Microphones and acceleration sensors can be used to "listen in" on machines and detect anomalies at an early stage, for example leaks from pneumatic actuators. Festo and partners are using the AI-Music4.0 research project to determine how innovative sensor data along with artificial intelligence (AI) could be used in industrial applications. This new quality of data processing directly at the components could lead to the secure, decentralised analysis and reliable forecasting of the status of components or systems.
As part of the research project "Microelectronics-based Universal Sensor Interface with Artificial Intelligence for Industry 4.0" (AI-MUSIC 4.0), the project partners are developing innovative approaches to process acoustic sensor data with AI. Among other things, the sensors can "hear" leaks and detect vibrations as structure-borne sound. This means that they can provide important data for assessing the machine status.
Once the sensor data has been compiled, it will be evaluated in innovative chips using signal processing and AI. Each sensor unit could then send condensed information on the status of the component to a controller or gateway. These would only, for example, receive the message that there is or isn't a leak. This would save bandwidth and speed up condition monitoring. This means that the intelligence would be even closer to the components.
The analysed data could form the basis for recommended actions, for example if the process quality is not being maintained and the system should be stopped. This opens the way to a new generation of intelligent, autonomous production systems.
Festo takes on the role of the user in the research laboratory. Using various test setups and integrated sensor technology, we collect the data needed to teach the AI to detect leaks.