Audi Calls Upon the Powers of AI To Generate Fresh and Limitless Rim Designs

If you work in anything creative-related, you may feel that inspiration comes and goes. Well, that’s certainly not the same when a bunch of ones and zeroes think for you. And it might just come up with solutions that we wouldn’t have ever thought about.

FelGAN was developed in-house in a collaboration between Audi’s IT and Design departments to gain fresh perspectives and evolve their work. The name is a mash-up between “Felge,” the German word for rim, and “GAN,” an acronym for Generative Adversarial Networks. GANs are a particular type of self-learning computer program which consists of two algorithms competing for better and better results in a so-called training process.

So, how does FelGAN work exactly? One of the algorithms acts like a “generator” by creating artificial images of vehicle rims. The other algorithm, the “discriminator,” analyzes a set of both real and generated images and decides which is which. The process is repeated until training is finished. FelGAN is essentially an idea hub where the design team plays with the shape, color, surface, structure, and other elements and comes up with new variations of rims.

As you’d expect, the algorithms learn from their mistakes and continuously adapt. After enough runs, there’s a low chance you’ll be able to distinguish what’s real from what’s machine-generated.

The project’s user interface is based on Streamlit tech, which creates short development cycles and a quick feedback response from both the departments involved. Moreover, the whole process requires a lot of processing power, so the software is run in the cloud.

Here’s where it gets even more interesting – FelGAN assigns a mathematical value for each of its designs, which the developers also refer to as “DNA.” The values can be used anytime to reproduce specific designs. Audi’s team can also feed the program with their photos and designs, and it will assign the appropriate DNA values for them.

Still, in most cases, only individual elements from FelGAN creations end up being selected and used by the designers. They are incorporated and refined into a harmonious design. Even though technology is rapidly advancing, experience in the industry plays a more critical role. They bring the virtual detail to life by producing a wheel prototype out of aluminum or plastic.

Audi is on its way to turning digital, and it’s demonstrating it has both the know-how and the tools to use software and AI to create value. For instance, late last month, Audi revealed it took part in an extensive drone research project that helped preserve the environment by thoroughly analyzing meadow orchard fields.

The company’s Head of Machine Learning & Data Science, Thomas Knispel, said, “In the modern age, data brings immense added value for companies and their employees. Audi has committed to the goal of becoming a data-driven company. To this end, we are going to use AI in many departments. So our data team is always on the lookout for new technologies.”

Looking ahead, Audi is working on developing an AI rating system that assesses each software-generated rim’s carbon balance, as the company is keeping sustainability in mind.

FelGAN technology might be used to create new AI design platforms meant for other Audi departments. Think about it, who knows how long it takes until we end up having our car interior designed by AI? For now, we’re definitely keeping an eye on the German automaker to see what they come up with next.