We vary from the AlexNet topology for object classification to the tiny YOLO network that can detect multiple objects on the scene and vary from the scene evaluation, face recognition and face expression analysis as well as behavior analysis. So it says here 28 nanometers of the SOI SOC for deep convolutional neural networks. Is this shaping or is it the prototype or what is it? This is an R&D prototype. First prototype of the SOC? Yes. The development of the product is ongoing. And eventually there will be microcontrollers with AI with neural networks. The AI is already implemented in this system. Eventually it can be connected to a sensor controller like a system based on an ARM sensor. How so many is this chipset available on the market? I don't have a specific deadline for... And this will be important for the mobile world? There will be a lot of security cameras with AI? This is meant for run smoothly on autonomous systems so for any mobile or battery operated system. Cool. All right. It is meant for object detection on a variety of applications. All right. And right here it says... So you have a lot of partners? This is an R&D project. Yes, that's R&D. A lot of interest at the show? Yes, a lot of people that is coming to us for many different questions. Finally, one last one. Precise positioning are extremely precise GNSS applications. So Louis can tell you a bit more about that. Yes, hello. Is this more than GPS? More than GPS. So it's GPS, GLONASS, Baidu, all the flavors, multi-constellation for GNSS others. So really targeting high accuracy with functional safety for automotive. How accurate? So we bring down to 20 centimeters accuracy to Sigma. So really high reliability behind this application for automakers here once. So we're not targeting mass markets really for automotive with security. And here it looks like there's a chip that says ESCAPE. No, ESCAPE is a project. It's a project. Yes, we're announcing today but ESCAPE is already a research project funded by European Union to really bring the best of European companies. So our silicon is inside. Yeah. And it's chipset. It's chipset, exactly. And these companies, what do they do? All of them. So this is a tier one integrating. And the group here know it's testing in their vehicles for autonomous driving. So it's really going to help the self-driving car? It will help? Yes, exactly. To locate them in the maps, in the HD maps, but after that also bringing security because maps might have some errors. So absolute positioning brings the best accuracy and reliability. So is it the STM32 or something else? No, it's different. Actually, you're lucky. So today, in about 10 minutes, we'll be announced. Worlds first. Dual frequency, multi-consolation, genoset silicon. It's a different product. It's a new silicon. It's a new silicon just targeting high accuracy, high safety applications. So my video is 90 minutes. So we have to wait until the 29th minute before you can say exactly what it is. No, I can show you. We have a silicon. Yes. So it's a new chipset. Is it a microcontroller? No, no. It's a chipset for GPS. So for positioning that can provide multi-constellation, GPS, GLONASS, Baidu for the Chinese market, Indian market, Japanese, and of course European, Galileo for Europe, bringing multi-constellation that brings down the accuracy to 20 centimeters. There is no other company doing this. Nobody do this on the chip. Nobody. Everybody else doing software or what? They bring different software and they cannot bring... FPGA? Yeah, FPGAs, but they cannot bring safety. So AISL and the silicon. What is AISL? So it stands for Functional Safety. So accuracy and safety for automotive applications. And can you say if this is an ARM Cortex-M? Yes. So our silicon actually have here the picture. You can see it. So it's being real here. It's this one. There's your five. You hold it up to this AISL. So it's this one. Which one? This one, T0-5. So it's an M7 base. M7, correct. What is a hyperbos? There's a different communication for this silicon. And if you want also the Functional Safety, then it's T0-PP, this one. So it's the same, multi-constellation, multi-frequency. And this is a differentiation for Functional Safety. So the name of it is T0. T0? What does T0 mean? T0 is actually a mythical Greek figure that helped people navigate to non-areas. This is like the guy on the boat on the ship. Exactly. It's something like that. And then we're using T0 as an old family already. We are actually very innovative since the 1990s, end of 1990s. Single frequency, single constellation. And now, all together, multi-constellation, multi-frequency for high accuracy. Why is a box this size? Is it because it's automatic? You need a big thing to make a smaller one too? Definitely. How can it be? The silicon is actually very small, just less than 1 cm by 1 cm. Can this be in a child tracker or something? Definitely. It could be. We're targeting automotive. It can be in a watch. Definitely. It can be anywhere. It can be in many different locations. A smart phone, for instance. Pet tracker? Pet tracker. But here, for high accuracy and safety, we're targeting mostly automotive. Automotive is most important? Very important. Any military? I don't know about the military, but for automotive, it definitely gets our target. And we're the first ones doing this. All right. And how soon is it mass production, mass deployment? This is all for very soon in automotive. You mean two, three years time. So SOP 2021. We work with all the German automakers tier one, so it's point production. All right. Okay. Thanks a lot. Thank you very much. All right. That was great. Okay. Thanks a lot for this. Very busy show. And hopefully we'll see you again next year as well. I'm going right now to the embedded world. So that's a whole other bunch of demos, right? You'll see lots, the Fan Zone, the STM32, some automotive demos, some NFC. There's lots of things to see over there. So have a good trip to Nuremberg. And nice to see you. Thanks a lot. Cheers.