Researchers develop AI chip for devices without cloud
The smartphone queries the cloud, the camera queries the server, the machine queries the data center. Almost all smart devices first send their data to the Internet before they can “think.” That’s set to change in the future. In the EdgeVision research consortium, researchers from the Zittau/Görlitz University of Applied Sciences (HSZG) are working with partners from academia and industry on a chip that brings AI directly into the device.
The project is called EdgeVision. “Edge” refers to the location where the data is generated – for example, in a camera, a sensor, or a measuring device. In the future, AI will perform computations right there, rather than first sending data to a data center or the cloud. This saves time, protects sensitive information, and makes devices less dependent on a stable internet connection.
A Chip for the Devices of Tomorrow
At the heart of the project is a new, energy-efficient chip with embedded AI. It is combined with optimized sensors and specialized optics. The result is a compact platform capable of independently analyzing images or measurement data, for example in mobile devices, smart metering systems, or technical equipment operating directly on-site.
Under the leadership of Prof. Dr.-Ing. Knut Meißner, the HSZG is contributing its expertise to the development of the chip and the testing of the platform. The university is also establishing a so-called Vision Lab, a space where new AI applications will be developed, tested, and presented to companies.
Saxony’s economy is set to benefit
EdgeVision is a collaborative project. Several companies and research institutions are collaborating. The project brings together three fields of the future: artificial intelligence, microelectronics, and digitalization. For the Zittau and Görlitz region, this means not only new research but also the opportunity to bring new technologies to businesses and products more quickly. What is developed in the lab is set to be put into practice soon.
However, the overarching goal of all partners is clear: In the future, AI should be used where the data is generated – quickly, energy-efficiently, and without a constant connection to the cloud.
