INDUSTRY

Host: Hello, and welcome to this training on Edge+ Solutions for the Future of AIoT. Today, we’ll be hearing from Thomas Kaminski from Advantech who will explore how embedded computing and AI are fueling innovations and how you can optimize your next AIoT project with Advantech’s next generation solutions. Let’s get started. Thomas, take it away.

Thomas Kaminski: Hello and welcome to my brief introduction of Advantech’s latest Intel products and our basic strategy and vision about the future of artificial intelligence and IoT. My name is Thomas Kaminski and I’m leading the product sales manager department of Advantech AIoT in Europe. The edge computing market is having huge potential for the next years in many different applications and ranges from basic research to business industry, transport, services and our personal daily life.

In the upcoming years, we will have billions of connected devices at the Edge. We will see more and more large amounts of data transferred from the Edge into the cloud, and the software will be provided as a service. Already today, we see new highly automated factories with highly efficient processes, driven by a lot of artificial intelligence. Looking on an example, autonomous mobile robots, getting goods on demand and just in time from its deposits to the production lines. They are watching carefully their environment to cooperate together on time to ensure a safe flow and delivery of the goods. That brings new challenges for real time computing, data transfer and image processing. Also safety aspects as an important function shouldn’t be ignored.

We at Advantech took this challenge as a strategy from the edge to the cloud with a lot of additional hard and software services. Following our Edge+ strategy, we develop all our platform, solutions and peripherals for dedicated vertical markets. In addition to that, we offer various software tools and services for artificial intelligence and other industrial applications, which can shorten your time to market and helps you to be focused on your business. The hardware remains as an important part, but the future requires more and more software skills.

Our target markets are medical, 5G and networking, automation, mission critical, and infotainment and retail. We are starting to approach this markets with our board-level portfolio as one solution. The portfolio covers all form factors like COM modules based on x86 or SCPUs following the new standard COM-HPC and mother or single boards. With our design and manufacturing services, we can offer customized solutions out of it. As the member of the ASCAT, we collaborates to define new standards for future demands. The COM-HPC standard is a good example. This new COM standard fulfills the demands for higher data transfer, higher computing power on COM modules, and supports more DRAM and more I/Os.

Beside of our industrial board-level products, we are offering a comprehensive portfolio of ready-to-use edge to cloud solutions. You don’t need to care about integration of components or server concepts or certifications of the computing devices, BSP and application deployment or other accessories like storage or memory modules. The key factor of an AI solution is the software, dedicated to the edge intelligence systems.

Our available Software apps and services can be easily used, deployed for the customer’s application. Just to highlight a few of the available apps. The ePaper can manage and control your installed ePaper displays with a integrated wireless connectivity in a warehouse or supermarket or in another environment. DeviceOn Kiosk+ offers grouping management, zero-touch onboarding, over-the-air updating, and remote control capabilities for peripherals of security device.

This SMARC kiosk solution optimizes network management, reduces costs and minimize employee overheads. Signage of content management software, our server-based content management software. You can simply upload your content on a server, schedule it for all connected signage players. Supported by the Intel VPU and our edge AI suite, FaceView is our recognition solution. The app has an international certified currency rate of 99.7%, and it’s suitable for various applications where you need to detect something. The app already offers functions like masked people or object detection, which can be really helpful in the retail environment today, especially in this difficult pandemic times. Recognition results can be obtained on the new Intel platforms in real time at the edge. An additional GPU card can accelerate the results in detecting many objects in real times from several cameras.

Advantech’s Edge AI on Intel platforms is a very powerful tool using the OpenVINO AI models, which enables you to develop various use cases and applications in a short time, or simply using your already finished apps and integrate them into your application. Edge AI can collect various machine data from the edge sensors. The app supports industrial protocols like OPC UA, Modbus or ODBC, and as the graphical user interface for the protocol configuration. The app also supports the data visualization. Integrated dashboard templates helps you to realize that initial time.

Looking on the news upcoming products based on the Intel x6000E Atom Series with the code name Elkhart Lake. Already, this class of new CPUs offers more computing power than the previous generations, and is prepared for you real-time computing. The new features of time-coordinated computing can be perfectly used to sync all your edge devices. And the integrated function Safety Island can manage redundant device information. We are launching several platforms and solution based on this new Atom platform.

The SMARC and Mini-ITX board on the left side, and the SDM Player, you can see on the right side, is already available for sampling and follows the mass production schedule of Intel. In Q2, we are going to launch the three and a half inch SBC MIO-5152 and COM Express Mini Module. And later on this year, the Fanless ARK computer inbox. The Mini-ITX board AIMB-218 will be integrated in our center compact system and will be launched mid of this year on a Celeron SKU first.

Based on the 11th generation of Intel’s core CPU, we are going to launch a lot of new models, versions and solutions. Up to four cores and the clock frequency of 4.4 gigahertz, this new CPU will start soon in mass production. Same as for the new Atom CPUs, we will have the time coordinated computing on selected SKUs, functional safety or the integrated error code correction. The new Iris Xe graphics supports four screens with 4K resolution and HDR, or two screens with 8K resolution at SDR.

USB 4.0, or you might know it as Thunderbolt will be supported. Four PCI Express Gen 4 lanes can be perfectly used to connect the high speed SSD. The Integrated Intel Deep Learning Boost supports artificial intelligence applications. Also here, we are starting with the COM Express Type 6, and the COM Express Mini Type 10 Module. Our product division for the single board computers will launch Pico in a three and a half inch version during this year. And we are going to have three finished devices for different vertical markets. So rugged ARK system, the Signage Player DS-085, and our EI system for various industrial applications supporting the EdgeX Foundry.

Altogether, there are a lot of product launches from us in 2021. You might need to know that every product category is driven by an own product division with its own R&D capacity and their vertical target markets. This leaves us in a strong position to develop that much product at the same time and target different vertical markets.

Summarizing the information from the previous slides, we are providing a lot of different platforms, mainly working at the edge, collecting and processing data. Our WISE-DeviceOn is a strong tool, which is available with different apps to collect and to process data. It also provides the connection into the different clouds, where you can get access to many additional services for cloud computing. For example, the Microsoft Azure Cloud. As a Microsoft partner, we can prepare the ground and get you connected to dedicated solution architects to finalize the journey from the edge to the cloud.

On this last slide, I want to show you a typical use case, collecting data, such like optical inspection results from several cameras connected to a VPU card in the production line, supported by all phase UI. Selected data from this edge device, it’s transferred for further analytics to the training server. Results can be analyzed and evaluated with an external on-prem software, looking on one application of an automatic glass inspection. Cameras are scanning the glass surface for optical failures and damages before it’s allowed to move into the further processing. The device at the edge is deciding if the glass is good or not good. Means, if it will move into the production. You can imagine that the detection of a non-good glass in the production process themself will cost a lot of efforts and costs. Therefore, we want to keep it outside.

But there will be always borderline cases in the decision-making of the edge device during the automatic inspection process. The on-prem training server is collecting the information about these cases, helps the user or an artificial intelligence to add another decision-making models into the process at the edge. Collecting and display the data with our iEdge app from the motion control about speed, machine temperature or load can be used for predictive maintenance or get synced with other devices. Having an edge to cloud server connected to Azure or other cloud providers, you can share production data with other facilities or locations, or for example, with sub-suppliers to get the supply chain and logistics synced with the production lines. So I’d like to thank you for your attention. And now we are going to open the Q&A session, and please don’t hesitate to raise your further questions. Thank you very much.