Repositories and Packages

Use this index to find the Isaac ROS packages you need and examples on how to use them.

Repository

Package(s)

Description

AprilTag

CUDA-accelerated AprilTag detection and pose estimation.

Argus Camera

ROS 2 packages based on NVIDIA libArgus library for NVIDIA-accelerated CSI camera support.

Benchmark

ROS 2 package which adds on top of ros2_benchmark package plugins, configuration files, and scripts to perform benchmarking on graphs of nodes using Isaac ROS.

Common

Common utilities, packages, scripts, Dockerfiles, and testing infrastructure for Isaac ROS packages.

Compression

NVIDIA-accelerated data compression

cuMotion

CUDA-accelerated motion generation for robot arms

Depth Segmentation

DNN-based depth segmentation and obstacle field ranging using Bi3D

DNN Inference

NVIDIA-accelerated DNN model inference ROS 2 packages using NVIDIA Triton/TensorRT for both Jetson and x86_64 with CUDA-capable GPU

DNN Stereo Depth

Deep learned stereo disparity estimation

Freespace Segmentation

DNN-based freespace segmentation for navigation

Jetson

Utility tools for NVIDIA Jetson

Image Pipeline

NVIDIA-accelerated ROS 2 packages for camera image processing.

Image Segmentation

NVIDIA-accelerated, deep learned semantic image segmentation

Manipulator

Use Isaac Manipulator for accelerated manipulator motion planning with camera-based obstacle avoidance

Map Localization

ROS 2 package for map localization

Mission Client

Receives tasks and actions from the fleet management service and update progress, state, and errors

NITROS

NVIDIA Isaac Transport for ROS package for hardware-acceleration friendly movement of messages

NITROS Bridge

Use Isaac ROS packages in your ROS 1 Noetic with NITROS-enabled performance.

Nova

Optimized Isaac ROS packages for Isaac Nova Orin.

Nvblox

NVIDIA-accelerated 3D scene reconstruction and Nav2 local costmap provider using nvblox

Object Detection

Deep learning model support for object detection including DetectNet

Perceptor

Use Isaac Perceptor for camera-based odometry and 3D reconstruction

Pose Estimation

Deep learned, NVIDIA-accelerated 3D object pose estimation

Visual SLAM

Visual odometry package based on NVIDIA-accelerated NVIDIA cuVSLAM library with world class quality and performance.