SimReady Assets Simplify Robot Pick‑and‑Place Simulations in Isaac Sim

How can a robotics team build a pick‑and‑place simulation without spending hours writing low‑level physics code for each component? Traditionally developers must export CAD models, create collision meshes, define joint limits and drive parameters, then manually verify that everything behaves correctly in the simulator—a tedious, error‑prone process that slows prototyping and increases iteration time significantly.

SimReady changes that workflow by supplying OpenUSD assets that already embed the required simulation properties. The library bundles robot arms, grippers and common objects with predefined mass, collision geometry and joint‑drive settings, so users can drop them into NVIDIA Isaac Sim and focus on task logic instead of physics plumbing during early development stages to accelerate testing.

Source: Hugging Face: a]:hidden"> Build a Robot Pick-and-Place Simulation with SimReady Assets ↗

Hugging Face’s SimReady Central provides the UR17G arm, the Robotiq 2F‑140 gripper and a beverage‑can object, each packaged with collision meshes, mass values and joint‑drive configurations needed for a pick‑and‑place use case. The assets are labeled “SimReady” because they satisfy the SimReady Foundation profile for robot bodies and grippers, ready for immediate integration into simulation pipelines, saving developers hours of setup.

Validation reports for the arm and gripper show they pass PhysX driven‑joint checks, confirming that joint‑drive settings meet the required thresholds. The example scene then runs the assembled robot, gripper and can together in Isaac Sim, testing whether the gripper can hold the can while the arm moves it to a target.

Source: Hugging Face: a]:hidden"> Build a Robot Pick-and-Place Simulation with SimReady Assets ↗

The demonstration runs under a fixed setup: the beverage can starts at a known position, the robot is anchored, and no vision system is used. The can model omits deformation and liquid dynamics, and the grasp is scripted with a 2 N·m drive‑effort cap chosen from simulation tests, which limits realism but speeds up validation. Consequently, the results cannot be directly extrapolated to real hardware.

Using SimReady also requires NVIDIA Isaac Sim 6.2, which has specific GPU, driver and licensing requirements. While the assets themselves are free on Hugging Face, the overall stack depends on licensed software and compute resources, which may be a barrier for small teams or hobbyists, especially when GPU upgrades are needed to meet performance targets. Consider cloud‑based access if local hardware is insufficient.

Begin by installing Isaac Sim 6.2 and enabling the remote Python server as described in the documentation. Then visit SimReady Central on Hugging Face, locate the UR17G arm, Robotiq 2F‑140 gripper and beverage‑can packages, and download the full USD folders to a local asset directory, preserving the folder structure so Isaac Sim can resolve all supporting files. After download, open each asset in Isaac Sim to verify there are no missing‑file errors.

Run the provided validation reports to confirm the assets pass the PhysX joint checks, then use the example Python script or an AI agent with Isaac Sim skills to generate the pick‑and‑place control sequence. Experiment with different target locations, share your scene and results in the Hugging Face comments, and consider contributing new assets following the five‑step CAD‑to‑SimReady guide.