© Adobe Stock / KPs Photography - stock.adobe.com
The research project ALERT is developing a novel sensor system for monitoring industrial plants – energy-autonomous, wireless, AI-driven, and maintenance-free. At its core lies an intelligent sensor node that detects vibrations and acoustic signals, harvests its own energy (Energy Harvesting), and analyzes the captured data locally in real time.
With ALERT, a modular toolkit for retrofit, self-powered sensor systems is being developed – specifically designed for machine monitoring in industrial plants, infrastructure facilities, and production environments. The focus is on energy efficiency, durability, and minimal installation effort.
Power Management ICs are microchips that control the energy flow within a system.
In the ALERT project, they ensure that even the smallest amounts of harvested energy are used as efficiently as possible.
The unique strength of ALERT lies in the combination of edge AI, energy harvesting, and adaptive wireless communication.
Adaptive communication means that the sensor automatically switches between different wireless modes depending on its energy status or operating phase – for example, between data-intensive setup mode and energy-efficient normal operation.
This enables the development of robust, retrofit-capable sensor systems – ideal for condition monitoring without the need for wiring or battery replacement.
Condition monitoring refers to the continuous monitoring of machines and industrial equipment to detect deviations or faults at an early stage.
The goal is to prevent failures and plan maintenance activities proactively and efficiently.
The project addresses several central challenges:
On-device training refers to the retraining of AI models directly on the sensor or device, without the need to transfer data to an external cloud.
This allows the model to continuously adapt and improve using new measurement data.
Advantages:
Fraunhofer Institute for Integrated Circuits IIS