Getting Started

Isaac ROS is a collection of NVIDIA® CUDA®-accelerated computing packages and AI models designed to streamline and expedite the development of advanced AI robotics applications.

NVIDIA Isaac ROS is built on open-source ROS 2. Through Isaac ROS, developers in the ROS community can take advantage of NVIDIA-accelerated libraries and AI models to fast track their AI robot development and deployment workflows.

System Requirements

Supported Platforms

The platforms defined in this table are the only hardware and software combinations that Isaac ROS tests and officially supports.

Users may be able to rely on backward and forward compatibility utilities like cuda-compat to use Isaac ROS on other platform versions.

Platform	Hardware	Software	Storage	Notes
Jetson	Jetson Thor (T5000 and T4000) and Jetson Orin	JetPack 7.2	128+ GB NVMe SSD	For best performance, ensure that power settings are configured appropriately for your platform: Jetson Thor power settings or Jetson Orin power settings.
x86_64	Ampere or higher NVIDIA GPU Architecture with 8 GB RAM or higher	Ubuntu 24.04	32+ GB disk space available	CUDA 13.2+ NVIDIA Driver 595+
DGX	DGX Spark	DGX OS 7.2.3	32+ GB disk space available	For best performance, use the supplied power adapter with the DGX Spark system.

ROS Support

All Isaac ROS packages are designed and tested to be compatible with ROS 2 Jazzy.

Initialize Compute

x86 Platforms

1. Install the latest NVIDIA GPU Driver using the official instructions.

2. Verify the NVIDIA GPU Driver is installed correctly:

   nvidia-smi
   

   Confirm that all relevant fields satisfy the requirements.

   • Driver Version

   • CUDA Version

   • GPU Memory

Jetson Platforms

Jetson AGX Thor

1. Complete the Jetson AGX Thor Quick Start Guide.

2. In the JetPack setup guide, complete only the Jetson AGX Thor Install all components section.

3. After boot, confirm that you have installed the correct version of Jetpack by running the following command. Confirm that the output has the terms R39 (release), REVISION: 2.0.

   cat /etc/nv_tegra_release
   

4. Set the power mode to MAXN using the Jetson Thor guide.

5. Set the GPU and CPU clock speeds to max using the Jetson Thor guide.

6. Consider the following recommendations for a better experience:

   • Jetson Storage Setup

   • Preempt Setup

Jetson AGX Orin

1. Complete the Jetson AGX Orin Quick Start Guide. Installing the BSP to an NVMe SSD is recommended when using Virtual Environment or Bare Metal modes.

2. In the JetPack setup guide, complete only the Jetson AGX Orin Install via apt section.

3. After boot, confirm that you have installed the correct version of Jetpack by running the following command. Confirm that the output has the terms R39 (release), REVISION: 2.0.

   cat /etc/nv_tegra_release
   

4. Set the power mode to MAXN using the Jetson Orin guide.

5. Set the GPU and CPU clock speeds to max using the Jetson Orin guide.

6. Consider the following recommendations for a better experience:

   • Jetson Storage Setup

   • Preempt Setup

DGX Spark Platforms

No additional compute initialization is required for DGX Spark in this section. Continue to Create a Workspace.

Create a Workspace

x86 Platforms

Create a ROS 2 workspace for experimenting with Isaac ROS:

mkdir -p  ~/workspaces/isaac_ros-dev/src
echo 'export ISAAC_ROS_WS="${ISAAC_ROS_WS:-${HOME}/workspaces/isaac_ros-dev/}"' >> ~/.bashrc
source ~/.bashrc

Jetson Platforms

Jetson AGX Thor

Create a ROS 2 workspace for experimenting with Isaac ROS:

mkdir -p  ~/workspaces/isaac_ros-dev/src
echo 'export ISAAC_ROS_WS="${ISAAC_ROS_WS:-${HOME}/workspaces/isaac_ros-dev/}"' >> ~/.bashrc
source ~/.bashrc

Jetson AGX Orin

If the BSP is installed on eMMC rather than NVMe, install an SSD using the Jetson Storage Setup instructions, then create the workspace on the SSD:

mkdir -p  /mnt/nova_ssd/workspaces/isaac_ros-dev/src
echo 'export ISAAC_ROS_WS="${ISAAC_ROS_WS:-/mnt/nova_ssd/workspaces/isaac_ros-dev/}"' >> ~/.bashrc
source ~/.bashrc

