Felix Pinkston
Oct 22, 2024 11:48
NVIDIA unveils generative AI tools, simulation, and perception workflows for ROS developers, enhancing robotics capabilities with new nodes and workflows for NVIDIA Jetson platform.
In a significant advancement for the robotics community, NVIDIA announced the introduction of generative AI tools, simulation, and perception workflows for developers using the Robot Operating System (ROS). The announcement was made at ROSCon in Odense, Denmark, a city renowned for its automation industry, according to NVIDIA Blog.
Generative AI Enhancements
NVIDIA’s new offerings include generative AI nodes and workflows designed for the NVIDIA Jetson platform, which supports edge AI and robotics. Generative AI empowers robots to better understand their environment, interact naturally with humans, and make autonomous decisions. This development is poised to significantly enhance the capabilities of ROS developers.
A standout feature is ReMEmbR, a system built on ROS 2 that leverages generative AI to improve robotic reasoning and interactions. By integrating large language models (LLMs), vision language models (VLMs), and retrieval-augmented generation, ReMEmbR allows robots to build and query long-term semantic memories, enhancing their ability to navigate and interact with their surroundings.
Simulation and Perception Workflows
NVIDIA’s Isaac Sim, a robotics simulation platform founded on OpenUSD, provides developers with a virtual testing environment for AI-enabled robots. This ‘sim-first’ approach is crucial for safely validating robotic operations prior to real-world deployment. NVIDIA also released a comprehensive guide to help developers understand the end-to-end workflow for robot simulation and testing with ROS 2 and Isaac Sim.
NVIDIA Isaac ROS, built on the open-source ROS 2 framework, offers a suite of accelerated computing packages and AI models. The upcoming 3.2 release is set to enhance robot perception, manipulation, and environment mapping, with improvements in reference workflows and visual SLAM capabilities.
Industry Partnerships and Implementations
The adoption of NVIDIA’s technology is evident in its partnerships with several robotics companies. Universal Robots, a Teradyne Robotics company, has launched an AI Accelerator toolkit to develop AI-powered collaborative robot applications. Miso Robotics leverages Isaac ROS to advance its AI-driven Flippy Fry Station, enhancing efficiency in food service automation.
Other companies, such as Wheel.me and Main Street Autonomy, are integrating NVIDIA’s Isaac Perceptor to create advanced autonomous mobile robots (AMRs) and streamline sensor calibration processes. Canonical has also introduced a fully certified Ubuntu environment for ROS developers, offering long-term support.
Community Engagement at ROSCon
At ROSCon, various industry leaders and partners, including Canonical, Foxglove, and Siemens, shared insights through workshops and sessions. Highlights included discussions on integrating AI into robot manipulation workflows and accelerating robot learning in simulation environments.
NVIDIA’s continued support of the Open Source Robotics Foundation (OSRF) underscores its commitment to advancing robotics technology. The collaboration with Teradyne Robotics and Siemens further demonstrates the industry-wide impact of NVIDIA’s innovations.
Image source: Shutterstock
