Why Kion relies on technology from Nvidia

At CES in Las Vegas, not only technical innovations around the vehicle itself are presented, but also the processes behind the creation of the automobile are in focus this year. Logistics specialist Kion, best known for its forklifts, has now presented new ways to set up and optimize new and existing warehouses in the Nevada desert.

What does Kion hope for from Nvidia?

To this end, the company from Frankfurt am Main has brought on board not only consultants from Accenture but also a company that is currently highly regarded by many players in the automotive industry in terms of AI and simulation: Nvidia. And this goes beyond the context of new, powerful vehicle architectures: The US tech giant's Omniverse Platform is intended, among other things, to transfer factory planning at BMW or Mercedes-Benz into the virtual world.

Kion is now also relying on the expertise of the Californians. Together with Accenture, an AI-optimized digital twin named Mega was designed, which is intended to continuously improve intelligent warehouses with automated forklifts, smart cameras, and current automation and robotics solutions. “At Kion, we use AI-driven solutions as an integral part of our strategy to optimize our customers' supply chains and increase their productivity,” says Kion CEO Rob Smith. “Together with Nvidia and the digitalization experts from Accenture, we are redefining the supply chain.”

The advantages of digital twins for warehouse planning

The Omniverse platform and Kion's in-house development, Mega, are intended to enable digital twins that allow industrial companies to find the most efficient and secure configuration of their warehouse - all without having to interrupt ongoing operations for testing. The ideal configuration includes the optimal number of robots, employees, and automation systems, according to the companies involved. The digital twin serves as a test environment for all aspects of warehouse operations, including the arrangement of facilities, the behavior of robot fleets, and personnel deployment.

“Industrial metaverse is in its infancy”

Virtual factories, digital twins, and immersive training: The metaverse offers great potential for efficiency gains in automotive production. Learn more in the interview with doctoral candidate Judith Brenner.

Beyond simulating and testing various configurations, the digital twin is intended to help robots adapt flexibly to changing warehouse conditions. This includes fluctuations in demand and inventory as well as changes in spatial conditions. In combination with Kion's warehouse management software, the twin assigns tasks to its virtual robot models, such as relocating goods between buffer zones and storage locations. The virtual robots use the latest AI technologies.

For the CES presentation, the consulting firm Accenture developed an interactive application that shows how this training can optimize key performance indicators of warehouses. These include throughput rates, processing times, as well as the reduction of safety incidents and error rates. In the next step, the partners want to integrate a precisely tuned generative AI model for image and video analysis into the digital twin. This model should be able to monitor processes in real-time to prevent bottlenecks, accidents, and other unexpected events. For this purpose, cameras and robots will be coupled with Nvidia NIM - a platform for deploying generative AI models on edge devices.