If the BSP is installed on NVMe, create the workspace in your home directory:

mkdir -p  ~/workspaces/isaac_ros-dev/src
echo 'export ISAAC_ROS_WS="${ISAAC_ROS_WS:-${HOME}/workspaces/isaac_ros-dev/}"' >> ~/.bashrc
source ~/.bashrc

DGX Spark Platforms

Create a ROS 2 workspace for experimenting with Isaac ROS:

mkdir -p  ~/workspaces/isaac_ros-dev/src
echo 'export ISAAC_ROS_WS="${ISAAC_ROS_WS:-${HOME}/workspaces/isaac_ros-dev/}"' >> ~/.bashrc
source ~/.bashrc

Configure Isaac ROS Apt Repository

The NVIDIA Isaac ROS Apt Repository hosts the following categories of packages for all users:

• Isaac ROS CLI package for managing Isaac ROS development environments

• Pre-built Isaac ROS packages for currently-supported platforms

• Auxiliary packages to improve developer experience or to support a legacy OS

To configure the Isaac ROS Apt Repository, follow the steps below:

1. Set the locale on your host system to UTF-8:

   locale  # check for UTF-8
   
   sudo apt update && sudo apt install locales
   sudo locale-gen en_US en_US.UTF-8
   sudo update-locale LC_ALL=en_US.UTF-8 LANG=en_US.UTF-8
   export LANG=en_US.UTF-8
   
   locale  # verify settings
   

2. Install the dependencies:

   sudo apt update && sudo apt install curl gnupg
   sudo apt install software-properties-common
   sudo add-apt-repository universe
   

3. Source the Isaac ROS Apt Repository:

   Select your platform to see the appropriate APT repository sourcing instructions.

   x86 (Ubuntu 24.04 Noble)

   US CDN

   Option 1: Specific Version (release-4.6)

   Use this option to pin to a specific minor version.

   k="/usr/share/keyrings/nvidia-isaac-ros.gpg"
   curl -fsSL https://isaac.download.nvidia.com/isaac-ros/repos.key | sudo gpg --dearmor \
       | sudo tee -a $k > /dev/null
   f="/etc/apt/sources.list.d/nvidia-isaac-ros.list"
   sudo touch $f
   s="deb [signed-by=$k] https://isaac.download.nvidia.com/isaac-ros/release-4.6 noble main"
   grep -qxF "$s" $f || echo "$s" | sudo tee -a $f
   
   sudo apt-get update
   

   Option 2: Latest Minor Version (release-4)

   Use this option to automatically receive updates for the latest release-4 minor version.

   k="/usr/share/keyrings/nvidia-isaac-ros.gpg"
   curl -fsSL https://isaac.download.nvidia.com/isaac-ros/repos.key | sudo gpg --dearmor \
       | sudo tee -a $k > /dev/null
   f="/etc/apt/sources.list.d/nvidia-isaac-ros.list"
   sudo touch $f
   s="deb [signed-by=$k] https://isaac.download.nvidia.com/isaac-ros/release-4 noble main"
   grep -qxF "$s" $f || echo "$s" | sudo tee -a $f
   
   sudo apt-get update
   

   China CDN

   Option 1: Specific Version (release-4.6)

   Use this option to pin to a specific minor version.

   k="/usr/share/keyrings/nvidia-isaac-ros.gpg"
   curl -fsSL https://isaac.download.nvidia.cn/isaac-ros/repos.key | sudo gpg --dearmor \
       | sudo tee -a $k > /dev/null
   f="/etc/apt/sources.list.d/nvidia-isaac-ros.list"
   sudo touch $f
   s="deb [signed-by=$k] https://isaac.download.nvidia.cn/isaac-ros/release-4.6 noble main"
   grep -qxF "$s" $f || echo "$s" | sudo tee -a $f
   
   sudo apt-get update
   

   Option 2: Latest Minor Version (release-4)

   Use this option to automatically receive updates for the latest release-4 minor version.

   k="/usr/share/keyrings/nvidia-isaac-ros.gpg"
   curl -fsSL https://isaac.download.nvidia.cn/isaac-ros/repos.key | sudo gpg --dearmor \
       | sudo tee -a $k > /dev/null
   f="/etc/apt/sources.list.d/nvidia-isaac-ros.list"
   sudo touch $f
   s="deb [signed-by=$k] https://isaac.download.nvidia.cn/isaac-ros/release-4 noble main"
   grep -qxF "$s" $f || echo "$s" | sudo tee -a $f
   
   sudo apt-get update
   

   Jetson (JetPack 7.2)

