 Hello everyone, my name is Thiago Hayes and I am ST's Product Marketing Engineer responsible for sensors and low power RF products. In today's webinar, I will highlight how to explore new applications using the Sensor Tile Box and its ecosystem. We are going to explore the benefits of using the Sensor Tile Box as a key hardware platform from entry to professional levels of development. As an agenda, I will quickly go through the MEMS and sensors landscape and how ST is addressing key market verticals as well as diving into the Sensor Tile Box hardware and its components, and specifically the operational modes that are supported, entry, expert and professional modes. We will also rely on Algo Builder and Unicly UI during this webinar. These tools are now part of the professional development mode of the Sensor Tile Box. Let's start by talking about the IoT movement, where familiar devices are evolving to become more connected and smarter by leveraging the massive computing capability of the cloud. And by the way, IoT comes in many factors from consumer, industrial, automotive, smart homes and smart cities as well. The IoT market movement is where almost any of those systems is able to leverage the Internet and benefit from its ecosystem of cloud computing. And for sure, sensors play a key role on this market trend as the need to acquire and understand data is growing significantly. And with the expansion of IoT, the variety of sensors deployed in the field is facing a significant growth, and ST is one of the key drivers towards best-in-class experience at application level. Not only the breadth of the portfolio is increasing towards consumer, industrial and automotive industries, but it is also key parameters in terms of products that are being optimized, such as accuracy, noise density, size, and power consumption. And here are some of the key sensor technologies that ST can support, motion sensors with our accelerometers, gyroscopes, six-axis IMUs, magnetometers, and e-compasses, environmental sensing with our pressure, humidity, and temperature sensing, acoustic sensing with our MEMS microphones, and ranging technologies through our time-of-flight product portfolio. Additionally, from a sensor product portfolio perspective, we can support our customers among three key market verticals, automotive, industrial, and consumer. Automotive with ACQ-100 certified products, in this case motion sensors, such as accelerometers, gyros, and IMUs. Industrial with our motion sensors with 10-year longevity plus our microphones as well, and last but not the least, the consumer grade portfolio, which is a growing product portfolio that features the latest motion, environmental, and acoustic sensors as well. Now transitioning to the sensor tile box and the ecosystem, we are going to explore in this webinar. The main goal of this webinar is to allow our customers to get familiar with new sensors and provide a path to speed up the development cycle of novel applications. The development ecosystem can involve in complexity together with the user by starting from the entry mode with simple sensor evaluation, data logging, and exploring different functionalities, passing through the expert mode, which allows the user to create a fast proof-of-concept algorithm, without writing a single line of code directly through the SCBLE sensor app, and finally through the professional working mode, where the user can explore rapid prototyping and have access to full customization capabilities through a dedicated function pack and on the STM32 open development environment. Also, the professional working mode enables the user to explore tools such as the Algo Builder and the Uniclio GUI. And as I mentioned before, one of the most important hardware platforms that SCBLE can enable this design flow is the sensor tile box, which is a flexible plug-and-play IoT node that allows users to jumpstart applications by relying on a user-friendly mobile app interface called STBLE Sensor App, that is, by the way, available on both Google Play and the App Store. Now let's take a look inside the sensor tile box. Its main board features from one side the latest motion sensors from ST, our environmental sensors, and our analog wideband MEMS microphone as well. On top of this, you have the SCM32 L4 microcontroller as the main core of this board and an ST Blue Energy-based BLE module as well. On top of those, you have a micro USB connector, and at its back, you have a micro SD card, a 500mAh battery, and our ST Link V3 connector. From a motion sensor's perspective, the sensor tile box includes our low-power 6-axis IMU with machine learning core, our stand-alone inklinometer, our ultra-low-power accelerometer, and a stand-alone magnetometer. And also, from the environmental sensor's perspective, the sensor tile box features a temperature and humidity sensor combo, a stand-alone digital temperature sensor, our absolute pressure sensor, ideal for altimeter applications, and from an acoustic side, we also have our analog wideband microphone on board. Now let's dive into the specific part numbers of the components of the sensor tile box. Starting with our motion sensors, the platform benefits from our latest 6-axis inertial measurement unit, or IMU, the LSM6-DSOX, which is best in class-in-power consumption, featuring high-performance accelerometer and gyroscope combo. The device also enables a new concept in smart features by implementing finite state machines and decision trees. The board also features two stand-alone accelerometers, the LIS3DHH, our high-performance inklinometer, based on a ceramic package, ideal for inclination measurement over temperature and time, and the LIS2DW12, which is our ultra-low-power accelerometer, featuring less than 0.5 microamps of power consumption in active mode. The platform also benefits from our stand-alone magnetometer, the LIS2MDL, the