BMW has taught artificial intelligence to hunt for defects in manufacturing
The BMW Group's Regensburg plant has launched the "GenAI4Q" pilot project to answer the question: can artificial intelligence make vehicle inspections faster, cheaper and better?
Here's What We Know
It looks like the answer is in the affirmative. BMW, in collaboration with startup Datagon AI, has developed a system that makes personalised recommendations for quality control for around 1,400 cars a day.
Every 57 seconds, a new car rolls off the assembly line in Regensburg. All cars are assembled for specific orders for the global market, which makes the inspection task more difficult. Petrol, hybrid and all-electric models are assembled on the same line, as well as a multitude of equipment variants.
A demonstration of the GenAI4Q in operation. Photo: BMW
The AI analyses mountains of data about a particular car: model, options, actual condition in production and automatically generates a list of individual checks. The checks are organised via a mobile app. Employees can also access additional functions, such as a microphone: a detected problem can be dictated immediately and the AI will translate the speech into text and encode the comment.
According to Armin Ebner, plant manager at the Regensburg plant, the implementation of "GenAI4Q" in quality control is part of BMW's transformation towards smart factories.
Source: BMW
