Bin choosing simulation with bodily engine. Supply: FANUC
FANUC Corp. final week stated it has strengthened the mixing between its ROBOGUIDE simulation software program and the NVIDIA Isaac Sim open robotic simulation reference framework. The corporate stated it’s going to allow intuitive operation inside a “digital manufacturing unit” and assist notice correct, built-in digital twins.
FANUC is utilizing imitation studying and the NVIDIA GR00T basis mannequin to allow considered one of its robots to fold T-shirts. The corporate famous that its system makes use of the NVIDIA Jetson Thor platform.
Editor’s observe: FANUC is a 2026 RBR50 Robotics Innovation Award winner for serving to Wilson Bohannan enhance U.S.-based manufacturing. See the RBR50 honorees on the present flooring and on the ticketed RBR50 gala dinner on the Robotics Summit & Expo subsequent week. Register now to attend.
‘Digital manufacturing unit’ runs on Isaac Sim and ROBOGUIDE
On the Worldwide Robotic Exhibition (IREX) final December in Tokyo, FANUC demonstrated know-how that imports robotic movement simulations created in ROBOGUIDE into NVIDIA Isaac Sim. It reproduced exact trajectories and cycle occasions in a digital setting utilizing the identical management algorithms because the precise robotic.
With the improved integration, FANUC stated the 2 techniques can now ship a extra sensible and environment friendly simulation setting for robotic testing and digital commissioning.
The primary mode of integration locations NVIDIA Isaac Sim on the forefront, with ROBOGUIDE working within the background to make sure correct robotic conduct within the digital area. ROBOGUIDE is tightly built-in with Isaac Sim, with steady direct communication.
“On this new setting, customers can intuitively function robots in Isaac Sim in actual time from digital or bodily educate pendants linked to ROBOGUIDE—simply as in the event that they had been controlling an precise robotic,” stated FANUC. “Customers can carry out jogging operations, educate robotic applications, execute the applications, and confirm outcomes straight inside Isaac Sim. This allows intuitive and environment friendly pre-installation research and course of design for robots inside GPU accelerated bodily correct sensor and setting simulations.”
The businesses asserted that the NVIDIA Isaac Lab open robotic studying framework and NVIDIA Omniverse libraries allow high-precision simulations of duties that had been historically troublesome to breed. Examples embody dealing with versatile elements like cables or performing insertion and meeting operations.
Robots working in Isaac Sim keep an identical trajectories and cycle occasions to actual machines by way of integration with ROBOGUIDE, eradicating the “sim-to-real hole,” claimed the companions. This mixed setting additionally helps reinforcement and imitation studying, accelerating the analysis and deployment of bodily AI techniques, stated FANUC and NVIDIA.
The second integration mode locations ROBOGUIDE on the forefront, whereas the NVIDIA PhysX physics engine permits superior simulation within the background. ROBOGUIDE can now use PhysX for correct simulation of complicated duties reminiscent of bin picking, which had been beforehand troublesome to simulate.
Randomly piled components might be realistically simulated utilizing physics-based modeling, whereas ROBOGUIDE’s 3D imaginative and prescient system identifies half positions for pick-and-place operations. Actual-world eventualities—reminiscent of figuring out when a robotic can’t extract a particular half and deciding on another—might be replicated nearly, stated FANUC.
“This development permits customers to finish feasibility research for bin-picking techniques in a digital setting, eliminating the necessity for in depth trial-and-error utilizing precise components,” asserted the corporate. “Mixed with NVIDIA PhysX, ROBOGUIDE considerably improves the effectivity of designing and deploying bin-picking techniques that beforehand required appreciable time and experience on website.”
The ROBOGUIDE educate pendant permits for the operation of robots in NVIDIA Isaac Sim. Supply: FANUC
GR00T N permits FANUC cobots to fold T-shirts
At an open home this month, FANUC demonstrated a dual-arm system utilizing two CRX collaborative robots to fold versatile objects — on this case, T-shirts — by imitation studying. It used cameras for visible recognition and the NVIDIA Isaac GR00T N open robotic basis mannequin.
Folding versatile objects like T-shirts requires steady adaptation of robotic movement to altering shapes, famous FANUC. Such folding is troublesome to realize with standard playback educating or vision-based robotic path compensation, it stated.
The corporate will exhibit an operator performing the folding activity utilizing CRX robots in actual time. The system will study from these examples by way of imitation studying. The twin-arm CRX can be educated to duplicate the duty and purchase the mandatory methods to finish the folding course of.
Historically, imitation-learned robotic motions tended to seem segmented and jerky. By combining FANUC’s superior movement management know-how with the NVIDIA GR00T N mannequin, the system can have clean, steady motion, stated the businesses.
A dual-armed robotic makes use of imitation studying to fold T-shirts. Supply: FANUC
NVIDIA Jetson Thor platform
At IREX, FANUC additionally launched an “AI robotic that avoids human” based mostly on its open platform. The system has been upgraded with NVIDIA’s newest robotic pc, Jetson Thor.
By changing NVIDIA Jetson AGX Orin module with Jetson T5000 module, the system’s AI compute has improved by greater than 7.5 occasions, asserted the businesses. Jetson Thor permits the robotic to avoids human motion extra successfully, stated FANUC.
The corporate’s new product exhibition this month additionally options the operation of robots in a digital area, high-precision simulations, physics simulations utilizing NVIDIA PhysX, demos of dual-arm robots educated by way of imitation studying, and cobots that includes the most recent NVIDIA edge computing platform.
A cobot utilizing AI powered by NVIDIA Jetson Thor. Supply: FANUC
The put up FANUC strengthens robotic integration with NVIDIA Isaac Sim appeared first on The Robotic Report.
