Isaac Sim

NVIDIA Omniverse™ Isaac Sim, is a scalable robotics simulation application and synthetic data generation tool, built on the NVIDIA Omniverse™ platform. It leverages Omniverse Kit powerful simulation technologies including advanced GPU-enabled physics simulation and photorealism with real-time ray tracing.

Isaac Sim is a physically accurate virtual environment to develop, test, and manage AI-based robots, and provides the essential features for building virtual robotics experiments including:

Tools for importing and simulating robot models, and simulating sensors such as RGB-D, PhysX-Lidar, RTX-Lidar, Radar, contact and IMU

Support for various robotics applications including navigation, manipulation, and AI applications through Python and ROS

Extensive tools for AI applications such as Isaac Replicator for synthetic data generation and domain randomization and Isaac Lab for robot learning

Extensions for building a warehouse such as sim-ready assets, conveyor tools, peoples and CuOpt for logistics and route optimization

Development tools, Isaac Core API is a robotics abstraction layer which hides the USD complexity

This release is offered as a container that runs locally or on RTX equipped Amazon Web Services (AWS), Microsoft Azure or Google cloud (GCP) with streaming to the user’s desktop. This cloud-based delivery provides the latest RTX graphics and performance to any desktop system without requiring local NVIDIA RTX GPUs.

What's New in Isaac Sim 5.1.0

General

- Updated to Kit 107.3.3

- DGX Spark support

- The Compatibility Checker is no longer a standalone application. It is now integrated into Isaac Sim in all its installation modalities.

- All deprecated extensions will be fully removed in the next major 6.0 release

PhysX

- Articulations

• New joint parameter tuning tutorial for robotic grippers.
• New feature for solving articulation collision contacts last to improve on gripper penetration issues, especially for cases with sub-optimally tuned joints.
• Fixed regression for interaction with deformables.

Synthetic Data Generation

- Perception Data Generation (Replicator)

• Tutorial for generating synthetic data for Cosmos

- Action and Event Data Generation

• Updated the Navigation Mesh API in Actor SDG
• Removed unused settings in Actor SDG
• Fixed the writer trigger bug in Object SDG

- Data Generation with MobilityGen

• Improved documentation for data generation script parameters
• Added documentation for configuring custom robot parameters

Robots

- New robots:

• Egk25
• Egu50
• Ezu35
• Svh hand (left, right)

- New Robot Booster T1

- Updated G1

• Updated left and right hand as variants, support three finger and inspire hand

- Added UR10e follow target examples

Sensors

- Added support for semantic segmentation (Object ID) using RTX Sensor

- Enabled non-visual materials for RTX Sensor using USD attributes, supported by new APIs

- Enabled device-side data processing for IMU sensor, improving performance

- Replaced OgnIsaacExtractRTXSensorPointCloud with per-frame output option on OgnIsaacCreateRTXLidarScanBuffer

- MotionBVH now disabled for RTX Sensor by default to improve baseline performance

- Performance improved for all RTX Sensor Annotators

- Fixed sensor return timestamp output from OgnIsaacCreateRTXLidarScanBuffer

- Fixed many small bugs in LidarRtx API

- Fixed specifying translation and orientation when creating sensors via command

- imageSize attribute now correctly specified for OpenCV lens distortion when using Camera API

- SingleViewDepthSensorAsset correctly sets translation and orientation when added to stage

ROS

- Add Simulation Interfaces Worlds support 1.1.0 to ROS 2 Simulation Control extension and updated service/action names. New services added: GetAvailableWorlds, GetCurrentWorld, LoadWorld, UnloadWorld.

- Optimized performance using rclpy executors within the ROS 2 Simulation Control extension.

- Simulation Control extension easily enabled at startup by setting:

       --isaac/startup/ros_sim_control_extension=True\n

- Updated internal ROS 2 Humble and Jazzy libraries to now include common_interfaces, tf2_ros, sensor_msgs_py, and simulation_interfaces v1.1.0

- Default ROS 2 bridge library setting set to system_default. On Ubuntu 22.04, ROS 2 Humble internal libs are loaded automatically. On Ubuntu 24.04, ROS 2 Jazzy internal libs are loaded automatically.

