 Hi, welcome to ST in Las Vegas at CS2020. My name is Roberto Sandino and I'm a system research manager. Here I will discuss sensors for industry 4.0. So with industry 4.0 smart factories achieve better knowledge and control over the entire production process through real-time monitoring techniques, while distributed control systems based on IoT architecture may take advantage of big data and IA technology in order to increase and enhance sustainability while reducing maintenance costs. So a key concept we want to show at this demo stage is the concept of a smart sensor node, which is a node, a little board that concentrates processing, sensing and connectivity features and capabilities. It is a smart node because it must adapt to the different system requirements which will change from case by case in a production factory. So with this demo we show several examples of modularity where we exploit a different kind of connectivity in order to achieve control of monitoring, constant monitoring of vibrations, acoustics and environmental parameters of the devices under investigation. In this demo set we are going to see several examples of connection of a smart sensor node to the cloud. In the specific case the cloud is hosted by the ST.com website and is based on Amazon AWS. So we see first of all in this example a sensor node that is connected via a typical industrial serial link that is called IO link. And here the system is measuring the vibration of this motor. So let's turn on the motor in order to start observing the vibrations that are shown here from the dashboard connected to the cloud. So while the motor runs we observe a number of harmonics in the frequency domain that are a typical signature of a good behavior of the motor. Now we are going to disturb the behavior of the motor with an external magnet that is placed close to it. Now this is going to generate an anomaly that will impose certain spools vibrations to the system that are detected by the board. The board senses this irregularity and sends an alarm into the gateway and then into the cloud. So if we check the dashboard we see that we receive the notification of an event which in this case is an alarm showing that pump one is malfunctioning. Now if we correct the problem by taking away the magnet from the motor we will see that the system, because of its continuous monitoring, checks and tells us that the problem is being collected. So in this case we use the Greengrass based architecture gateway where that manages connection to the cloud in a seamless way. Even if there are connectivity instabilities the gateway can solve any problem and keep synchronization between the local information and the cloud based information. So let's now consider what we can do with a wireless sensor node which is shown here in this application. So in this case we have this board that is called ST-WIN that is a wireless industrial node, smart node connected to a bearing. Now this system being battery powered can work in low power mode which means by pushing this button we are going to cut out the constant communication to the cloud so that the system can go on and analyze the data. While the motor has a good behavior we don't need to continuously send the same data over to the cloud but we can optimize the process. Now the system in low power mode checks the vibrations, inertial and acoustic of these systems and runs it continuously. Now let's inject a problem into this bearing by simulation and see what happens when we do so. We have injected the problem and the system has just waked up from the low power mode. A warning has been raised as you can see in the cloud interface and through this warning the system is doing two things. It's notifying that something strange is happening to the system and it's also beginning to send actually data to the cloud so that they can be further analyzed. If we inject an even stronger problem into this bearing, we're simulating a worse damage, the system can also autonomously raise an alarm that is more important calling for help immediately into the system. This is something that can happen using an autonomous Wi-Fi connection to the provided cloud. The SD-WIN is also meant to be a tool to be used by data scientists because it can stream all its sensors and we're talking about 22 sensors at full resolution in real time over USB onto a host. In this case you see this PC receiving all the 22 sensors which include vibrational, inertial, environmental, acoustics and ultrasound or in real time it's synchronized onto a PC. This is basically what is needed by a data scientist in order to do his job and generate advanced algorithm in order to process the application data. Here we see two other examples of sensor nodes implemented as an ODE stack of boards based on the STM32WB or a sensor tile dot box system. They are connected via Bluetooth to a gateway implementing a green grass architecture by Amazon. This part of the demo can be done through a third-party company, an STI partner called Kirka Technology and it connects autonomously to the web onto a dashboard as shown in this picture. Everything that has been shown here can be easily replicated because all the hardware and software devices we have shown are available at STI.com can be used and so every part of this demo can be replicated in your labs. In particular we consider the fundamental to lower the barriers of entry into this interesting application field and for that reason we have prepared this kind of sensor node boards as available through kits like this one where they can be used, connected to battery, plugged and closed into a case and deployed on the field. So the cloud is available through STI.com giving exactly these results which can be customized depending on your needs. So thank you for your attention and for further information please visit www.tasty.com. Thank you.