AI-based assistant detects noise

December 14, 2020 //By Christoph Hammerschmidt
AI-based assistant detects noise
Porsche Digital, the sports car manufacturer's subsidiary, has developed a digital assistant for detecting acoustic noise in vehicle component tests. The aim is, among other things, to improve development and production of components in the automotive industry.

Porsche's technology, which is based on Deep Learning methods, detects noise more reliably than human ears during endurance tests, for example. This relieves the burden on development engineers in particular, who have to be present during such tests. The assistant documents errors and simplifies root cause analysis.

"With our 'Sounce' system, we can reliably check the noise development of components under load around the clock and under a wide range of conditions. This enables us to improve the analysis options in early-phase component tests," explains Patricia Rennert, Head of Industry Solutions at Porsche Digital . Using artificial intelligence (AI)-based acoustic testing, quality is increased and costs are reduced across a wide range of use cases.

The technology is based on the principle of anomaly detection and can be applied in various areas. Its use is particularly conceivable in testing situations where a large number of different acoustic signals make it difficult to analyze by hearing alone.

The Industry Solutions division of Porsche Digital designed and developed the software-as-a-service (SaaS) solution together with the development department of Porsche AG. Porsche Digital is responsible for operating the technology.

Following pilot deployment, the solution is now also being offered to external customers.

More information: www.sounce.io

Related articles:

Radiator grille generates artificial engine noise

After eyes, cars get ears for the environment

Israeli startup develops noise-resistant voice control for cars

 

 


Vous êtes certain ?

Si vous désactivez les cookies, vous ne pouvez plus naviguer sur le site.

Vous allez être rediriger vers Google.