Building robots that can operate alongside humans requires advanced skills. A “sim-first” approach using OpenUSD allows training of hundreds of robots simultaneously in simulation environments. NVIDIA introduced groundbreaking advances in physics simulation and open-source models. Newton Physics Engine, managed by the Linux Foundation, is a GPU-accelerated engine enabling robots to learn complex tasks precisely with frameworks like MuJoCo Playground and NVIDIA Isaac Lab.
NVIDIA announced Newton Physics Engine and Isaac GR00T N1.6 to enhance robot learning. Newton, co-developed by Google DeepMind, Disney Research, and NVIDIA, is an open-source physics engine for robot learning. Isaac GR00T N1.6 integrates NVIDIA Cosmos Reason for physical AI tasks. NVIDIA Isaac Lab, an open-source robot learning framework, offers advanced features for robotics research and development.
Agility Robotics, Lightwheel, Mentee, Universal Robots, and Wandelbots are using simulation technologies to accelerate physical AI development. NVIDIA Isaac Lab is used by Agility Robotics for Digit robot training. Lightwheel developed a simulation platform on NVIDIA Omniverse. Mentee Robotics uses NVIDIA’s three-computer architecture with OpenUSD for synthetic data generation. Universal Robots and Wandelbots are leveraging NVIDIA’s platforms for robot simulation and learning.
Developers in the robotics community are adopting NVIDIA’s open frameworks and libraries for innovation. Dylan Tobin created an AI chatbot trained on Isaac Sim workflows. Livestreams demonstrate how developers are using Isaac Sim and Isaac Lab for robotics innovation. NVIDIA Robotics office hours sessions show how Brev simplifies running Isaac Sim and Isaac Lab on Omniverse.
To learn more about robot learning with OpenUSD and NVIDIA’s robotics technologies, explore NVIDIA Omniverse news and join the community. Stay updated by following Omniverse on social media platforms. Explore the Alliance for OpenUSD forum and the AOUSD website for more information.
Read more at NVIDIA: Open-Source Physics Engine and OpenUSD Advance Robot Learning