Techtalk

Special

More intelligence into the device

AI and ML do not stop at embedded systems. Ginzinger electronic systems now offers the robust platform for different industry segments to use the added value of artificial intelligence directly in the devices.

Andreas Pfeiffer Portrait
by Andreas Pfeiffer

Artificial intelligence (AI) and machine learning (ML) permeate all areas of our lives. In our everyday life with Apple, Google and Co, we are constantly confronted with these systems. AI and ML do not stop at embedded systems. Ginzinger electronic systems now offers the robust platform for different industry segments to use the added value of artificial intelligence directly in the devices.

What can artificial intelligence and machine learning do in embedded systems? The best-known examples are voice control and image recognition.  Being able to operate a device error-free without having to touch it is not only a great advantage in environments with increased hygienic requirements. Image recognition using AI supports users and helps to avoid operating errors. The evaluation of large amounts of data and continuous data streams using AI allows patterns to be recognized and suitable actions to be triggered.

These systems have long been established for the maintenance of machines and plants. Vibrations that deviate from normal operation are identified, and the user is alerted to anomalies or potential damage in good time. In many highway tunnels, microphones are now installed that use AI to alert the tunnel control room staff to suspicious noises.

Visualisierung Sprachsteuerung

What all these use cases have in common is that very large amounts of data need to be processed with high computing power in a short time. This is why most AI and ML applications run in the cloud. The voice assistants of cell phones are the best known example. After identifying a keyword, the voice data is sent to the cloud and analyzed there.

In many application areas, processing the data in the cloud is not possible or desired. Data protection regulations speak against it, or it is not an option to network the device with the cloud. Running AI and ML algorithms directly in the embedded system therefore opens up completely new possibilities. More and more application processors and microcontrollers today offer the necessary functions and performance for this.

AI Embedded Platform

unimagined possibilities

Ginzinger electronic systems now provides the robust embedded platform for artificial intelligence directly in the device. This opens a wide field of unimagined possibilities for new business areas. The platform consists of powerful hardware modules paired with the GELin Embedded Linux Suite, proven in numerous industrial applications. Powerful machine learning tools are available to developers to realize AI and ML applications including model building in a short time.

Ginzinger Embedded NXP iMX6 DevBoard

Ginzinger i.MX 8M Plus Evaluation Kit

  • NPU - hardware acceleration for neural network computation
  • Numerous sensors on board
  • Processing of video streams
  • Display Extension and HDMI
  • Connectivity with Gigabit Ethernet, CAN FD and USB 3.0
  • Robust Ginzinger Embedded Linux Suite GELin
  • Customized kernel, extensive tools and libraries
  • Numerous sample projects

The hardware platform with Neural Processing Unit (NPU) allows demanding applications offline and directly in the device without depending on the availability of network connections or cloud computing. Sensitive data is protected to the maximum and remains local.

For a quick start into the world of AI and Machine Learning, Ginzinger electronic systems offers turnkey evaluation kits with sample projects. Ginzinger webinars on the topic and tips and tricks from Ginzinger developers allow you to quickly try out the new possibilities. The Ginzinger experts are also happy to advise device manufacturers on the potential, usage limits and cost-effectiveness of AI in their applications.

DevKit

More about AI in Embedded Systems