Skip to main content

Navigation

For ROS 2 and Zenoh communication, see Software Architecture: Middleware.

This guide shows how to bring up the AI Worker's navigation stack so the robot can move autonomously after basic setup.

1. Bringup the Robot​

Complete the Bringup Guide to start the Zenoh daemon and bring up the robot before proceeding.

warning

Place the robot on level ground with sufficient clearance for autonomous driving. For battery-powered models, check that the battery is charged.

2. Set Up the User PC for RViz2​

On the User PC, prepare Git, Docker Engine, and the Docker Compose plugin. Use a graphical desktop session so RViz2 can display its window. See the Docker installation guide if Docker is not installed.

Clone the AI Worker repository on the User PC. If you already have a checkout, use the existing directory instead:

cd ~
git clone -b jazzy https://github.com/ROBOTIS-GIT/ai_worker.git

Start and enter the AI Worker Docker container:

cd ~/ai_worker/docker
./container.sh start
./container.sh enter

In the container terminal, follow External ROS 2 Communication to configure Zenoh and connect the User PC to the robot. Apply the same settings in each User PC terminal used for ROS 2.

3. Create and Save a Map​

1. Start Mapping​

In a new terminal, connect to the Robot PC via SSH, then enter the AI Worker container:

cd ~/ai_worker/docker
./container.sh enter

Start Navigation with SLAM enabled:

ros2 launch ffw_navigation navigation.launch.py use_slam:=true

Keep robot bringup, the Zenoh daemon, and this command running while mapping.

2. Open RViz2​

On the User PC, use the AI Worker container terminal with the Zenoh configuration applied in step 2 to open RViz2 with the mapping configuration:

rviz2 -d ~/ros2_ws/src/ai_worker/ffw_navigation/rviz/mapping.rviz

3. Explore the Environment​

Drive the robot through the area while monitoring the map in RViz2. Choose one of the following methods:

  • RViz2 goal: Use 2D Goal Pose to send a goal in the explored area.

  • Keyboard teleoperation: In another Robot PC terminal, enter the AI Worker container and run:

    cd ~/ros2_ws/src/ai_worker/ffw_teleop/ffw_teleop
    python3 mobile_teleop.py

    Use W / S to move forward or backward, A / D to turn, and Space to stop. Press Ctrl+C to exit.

SLAM mapping (Nav2 + SLAM Toolbox)

In the map, white represents free space, black represents obstacles, and gray represents unexplored space.

4. Save the Map​

On the Robot PC, open another terminal, enter the AI Worker container, and save the map to the host-mounted workspace:

mkdir -p /workspace/maps
ros2 run nav2_map_server map_saver_cli -f /workspace/maps/my_map

This creates my_map.yaml and my_map.pgm in /workspace/maps. Choose another name if you want to keep an existing map with that name.

After saving the map, stop the SLAM-enabled Navigation launch with Ctrl+C in the terminal where it is running before starting Navigation with the saved map.

4. Navigate with a Saved Map​

1. Start Navigation​

On the Robot PC, enter the AI Worker container and start Navigation with the saved map:

cd ~/ai_worker/docker
./container.sh enter
ros2 launch ffw_navigation navigation.launch.py map:=/workspace/maps/my_map.yaml

Replace /workspace/maps/my_map.yaml with the path to your map YAML file if you saved it elsewhere.

2. Open RViz2​

On the User PC, open a terminal and enter the AI Worker container:

cd ~/ai_worker/docker
./container.sh enter

Apply the same Zenoh settings from step 2 in this terminal, then open RViz2 with the navigation configuration:

rviz2 -d ~/ros2_ws/src/ai_worker/ffw_navigation/rviz/navigation.rviz

3. Set the Initial Pose​

In RViz2, select 2D Pose Estimate and set the robot's position and orientation on the map so AMCL can localize it.

Illustrative animation of selecting the robot's position and dragging to set its initial orientation in RViz2 (composited)

4. Send a Goal​

In RViz2, select 2D Goal Pose and set the target position and orientation. The robot plans and follows a path to the goal.

RViz goal and path with saved map

Real robot following planned path

Troubleshooting​

For common issues, see Navigation Troubleshooting.

Tuning Guide​

  • AMCL localization: Adjust parameters in the amcl section of ffw_navigation/config/navigation.yaml.
  • Nav2 stack (costmaps, BT navigator, controller, planner): Adjust parameters in ffw_navigation/config/navigation.yaml.
  • For detailed parameter descriptions across Nav2, see https://docs.nav2.org/index.html.

Safety and Usage Tips​

  • Keep people and obstacles clear while the robot is localizing or moving.
  • If you need to pause, use RViz to clear goals or stop the navigation nodes with Ctrl+C in the running terminals.
  • For simulation-based testing (recommended before field deployment), see the Nav2 section in Gazebo.