Intrinsic, an AI robotics company that emerged from Alphabet, is collaborating with NVIDIA to enhance autonomous robotic manipulation. Leveraging NVIDIA’s AI and Isaac platform technologies, the partnership aims to introduce cutting-edge dexterity and modular AI capabilities for robotic arms.
NVIDIA’s recent release, Isaac Manipulator, offers foundation models and GPU-accelerated libraries to expedite AI model training and task reprogramming, boasting an impressive 80x acceleration in path planning.
Foundation models, rooted in transformer deep learning architecture, enable robots to perceive and make decisions based on vast datasets, similar to how ChatGPT processes text. NVIDIA’s Project GROOT, a foundation model tailored for humanoids, underscores the industry’s pursuit of advanced robotics technology.
Intrinsic and NVIDIA are addressing the longstanding challenge of robotic grasping, which has historically been time-consuming and costly to program. By harnessing NVIDIA Isaac Sim and generating synthetic data for vacuum grasping, Intrinsic developed a prototype for Trumpf Machine Tools, facilitating visualized processes and motion planning through Intrinsic Flowstate.
This workflow, augmented by Isaac Manipulator and CUDA-accelerated robot motions, enables evaluation in simulation before real-world deployment.
Deepu Talla, NVIDIA’s vice president of robotics and edge computing, emphasizes the potential for AI foundation models to revolutionize industrial manufacturing, streamlining operations and enabling robot interaction with diverse objects.
Intrinsic’s development of Flowstate, initially targeting industrial robots, aims to extend software skills to other robotic classes, building upon its acquisition of Open Source Robotics Corporation in December 2022, the for-profit entity behind the Robot Operating System (ROS).
As the collaboration between Intrinsic and NVIDIA progresses, the integration of advanced AI technologies promises to scale automation in industrial settings.
Filed in AI (Artificial Intelligence) and Robotics.
. Read more about