Machine listening for smart decisions

In the automotive sector, we are researching innovative acoustic functions. These include scene recognition, for example for the siren of an approaching emergency vehicle or a nail in a tire. Acoustic sensors at various points in and on the car send data to an analysis unit. Recordings from the wheel arch, for example, can provide information about the condition of the road and enable automatic chassis adjustments to be made to the suspension or brake system. Continuous acoustic monitoring of the car's operating noises also allows damage or wear to be detected. An “uneven” engine running or worn brakes can be identified in advance, which is referred to as predictive maintenance. This enables smart maintenance intervals and repairs before breakdowns or even accidents occur. The Hearing Car thus makes an important contribution to research into autonomous driving.

Bodily sounds can provide clues about someone's health. That is why we are conducting research into the long-term monitoring of bowel sounds using a multi-channel acoustic sensor. The aim is to gather information that could enable the early detection of post-operative complications, for example. The sensor is intended to be used during and after surgery to record body sounds in general and intestinal sounds in particular over a long period of time.

In addition, acoustic monitoring systems from Fraunhofer IDMT-HSA detect machine damage in industry and plant engineering before it occurs. This prevents downtime and makes maintenance more predictable (predictive maintenance). In addition, noises occurring in many areas can be assigned to specific events. In quality assurance and end-of-line inspection in industrial production, audible damage, such as the smallest material fractures in computer chip manufacturing, can be detected.