Through edge-to-cloud integration framework, Advantech helps enterprises realize an AIoT blueprint

The growth of AIoT (Artificial Intelligence IoT) is at an all-time high. High-tech manufacturing industries and traditional manufacturing industries, are eager to establish an industrial IoT with AI as the core. However, the ultimate goal of AIoT is to realize the value of data, which entails a complex process from edge devices to the cloud. Alan Kao, Product Manager, Advantech Embedded-IoT Group, pointed out that the starting point of AIoT is data collection and integration, followed by visual presentation of big data on screen. These steps might seem easy to complete, but in fact, these steps are full of complexity. This includes consideration for different fields, equipment and needs; which means different data integration modules are needed, not to mention the following requirements of visualization such as; asset performance management, analytic models and algorithms, and edge computing.

In order to establish the AIoT link, enterprises need to invest a lot of resources, manpower and time in order to carry out complex integration and development. In order to simplify the process of edge-to-cloud AI-powered industrial IoT application, Advantech integrates its WISE-PaaS, edge intelligence, AI and other solutions into an overall framework.

EIS+ AI + DeviceOn, ensure the stable and intelligent operation of the devices

According to Alan Kao, the edge of the device includes several key components. Firstly, the hardware part is based on Edge Intelligence Server (EIS). It not only provides common RISC, Intel Atom and other architectural options, but also models with Intel Core I 3 / 5 / 7, Xeon and other high-level processors; as well as related software components of data integration and device management. Thus, making it the first set of edge computing systems in the market that support an advanced x86 architecture. It is also worth mentioning that EIS supports Modbus, OPC UA, ODBC and other communication protocols by default, meeting 70-80% of equipment data integration requirements. For other communication formats, it provides SDK (example code and original code) to help user’s quickly complete development and integration.

Another important software component is WISE-PaaS/DeviceOnfor collecting and visualizing the operating status of each EIS device, and use OTA upgrades for EIS software, firmware and BIOS. In addition, EIS is embedded with the Node-RED logic engine, which is used to edit the edge devices effectively. As for the introduction of artificial intelligence into the edge of the device, Advantech has cooperated with Intel and loaded Intel's Movidius X VPU into Advantech's VEGA module. This can be embedded in EIS, to produce more powerful AI acceleration effects combined with Edge AI Suite. Edge AI Suite is a software tool for AI development. Directing the users to quickly realize AI functions in application scenarios.

One-stop service for edge to cloud and the implementation of AI applications

Sam Chuo, the business development manager of Advantech WISE-PaaS, added that Edge Intelligence can further connect WISE-PaaS Cloud, forming a one-stop AI service from edge to cloud. For example, Everest Textile Co., Ltd. at the beginning, it only introduced EIS and WISE-PaaS/DeviceOn to collect the data of membrane manufacturing machine. Which met the demands of production line management. When Everest learned that the WISE-PaaS industrial IoT cloud platform not only provided the WISE-PaaS/Dashboard to meet the demand of data visualization, but also the WISE-PaaS/APM for I/O point management of machine parameters. The WISE-PaaS/AFS for AI prediction and analysis, it cooperated with Advantech in POC. Which can upload the data of the membrane manufacturing machine to WISE-PaaS to monitor the key parameters, master the change of yield, and even develop AI predictive quality analysis model. Then it can put it in EIS for real-time computing to predict whether the quality of each batch of membrane meets the yield target. This example fully demonstrates the application value of the framework of EIS+DeviceOn+WISE-PaaS=Edge-to-Cloud Intelligence & AI.

So far, "EIS+DeviceOn+WISE-PaaS" has not only been used for yield prediction within the plant, but also for the production end of the device, said Alan Kao. Tong Cheng, whose reputation in the industry for its SWAN air compressor, has embedded EIS and WISE-PaaS on its sold machines. This enables its customers to read the air compressor status information, as well as external data such as; voltage, current, temperature, air tight valve pressure, etc. Thus, achieving the first stage of software service added-value. For the second and third stages, the company expects to upgrade from passive repair to active service. Then gradually promote business mode transformation, from an equipment provider to a service provider.

Now, semiconductor manufacturers begin to introduce Intelligence & AI integrated solutions. The main reason for this is that the semiconductor factory needs to test the components, and then send the test data to the background for analysis. Originally, the commercial computers were used to collect data, but there was a problem with the stability. Then, the EIS was introduced to collect the data, and with WISE-PaaS/DeviceOn to monitor its mainframe, network transmission and software operation. All of these make the semiconductor factory the group with the greatest demand for WISE-PaaS/DeviceOn.

To sum up, by combining WISE-PaaS, edge intelligence and AI, Advantech forms the most complete edge-to-cloud AIoT solution framework at present. And it even uses high-level EIS models to handle complex edge computing and AI reasoning tasks. So now enterprises can quickly apply AI into the field of intelligent manufacturing.