BMW’s virtual factory uses AI to refine assembly line


German car maker BMW plans to start manufacturing transmissions for electric vehicles at a large plant in Regensburg, Bavaria, later in 2021. Long before new parts roll off the production line, the entire manufacturing process will take place in incredibly realistic details in a virtual version of the factory.

The simulation allows managers to plan the production process in more detail than before, says Markus Grüeneisl, who heads production strategy at BMW. “We now have a perfect digital twin of our real-time production,” he says.

The simulation is part of BMW’s plan to use more artificial intelligence in the making. Grüeneisl says machine learning The algorithms can simulate robots performing complex maneuvers to find the most efficient process. Over time, BMW wants to use simulation to teach robots to perform increasingly complex tasks.

BMW used a software platform called Omniverse, developed by the chipmaker Nvidia, to recreate the production line in Regensburg. Last year, BMW said it was using an AI platform from Nvidia called Isaac to train robots for some new tasks.

“In the future, I’m very confident that we can just install a new robot in this facility and say, ‘OK, talk to the other robots and find the best way to produce this body,’” Grüeneisl says.

Manufacturers have been using computer simulations to refine their assembly lines for some time. But Omniverse makes it possible to simulate the entire production process with photorealistic details and physical properties such as gravity and different materials. It is possible to present the production process from start to finish and see how changes made to one part can have spillover effects on another. It is easier to create a more complex virtual environment because different 3D models can be imported into the system. Omniverse uses an open file standard compatible with many computer aided design packages.

The software will also simulate avatars of human workers grabbing parts and tools, and assembling components, to find the best procedure and minimize ergonomic issues. It could also allow fewer workers to complete a particular job, says Grüeneisl.

“We are AI-simulating the way people move around the factory,” explains Richard Kerris, general manager of Omniverse at Nvidia. He calls the project “one of the most complex simulations ever made”.

There is growing interest in using AI to control the robots and other industrial machinery. Encouraged by recent advances in AI, some startups are focusing on robots learning in simulation how to perform wickedly difficult tasks such as grab irregular objects, a technology that could eventually help automate much of e-commerce and logistics work. This often uses an AI approach called reinforcement learning, which involves an algorithm experimenting and learning, from positive feedback, how to achieve a specific goal.

“This is definitely the way to go,” says Ding zhao, a professor at Carnegie Mellon University who focuses on AI and numerical simulations. Zhao says simulations are crucial for using AI for industrial applications, in part because it’s impossible to run machines over millions of cycles to collect training data. Also, he says, it’s important for machine learning models to learn by experiencing dangerous situations, such as the collision of two robots, which cannot be done with real hardware. “Machine learning is data intensive and collecting it in the real world is expensive and risky,” he says.

Willy shih, a Harvard Business School professor who specializes in manufacturing technology, says the sophistication of simulation has steadily increased, and he says simulation primarily saves time and money by preventing future manufacturing problems.

Shih says there is a lot of hype around AI for manufacturing, but adds, “There are many, many applications” for the technology, too.

Nvidia CEO Jensen Huang discussed BMW’s use of the Omniverse during his speech at the company’s annual GTC conference, which was held virtually Monday. Nvidia initially designed graphics chips for games, but widened its scope when these chips proved suitable for driving AI programs. The company has since branched out into several other industries where AI is important, including automotive and medical imaging.


More WIRED stories

Leave a Reply

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