On September 24, 2026, the transformation network transform.r hosted a working group meeting on embedded AI in manufacturing at TechBase Regensburg.
The highlight of the event was a presentation by Andreas de Jonge, CEO of de Jonge GmbH, on 3D-vibration analysis using EdgeBrain. Using a compact demonstrator, he showed how vibrations are captured via a triaxial vibration sensor with an IO-Link interface and transmitted to the EdgeBrain module via MQTT.
EdgeBrain is designed as a plug-and-play inline AI module that can be integrated into existing interfaces between sensors and hosts and processes data streams locally in real time. Thanks to its “offline-first” architecture, no cloud connection is required for initial analyses.
The demonstrator illustrated how edge AI can be integrated into existing infrastructures via established industry standards such as IO-Link and MQTT. During operation, deviations in vibration behavior—such as those caused by intentionally introduced imbalances—were displayed in a clear and traceable manner.
During the discussion that followed, participants discussed potential applications, technical requirements, and possible use cases in production. The group also looked at possible next steps for the Embedded AI working group.