 Hello everyone and thank you for joining us today, I'm Jussie and I'm going to give you some quick demonstration around industry 4.0 application for condition-based monitoring and predictive maintenance. First a few words about the contests we are in today. We are really seeing a rapid evolution in industry where we are leveraging a wide variety of solutions to render production more efficient. The terms we hear a lot about are industry 4.0 or digital manufacturing for example the driving forces behind their clear energy efficiency, secure and safe production, and maintenance optimization. Being capable to monitor and detect equipment malfunction and implement fast corrective actions has a dramatic impact on safety, quality control, operational costs and finally if you refer to electric motors on energy efficiency as well. Technology breakthrough like sensor miniaturization, high performance and low cost microcontroller are tanky connectivity are essential as they are low real-time acquisition and sophisticated processing like AI at the hedge and digital twins. In this domain one of the key applications is indeed condition-based monitoring and predictive maintenance which brings me to the main points of this video. We believe that developing condition monitoring is the first and fundamental step towards implementing a predictive maintenance strategy. In this contest the field of application is very wide going from factory automation to transportation to energy and oil industries. We can summarize the main challenges follows. First data management architecture based on the physical system we need to define the architecture that enables the gathering of the right data at the right time which means defining the data logging capabilities the latency to log the data and therefore the connectivity. So we need to build up the right sensor nodes by choosing the products that fit better in terms of performance for example the microcontroller or the sensor or the analog device. Then we need to implement processing strategies that guarantee power consumption, security and therefore the right partitioning between edge and high-level processing happening in the cloud or on company premise. In this demo we are going to showcase some of the possible implementation that helped the designer to solve those challenges. The demo is composed indeed by three elements that are driven by a touchscreen display based on STM32 H7 Discovery Kit with touch GFX software utilities. The three scenarios represents uskated enabled by vibration analysis, ultrasound and environmental sensor. Let's see the demo working. This part of the demo will show how it is possible to stream all the data from the smart sensor node that we call STWIN. STWIN stands for industrial wireless sensor tile and it's equipped with nine sensors for 16 degrees of freedom. The kit features a battery power core system board with ultra-low power STM32 L4MCU that receives data reflecting the operating condition of an industrial machinery from a range of advanced STM environmental and motion sensors including the state-of-art ISM330 DHC XE inertial measurement unit with machine learning capability and vibrometer IIS D3WB. STWIN comes with a complete ecosystem including software for high-speed data logger and interoperability with Unicode GUI. Let's turn on the device and here we see the GUI showing all the data streaming. This tool is great for helping engineers in gathering size and test their own algorithms. The GUI is available also as open source. I contain some famous example and high-speed sensor data login to SD card or via USB. Let's move to the second use case. Here we show the STM32 MP1-7 DK2 acting like a gateway with process capability that's collected wireless smart sensor node connected to a motor. To emulate a workflow where anomaly detection works on the edge we are using IO-Link serial communication and the TINIS smart sensor node STVAL BFA001B2. Data is streaming to the MP1 which detects the anomalies and sends the event and the data in the cloud-based application. The event is communicated and displayed in real time. And for the analysis or algorithm development could be done to classify or implement action on the edge also by using AWS Greengrass. Let's turn on the device. Here we see the normal behavior of the motor. Let's simulate a fault. Now you see the event that is shown and data are streamed as well. Finally, let's move to the third use case where the anomaly detection capability for upbearing. Inference works on the edge in DEL4MCU of the STWIN in an ultra low power consumption mode. To emulate a workflow where anomaly detection works on the edge, in this case we are using a Wi-Fi connection that is enabled when an anomaly is detected to stream the event. And data in the cloud based on application developed always by using AWS service. The event is communicated in real time and further analysis or algorithm development could be done to classify or implement actions on the edge. And the event is shown in the dashboard and the data are streamed successfully. The DEL4MCU has different inference capabilities. STWIN has developed partnership with software companies and at the same time enabled the programming on neural networks through CUBE.AI software that is part of STM32 CUBE MX. All the demo components, the hardware, the software and the cloud application are available at st.com. Please visit our webpage dedicated to condition monitoring and predictive maintenance. Thank you for your time and start to discover all the future ST ecosystem author. Stay tuned and see you soon with Let Us New. Bye.