K3AI Collaborates with Siemens and KETI to Develop Physical AI Manufacturing Infrastructure Integrated with NVIDIA Omniverse

K3AI is set to actively implement a physical AI-based smart manufacturing infrastructure through ongoing international joint research with Siemens, the Korea Electronics Technology Institute (KETI), and other partners.

On June 29, K3AI announced that it is pursuing the development of a digital twin environment that enables manufacturing sites to be replicated and verified in a virtual space, centered around its proprietary platform “Neuro Twin X” and utilizing NVIDIA Omniverse integration technology.

This research began after K3AI was selected last year as a joint research institution for a national project led by the Korea Electronics Technology Institute, which aims to develop AI-based digital twin technology. The project includes participation from Germany’s largest industrial automation company, Siemens, as well as German research organizations and enterprises.

During the first year of research, K3AI completed the simulation design phase. From this year, the company plans to focus on developing a physical AI manufacturing infrastructure that can virtually replicate various elements of a manufacturing site and verify them before actual process application.

The physical AI manufacturing infrastructure is a technology that virtually places workers, robots, and manufacturing equipment in a simulated environment, allowing the company to preemptively check process flows and equipment movements. This approach reduces trial-and-error during actual facility construction and enhances both the efficiency and safety of manufacturing processes.

K3AI cites its ability to build high-precision 3D spatial information as a core competitive advantage in digital twin technology. The company has experience in precisely replicating secondary battery manufacturing processes in virtual space using CAD drawings and real measurement data, and has improved digital twin accuracy by recreating facility layouts and work paths to closely match actual environments.

Additionally, K3AI is developing a platform that integrates factory digital twin construction and robot simulation, based on Neuro Twin X, which is compatible with NVIDIA Omniverse. Precisely built virtual factories allow for deployment and testing of various types of robots in environments similar to real-world conditions, providing a foundation for physical AI learning.

The company intends to expand the application of its technology beyond manufacturing simulation to include robot verification and the field of physical AI learning, leveraging its proprietary 3D spatial information engine and digital twin experience.

A representative from K3AI stated, “The second year of research involves concretizing a prototype of the physical AI manufacturing infrastructure based on simulation design. We are advancing smart manufacturing digital twin technology that enables deployment and verification of workers, manufacturing equipment, and robots within a virtual environment, with Neuro Twin X at the center.”

The representative added, “K3AI has verified actual robot control and virtual factory simulation through its proprietary Omniverse-linked platform and has secured technology for structuring physical AI data. Through this national project, we aim to further advance digital twin-based manufacturing simulation environments applicable to manufacturing sites and to expand our presence in the physical AI manufacturing infrastructure market.”

K3AI is moving to lead the next-generation smart manufacturing infrastructure market by combining digital twin and physical AI technologies, based on global industry-academia-research collaborations with organizations such as Siemens and KETI.

This content was produced with the assistance of AI translation services.