   US CDN

   Option 1: Specific Version (release-4.6)

   Use this option to pin to a specific minor version.

   k="/usr/share/keyrings/nvidia-isaac-ros.gpg"
   curl -fsSL https://isaac.download.nvidia.com/isaac-ros/repos.key | sudo gpg --dearmor \
       | sudo tee -a $k > /dev/null
   f="/etc/apt/sources.list.d/nvidia-isaac-ros.list"
   sudo touch $f
   s="deb [signed-by=$k] https://isaac.download.nvidia.com/isaac-ros/release-4.6 noble-jetpack main"
   grep -qxF "$s" $f || echo "$s" | sudo tee -a $f
   
   sudo apt-get update
   

   Option 2: Latest Minor Version (release-4)

   Use this option to automatically receive updates for the latest release-4 minor version.

   k="/usr/share/keyrings/nvidia-isaac-ros.gpg"
   curl -fsSL https://isaac.download.nvidia.com/isaac-ros/repos.key | sudo gpg --dearmor \
       | sudo tee -a $k > /dev/null
   f="/etc/apt/sources.list.d/nvidia-isaac-ros.list"
   sudo touch $f
   s="deb [signed-by=$k] https://isaac.download.nvidia.com/isaac-ros/release-4 noble-jetpack main"
   grep -qxF "$s" $f || echo "$s" | sudo tee -a $f
   
   sudo apt-get update
   

   China CDN

   Option 1: Specific Version (release-4.6)

   Use this option to pin to a specific minor version.

   k="/usr/share/keyrings/nvidia-isaac-ros.gpg"
   curl -fsSL https://isaac.download.nvidia.cn/isaac-ros/repos.key | sudo gpg --dearmor \
       | sudo tee -a $k > /dev/null
   f="/etc/apt/sources.list.d/nvidia-isaac-ros.list"
   sudo touch $f
   s="deb [signed-by=$k] https://isaac.download.nvidia.cn/isaac-ros/release-4.6 noble-jetpack main"
   grep -qxF "$s" $f || echo "$s" | sudo tee -a $f
   
   sudo apt-get update
   

   Option 2: Latest Minor Version (release-4)

   Use this option to automatically receive updates for the latest release-4 minor version.

   k="/usr/share/keyrings/nvidia-isaac-ros.gpg"
   curl -fsSL https://isaac.download.nvidia.cn/isaac-ros/repos.key | sudo gpg --dearmor \
       | sudo tee -a $k > /dev/null
   f="/etc/apt/sources.list.d/nvidia-isaac-ros.list"
   sudo touch $f
   s="deb [signed-by=$k] https://isaac.download.nvidia.cn/isaac-ros/release-4 noble-jetpack main"
   grep -qxF "$s" $f || echo "$s" | sudo tee -a $f
   
   sudo apt-get update
   

   DGX Spark (FastOS)

   US CDN

   Option 1: Specific Version (release-4.6)

   Use this option to pin to a specific minor version.

   k="/usr/share/keyrings/nvidia-isaac-ros.gpg"
   curl -fsSL https://isaac.download.nvidia.com/isaac-ros/repos.key | sudo gpg --dearmor \
       | sudo tee -a $k > /dev/null
   f="/etc/apt/sources.list.d/nvidia-isaac-ros.list"
   sudo touch $f
   s="deb [signed-by=$k] https://isaac.download.nvidia.com/isaac-ros/release-4.6 noble-fastos main"
   grep -qxF "$s" $f || echo "$s" | sudo tee -a $f
   
   sudo apt-get update
   

   Option 2: Latest Minor Version (release-4)

   Use this option to automatically receive updates for the latest release-4 minor version.

   k="/usr/share/keyrings/nvidia-isaac-ros.gpg"
   curl -fsSL https://isaac.download.nvidia.com/isaac-ros/repos.key | sudo gpg --dearmor \
       | sudo tee -a $k > /dev/null
   f="/etc/apt/sources.list.d/nvidia-isaac-ros.list"
   sudo touch $f
   s="deb [signed-by=$k] https://isaac.download.nvidia.com/isaac-ros/release-4 noble-fastos main"
   grep -qxF "$s" $f || echo "$s" | sudo tee -a $f
   
   sudo apt-get update
   

   China CDN

   Option 1: Specific Version (release-4.6)

   Use this option to pin to a specific minor version.