perfect companion for the LSM6-DSOX, moving towards a full 9 degrees of freedom solution. The part benefits specifically from a high dynamic range, increasing the robustness of any magnetometer subsystem. From the environmental sensors perspective, the sensor tile box includes the latest high-performance pressure sensor, the LPS32HH, which offers high accuracy, low noise, and low power consumption. In addition to pressure, measuring temperature is one of the most common implementations in the industry, and the STTS751 can bring high accuracy and low power consumption to IoT devices. We also included our small form factor and low power humidity and temperature sensor combo, STTS221, featuring 3.5% relative humidity accuracy and 0.5 degrees Celsius temperature accuracy. And finally, but not the least, our MP23 ABS1 analog wideband MEMS microphone, capable of capturing not only voice, but also ultrasonic sounds generated by the environment. From a processing perspective, all the sensors are managed by our low power STM32L4 microcontroller, in conjunction with the SPBTLE-1S, our low power Bluetooth low energy module, based on Blue Energy-1 system on chip. Additionally, as I mentioned earlier, the sensor tile box is battery-powered, and it features our STBC02 battery charger technology. Now let's talk about the different modes supported by the sensor tile box. In this case, we'll start with the entry mode that is the out-of-the-box experience of the sensor tile box. There is no need for programming. You just need to connect the boards to the smartphone mobile app and evaluate one of the available functions, such as pedometer, human activity recognition, sensor fusion, compass and level, data recorder, vibration monitor, barometer, and two AI-based implementations, in-vehicle baby alarm, and the baby crying detector. The next mode is the expert working mode that increases complexity and flexibility of the mobile app interface without writing a single line of code. In this case, the user benefits from the STBLE sensor app graphical user interface to create and customize new applications. It's possible to define additional applications by selecting the necessary building blocks and allocating the functions that does make sense for each of those specific use cases. Some examples are power optimization by selecting which sensors are activated or shut down, vibration analysis through FST, and even pattern recognition by benefiting from algorithms such as ST sensor fusion that can be enabled as needed. The last working mode of the sensor tile box is the PRO mode. It is not only fully compatible with the STM32 open development environment, but it is also fully compatible with the STM32 QBMX, allowing the user to properly set up the hardware and specifically the microcontroller that is as needed. The platform also supports a dedicated function pack featuring a wide range of examples that benefit from the sensor tile box hardware. In this case, the function pack is called SP-SNS-STBOX1V1. In addition to these features, the PRO working mode now also supports the AlgoBuilder and Uniquely UI that will be utilized during one of the examples of this webinar. Additionally, this is the mode where the user can also benefit from the embedded ST-LinkV3 connector for programming and debugging. So now let's get started. We are going to explore the sensor tile box ecosystem benefiting from entry expert and professional modes targeting a single application, in this case, vibration analysis. From simply playing with ST sensors to customizing my sensor tile box without writing a single line of code, and finally leveraging tools such as the AlgoBuilder and Uniquely UI to enable a high level of customization and also generate my own C code. And for this purpose, we selected vibration analysis as the example to be explored in today's webinar. The range of applications that can be developed with our sensors in predictive maintenance is very wide and can be implemented within consumer, industrial, and even automotive products. For instance, being able to measure vibration of an industrial motor is critical as it allows preventing failure in the field and it also enables saving a significant amount of time and money. We're starting with the entry mode of the sensor tile box. We'll explore the vibration monitoring example present in the ST BLE sensor mobile app. And we'll use it to evaluate the SFT transform of data coming from a wide band accelerometer. In this case, the LSM6DSOX, which by the way is part of the sensor tile box. To complete the example in vibration monitoring within entry mode, there are two steps required. The first is to train our hardware to understand what is a typical level of vibration and this is based on SD card data logging. And the second step is to compare real-time data versus the good baseline defined in the previous step and then trigger a warning. Once the vibration reaches a value that is outside of the expected range, the LED present in the mobile app will turn on. And in a practical scenario, this is a great example of an early failure warning in the field. When opening the ST BLE sensor app, the first step in order to train the sensor tile box to recognize a good vibration pattern is to click on create apps and from the list select the vibration monitor dash training example. This is a preloaded app example that will allow you to collect FFT vibration data within the SD card of the sensor tile box. Once the example is selected, just click play and select your sensor tile box hardware. After uploading the vibration monitor dash training app into your sensor tile box, simply connect to it using the ST BLE sensor app and once connected, the mobile app screen will feature the start logging button which will trigger the baseline vibration data log into your SD card. Once done, just click stop logging. The baseline vibration can be not