Felix Pinkston
January 25, 2025 04:20
NVIDIA uses OpenUSD and Omniverse to accelerate physical AI development, focusing on robotics and autonomous vehicles. The new COSMOS platform promises exponential growth of synthetic data creation.
NVIDIA is at the forefront of developing physical AI through the Omniverse platform driven by Openusd reported on the NVIDIA blog. This technical leap provides a powerful framework for creating a realistic simulation for training the AI model by revolutionizing the development of robotics and autonomous vehicles.
Development of physical AI
Physical AI is emerging as a pivotal component in the development of autonomous machines such as robots and autonomous vehicles. Like the large language model process text, it includes teaching the AI model to understand and interact with the real world. This model is polished in a simulated environment that replicates physical epidemiology and spatial relationships essential for actual AI deployment.
Omniverse and Cosmos: Synergy of AI
In the CES event, NVIDIA has introduced a new creative AI model and blueprint that expands the function of the Omniverse platform to physical AI applications. The main highlights are the NVIDIA Cosmos platform. The platform has a video processing pipeline designed to simplify the development of physical AI by creating a high -quality model and a vast amount of synthetic data.
Cosmos’ World Foundation Models (WFM) play an important role in this process. They can predict future world countries, which can be converted into videos based on multimodal input. This feature allows developers to produce a wide range of photographs and physical -based synthetic data that are essential for efficient education on AI systems.
Industry adoption and influence
Major technology companies, including Uber and various robot companies, have begun to integrate space into the development process. By using this platform, these companies can greatly improve the AI model to increase the development speed and accuracy of autonomous technology.
The universe also acts as a powerful engine that multiplys synthetic data when combined with Omniverse. Developers can create a lot of 3D scenarios to create a controlled output array by treating the following cosmos. This approach provides more wider and more educational data to accelerate the development of systems such as autonomous vehicles.
Actual application
COSMOS WFMS is applied to various industries to provide an integrated framework for distributing large AI models. In robotics, NVIDIA ISAAC GR00T blueprints help to create synthetic motion data sets, while autonomous vehicles enable comprehensive simulation and testing through Omniver’s sensor RTX API. Industrial applications are benefited from MEGA OMNIVERSE BLUEPRINT so that they can simulate high -faithful sensors that can optimize robot work in factories.
As NVIDIA continues to pursue the boundaries of AI technology, the integration of Openusd and Omniverse is ready to improve both functions and applications in various fields by setting new standards for sophisticated AI systems.
Image Source: Shutter Stock