Build a Robot Pick-and-Place Simulation with SimReady Assets

Hugging Face’s SimReady Central provides OpenUSD assets with predefined simulation properties such as collisions, mass, joints, and joint drives, enabling reuse across robotics workflows. The library includes assets like the UR17G robotic arm, Robotiq 2F-140 gripper, and beverage can, each pre-configured for simulation tasks in Isaac Sim. Users can download these assets directly from SimReady Central or via the Hugging Face command-line interface, ensuring all supporting files remain intact for proper simulation execution.

The workflow demonstrates assembling a pick-and-place task where the UR17G arm uses the Robotiq 2F-140 gripper to grasp, lift, move, and release a beverage can within Isaac Sim. Before running the task, users must validate joint settings, movement, and contact behavior to ensure correct interaction between the arm, gripper, and can. Validation reports for the arm and gripper confirm passing PhysX-driven joint checks against SimReady profiles Robot-Body 4.0.0 and Robot-Gripper 3.0.0, respectively.

To execute the task, users start Isaac Sim with a remote Python server enabled, allowing an agent to interact with the simulation. The agent, such as GPT-6 Sol, can assemble the robot, validate physics settings, and generate controls for the pick-and-place operation. Alternatively, users can manually assemble the arm and gripper using Isaac Sim’s GUI or follow the Robot Assembler tutorial to ensure proper articulation and joint validation.

Once the robot and can are assembled in the scene, the agent runs the pick-and-place task while evaluating simulation behavior. Users are advised to inspect generated code and resulting behavior before adapting the task for their specific needs. SimReady Central’s Space search service can also help locate additional OEM-curated content for expanding simulation scenarios.