   k="/usr/share/keyrings/nvidia-isaac-ros.gpg"
   curl -fsSL https://isaac.download.nvidia.cn/isaac-ros/repos.key | sudo gpg --dearmor \
       | sudo tee -a $k > /dev/null
   f="/etc/apt/sources.list.d/nvidia-isaac-ros.list"
   sudo touch $f
   s="deb [signed-by=$k] https://isaac.download.nvidia.cn/isaac-ros/release-4.6 noble-fastos main"
   grep -qxF "$s" $f || echo "$s" | sudo tee -a $f
   
   sudo apt-get update
   

   Option 2: Latest Minor Version (release-4)

   Use this option to automatically receive updates for the latest release-4 minor version.

   k="/usr/share/keyrings/nvidia-isaac-ros.gpg"
   curl -fsSL https://isaac.download.nvidia.cn/isaac-ros/repos.key | sudo gpg --dearmor \
       | sudo tee -a $k > /dev/null
   f="/etc/apt/sources.list.d/nvidia-isaac-ros.list"
   sudo touch $f
   s="deb [signed-by=$k] https://isaac.download.nvidia.cn/isaac-ros/release-4 noble-fastos main"
   grep -qxF "$s" $f || echo "$s" | sudo tee -a $f
   
   sudo apt-get update
   

Select Environment Isolation Mode

Based on your use case, decide on an appropriate level of isolation between Isaac ROS and your host system:

Tip

Docker is the recommended option for most users. It provides the highest level of isolation from your host system.

In Virtual Environment or Bare Metal mode, you may need to remove existing packages from your host system to avoid version conflicts with Isaac ROS and its dependencies.

Important

On Jetson AGX Orin, we recommend placing the root filesystem on an NVMe SSD when using Virtual Environment or Bare Metal modes. These modes install Isaac ROS dependencies, including ML frameworks, on the root filesystem and can exceed the default eMMC capacity.

Mode	Debian packages	Python pip packages	Isolation
Docker (recommended)	Installed inside the container (host unaffected)	Installed inside the container (host unaffected)	High (full container isolation)
Virtual Environment	Installed on host via apt	Installed into Isaac ROS virtual environment (isolated from system Python)	Medium
Bare Metal	Installed on host via apt	Installed into system Python site-packages	None

This selection will inform subsequent steps in the setup process.

Initialize Isaac ROS CLI

1. Install the Isaac ROS CLI:

   sudo apt-get install isaac-ros-cli
   

2. Configure the environment isolation mode you selected earlier:

   Docker

   1. Complete platform-specific Docker setup:

      x86 Platforms

      1. Complete the Docker Engine setup using the official Ubuntu instructions.

      2. Install and configure nvidia-container-toolkit for Docker using the official instructions:

         • Install NVIDIA Container Toolkit

         • Configure Docker runtime

      3. Add your user to the docker group:

         sudo usermod -aG docker $USER
         newgrp docker
         

      Jetson Platforms

      1. Complete the Jetson AGX Thor Docker Setup guide or Jetson AGX Orin Docker Setup guide.

      DGX Spark Platforms

      1. Docker and NVIDIA Container Toolkit are already configured on DGX Spark. See the DGX Spark container runtime documentation.

      2. Add your user to the docker group:

         sudo usermod -aG docker $USER
         newgrp docker
         

   2. Restart Docker:

      sudo systemctl daemon-reload && sudo systemctl restart docker
      

   3. Run Docker pre-flight checks:

      docker info | grep -E "Runtimes|Default Runtime"
      docker run --rm hello-world
      docker run --rm --gpus all ubuntu:24.04 bash -lc 'echo "NVIDIA runtime OK"'
      

      Confirm that Docker lists available runtime entries, hello-world exits successfully, and the GPU runtime check prints NVIDIA runtime OK.

   4. Initialize the Isaac ROS CLI:

      sudo isaac-ros init docker
      

   
   

   Customizing the Docker Image

   The default Isaac ROS Docker image contains a core set of dependencies for basic Isaac ROS usage.

   To add additional dependencies (e.g., RealSense camera support), refer to the Docker Mode Configuration documentation.

   Virtual Environment

   1. Configure your host environment for Isaac ROS:

      1. Configure the CUDA apt repository:

         x86 Platforms

         wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2404/x86_64/cuda-keyring_1.1-1_all.deb
         sudo dpkg -i cuda-keyring_1.1-1_all.deb
         rm cuda-keyring_1.1-1_all.deb
         sudo apt update
         

         Jetson Platforms

         wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2404/sbsa/cuda-keyring_1.1-1_all.deb
         sudo dpkg -i cuda-keyring_1.1-1_all.deb
         rm cuda-keyring_1.1-1_all.deb
         sudo apt update
         

         DGX Spark Platforms

         No CUDA apt repository setup is required on DGX Spark platforms.

