Docker Setup with Zenoh Bridge (Cross-Distro)

This tutorial explains how to deploy EasyNav using Docker while maintaining interoperability with a robot running a different ROS 2 distribution.

The architecture described here is ideal for scenarios where the robot hardware requires an older LTS version (like Humble) but you want to run the latest EasyNav algorithms on a newer distribution (like Rolling or Jazzy). The Zenoh Bridge acts as the middleware glue that enables seamless communication between these containers and the physical hardware.

Overview

We will set up a system composed of three main parts:

  1. The Robot Host: Runs the hardware drivers on its native ROS 2 distribution (e.g., Humble).

  2. The Docker Container: Runs EasyNav on ROS 2 Rolling/Jazzy using the provided image.

  3. The Remote PC: Used for visualization (RViz2) and control.

Note

Ensure that all machines (Robot, Docker Container, and PC) use the same Zenoh version to avoid protocol mismatches. The provided Dockerfile uses version 1.7.2.

β€”

Step 1: Building the Docker Image (Rolling)

First you must create the Docker image containing the EasyNav stack and the Zenoh bridge.

Create a file named Dockerfile with the following content. This configuration installs the necessary dependencies and compiles the workspace.

# Use the official ROS 2 Rolling image as base
FROM ros:rolling-ros-base-noble

# Prevent interactive prompts during installation
ENV DEBIAN_FRONTEND=noninteractive
# Define the workspace path
ENV WORKSPACE=/root/easynav_ws

# Install essential build tools, git, and networking utilities
RUN apt-get update && apt-get upgrade -y && apt-get install -y \
    git \
    python3-colcon-common-extensions \
    python3-rosdep \
    python3-vcstool \
    wget \
    nano \
    unzip \
    && rm -rf /var/lib/apt/lists/*

# Install Zenoh Bridge for ROS 2 DDS
# Using version 1.7.2 as requested by the user
RUN wget https://github.com/eclipse-zenoh/zenoh-plugin-ros2dds/releases/download/1.7.2/zenoh-plugin-ros2dds-1.7.2-x86_64-unknown-linux-gnu-standalone.zip && \
    unzip zenoh-plugin-ros2dds-1.7.2-x86_64-unknown-linux-gnu-standalone.zip && \
    chmod +x zenoh-bridge-ros2dds && \
    mv zenoh-bridge-ros2dds /usr/local/bin/ && \
    rm zenoh-plugin-ros2dds-1.7.2-x86_64-unknown-linux-gnu-standalone.zip

# Set up the workspace structure
RUN mkdir -p $WORKSPACE/src
WORKDIR $WORKSPACE

# Clone the Easynav project and research repositories
RUN cd src && \
    git clone -b rolling https://github.com/EasyNavigation/EasyNavigation.git  && \
    git clone -b rolling https://github.com/EasyNavigation/NavMap.git  && \
    git clone -b rolling https://github.com/EasyNavigation/easynav_plugins.git  && \
    git clone -b rolling https://github.com/fmrico/yaets.git

# Initialize and update rosdep for dependency management
RUN rosdep update && \
    apt-get update && \
    rosdep install --from-paths src --ignore-src -r -y && \
    rm -rf /var/lib/apt/lists/*

# Compile the workspace using colcon
RUN . /opt/ros/rolling/setup.sh && \
    colcon build --symlink-install

# Configure the environment to source setups automatically
RUN echo "source /opt/ros/rolling/setup.bash" >> ~/.bashrc && \
    echo "source $WORKSPACE/install/setup.bash" >> ~/.bashrc

# Create an entrypoint script to launch Zenoh in the background
RUN echo '#!/bin/bash\n\
set -e\n\
# Start Zenoh bridge in the background and redirect output to a log file\n\
zenoh-bridge-ros2dds > /var/log/zenoh.log 2>&1 &\n\
echo "Zenoh bridge started in the background."\n\
# Execute the original command (usually bash)\n\
exec "$@"' > /usr/local/bin/entrypoint.sh && \
chmod +x /usr/local/bin/entrypoint.sh

# Use the custom script as the entrypoint
ENTRYPOINT ["/usr/local/bin/entrypoint.sh"]
CMD ["bash"]

Build the image with the tag easynav_rolling:

docker build -t easynav_rolling .

β€”

Step 2: Preparing the Robot (Host)

The robot needs to run the zenoh-bridge-ros2dds to expose its topics to the network efficiently. Since the robot might be running an older distro like Humble, the bridge ensures compatibility with the Rolling container.

  1. Download the Bridge: Ensure you download the same version used in the Dockerfile (v1.7.2).

wget https://github.com/eclipse-zenoh/zenoh-plugin-ros2dds/releases/download/1.7.2/zenoh-plugin-ros2dds-1.7.2-x86_64-unknown-linux-gnu-standalone.zip
unzip zenoh-plugin-ros2dds-1.7.2-x86_64-unknown-linux-gnu-standalone.zip
chmod +x zenoh-bridge-ros2dds
sudo mv zenoh-bridge-ros2dds /usr/local/bin/
rm zenoh-plugin-ros2dds-1.7.2-x86_64-unknown-linux-gnu-standalone.zip

  1. Run the bridge on the robot:

If you use a specific ROS_DOMAIN_ID, export it before running the bridge.

# Example for a robot using Domain ID 23
export ROS_DOMAIN_ID=23
./zenoh-bridge-ros2dds

This command will start the bridge in peer discovery mode.

β€”

Step 3: Running the Container

Now we launch the EasyNav container on the robot (or on a computer connected to the robot). We use --net=host and --ipc=host.

# Run the container with host networking and shared memory
docker run -it --net=host --ipc=host --name easynav_test -e ROS_DOMAIN_ID=23 easynav_rolling

Inside the container you can now start the navigation system. The entrypoint script automatically starts the Zenoh bridge in the background.

# Inside the container
ros2 launch easynav_system system_main ...

Tip

If you exited the container and want to restart it later use the following commands:

# Start the container
docker start easynav_test

# Open a new terminal in the container
docker exec -it easynav_test bash

β€”

Step 4: Visualization from a Remote PC

To control the robot and visualize data from your laptop you also need the Zenoh bridge.

  1. Install the Bridge on your PC (Check Step 2).

  2. Connect to the Robot: Run the bridge in client mode connecting to the robot IP address.

# Replace <ROBOT_IP> with the actual IP address of the robot
# Port 7447 is the default Zenoh peering port
zenoh-bridge-ros2dds -e tcp/<ROBOT_IP>:7447

  1. Open RViz: Open a new terminal and set the Domain ID to match the one used in the bridge.

export ROS_DOMAIN_ID=23
ros2 run rviz2 rviz2

You should now be able to see topics from both the Humble robot (Hardware) and the Rolling Docker container (EasyNav) seamlessly integrated in RViz.