Intrinsic is also partnering with Alphabet-owned DeepMind to tackle pose estimation and path planning, two other essential aspects of automation. The system, trained on over 130,000 objects, can determine the orientation of objects in a few seconds, a vital part of being able to pick them up. Intrinsic's work with DeepMind also includes the ability to operate multiple robots simultaneously. The team is developing systems that use two arms at once, opening up a range of additional applications for these systems.
Key takeaways:
- Intrinsic announced at the Automate conference that it is incorporating several Nvidia offerings into its Flowstate robotic app platform, including Isaac Manipulator, a collection of models designed to create workflows for robot arms.
- The collaboration is focused on grasping (grabbing and picking up objects), a key modality for both manufacturing and fulfillment automation. The systems are trained on large datasets to execute tasks across different hardware and objects.
- Intrinsic is also working with Alphabet-owned DeepMind to crack pose estimation and path planning, two other key aspects of automation. The system was trained on more than 130,000 objects and can determine the orientation of objects in a few seconds.
- Intrinsic and DeepMind are also developing systems that use two arms at once, a setup more in line with the emerging world of humanoid robots. This opens up a whole world of additional applications for these systems.