Stuttgart  /  May 09, 2023  -  May 12, 2023

35. Control - International Trade Fair for Quality Assurance

Meet us at the Fraunhofer-Vision booth, hall 7, booth 7301 and learn about our research topics in the field of "Acoustic Monitoring for Production".

At the trade fair, we will demonstrate how acoustic methods based on airborne sound analysis and artificial intelligence can be used for quality assurance of products and processes in industrial manufacturing.

Assuring quality, stabilizing processes, extending test method with acoustics

AI-based acoustic monitoring (amo) can provide great added value in process monitoring, in-line and end-of-line quality control of products, and predictive maintenance applications. In production, problems regarding unexpected machine downtime, production of bad or destroyed products, and low automation rates can be solved using amo.

With amo, we want to optimize the user's production, and with this non-destructive and non-contact testing method, stabilize processes, avoid recourse, and reduce faulty production. AI-based acoustic monitoring aims to be integrated into existing facilities in a minimally invasive way, adding value where optical monitoring methods, for example, reach their limits. 


AI hears and classifies faults correctly

The successful application of this innovative test method has already been demonstrated in the in-line quality control of weld seams. Our scientists are currently conducting research on approaches for monitoring different manufacturing processes, including welding and machining. Expert knowledge of suitable sensor setups, sensor data fusion, processing of sensor data without connection to an external cloud, as well as energy-efficient AI models, play a significant role in all their approaches.

With our research, we want to shape the production of the future, avoid errors, use resources efficiently, and protect our environment.

Our principle demonstrator for acoustic event detection (AED)

The new acoustic monitoring is demonstrated on an air-hockey table. Three pucks are used, which are made of different materials and cause different "pling" sounds as soon as they hit the rail of the game table. During play, these acoustic signals occur frequently and irregularly and can be analyzed and classified using machine learning techniques. This entertaining principle demonstrator highlights the potential of acoustic monitoring: providing near-real-time, reliable analysis results - even in challenging acoustic environments - without direct contact to the measured object (pucks).

We are looking forward to meeting you at our booth!