Alphabet’s robot division Intrinsic using NVIDIA’s Isaac platform to bring easy, flexible industrial automation to the masses.
Intrinsic, a software and AI robotics company at Google’s parent company Alphabet, has announced a collaboration with NVIDIA to leverage its AI and Isaac platform technologies to make the complex field of autonomous robotic manipulation accessible to everyone.
The announcement came at the Automate 2024 trade show in Chicago, where Intrinsic is showcasing its advances in robotic grasping and industrial scalability assisted by foundation models enabled by NVIDIA’s Isaac Manipulator.
The Isaac Manipulator is a collection of foundation models and modular GPU-accelerated libraries that help industrial automation companies build scalable and repeatable workflows for dynamic manipulation tasks by accelerating AI model training and task reprogramming.
Foundation models are based on a transformer deep learning architecture that allows a neural network to learn by tracking relationships in data. They’re trained on huge datasets to process and understand sensor and robot information similar to the way ChatGPT does for text.
This enables enhanced robot perception and decision-making and facilitates zero-shot learning—the ability to perform tasks without prior examples—and the potential for a universally applicable robotic-grasping protocol to work across grippers, environments and objects.
“For the broader industry, our work with NVIDIA shows how foundation models can have a profound impact, including making today’s processing challenges easier to manage at scale, creating previously infeasible applications, reducing development costs, and increasing flexibility for end users,” said Wendy Tan White, CEO at Intrinsic, in a blog post announcing the collaboration with NVIDIA.
Developing Better Robot Grip
Grasping is the most sought after robotics functionality, but it’s time-consuming, expensive to program and difficult to scale. Because of this, many companies still use workers to handle repetitive pick-and-place tasks.
Simulation is changing that. Using NVIDIA Isaac Sim on the NVIDIA Omniverse platform, Intrinsic generated synthetic data for vacuum grasping using computer-aided design models of sheet metal and suction grippers, creating a prototype for Trumpf Machine Tools, an industrial machine tools manufacturer.
The prototype uses Intrinsic Flowstate, a developer environment for AI-based robotics solutions, for visualizing processes, associated perception and motion planning. With a workflow that includes Isaac Manipulator, a user can generate grasp poses and Compute Unified Device Architecture (CUDA)-accelerated robot motions, which can first be evaluated in simulation with Isaac Sim before deployment in the real world with the Intrinsic platform.
This grasping skill, trained with fully synthetic data generated by NVIDIA Isaac Sim, can be used to build sophisticated solutions that perform adaptive and versatile object grasping tasks in simulations and the real world. Instead of hard-coding specific grippers to grasp specific objects in a certain way, efficient code for a particular gripper and object is auto-generated to complete the task using the foundation model and synthetic training data.
“With the latest AI foundation models, companies can program a variety of robot configurations that are able to generalize and interact with diverse objects inside real-world environments,” said Deepu Talla, Vice President of Robotics and Edge Computing at NVIDIA.