      2. Add NVIDIA binary installation locations to your PATH:

         x86 Platforms

         echo 'export PATH="/usr/local/nvidia/bin:/usr/local/cuda/bin:/usr/src/tensorrt/bin:${PATH}"' >> ~/.bashrc
         source ~/.bashrc
         

         Jetson Platforms

         Nothing required on Jetson platforms.

         DGX Spark Platforms

         Nothing required on DGX Spark platforms.

      3. Configure the Jetson apt repository:

         x86 Platforms

         curl -fsSL https://repo.download.nvidia.com/jetson/jetson-ota-public.asc | sudo gpg --dearmor | sudo tee /etc/apt/keyrings/nvidia-jetson.gpg > /dev/null
         echo 'deb [signed-by=/etc/apt/keyrings/nvidia-jetson.gpg] https://repo.download.nvidia.com/jetson/x86_64/noble r39.2 main' | sudo tee /etc/apt/sources.list.d/nvidia-jetson-apt-source.list
         

         Jetson Platforms

         Jetson platforms already include the NVIDIA Jetson APT repository by default, so no additional setup is required for this step.

         DGX Spark Platforms

         No Jetson apt repository setup is required on DGX Spark platforms.

      4. Install ROS 2 Jazzy by following the official ROS 2 documentation:

      5. Install ROS 2 development tools:

         sudo apt-get update
         sudo apt-get install -y python3-colcon-common-extensions python3-rosdep
         

      6. For convenience, source your ROS 2 installation and enable colcon auto-completion in your .bashrc file:

         These instructions are for bash shells; for other shells, configure accordingly.

         echo "source /opt/ros/jazzy/setup.bash" >> ~/.bashrc
         echo "source /usr/share/colcon_argcomplete/hook/colcon-argcomplete.bash" >> ~/.bashrc
         source ~/.bashrc
         

      7. Initialize rosdep and configure the Isaac ROS rosdep definitions:

         sudo rosdep init
         sudo curl -o /etc/ros/rosdep/sources.list.d/nvidia-isaac.yaml \
            https://raw.githubusercontent.com/NVIDIA-ISAAC-ROS/isaac-ros-cli/release-4.6/docker/rosdep/extra_rosdeps.yaml
         echo "yaml file:///etc/ros/rosdep/sources.list.d/nvidia-isaac.yaml" | sudo tee /etc/ros/rosdep/sources.list.d/00-nvidia-isaac.list
         rosdep update
         

      8. Remove potentially conflicting packages from your host system.

         As specified in the Isaac ROS CLI apt preference files, Isaac ROS has been tested with specific versions of the following packages:

         • CUDA 13.2 (Jetson / x86) / CUDA 13.0 (DGX Spark)

         • OpenCV 4.6.0

         • TensorRT 10.16.*

         If you have installed different versions of these packages on your host system, you may need to remove them to avoid conflicts.

         Note

         On Jetson with JetPack 7.2, the OpenCV 4.8.0 packages are installed by default.

         Remove the conflicting packages with the following command:

         sudo apt-get remove -y libopencv* opencv*
         

         OpenCV 4.6.0 will then be automatically installed by the Isaac ROS packages that require it.

         Note

         On DGX Spark platforms bare-metal and virtual environment workflows, rosdep install commands in later steps may fail with an error like:

         E: Packages were downgraded and -y was used without --allow-downgrades.

         This occurs because some DGX OS system packages are at newer versions than those pinned by Isaac ROS, causing apt to reject the entire installation transaction. This means even unrelated packages fail to install.

         See rosdep install fails on DGX Spark with package downgrade error for resolution options.

   2. Initialize the Isaac ROS CLI:

      sudo isaac-ros init venv
      

   Bare Metal

   Caution

   The baremetal mode is only recommended for advanced users.

   This mode installs Isaac ROS packages and their dependencies into your system Python environment.

   Without proper care, this has the potential to break your host system.

   1. Configure your host environment for Isaac ROS:

      1. Configure the CUDA apt repository:

         x86 Platforms

         wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2404/x86_64/cuda-keyring_1.1-1_all.deb
         sudo dpkg -i cuda-keyring_1.1-1_all.deb
         rm cuda-keyring_1.1-1_all.deb
         sudo apt update
         

         Jetson Platforms

         wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2404/sbsa/cuda-keyring_1.1-1_all.deb
         sudo dpkg -i cuda-keyring_1.1-1_all.deb
         rm cuda-keyring_1.1-1_all.deb
         sudo apt update
         

         DGX Spark Platforms

         No CUDA apt repository setup is required on DGX Spark platforms.

