Artificial intelligence as a quality booster

Can artificial intelligence (AI) make quality control in vehicle production more efficient, faster and more reliable? A software solution developed at BMW Group Plant Regensburg as part of the “GenAI4Q” pilot project is making this a reality. At its core is an AI that delivers tailored inspection recommendations for the approximately 1,400 vehicles manufactured each day.

“The use of artificial intelligence supports the digital transformation of BMW Group production towards an intelligently connected factory – for example, we are using AI for quality control in vehicle assembly. In this way, we are optimizing our production processes and creating added value for our products and, ultimately, for our customers” says Armin Ebner, head of BMW Group Plant Regensburg.

BMW vehicles meet the highest standards of product quality. To ensure this, Plant Regensburg conducts numerous quality tests throughout the production process. Shortly before leaving the plant, vehicles undergo a final inspection: Trained specialists examine each finished vehicle once more to ensure it meets premium quality standards.

A new vehicle rolls off the Regensburg assembly line every 57 seconds. Each vehicle is built to individual customer specifications for the global market. Different types of drive trains are flexibly manufactured on one production line – ranging from vehicles with internal combustion engines to plug-in hybrids and fully-electric models, as well as countless model and equipment variants. Virtually no two vehicles are alike. “Our AI tool generates an individual inspection catalogue for each specific customer vehicle,” explains Rüdiger Römich, responsible for Test Floor and Finish in Regensburg’s vehicle assembly.

The AI analyses vast amounts of data to create these custom specifications. This includes not only vehicle data, such as model and equipment variant, but also real-time production data for each specific vehicle. By recognising patterns and correlations, the AI system quickly and automatically determines the scope of the inspection and organises it intelligently, in the right order, within a smartphone app. Römich: “Intuitive usability makes it easy to record findings. Employees can also access additional functions when needed – for example, a microphone icon in the app activates voice recording.” The AI also manages speech recognition and transcription using standardised coding.

AI-based quality inspection was developed at BMW Group Plant Regensburg in collaboration with Munich startup Datagon AI.