.Ted Hisokawa.Aug 08, 2024 13:53.Figure has actually introduced its next-gen Body 02 humanoid robotic, utilizing NVIDIA Omniverse as well as GPUs for improved artificial intelligence capacities and autonomous procedures. Silicon Valley-based Body has unveiled its own next-generation Figure 02 humanoid robotic, which leverages NVIDIA Omniverse as well as NVIDIA GPUs to conduct totally autonomous duties. According to the NVIDIA Blog Site, the Figure 02 model has been tested for information compilation and use-case instruction at BMW Team’s Spartanburg production line in South Carolina.Technological Advancements.Number 02 is a considerable upgrade from the initial version released merely ten months earlier.
The new version incorporates a 2nd NVIDIA RTX GPU-based element, supplying 3 times the assumption gains matched up to its own forerunner. This innovation helps with dealing with totally independent real-world AI tasks. The firm made use of NVIDIA Isaac Sim, a referral request improved the NVIDIA Omniverse system, to design, train, and exam AI-based robotics using synthetic data.Market Treatments.Number strives to commercialize industrial humanoid robotics to resolve effort shortages throughout a variety of markets.
The provider also intends to create consumer models. Brett Adcock, CEO of Figure, focused on the rapid improvement in pep talk, outlook, dexterity, as well as computational energy, which delivers them closer to delivering humanoid robots for multiple markets.Relationships and also Financing.Founded in 2022, Figure has actually partnered along with OpenAI to create customized artificial intelligence versions qualified on NVIDIA H100 GPUs. These models boost the robots’ informal AI capabilities.
The startup lately safeguarded $675 million in financing from leading innovation companies, featuring NVIDIA.Deepu Talla, vice president of robotics and upper hand computing at NVIDIA, highlighted the significance of the NVIDIA robotics stack, from advantage to cloud, in steering technology in humanoid robotics. “Creating autonomous humanoid robotics needs the fusion of 3 pcs: NVIDIA DGX for artificial intelligence instruction, NVIDIA Omniverse for simulation, as well as NVIDIA Jetson in the robotic,” Talla specified.Robot Palms Capable of Handling Real-World Tasks.Figure 02 components new human-scale palms, 6 RGB video cameras, as well as understanding AI models trained with synthetic information produced in Isaac Sim. These capabilities allow the robotic to carry out high-precision pick-and-place activities needed for smart manufacturing functions.Body is among the preliminary members to participate in the brand-new NVIDIA Humanoid Robotic Developer System, which gives early access to the latest resources as well as calculating technologies for humanoid robotic progression.
This includes the most up to date releases of NVIDIA Isaac Sim, Isaac Laboratory, NIM microservices (RoboCasa and MimicGen), OSMO, Jetson Thor, and Project GR00T general-purpose humanoid foundation models.Image source: Shutterstock.