      2. Add NVIDIA binary installation locations to your PATH:

         x86 Platforms

         echo 'export PATH="/usr/local/nvidia/bin:/usr/local/cuda/bin:/usr/src/tensorrt/bin:${PATH}"' >> ~/.bashrc
         source ~/.bashrc
         

         Jetson Platforms

         Nothing required on Jetson platforms.

         DGX Spark Platforms

         Nothing required on DGX Spark platforms.

      3. Configure the Jetson apt repository:

         x86 Platforms

         curl -fsSL https://repo.download.nvidia.com/jetson/jetson-ota-public.asc | sudo gpg --dearmor | sudo tee /etc/apt/keyrings/nvidia-jetson.gpg > /dev/null
         echo 'deb [signed-by=/etc/apt/keyrings/nvidia-jetson.gpg] https://repo.download.nvidia.com/jetson/x86_64/noble r39.2 main' | sudo tee /etc/apt/sources.list.d/nvidia-jetson-apt-source.list
         

         Jetson Platforms

         Jetson platforms already include the NVIDIA Jetson APT repository by default, so no additional setup is required for this step.

         DGX Spark Platforms

         No Jetson apt repository setup is required on DGX Spark platforms.

      4. Install ROS 2 Jazzy by following the official ROS 2 documentation:

      5. Install ROS 2 development tools:

         sudo apt-get update
         sudo apt-get install -y python3-colcon-common-extensions python3-rosdep
         

      6. For convenience, source your ROS 2 installation and enable colcon auto-completion in your .bashrc file:

         These instructions are for bash shells; for other shells, configure accordingly.

         echo "source /opt/ros/jazzy/setup.bash" >> ~/.bashrc
         echo "source /usr/share/colcon_argcomplete/hook/colcon-argcomplete.bash" >> ~/.bashrc
         source ~/.bashrc
         

      7. Initialize rosdep and configure the Isaac ROS rosdep definitions:

         sudo rosdep init
         sudo curl -o /etc/ros/rosdep/sources.list.d/nvidia-isaac.yaml \
            https://raw.githubusercontent.com/NVIDIA-ISAAC-ROS/isaac-ros-cli/release-4.6/docker/rosdep/extra_rosdeps.yaml
         echo "yaml file:///etc/ros/rosdep/sources.list.d/nvidia-isaac.yaml" | sudo tee /etc/ros/rosdep/sources.list.d/00-nvidia-isaac.list
         rosdep update
         

      8. Remove potentially conflicting packages from your host system.

         As specified in the Isaac ROS CLI apt preference files, Isaac ROS has been tested with specific versions of the following packages:

         • CUDA 13.2 (Jetson / x86) / CUDA 13.0 (DGX Spark)

         • OpenCV 4.6.0

         • TensorRT 10.16.*

         If you have installed different versions of these packages on your host system, you may need to remove them to avoid conflicts.

         Note

         On Jetson with JetPack 7.2, the OpenCV 4.8.0 packages are installed by default.

         Remove the conflicting packages with the following command:

         sudo apt-get remove -y libopencv* opencv*
         

         OpenCV 4.6.0 will then be automatically installed by the Isaac ROS packages that require it.

         Note

         On DGX Spark platforms bare-metal and virtual environment workflows, rosdep install commands in later steps may fail with an error like:

         E: Packages were downgraded and -y was used without --allow-downgrades.

         This occurs because some DGX OS system packages are at newer versions than those pinned by Isaac ROS, causing apt to reject the entire installation transaction. This means even unrelated packages fail to install.

         See rosdep install fails on DGX Spark with package downgrade error for resolution options.

   2. Initialize the Isaac ROS CLI:

      sudo isaac-ros init baremetal
      

      Acknowledge the prompt, then confirm to continue.

      Unattended/Non-Interactive Installation

      In some cases, it may be desirable to skip the prompt when initializing the baremetal mode.

      To perform a non-interactive installation, use the --yes flag:

      sudo isaac-ros init baremetal --yes
      

Configuring the Isaac ROS Development Environment

The Isaac ROS CLI uses layered configuration files to manage settings across all isolation modes.

To use these configuration files and customize your development environment, refer to the Isaac ROS Development Environment page.

