|
Hearken to this text  |
NVIDIA CEO Jenson Huang ended his GTC 2024 keynote backed by life dimension photos of the entire varied humanoids in growth and powered by the Jetson Orin pc. | Credit score: Eugene Demaitre
SAN JOSE, Calif. — The NVIDIA GTC 2024 keynote kicked off like a rock live performance yesterday on the SAP Enviornment. Greater than 15,000 attendees crammed the world in anticipation of CEO Jensen Huang’s annual presentation of the most recent product information from NVIDIA.
To construct the thrill, the ready crowd was mesmerized by an interactive and real-time generative artwork show working reside on the principle stage display screen, pushed by the prompts of artist Refik Anadol Dustio.
New basis for humanoid robotics
The massive information from the robotics facet of the home is that NVIDIA launched a brand new general-purpose basis mannequin for humanoid robots known as Undertaking GR00T. This new mannequin is designed to convey robotics and embodied AI collectively whereas enabling the robots to know pure language and emulate actions by observing human actions.
Undertaking GR00T coaching mannequin. | Credit score: NVIDIA
GR00T stands for “Generalist Robotic 00 Expertise,” and with the race for humanoid robotics heating up, this new expertise is meant to assist speed up growth. GR00T is a big multimodal mannequin (LMM) offering robotics builders with a generative AI platform to start the implementation of enormous language fashions (LLMs).
“Constructing basis fashions for common humanoid robots is without doubt one of the most fun issues to unravel in AI right now,” stated Huang. “The enabling applied sciences are coming collectively for main roboticists around the globe to take big leaps in direction of synthetic common robotics.”
GR00T makes use of the brand new Jetson Thor
As a part of its robotics bulletins, NVIDIA unveiled Jetson Thor for humanoid robots, based mostly on the NVIDIA Thor system-on-a-chip (SoC). Vital upgrades to the NVIDIA Isaac robotics platform embrace generative AI basis fashions and instruments for simulation and AI workflow infrastructure.
The Thor SoC features a next-generation GPU based mostly on NVIDIA Blackwell structure with a transformer engine delivering 800 teraflops of 8-bit floating-point AI efficiency. With an built-in useful security processor, a high-performance CPU cluster, and 100GB of Ethernet bandwidth, it will possibly simplify design and integration efforts, claimed the firm.
Undertaking GR00T, a general-purpose multimodal basis mannequin for humanoids, allows robots to study completely different expertise. | Credit score: NVIDIA
NVIDIA confirmed humanoids in growth with its applied sciences from corporations together with 1X Applied sciences, Agility Robotics, Apptronik, Boston Dynamics, Determine AI, Fourier Intelligence, Sanctuary AI, Unitree Robotics, and XPENG Robotics.
“We’re at an inflection level in historical past, with human-centric robots like Digit poised to alter labor endlessly,” stated Jonathan Hurst, co-founder and chief robotic officer at Agility Robotics. “Fashionable AI will speed up growth, paving the way in which for robots like Digit to assist individuals in all features of every day life.”
“We’re excited to associate with NVIDIA to put money into the computing, simulation instruments, machine studying environments, and different mandatory infrastructure to allow the dream of robots being part of every day life,” he stated.
NVIDIA updates Isaac simulation platform
The Isaac instruments that GR00T makes use of are able to creating new basis fashions for any robotic embodiment in any atmosphere, based on NVIDIA. Amongst these instruments are Isaac Lab for reinforcement studying, and OSMO, a compute orchestration service.
Embodied AI fashions require huge quantities of actual and artificial information. The brand new Isaac Lab is a GPU-accelerated, light-weight, performance-optimized utility constructed on Isaac Sim for working 1000’s of parallel simulations for robotic studying.
NVIDIA software program — Omniverse, Metropolis, Isaac and cuOpt — mix to create an ‘AI gymnasium’
the place robots, AI brokers can work out and be evaluated in advanced industrial areas. | Credit score: NVIDIA
To scale robotic growth workloads throughout heterogeneous compute, OSMO coordinates the information technology, mannequin coaching, and software program/hardware-in-the-loop workflows throughout distributed environments.
NVIDIA additionally introduced Isaac Manipulator and Isaac Perceptor — a set of robotics-pretrained fashions, libraries and reference {hardware}.
Isaac Manipulator provides dexterity and modular AI capabilities for robotic arms, with a strong assortment of basis fashions and GPU-accelerated libraries. It may well speed up path planning by as much as 80x, and zero-shot notion will increase effectivity and throughput, enabling builders to automate a larger variety of new robotic duties, stated NVIDIA.
Amongst early ecosystem companions are Franka Robotics, PickNik Robotics, READY Robotics, Solomon, Common Robots, a Teradyne firm, and Yaskawa.
Isaac Perceptor offers multi-camera, 3D surround-vision capabilities, that are more and more being utilized in autonomous cellular robots (AMRs) adopted in manufacturing and success operations to enhance effectivity and employee security. NVIDIA listed corporations resembling ArcBest, BYD, and KION Group as companions.
Be taught from Agility Robotics, Amazon, Disney, Teradyne and plenty of extra.
‘Simulation first’ is the brand new mantra for NVIDIA
A simulation-first method is ushering within the subsequent section of automation. Actual-time AI is now a actuality in manufacturing, manufacturing facility logistics, and robotics. These environments are advanced, usually involving a whole bunch or 1000’s of transferring elements. Till now, it was a monumental process to simulate all of those transferring elements.
NVIDIA has mixed software program resembling Omniverse, Metropolis, Isaac, and cuOpt to create an “AI gymnasium” the place robots and AI brokers can work out and be evaluated in advanced industrial areas.
Huang demonstrated a digital twin of a 100,000-sq.-ft, warehouse — constructed utilizing the NVIDIA Omniverse platform for growing and connecting OpenUSD functions — working as a simulation atmosphere for dozens of digital employees and a number of AMRs, imaginative and prescient AI brokers, and sensors.
Every cellular robotic, working the NVIDIA Isaac Perceptor multi-sensor stack, can course of visible info from six sensors, all simulated within the digital twin.
Picture depicting AMR and a manipulator working collectively to
allow AI-based automation in a warehouse powered by NVIDIA Isaac. | Credit score: NVIDIA
On the similar time, the NVIDIA Metropolis platform for imaginative and prescient AI can create a single centralized map of employee exercise throughout the complete warehouse, fusing information from 100 simulated ceiling-mounted digital camera streams with multi-camera monitoring. This centralized occupancy map might help inform optimum AMR routes calculated by the NVIDIA cuOpt engine for fixing advanced routing issues.
cuOpt, an optimization AI microservice, solves advanced routing issues with a number of constraints utilizing GPU-accelerated evolutionary algorithms.
All of this occurs in real-time, whereas Isaac Mission Management coordinates the complete fleet utilizing map information and route graphs from cuOpt to ship and execute AMR instructions.
NVIDIA DRIVE Thor for robotic axis
The corporate additionally introduced NVIDIA DRIVE Thor, which now supersedes NVIDIA DRIVE Orin as a SoC for autonomous driving functions.
A number of autonomous autos are utilizing NVIDA architectures, together with robotaxis and autonomous supply autos from corporations together with Nuro, Xpeng, Weride, Plus, and BYD.