only a vibration pattern but also in many cases, the absence of movement at all on the sensor node. There is a wide range of applications that are focused on understanding if certain equipment are in movement or without any movement. So active or inactive use cases. Once you finalize your SD card logging, you can always connect your sensor tile box to your PC via USB cable. The PC will recognize the sensor tile box as a USB drive and in there, you will find the vibration monitoring baseline CSV file. This capability allows the user to further analyze and understand the collected vibration data that will be utilized as a reference to be compared against real-time vibration data as well. The second step of this process is to utilize the vibration monitor dash compare example present in the ST BLE sensor app. Following the same steps as the previous app example, simply click on create apps and click on vibration monitor dash compare. From there, just click the play button to upload the mobile app example into your sensor tile box. After you click the play button, the mobile app will ask you if you want to override the board. Just click okay and it will upload the new application into your sensor tile box. The upload should happen successfully and once done, you're ready to prepare and play with the example. In order to utilize the vibration monitoring dash compare example, the first step is to connect to your sensor tile box. Once done, you will notice that the mobile app interface has changed and it now displays an LED symbol that is turned off. The off LED represents a healthy vibration status. We find the vibration guideline defined in the first step of this example. As soon as the device starts vibrating differently, showing an atypical vibration, the LED will turn on. This means that the vibration measured by the sensor tile box is now outside of the specified baseline. Additionally, the vibration monitor dash compare example also allows the user to start and stop data logging. This procedure can enable the user to collect a CSV file that will display the LED status over time, allowing the user then to identify when the fault has actually occurred. Now that we completed the vibration monitoring example in the entry mode of the sensor tile box, we can now move to the expert mode. The expert mode of the sensor tile box enables the creation of a customized application for vibration monitoring. It allows selecting the desired inputs in this case, the wideband accelerometer or vibrometer, the LSM6 DSOX. It also allows selecting a wide range of functions and also for this example, the FFT transformation will be utilized. Additionally, it also allows the user to select the necessary output. In this example, we will rely on streaming sensor data via Bluetooth flow energy to the STBLE sensor app. To get started with the expert mode, when opening the STBLE sensor app, simply click on create apps at the bottom left of the main screen of the app. And from there, scroll all the way down through all the examples of the entry level mode and click on expert view. The expert view will allow you to create a customized application for your sensor tile box. By clicking on new app, the mobile app will initiate a design flow starting with the selection of an input. The input list includes all the sensors available within the sensor tile box. This allows the user to evaluate specific sensors and functionalities, is still without writing a single line of code and relying on the STBLE sensor app. Once the vibrometer input is selected, the main flow of the expert mode includes a gear button together with the selected input. This allows the user to select and configure a wide range of parameters related to the sensor being utilized. The user can configure parameters such as performance modes, output data rates, full scale, filtering chain, and more. In this example, we are looking into a low frequency vibration and for this reason, 52 hertz is the selected output data rate. The next step is now to choose the function to be applied on top of the sensor data. When clicking on the functions button, the mobile app will load the list featuring all the available functions in expert mode. For this example, in specific, we are looking into the SST function and we'll configure it by clicking on the gear button present in the main flow window to consider 256 samples as the FST length. The next step is now to select the desired output of the custom flow created in this STBLE sensor app. Within the output selection tab, the mobile app lists a few different options that can be explored by the end user, saving the file to the SD card, streaming data via USB, saving the output of the custom algorithm as an input, and also streaming data over Bluetooth low energy. In this case, BLE is the option we will select and then click on continue. From there, it's possible to save the app and set a customized name for it. The vibration monitor application in expert mode is finally ready to be tested. To initiate the evaluation of the custom application, just tap the upload button and from there, the SensorTile box firmer will be updated by the STBLE sensor app. The app loading should happen successfully and you'll be ready to test, as we did in the entry mode, your custom application. On the main screen of the STBLE sensor app, click the connect to a device button and connect to your SensorTile.box hardware. From there, the mobile app will initiate vibration data acquisition. And one of the tests that can be done is applying different level of vibrations, for instance, slow and fast vibration, and directly to the mobile app interface, visualize the SST output chart. This allows the developer to quickly mount a wireless sensor node into a vibrating system and not only collect, but also visualize data directly to the mobile app interface. With that, we just completed the expert