Integrate External Data Sources (Optional)

Many Isaac ROS demos include pre-recorded rosbag data to help you get started as quickly as possible.

Optionally, these demos can also be integrated with external data sources, including cameras and the Isaac Sim simulation platform.

Sensors

Isaac ROS provides setup instructions for specific camera models from the following product lines:

• RealSense

• StereoLabs ZED

• Leopard Imaging Eagle

• Leopard Imaging Hawk

Calibrate your cameras depending on your type of camera and the object it is on:

• Monocular Camera

• Stereo Camera on Robot Arm

• Stereo Camera on Mobile Robot

Isaac Sim

Isaac Sim is NVIDIA’s robotics simulation platform powered by Omniverse.

Isaac Sim can be used to run and test the ROS 2 applications. There are two ways we can configure the system:

• Software in the Loop (SIL) refers to a configuration where the software being tested is not running on the target hardware platform. For example, Isaac ROS packages being tested on x86 before deployment on a Jetson device is SIL.

• Hardware in the Loop (HIL) refers to a configuration where the software is being tested on the target hardware platform. For example, Isaac ROS packages being tested on a Jetson device before deployment is HIL.

In both the configurations, Isaac Sim runs on a x86_64 machine providing sensor data and world information.

Isaac Sim 6.0

1. Install Isaac Sim, choosing the appropriate working environment:

   • Native

   • Docker

   • Cloud

   Note

   This tutorial assumes the Native environment for Isaac Sim. The steps below assume Isaac Sim was extracted to an installation directory referred to as <isaac-sim-root> (for example, ~/isaacsim).

2. Set up ROS 2 for Isaac Sim.

   Isaac Sim 6.0 ships with full Python 3.12 support and includes the ROS 2 Bridge. You can drive the bridge using either a system-level ROS 2 installation or the internal ROS 2 libraries bundled with Isaac Sim. For additional details, refer to the Running Native ROS instructions.

   System-level ROS 2 (Jazzy)

   Because Isaac Sim 6.0 supports Python 3.12, a system-level ROS 2 Jazzy installation can be sourced directly in the same terminal as Isaac Sim. Follow the steps here under Ubuntu 24.04 > Jazzy, then source your ROS 2 installation before launching Isaac Sim:

   source /opt/ros/jazzy/setup.bash
   

   Internal ROS 2 libraries

   If you do not have a system-level ROS 2 installation, export the following environment variables so that the ROS 2 Bridge can load the internal libraries bundled with Isaac Sim. If these variables are not set, the bridge fails to start with an error such as librcutils.so: cannot open shared object file; Isaac Sim prints these same instructions to its console log when this happens.

   export ROS_DISTRO=jazzy
   export RMW_IMPLEMENTATION=rmw_fastrtps_cpp
   export LD_LIBRARY_PATH=${LD_LIBRARY_PATH}:<isaac-sim-root>/exts/isaacsim.ros2.core/jazzy/lib
   

   To use CycloneDDS instead of the default FastDDS, set RMW_IMPLEMENTATION=rmw_cyclonedds_cpp.

   Note

   Isaac Sim provides a setup_ros_env.sh helper at the root of the installation that exports these variables for you:

   source <isaac-sim-root>/setup_ros_env.sh
   

   This helper only updates LD_LIBRARY_PATH when ROS_DISTRO is not already set, so run it from a fresh terminal (or export the variables manually) to avoid silently skipping the library path.

   Optionally, if ROS_DOMAIN_ID is needed to create a separate logical network, run the following command:

   export ROS_DOMAIN_ID=<your_domain_id>
   

3. Launch Isaac Sim using the launch script from the root of your Isaac Sim installation:

   cd <isaac-sim-root>
   ./isaac-sim.sh
   

   Note

   (Optional) NITROS bridge provides zero-copy GPU data transfer between Isaac Sim and Isaac ROS. This feature pre-allocates IPC GPU memory regardless of utilization. To enable NITROS bridge on the Isaac Sim side, run:

   ./isaac-sim.sh --/exts/isaacsim.ros2.bridge/enable_nitros_bridge=true
   

4. Open Isaac ROS Sample scene as shown below.

   First, enable Robotics Examples menu by clicking Window > Examples > Robotics Examples

   

   Then, in the Robotics Examples tab, select ROS2 > Isaac ROS > Sample Scene and click Load Sample Scene

   

5. Continue with the next steps in your specific Isaac ROS package tutorial.

   Note

   Make sure to set the ROS_DOMAIN_ID environment variable (if set earlier) before running the sample application.

