Any machine with Docker installed. Learn more about Docker and container here.

Open a Terminal/Powershell

For x64 machine, run make init will download and initalize the container with ROS Foxy/Noetic.

If you don't have make, just copy and run line 6 in the Makefile.

For arm64 machine, run make init-arm.

A CLI to the container will be granted after this step.

For GUI access, use your broswer to visit this page with [password].

Other operations on the container can be found in the Makefile.

1. Download a bag file from here and move it to the shared folder rootfs/. Learn more about ROS bag here.

2. In the container CLI

   • run source /root/rootfs/foxy_setup.sh to initiate ROS 2 environment
   • run rviz2 & to start RViz
   • run ros2 bag play -r 10 -s rosbag_v2 --loop rootfs/16-mcity1.bag to play the downloaded bag file

3. In the container GUI

   • open config file demo-ROS2.rviz in folder /root/rootfs/ from RViz, it should show the images and point clouds now

1. Image ROS Node (view code | demo video)

   In the container GUI, open mutiple terminals

   • run source /root/rootfs/foxy_setup.sh; ros2 bag play -r 10 -s rosbag_v2 --loop rootfs/16-mcity1.bag in terminal 1
   • run source /root/rootfs/foxy_setup.sh; (ros2 run rqt_image_view rqt_image_view &); python3 /root/rootfs/image_subscriber-ros2.py in terminal 2
   • select image topic /front_camera/image_raw and /img_gray to see the original image and the gray image produced by the ROS node.

2. PointCloud ROS Node (view code | demo video)

   In the container GUI,

   • run apt install -y python3-pcl ros-foxy-sensor-msgs-py; python3 /root/rootfs/point_cloud_subscriber-ros2.py in terminal 2
   • pointcloud file will be exported to shared folder rootfs/ by the ROS node

   In the host machine, use CouldCompare or Matlab to view the exported pointcloud file

3. GPS ROS Node (view code)

   In the container GUI,

   • run pip install utm; python3 /root/rootfs/gps_subscriber-ros2.py in terminal 2
   • the ROS node will print the gps coordinates

1. Create a workspace with mkdir -p /root/rootfs/demo_workspace/src and initialize it following the tutorials [ROS 1, ROS 2]

2. Follow the rest of tutorials to create your ROS package and ROS node [ROS 1, ROS 2]

3. Add the dependencies of your package to Dockerfile.template and build your image with docker build -f Dockerfile.template --tag=IMAGE_NAME

1. No topic showes up in rqt_image_view
   • Try to run ros2 topic list before running rqt_image_view
2. Docker does not work on my computer
   • You can use a VM/WSL as an alternative, follow the setup in the Dockerfile (install fresh Ubuntu 20.04 and all the dependencies)
3. make init or line 6 in the Makefile does not work
   • Replace `pwd` with the absolute path of this tutorial folder
4. ros2: command not found or No module named 'rclpy'
   • run source /root/rootfs/foxy_setup.sh in your terminal and try again
5. Get Aborted (core dumped) when running rviz2
   • Restart your docker (exit container then make start) and try again, DO NOT source noetic_setup.sh after sourcing foxy_setup.sh
6. How to skip the source /root/rootfs/foxy_setup.sh step
   • Add source /root/rootfs/foxy_setup.sh to the end of /root/.bashrc, restart the terminal
7. Copy and paste to/from noVNC is annoying
   • Add -localhost no the end of the second line in /root/startVNC.sh, restart docker and use ifconfig to show DOCKER_IP, then connect to DOCKER_IP:5901 with an VNC client (CAUTION! This will expose your docker VNC to public IP instead of localhost)