 Please ask about problems and questions regarding this tutorial on answers.ros.org. Don't forget to include in your question the link to this page, the versions of your OS & ROS, and also add appropriate tags. |
Installation of a ROS4HRI environment
Description: This tutorial will take you through the steps to install a working 'ROS4HRI' sandbox environment to explore the frameworkKeywords: hri
Tutorial Level: BEGINNER
Next Tutorial: Getting Started With ROS4HRI
Base system
Many of the ROS4HRI packages require Python3. As such, ROS noetic is strongly encouraged.
Please follow these instructions to install a Ubuntu 20.04/ROS noetic environment.
Choose the desktop-full version of ROS, to have access to visual tools like rviz.
Base HRI packages
Install the following packages to get the basic ROS4HRI packages:
sudo apt install ros-noetic-hri-msgs: the ROS4HRI ROS messages
sudo apt install ros-noetic-hri: libhri, a C++ library to ease integration of ROS4HRI in your C++ nodes
sudo apt install ros-noetic-pyhri: pyhri, a Python library to ease integration of ROS4HRI in your Python nodes
sudo apt install ros-noetic-human-description: a parametric kinematic model of a human, in URDF format
sudo apt install ros-noetic-hri-rviz: rviz plugins to display detected humans (faces and bodies) and the humans' 3D skeletons (if you run a skeleton tracker)
Additional ROS4HRI nodes
some of these nodes are not packages in noetic and require manual installation.
To get started, you might want to install:
hri_face_detect: a Google MediaPipe-based multi-people face detector.
hri_fullbody: a Google MediaPipe-based 3D full-body pose estimator
Useful nodes to play with ROS4HRI
To explore ROS4HRI without eg a robot, you might want to use your webcam instead of the robot's cameras. ros-noetic-usb-cam is a popular choice.