mode example and we are now ready to take a step ahead and utilize the professional working mode of the SensorTile.box. The main advantage of the professional working mode of the SensorTile.box is that it is fully compatible with our STM32 open development environment. It allows the user to utilize powerful hardware platform as a baseline to program directly into the STM32 core and to benefit from a dedicated STM32Cube function pack. The function pack is called fp-sns-stbox1v1. And by the way, the professional working mode also features full compatibility with our powerful PC-based graphical user interfaces, such as the AlgoBuilder and UniqueLeo GUI. Our next vibration analysis implementation will be based on the professional working mode and will rely on AlgoBuilder. And by the way, I would like to take this chance to recap some of the key aspects of the powerful graphical user interface. The AlgoBuilder is an application for the graphical design and testing of algorithms. It relies on the existing algorithms and libraries provided by ST, and also user-defined data processing blocks. The graphical user interface also allows the user to quickly elaborate prototype or proof-of-concept applications for sensors and STM32 microcontrollers. And now, the SensorTile.box is one of the key hardware platforms supported by the tool. The key advantage of the AlgoBuilder is to ease the process of implementing such proof-of-concepting algorithms without writing a single line of code. From a more detailed perspective, the AlgoBuilder is based on merging into a single firmer project three key building blocks, a baseline firmer template that allows the utilization of the GUI and its generated C code, a wide range of ST libraries, which for this example will utilize the FFT building block, and the C code generated by AlgoBuilder through a conversion engine that takes the XML block diagram design and converts it into C code. From there, the firmer project is then compiled by AlgoBuilder by referencing an external compiler such as the STM32Q by DE, IR, and Kile. And finally, the project is then uploaded to the sensor tab box via USB connection. Once the firmer project is running, the AlgoBuilder connects to Uniclio GUI in order to display and plot any relevant data that is part of the customized proof-of-concept implementation. Since in this example, we visualize data through Uniclio GUI, I wanted to highlight some of the key aspects of this powerful software tool for sensory evaluation. One of the main aspects of Uniclio GUI is that it allows the user to configure sensor output data rate and full scale, as well as direct read and write the individual registers of each of the sensors on board. The Uniclio GUI also has a wide range of display options for connected sensors such as plot over time, 3D scatter, interrupt generation to a logic analyzer, sensor fusion, and more. It also allows the user to perform data logging into CSV and TSV files. It also allows the user to explore and evaluate a wide range of libraries provided by SC, in this case, part of the XCubeMEMS1 package. And when it comes to AlgoBuilder, Uniclio GUI is the tool that we utilize to visualize and display the outputs of the algorithms created by the tool. So in order to get started with the vibration monitoring example in the pro mode of the sensor tile dot box and AlgoBuilder, the first step is to click on new design on the upper left corner of the GUI. From there, a window will pop up and you have the firmer location where you'll be able to save your design. There are also two drop down menus in this window, one to select the IDE that AlgoBuilder will utilize. In this case, I'm using SCM32Cube IDE and the other that will allow you to select the desired hardware platform. In this case, the sensor tile dot box. Once all the selections are done, just click OK and you're ready to go to the next steps. The first step is to expand the sensor hub section of AlgoBuilder's library and from there, place the sensor hub block into the workspace. The sensor hub block is responsible for the initialization of the sensors that will be utilized in this design. And in this case, we will be using the accelerometer. For this specific vibration analysis use case, we are going to reference a full scale of 4G and an output data rate of 100 hertz. And these values can be customized on the right hand side properties tab. The next step is to include the Acceleration G function block into the block diagram. This is exactly the block responsible for retrieving the accelerometer data from the available sensors in the hardware. From there, just connect the sensor hub outputs with the accelerometer's input pins and you're ready to go. One of the key value propositions of AlgoBuilder is the ability to allow the user to monitor data within the block diagram structure. In this case, it's important to monitor and visualize the acceleration data coming from x, y, and z axis. The graph's building block is available as a part of the display library and can be connected to the output pin of the accelerometer. But before making the connection, it's important to select and adjust the properties of the graph's building block. For instance, the user can give it a name and adjust the quantity of waveforms that will be displayed. In this case, the name is going to be Acceleration and there are three waveforms that will be visualized for each of the axes, x, y, and z axes. The next step is to add the demux float function blocks from the other library field. The goal of this function is to separate the three acceleration axes into independent outputs. In this way, when I move forward in the block diagram, I can run a FFT analysis on x, y, and z axes independently. And as I mentioned, after I separated the inputs, we'll have three FFT blocks loaded into our block diagram. The three FFT function blocks will come from