When a machine combines different functions, engineers must understand how these functions work together before the system is installed. Virtual models can help evaluate concepts, optimize machine processes, and resolve integration issues as early as the development phase. Festo has already utilized NVIDIA Isaac Sim, an open simulation framework, and synthetic data for robotic perception and automation capabilities. This experience, combined with the existing ability to export USD files, serves as the starting point for expanding models to include motion behavior and application-relevant properties, thereby enabling the reliable evaluation of components’ real-world behavior in simulations.

“The collaboration between NVIDIA and Festo is a milestone”

“The collaboration between NVIDIA and Festo is a milestone for the future of industrial automation,” says Prof. Dr. Tomas Smetana, Chief Technology Officer (CTO) at Festo. “By combining NVIDIA Omniverse libraries with Festo’s decades of automation expertise, we are creating a learning virtual environment for the autonomous production of the future.”

Festo’s Quality Standards for Digital Product Models

“Our digital models must be just as trustworthy as our hardware,” says Jan R. Seyler, Head of Physical AI and Digital Thread at Festo. “Our goal is to provide product models that meet Festo’s high quality standards and enable the development and validation of automation solutions that are as close to reality as possible.”

To systematically evaluate model quality in the future, Festo plans to compare measurements from its own laboratories with simulation results in NVIDIA Isaac Sim. The open simulation framework, based on NVIDIA Omniverse libraries, enables evaluation within complete automation scenarios. Under defined operating conditions, simulated movements and dynamic behavior will be compared with corresponding measurements on real components, and the models will be specifically improved.

Festo and ABB Robotics Are Collaborating with NVIDIA on a Virtual Handling Application

A concrete step in this development is the collaboration with ABB Robotics and NVIDIA. In a joint handling application, the partners are integrating digital twins of a three-axis Festo handling system with ABB RobotStudio. This is intended to help machine builders clarify processes and integration issues before physical setup, accelerate commissioning, and shorten time to market. The RobotStudio HyperReality extension connects RobotStudio with NVIDIA Omniverse libraries for photorealistic, physics-based simulations. The ABB Virtual Controller makes it possible to virtually test robot programs using the same control software as on real ABB robots.

“Performing more tasks in more locations faster, safer, and smarter”

“Together with our partner network, ABB Robotics is developing more autonomous and versatile robotic solutions that can perform more tasks in more locations faster, safer, and smarter,” says Boris Fiedler, Head of Operations and Portfolio, Software and Digital, ABB Robotics. “The collaboration with Festo and NVIDIA will advance engineering and virtual commissioning for machine builders and help drive the widespread adoption of industrial-grade Physical AI in companies.”

A Reliable Digital Foundation for Physical AI

For Festo, this collaboration represents a concrete step forward in developing a reliable digital foundation for Physical AI. In the long term, the company aims to narrow the gap between simulation and reality to the point where automation solutions developed virtually can be transferred to real machines with minimal adaptation effort.

Video