- CameraInfo message corrected to not always set fy equal to fx.

- General performance and memory management improvements with message generation and ROS 2 publishing.

- Fixed duplicate and missing timestamps when publishing ROS 2 messages.

Live-streaming

- Live-streaming is not yet supported on aarch64.

Docker

- Added support for multi-arch.

- Container runs as a rootless user by default.

Prerequisites

Requirements for local deployment:

• Ubuntu 22.04/24.04 Operating System
• NVIDIA GPU (GeForce RTX 4080 or higher)
• NVIDIA GPU Driver (recommended version HERE )
• Docker
• NVIDIA Container Toolkit (minimum version 1.17.0)

Running Isaac Sim

Steps:

1. Check if the host meets the System Requirements for running Isaac Sim.
2. Run the command below to confirm your GPU driver version.$ nvidia-smi
3. Pull the Isaac Sim container.$ docker pull nvcr.io/nvidia/isaac-sim:5.1.0
4. Create the cached volume mounts on host:$ mkdir -p ~/docker/isaac-sim/cache/main/ov $ mkdir -p ~/docker/isaac-sim/cache/main/warp $ mkdir -p ~/docker/isaac-sim/cache/computecache $ mkdir -p ~/docker/isaac-sim/config $ mkdir -p ~/docker/isaac-sim/data/documents $ mkdir -p ~/docker/isaac-sim/data/Kit $ mkdir -p ~/docker/isaac-sim/logs $ mkdir -p ~/docker/isaac-sim/pkg $ sudo chown -R 1234:1234 ~/docker/isaac-sim
5. Run the Isaac Sim container with an interactive Bash session.$ docker run --name isaac-sim --entrypoint bash -it --runtime=nvidia --gpus all -e \"ACCEPT_EULA=Y\" --rm --network=host \\ -e \"PRIVACY_CONSENT=Y\" \\ -v ~/docker/isaac-sim/cache/main:/isaac-sim/.cache:rw \\ -v ~/docker/isaac-sim/cache/computecache:/isaac-sim/.nv/ComputeCache:rw \\ -v ~/docker/isaac-sim/logs:/isaac-sim/.nvidia-omniverse/logs:rw \\ -v ~/docker/isaac-sim/config:/isaac-sim/.nvidia-omniverse/config:rw \\ -v ~/docker/isaac-sim/data:/isaac-sim/.local/share/ov/data:rw \\ -v ~/docker/isaac-sim/pkg:/isaac-sim/.local/share/ov/pkg:rw \\ -u 1234:1234 \\ nvcr.io/nvidia/isaac-sim:5.1.0
6. Start Isaac Sim.$ ./runheadless.sh -v

This will download and launch the Isaac Sim container in headless mode.

Note:

• By using the -e \"ACCEPT_EULA=Y\" flag, you are accepting the NVIDIA Isaac Sim Additional Software and Materials License of the image.
• By using the -e \"PRIVACY_CONSENT=Y flag, you opt-in to the data collection agreement found at Omniverse Data Collection & Use FAQ. You may opt-licenseut by not setting this flag.
• The -e \"PRIVACY_USERID= flag can optionally be set for tagging the session logs.
• Livestreaming is not supported on aarch64 systems like DGX Spark for Isaac Sim 5.1.0.

Refer to our Isaac Sim User Guide and Isaac Sim Setup FAQ for more information.

Livestreaming Isaac Sim Remotely

See Livestream Clients in our User Guide to connect to Isaac Sim using your desktop. The default livestream client is Isaac Sim WebRTC Streaming Client.

License

By pulling and using the container, you accept the terms and conditions of the NVIDIA Isaac Sim Additional Software and Materials License.

Sources for OSS packages used in this container can be found here.

Suggested readings:

• Release Notes
• Documentation
• NVIDIA Forums