Boston Dynamics, a developer of mobile robotics, has expanded its collaboration with Nvidia to build the next generation of AI capabilities for humanoid robots using the Nvidia Jetson Thor computing platform.
According to Boston Dynamics, the compact size, high performance and efficiency of Jetson Thor enables Atlas to run complex, multimodal AI models for Boston Dynamics’ whole-body and manipulation controllers.
The two companies are collaborating to define specific platform parameters including functional safety and security architectures, as well as key learning and computer vision pipelines using Nvidia’s training and simulation platforms.
Developers at Boston Dynamics and its research partners also note that they are making breakthroughs in learned dexterity and locomotion AI policies using Nvidia’s Isaac Lab, an open-source, modular framework for robot learning in physically accurate virtual environments. Isaac Lab is built on Nvidia Isaac Sim and Nvidia Omniverse technologies.
In addition to the ongoing work on Atlas, Boston Dynamics has also continued introducing new AI capabilities for Spot, the company’s quadruped, and Orbit, its robot fleet management and data analysis software. New reinforcement learning tools are improving the Spot’s locomotion control and advanced foundation models are helping it avoid specific kinds of hazards that might appear in its path.
