備註

您正在閱讀開發版本的文件。對於最新發行的版本，請參見 Kilted。

NVIDIA ROS 2 專案

NVIDIA Jetson is working towards developing ROS 2 packages to ease the development of AI applications for robotics.

ROS 專案

• Isaac ROS Nvblox : Hardware-accelerated 3D scene reconstruction and Nav2 local costmap provider using nvblox.

• Isaac ROS Object Detection : Deep learning model support for object detection including DetectNet.

• Isaac ROS DNN Inference : This repository provides two NVIDIA GPU-accelerated ROS 2 nodes that perform deep learning inference using custom models. One node uses the TensorRT SDK, while the other uses the Triton SDK.

• Isaac ROS Visual SLAM : This repository provides a ROS 2 package that estimates stereo visual inertial odometry using the Isaac Elbrus GPU-accelerated library.

• Isaac ROS Argus Camera : This repository provides monocular and stereo nodes that enable ROS developers to use cameras connected to Jetson platforms over a CSI interface.

• Isaac ROS image_pipeline : This metapackage offers similar functionality as the standard, CPU-based image_pipeline metapackage, but does so by leveraging the Jetson platform’s specialized computer vision hardware.

• Isaac ROS Common : Isaac ROS common utilities for use in conjunction with the Isaac ROS suite of packages.

• Isaac ROS AprilTags : ROS 2 node uses the NVIDIA GPU-accelerated AprilTags library to detect AprilTags in images and publish their poses, ids, and additional metadata.

• ROS and ROS 2 Docker Images : Docker images for easy deployment on the NVIDIA Jetson platform, consisting of ROS 2, PyTorch, and other important machine learning libraries.

• ROS and ROS 2 DockerFiles: Dockerfiles for ROS 2 based on l4t which all you to build your own Docker image.

• ROS 2 Packages for PyTorch and TensorRT: ROS 2 packageis for classification and object detection tasks using PyTorch and NVIDIA TensorRT. This tutorial is a good starting point AI integration with ROS 2 on NVIDIA Jetson.

• ROS / ROS 2 Packages for Accelerated Deep Learning Nodes: Deep learning image recognition, object detection, and semantic segmentation inference nodes and camera/video streaming nodes for ROS/ROS 2 using the jetson-inference library and NVIDIA Hello AI World tutorial.

• ROS 2 Package for Human Pose Estimation: A ROS 2 package for human pose estimation.

• ROS 2 Package for Hand Pose Estimation and Gesture Classification: A ROS 2 package for real-time hand pose estimation and gesture classification using TensorRT.

• GPU accelerated ROS 2 Packages for Monocular Depth Estimation: ROS 2 package for NVIDIA GPU-accelerated torch2trtxb examples such as monocular depth estimation and text detection.

• ROS 2 Package for Jetson stats: ROS 2 package for monitoring and controlling your NVIDIA Jetson [Xavier NX, Nano, AGX Xavier, TX1, TX2].

• ROS 2 Packages for DeepStream SDK: ROS 2 package for NVIDIA DeepStream SDK.

Simulation Projects

• Isaac Sim Nav2 : In this ROS 2 sample, we are demonstrating Omniverse Isaac Sim integrated with the ROS 2 Nav2 project.

• Isaac Sim Multiple Robot ROS 2 Navigation : In this ROS 2 sample, we are demonstrating Omniverse Isaac Sim integrated with the ROS 2 Nav2 stack to perform simultaneous multiple robot navigation.

References

More updates on NVIDIA Jetson ROS 2 can be found here.