Škoda has an application that recognizes car problems by sound

The Sound Analyzer application, designed for smartphones and tablets, records the sound of the car and compares it with the already available acoustic recordings. They will then recommend the next steps to the technicians.

The After Sales Department of Škoda Auto and Škoda Auto Digilab is testing a new Sound Analyzer application, which should significantly facilitate and speed up service with the help of artificial intelligence, which should in itself contribute to higher customer satisfaction.

The Sound Analyzer application works with sophisticated artificial intelligence, thanks to which it can reliably, unambiguously and quickly assess the condition of worn parts and alert you to necessary service interventions. To do this, the program uses various specific parameters of the car and analyzes the use profile of the car. It may sound complicated, but technicians only need to upload the sound of the car to an application on a smartphone or tablet.

If the application detects inconsistencies, it will start using algorithms to determine what could be causing them and then provide a specific description of the audio source. Currently, the technology recognizes ten patterns with an accuracy of over 90 percent. These are complex components, such as the control system, the air conditioning compressor or the DSG dual-clutch gearbox. However, work is still underway to recognize other acoustic patterns.

The technical basis of the application is a neural network algorithm, which first converts the audio recording into a spectogram and thus displays the acoustic signals graphically. Artificial intelligence then compares this image with already stored values ​​and detects deviations. According to predefined models, it then warns of the need for maintenance or repair.

The carmaker has been testing the new application since June last year in 14 countries, including important markets such as Germany, Russia, Austria and France. A total of 245 dealers participate in the pilot program, contributing to software learning and further program development by providing audio recordings.


Share on facebook
Share on pinterest
Share on twitter
Share on linkedin
Share on email


Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.