ASSIST Software Wins Volkswagen Challenge for Automated Quality Inspection

AUGSBURG, Germany - PrZen -- The Volkswagen Automated Inspection Challenge tasked participating teams with automating final quality inspection at Volkswagen's engine plant in Chemnitz. Organized by Volkswagen Group Innovation and Startup Factory boOst, the competition required inspection of 157 drill holes and 24 surfaces across all six sides of a cylinder head.

Teams demonstrated detection performance using real component samples. Evaluation considered detection performance, processing speed, and cost-effectiveness, alongside requirements for integration into a production line.

Applying Expert Judgment to Automated Quality Inspection

ASSIST's engineers approached the task by examining how experienced inspectors assess a component. A mark on reflective metal may indicate damage, an acceptable variation, or a reflection. Distinguishing these requires clear images and an understanding of the feature being inspected.

"Our goal was to build a system that approaches inspection as an expert would. An experienced inspector considers the context, checks the component from the necessary angles, and knows when something needs a closer look. We wanted to account for those details, including the ones a machine might otherwise overlook," said Alexandru Boca, Robotics Expert and Head of the Mobile Department at ASSIST Software.

Recessed features and reflective surfaces complicate industrial computer vision. Lighting and positioning affect what cameras capture, while additional viewpoints increase inspection time. Teams had to obtain enough detail for reliable assessment within the practical limits of the production process.

Meeting Automotive Manufacturing Requirements

According to the challenge organizers, hundreds of thousands of cylinder heads undergo manual inspection at the Chemnitz plant annually. The competition sought an automated solution that could meet quality requirements while allowing experienced employees to focus on other tasks.

Teams also had to consider the consequences of incorrect decisions. A missed defect can allow an unsuitable component to continue through production. A false alarm sends an acceptable part for additional checking, adding work for factory staff.

The program included problem analysis, prototype development, and testing before a final demonstration and jury evaluation. Volkswagen Group Innovation provided feedback and assessed results, while Startup Factory boOst coordinated the process.

ASSIST's delegation included Gabriel Tironeac, VP of Innovation; Alexandru Boca, Robotics Expert and Head of the Mobile Department; Tudor Andronic, Deputy of AI; Cristina Tironeac, Communication Manager; and Teodor Drobotă, Business Operations Specialist.

The recognition gives ASSIST a concrete automotive reference for its automated inspection capabilities. Evaluated against a defined manufacturing task, the team's work demonstrates the relevance of its engineering expertise to industrial quality control and supports the company's focus on Physical AI.

About ASSIST Software

ASSIST Software is a software engineering company. Its capabilities include industrial computer vision, robotics, digital twins, and artificial intelligence.