Artificial intelligence will check the quality of AudiOctober 15, 2018
Audi began testing a neural network that detects cracks in sheet metal
Testing of an automatic quality control system using the capabilities of artificial intelligence began in the press shop at the Audi plant in Ingolstadt. This is a special software that runs under the control of a neural network and is able to detect the smallest cracks in sheet metal.
The platform program is based on machine learning algorithms that allow working with large amounts of unstructured data. For example, images. For several months, a team of specialists in artificial intelligence taught the neural network to recognize defects, while at the same time creating a base of examples. On their basis, the software has learned to identify and mark cracks independently and with high accuracy even on new images.
Now all the metal components leaving the press shop undergo a mandatory visual inspection. Additionally, they are checked using the program of recognition and image processing. In the future, they will be able to completely replace artificial intelligence.
In February, Porsche spoke about the new system, which will allow to detect problems with the car at an early stage. This is a neural network based on artificial intelligence that can recognize noise and vibrations, and use any deviations from the norm to interpret the nature of the breakdown.