Intelligent Quality Analysis Using AI

CASE STUDY:

AI-Powered Intelligent Quality Analysis

COMPANY: Client from the Industrial Automation Sector
INDUSTRY: Automotive
CHALLENGE: Lack of qualified personnel with the required education and expertise.

In the modern automotive industry, quality control of components must be fast, repeatable, and resistant to human error. One of the key challenges in the production of automotive seats is the analysis of sounds generated during the movement of seat mechanisms. Even minor acoustic irregularities may indicate assembly defects or quality issues that can affect user comfort and safety.
To address these challenges, an intelligent tool for the automated analysis of automotive seat sounds was developed using artificial intelligence and advanced audio signal processing technologies.

Challange:

The objective of the project was to create a modern application enabling automatic quality analysis of automotive seats, detection of acoustic anomalies using AI models, visualization and playback of measurement recordings, operation both offline and online, the ability for users to independently train models, integration with industrial automation systems, and integration with measurement stations.

Project objective:

The goal of the project was to develop a modern application capable of automatically performing qualitative analysis of car seats, detecting acoustic anomalies using artificial intelligence models, visualizing and playing back measurement recordings, operating in both offline and online modes, allowing users to train models on their own, and integrating with industrial automation systems and measurement stations.

The scope of work included:

Development of a desktop application for the analysis of .WAV audio files.

Implementation of signal processing algorithms and spectrogram generation.

Integration of previously trained artificial intelligence models.

Creation of a measurement data visualization module.

Implementation of audio recording playback functionality.

Development of a module for training custom AI models.

Implementation of communication with PLC systems.

Deployment of an online operating mode for measurement stations.

The result was a unified control environment that integrated all key elements of the technological process.

Applied Technology:

The system analyzes .WAV files containing recordings of sounds captured in the noise measurement cabin. The audio data is automatically processed into spectrograms, which are then analyzed by an artificial intelligence model.

The user has the ability to view the generated spectrogram, listen to the recording, compare analysis results, and add new seat references without the need to modify the application.

An important part of the project was also the implementation of online analysis. The system monitors files appearing at the measurement station and automatically performs real-time analysis. The test result is transmitted directly to the PLC controller, which can approve the measurement or require the test to be repeated.

Summary – Implementation Results:

The implemented solution enabled a significant reduction in quality analysis time, increased repeatability of results, elimination of subjective operator assessments, automation of the quality control process, easy expansion of the system to support new seat models, reduction of costs associated with software modifications, and integration of the quality control process with industrial automation systems.

Thanks to the use of artificial intelligence and acoustic analysis, it was possible to create a scalable and modern system that supports quality assurance processes in the automotive industry.

At MJ Group, we build robust systems, design them with the future in mind, and consistently develop them in line with market needs.