the FFT library available within Algo Builder. You set the properties to 256 data length for all the three blocks, and you connect the blocks to each of the outputs of the demux function block. From there, now that we have loaded the FFT function blocks into the vibration monitor example, the next step is to include three FFT plot function blocks from the display library. This display block will allow the user to visualize FFT data through Uniclio GUI. But it's also important to configure it properly. In this case, configure the block diagram by setting the data length at 256 and full scale one for all function blocks. From there, the FFT plot blocks can be connected to the first output ping of the FFT building block. The next step is to add the FFT peak function block from the FFT library. The FFT peak block allows the user to detect the peak amplitude and its frequency in the FFT analysis. In this example, since we separated x, y, and v-axis, the custom algorithm created within Algo Builder will be able to identify the peak acceleration on each of the axes separately. From there, it's important to set once more the data length of the three FFT peak blocks that were added in this diagram. In this case, value is 256. Once the settings were adjusted, the FFT peak can also be connected to the first output ping of the FFT building block. Now the next step is adding a max load function from the other library. In this case, we'll select six inputs that will receive the FFT peak data, so in terms of amplitude and frequency, and then connect the blocks into the max function. After implementing the max function, the next step is to add the value function block from the display library. The value function block will receive the output of the max building block and will include six values and labels, representing the peak amplitude and frequencies of the FFT-based vibration monitor. And finally, we have successfully completed the vibration monitor implementation using Algo Builder and the Pro mode of the sensor tile box. This block diagram is a great example of the flexibility that can be explored by a developer when using Algo Builder. And it also highlights an ecosystem that did not require writing a single line of code by the end user. In order to test this algorithm implementation, there are now three steps to be taken by the user. The first is compiling the design using Algo Builder. The second is programming the sensor tile box via USB cable. And the third is launching a unique UI for data visualization. But before moving to unique UI, I just wanted to highlight that Algo Builder can also generate the necessary C code of the algorithm implemented in the UI's block diagram. The conversion engine takes the XML block diagram data and converts it into C code. The C code then is visualized by clicking on the generate C code button and can be utilized by the end customer. Now opening unique UI to visualize the resulting vibration monitoring application, there are four display charts available featuring the XYZ acceleration data in the time domain and the FFT analysis for each of those three axes. The UI will also display the maximum acceleration amplitude and its respective frequency for each of the axes. And by the way, this example highlights the flexibility of Algo Builder and the professional working mode of the sensor tile box hardware platform. Note that we were able to explore vibration analysis in three different complexity levels, from entry to expert and professional levels of development. We will now go through a few additional points that are relevant to the sensor tile box professional working mode. Now specifically talking about the dedicated function pack for the sensor tile box, the FP SNS ST box one. It is an STM32 cube function pack dedicated for the pro mode of the sensor tile box. In this case, the package benefits from all the key components available within the hardware platform and it allows all the example software to run on the STM32 microcontroller and it also includes all the necessary drivers for the sensor tile box evaluation kit. The function pack features a complete set of examples of applications showing how to, for example, use ultra low-powered implementation based on an RTOS for transmitting the data via BLE connectivity to create a bootloader and an application for firmware over-the-air update to program the LSM-6DSOX machine learning core for activity recognition or vibration monitoring to easily send the data via BLE, to save the sensor data to the SD card and to visualize sensor data with the unique UI via PC serial terminal. And just to highlight, one of the key examples of the function pack is the data log extended. The data log extended streams data via USB or BLE to unique UI and also can be utilized directly with AlgoBuilder UI as well. The beauty of the data log extended binary that can be loaded into the sensor tile box is that it gives full access to each and every one of the registered settings of the sensors on board and it allows the user to take the maximum functionality provided by the unique UI. On top of the data log extended example, I would also like to spend a few minutes and highlight the BLE MLC application example. This implementation consists of a simple application example using the machine learning core functionality embedded in the LSM-6DSOX. I will detail the MLC in the next slide, but in this case, the machine learning core example is utilized to perform human activity recognition. The necessary settings and configurations to leverage the machine learning core are generated by our graphical user interface called Unico-GUI. It also requires a third-party software tool in order to design and customize the necessary decision tree. Now going to more details on the LSM-6DSOX smart sensor capabilities, the device offers two very flexible processing blocks. One is the finite