Introduction: Nvidia Project GRooT
Nvidia, the chipmaker that has turned into an artificial intelligence powerhouse. On March 18th 2024, it announced the new platform called Project GR00T for humanoid robots.
Project GR00T introduces a general-purpose foundation model designed to propel the field of humanoid through its pretrained AI models.
Project GR00T stands for Generalist Robot 00 Technology. It’s basically an AI platform designed specifically for humanoid robots. The goal? Robots that can understand natural language and learn from observing humans. This will supposedly allow them to develop dexterity, coordination, and other skills to navigate and interact with the real world. NVIDIA sees this as a major leap towards artificial general robotics. They’ve also released a new computer, Jetson Thor, specifically for these humanoid robots.
Let’s delve into what Project GR00T entails and the transformative impact it promises to deliver.
Follow us on Linkedin for everything around Semiconductors & AI
Understanding Nvidia Project GR00T
At its core, Project GR00T short for ‘Generalist Robot 00 Technology’, signifies a pivotal step forward in the development of humanoid robots.
This ambitious initiative from NVIDIA aims to equip robots with the ability to understand natural language and emulate movements by observing human actions — quickly learning coordination, dexterity and other skills in order to navigate, adapt and interact with the real world.
NVIDIA has built this AI platform for leading humanoid robot companies such as 1X Technologies, Agility Robotics, Apptronik, Boston Dynamics, Figure AI, Fourier Intelligence, Sanctuary AI, Unitree Robotics and XPENG Robotics, among others. Notably, Tesla is not listed by the company.
Read More: 5 Tech Areas Where US and China Might Clash in the Coming Decade – techovedas
NVIDIA Jetson Thor – Processor for robots
As part of Project GR00T, NVIDIA unveiled Jetson Thor, a new computer for humanoid robots. Jetson Thor is based on the NVIDIA Thor system-on-a-chip (SoC).
An SoC is an integrated circuit that combines all components of a computer or other system into a single chip. It improves performance and reduces power consumption.
Jetson Thor includes a next-generation GPU based on the NVIDIA Blackwell architecture. A GPU (Graphics Processing Unit) is a specialized electronic circuit designed to rapidly manipulate and alter memory to accelerate the creation of images in a frame buffer intended for output to a display device. The Blackwell architecture enhances the GPU’s performance.
The SoC also features a transformer engine delivering 800 teraflops of 8-bit floating point AI performance. A teraflop is a unit of computing speed equal to one trillion (10^12) floating-point operations per second. This high performance allows the SoC to run multimodal generative AI models like GR00T.
Read more Top 6 Humanoid Companies to Shape Our Future – techovedas
Pretrained AI Models: Isaac Manipulator and Isaac Perceptor
NVIDIA has developed a collection of pretrained models for robotics, which are part of the Isaac Manipulator and Isaac Perceptor. These models are designed to assist humanoid robots in navigating physical environments autonomously without pre-programmed routes.
The Isaac Manipulator
Issca manipulator helps robotic arms be very accurate and smart. It comes with lots of helpful tools and software that make it work better. With it, planning the best path for the robot to move is 80 times faster, and it can also understand new things without being taught beforehand, making it faster and more efficient. This allows developers to teach the robot to do more tasks automatically.
Among early ecosystem partners are Yaskawa, Universal Robots, a Teradyne company, PickNik Robotics, Solomon, READY Robotics and Franka Robotics.
Zero-shot perception refers to the ability of an artificial intelligence (AI) system to recognize or understand objects, concepts, or situations without any prior training or exposure to them
Isaac Perceptor
Preceptor increasingly integrates multi-camera, 3D surround-vision capabilities, empowering autonomous mobile robots adopted in manufacturing and fulfillment operations to enhance efficiency, ensure worker safety, and diminish error rates and costs. Early adopters include ArcBest, BYD and KION Group.
Isaac Sim
Successful development training and testing of complex robots for real world applications demands high fidelity simulation and accurate physics.
NVIDIA Isaac Sim™ is an extensible robotics simulation platform that gives a faster, better way to design, test, and train AI-based robots.
It combines immersive physically accurate photorealistic environments for training virtual robots. These virtual robots are a digital twin of your real robots. Isaac sim thus provides a metaverse for training your robots.
Read More: Year of the Humanoid – How 2024 will be the year for robotics – techovedas
Conclusion
The Nvidia CEO is not the only one that believe that humanoid robots have the potential to transform manufacturing, Tesla CEO Elon Musk also seems to think the same. Tesla is currently testing its Tesla Bot or Optimus, to make it “capable of performing tasks that are unsafe, repetitive, or boring.” Musk has previously posted videos showing that the Tesla Bot can walk forward, bend, and move its arms and legs. It can even pick up and sort objects.
While Musk produces robots. Jensen is on the way to make robots smarter. NVIDIA Project GR00T will help these machines possess the cognitive abilities and adaptability to thrive in the complex realities of the real world.
