Isaac Sim 6.0 Early Developer Release

[rthaker01]

The open-source early developer preview of NVIDIA Isaac Sim 6.0 is now available. This release focuses on higher-fidelity sensor simulation, simpler and more powerful synthetic data workflows, and clearer guidance for common robotics simulation pitfalls—so you can iterate faster and with more confidence.
This is a developer preview intended for evaluation and early integration feedback. APIs, asset behaviors, and performance characteristics may change before general availability, including a planned version change to Omniverse Kit, Isaac Sim’s underlying framework.

Data generation:

• Replicator Functional API: Synthetic data generation tutorials are transitioning to Replicator’s Functional API for randomization behavior. Instead of constructing OmniGraph graphs and executing them, the Functional API provides developers with more explicit and composable randomization control.

Sensors:

• Realtime 2.0 renderer: Isaac Sim now enables the Realtime 2.0 renderer, delivering visual fidelity approaching that of RTX - Interactive (Path Tracing), while maintaining realtime performance for improved RGB camera realism.
• LiDAR point cloud deskewing and motion compensation: Isaac Sim now supports LiDAR deskewing and motion compensation and provides a faster, simpler way to capture intensity, timestamp, material, and semantic data for individual point cloud hits, providing greater flexibility in post-processing pipelines.
• Depth sensor migration: All depth sensor assets have been migrated to the previously released depth sensor model, allowing disparity artifacts to be captured from existing stereo depth camera models for more realistic sensor simulation.

PhysX and robot simulation:

• Grasping documentation: New documentation explains common grasp simulation pitfalls and workarounds (related to physics materials, contact settings, collision approximations, simulation timestep, masses, and gains) to help you achieve stable, realistic grasps faster.
• Robot rigging documentation: Updated documentation covers the top robot-rigging issues and provides step-by-step troubleshooting tips, plus gain-tuning guidance for position-, velocity-, and torque-controlled robots.

Accessibility:

• Benchmarking documentation: Benchmark documentation now more clearly explains what each benchmark measures and its significance to critical workflows. This includes benchmark-specific optimization guidance, GPU scaling notes, and system configuration considerations.

Get started:

• Download the source code for the early developer preview of Isaac Sim 6.0 from Github.

• Connect with the community and submit feedback on Isaac Sim 6.0 on Discord or Github issues.