state machines. Those are composed by a series of states, configurable parameters, resources and variables. In terms of flexibility, it is possible to implement up to 16 independent FSMs, and those can be executed simultaneously or sequentially. Additionally, the FSM can use data from accelerometer, gyro or an external sensor managed by the sensor hub functionality of the LSM-6DSOX. And the other processing block that is available within the LSM-6DSOX is the machine learning core. The MLC is a programmable logic in this case embedding a decision tree logic composed by a series of configurable nodes and each node is characterized by one if-then-else condition. The MLC allows up to eight decision trees that can be configured to run within the LSM-6DSOX. And the key advantage of using FSM and MLC is the significant reduction in power consumption at system level. You are now offloading your microcontroller and performing pre-processing within the IMU. Now specifically talking about the BLE MLC example, when programmed to the sensor tile.box, it allows the user to visualize through the mobile app interface different activities, in this case, standing up, walking, fast walking, and running. And by the way, if you think about it, this is a great example to show the capabilities of the device as the activity recognition is running inside the LSM-6DSOX in specific the machine learning core. And in order to give you more details on the machine learning core, we have prepared a series of YouTube videos highlighting a comprehensive design flow tutorial passing through the key steps of when utilizing the MLC. The key steps are data collection, labeling and selecting features, decision tree generation, and register configuration. And by the way, this is a great example and a great way to learn more about the LSM-6DSOX at your own pace. Additionally, there are also a few other machine learning core sample implementations available in our dedicated GitHub repository. We keep it updated with every new release and the link to access it is available in this slide. Another key element of the promode of the SensorTel.box is the dedicated function pack targeting artificial intelligence. The function pack is called FPAI Sensing One. And the package enables advanced applications such as human activity recognition or audio scene classification. And those are based on outputs generated by neural networks. The neural networks are implemented by a multi-network library supporting both floating and fixed-point arithmetic. And those are generated by the X-QBAI extension for the SCM32 QBMX tool. The neural networks in this package, by the way, are just examples of what can be achieved by combining the output of X-QBAI with connectivity and sensing components from ST. And in this case, benefiting from the high integration of the SensorTel.box hardware. I also would like to take this opportunity and introduce an upcoming function pack that will be part of the SensorTel.box ecosystem. The SPATR BLE One is a dedicated package targeting low-power asset tracking devices that rely on BLE connectivity and the hardware available within the SensorTel.box. The package includes a complete sample application on how to create an asset tracking application controlled by BLE connectivity, and it also allows the user to benefit from microSD card data logging. Additionally, the SPATR BLE One also takes benefit from the low-power accelerometer present in the SensorTel.box, for instance, detecting wake-up, field, and device orientation. After going through three very different operational modes of the SensorTel.box, I would like to recap a few of the key messages we explore during this webinar, especially when evolving with the tools from entry to professional working modes. The first point is that the SensorTel.box hardware allows the user to explore entry expert and professional modes, entry mode, with simple applications where the user can log and explore new sensor functionalities, expert mode, where the user has a little bit more knowledge and can test and validate ideas by customizing the mobile app. And third, the professional working mode, where the user can customize write code and make their own application based on tools such as the algo-builder and the dedicated function pack. It's also important to highlight that the SensorTel.box hardware not only benefits from the three different working modes at software level, as it also benefits from the open-source platform in terms of hardware documentation available on sp.com. For instance, the Gerber's Schematics and BioMaterials are available within the SensorTel.box web page. For more information, you can also access the dedicated YouTube videos on SensorTel.box. These videos walk through entry, expert, and professional modes in detail. On top of this content, for mass market support, please rely on our online support environment at myst.com slash ols. You can also learn more about our latest sensors on st.com slash sensors. And by the way, you can also visit our longevity web page highlighting the industrial-grade sensor product portfolio. And finally, we also have a dedicated repository for Android, Linux, and platform-independency drivers available for all of our MEMS and sensors. As takeaways, the Internet of Things brings many opportunities to enable sensors to cloud connectivity. And ST plays a key role in this marketplace by bringing the necessary building blocks, not only in terms of products ranging from consumer, industrial, and automotive, but also by providing development tools that jumpstart your understanding, evaluation, and development when targeting a new application. We always keep our commitment on MEMS innovation and we are constantly expanding our development ecosystem. And with that, I finalize this presentation. I would like to thank you very much for your time and attention during today's webinar.