Isaac Sim 5.0 and Isaac Sim 5.1

1. Install Isaac Sim, choosing the appropriate working environment:

   • Native

   • Docker

   • Cloud

   Note

   This tutorial assumes the Native environment for Isaac Sim.

2. Follow the Running Native ROS instructions to ensure that ROS 2 and the required packages are installed to successfully enable the ROS 2 Bridge in Omniverse Isaac Sim.

3. Follow the instructions to launch Isaac Sim App Selector window and then click Open in Terminal button.

   

4. Using Isaac Sim with ROS 2

   Since Isaac Sim 5.1 only supports Python 3.11, follow the steps here under Ubuntu 24.04 > Jazzy. Ensure that you only run Isaac Sim with the internal ROS 2 libraries. If you have a system level ROS 2 installation, ensure to run Isaac Sim in a clean terminal and run ROS 2 nodes from a second system level ROS-sourced terminal.

   Sourcing a system level ROS 2 installation in the Isaac Sim terminal will only work with Isaac Sim if it was built from source with Python 3.11.

   Optionally, if ROS_DOMAIN_ID is needed to create a separate logical network, run the following commands.

   export ROS_DOMAIN_ID=<your_domain_id>
   

5. Launch Isaac Sim.

   ./isaac-sim.sh
   

   Note

   (Optional) NITROS bridge provides zero-copy GPU data transfer between Isaac Sim and Isaac ROS. This feature pre-allocates IPC GPU memory regardless of utilization. To enable NITROS bridge on the Isaac Sim side, run:

   ./isaac-sim.sh --/exts/isaacsim.ros2.bridge/enable_nitros_bridge=true
   

6. Open Isaac ROS Sample scene as shown below.

   First, enable Robotics Examples menu by clicking Window > Examples > Robotics Examples

   

   Then, in the Robotics Examples tab, select ROS2 > Isaac ROS > Sample Scene and click Load Sample Scene

   

7. Continue with the next steps in your specific Isaac ROS package tutorial.

   Note

   Make sure to set the ROS_DOMAIN_ID environment variable (if set earlier) before running the sample application.

Isaac Sim Tutorial Links

Isaac ROS includes a number of tutorials that demonstrate integration with Isaac Sim.

Each of these links takes you to a tutorial that is listed under Repositories and Packages. They are linked here as a reference and to help you find them faster.

Note

These tutorials were last validated with Isaac Sim 6.0 (recommended). They are also compatible with Isaac Sim 5.0 and Isaac Sim 5.1 releases.

• Tutorial for AprilTag Detection with Isaac Sim

• Tutorial for Visual SLAM with Isaac Sim

• Tutorial for Nvblox with Isaac Sim

• Tutorial for DNN Object Detection with Isaac Sim

• Tutorial for SGM Stereo Disparity with Isaac Sim

• Tutorial for DNN Stereo Depth Estimation with Isaac Sim

• Tutorial for Occupancy Grid Localizer with Isaac Sim

• Tutorial for DNN Image Segmentation with Isaac Sim

• Tutorial for Isaac ROS Mission Client

• Tutorial for RT-DETR with Isaac Sim

• Tutorial for FoundationPose with Isaac Sim

• Tutorial for NITROS Bridge with Isaac Sim

• Tutorial for cuMotion MoveIt Plugin with Isaac Sim

• Tutorials for Isaac for Manipulation Reference Workflows with Isaac Sim

Run an Isaac ROS Demo

You’re now ready to run an Isaac ROS demo!

Navigating the Isaac ROS Site

• Isaac ROS offers a collection of GitHub Repositories, each focused on a general robotics problem area. Every repository has its own page under the repository index.

• Isaac ROS repositories are further organized into ROS 2 Packages, each focused on a specific solution implementation or use case. Each package has its own page under the repository-level page.

• Most packages contain a Quickstart, which is a self-contained example that allows you to quickly test out the package’s core functionality. Quickstarts are located on each package’s page.

• Additionally, many packages contain more advanced Tutorials that dive deeper into the package’s functionality. When available, tutorials are listed below the Quickstart on each package’s page.

• Package pages also contain Troubleshooting and API sections to help you use Isaac ROS in your own applications.

• Robotics fundamentals, NVIDIA technology, and other Concepts are explained under the concepts index.

Browse the available Isaac ROS Repositories and Packages to find your first quickstart.

Troubleshooting

See the troubleshooting section for help with setting up your development environment.