Earth logo

NVIDIA Releases a Major Update to the Isaac Robotics Platform and the Project GR00T Foundation Model for Humanoid Robots.

NVIDIA's Latest Isaac Update and Project GR00T Model Pave the Way for Advanced Humanoid Development.

By Mohammad Shakhawat HossainPublished 2 months ago 3 min read
Like

Developers Can Now Access the Jetson Thor Robot Computer, CUDA-Accelerated Perception and Manipulation Libraries, Generative AI Foundation Models, and Robot Training Simulator through the Isaac Robotics Platform.

GTC—NVIDIA today unveiled Project GR00T, a general-purpose foundation model for humanoid robots intended to advance the company's efforts to propel advances in embodied AI and robotics.

In conjunction with the initiative, the company also unveiled major enhancements to the NVIDIA IsaacTM robotics platform, including generative AI foundation models, tools for simulation, and AI workflow infrastructure, as well as a new computer, Jetson Thor, for humanoid robots based on the NVIDIA Thor system-on-a-chip (SoC).

NVIDIA founder and CEO Jensen Huang stated, "One of the most exciting problems to solve in AI today is building foundation models for general humanoid robots." "World-class roboticists are making enormous strides toward artificial general robotics as enabling technologies come together."

GR00T (Generalist Robot 00 Technology)-enabled robots will be built to mimic human movements and comprehend natural language. They will pick up coordination, dexterity, and other skills fast, enabling them to navigate, adapt, and engage with the real world. Huang gave a demonstration of multiple such robots performing a range of activities during his GTC keynote.

Designed to Fit Humanoid Robots

A new computing platform called Jetson Thor was developed with the ability to carry out complicated tasks and engage in safe, natural interactions with both humans and other machines. Its modular architecture is maximized for power, size, and performance.

In order to run multimodal generative AI models like GR00T, the SoC has a next-generation GPU based on the NVIDIA Blackwell architecture with a transformer engine that delivers 800 teraflops of 8-bit floating point AI capability. It greatly streamlines design and integration efforts with its integrated functional safety processor, 100GB of Ethernet bandwidth, and high-performance CPU cluster.

For top humanoid robot firms including 1X Technologies, Agility Robotics, Apptronik, Boston Dynamics, Figure AI, Fourier Intelligence, Sanctuary AI, Unitree Robotics, and XPENG Robotics, among others, NVIDIA is developing a comprehensive AI platform.

Our time is unique because human-centered robots like Digit have the potential to completely alter the nature of work. "Modern artificial intelligence will spur progress, enabling robots such as Digit to assist humans in all facets of daily existence," stated Jonathan Hurst, chief robot officer and cofounder of Agility Robotics. "We are thrilled to work with NVIDIA to fund the infrastructure, simulation tools, machine learning environments, and computing power required to make the dream of robots in everyday life a reality."

Geordie Rose, cofounder and CEO of Sanctuary AI, stated, "Embodied AI will not only help address some of humanity's biggest challenges, but will also create innovations which are currently beyond our reach or imagination." Because we believe that such vital technology shouldn't be developed in silos, we give priority to enduring partners like NVIDIA.

Significant Improvements to the Isaac Platform

GR00T may generate new foundation models for any robot embodiment in any environment by using the Isaac tools. These resources include the computational orchestration service OSMO as well as the reinforcement learning tool Isaac Lab.

Large volumes of synthetic and real data are needed to build embodied AI models. Built on top of Isaac Sim, the new Isaac Lab is a lightweight, performance-optimized, GPU-accelerated program designed to execute thousands of concurrent simulations for robot learning.

OSMO coordinates the data collection, model training, and software/hardware-in-the-loop workflows across distributed settings to scale robot development workloads across heterogeneous computation.

Additionally, NVIDIA unveiled Isaac Manipulator and Isaac Perceptor, a set of reference hardware, libraries, and pretrained robotics models.

With a large library of foundation models and GPU-accelerated libraries, Isaac Manipulator provides robotic arms with cutting edge dexterity and customizable AI capabilities. Path planning can be accelerated by up to 80x, and zero-shot perception boosts productivity and throughput, allowing developers to automate more novel robotic jobs. Yaskawa, Universal Robots, a Teradyne business, PickNik Robotics, Solomon, READY Robotics, and Franka Robotics are a few of the early ecosystem partners.

The Isaac Perceptor is a multi-camera 3D surround-vision device that is being used more and more in autonomous mobile robots used in fulfillment and manufacturing processes to lower costs and increase worker safety while also enhancing efficiency. ArcBest, BYD, and KION Group are among the early adopters who want to attain unprecedented degrees of autonomy in material handling operations and other areas.

It is anticipated that the new Isaac platform features would be accessible in the upcoming quarter. Study up on Project GR00T.

Science
Like

About the Creator

Mohammad Shakhawat Hossain

Reader insights

Be the first to share your insights about this piece.

How does it work?

Add your insights

Comments

There are no comments for this story

Be the first to respond and start the conversation.

Sign in to comment

    Find us on social media

    Miscellaneous links

    • Explore
    • Contact
    • Privacy Policy
    • Terms of Use
    • Support

    © 2024 Creatd, Inc. All Rights Reserved.