Navigation

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

Prerequisites

• Complete the hardware and setup steps in the Setup Guide.
• Ensure the robot is on level ground with sufficient clearance for safe autonomous driving.
• Connect to the robot PC (keyboard/monitor or SSH) and make sure the battery is charged.

Bring Up the Robot

Before launching Nav2, start the SG2 follower using the commands from the Teleoperation Guide. Run these inside the Docker container:

cd ~/ai_worker
./docker/container.sh enter
ros2 launch ffw_bringup ffw_sg2_follower_ai.launch.py

Keep these terminals running. use a new terminal for the commands below.

View RViz via noVNC (headless/SSH)

RViz cannot be displayed over plain SSH. Use noVNC to launch RViz and view it from a browser.

TIP

• ./install.sh is only needed once. If the noVNC service is already set up, skip step 3.
• If 192.168.6.2:8090 is unreachable, confirm the host IP and the forwarded port in your environment.

1. Enter the AI Worker Docker container:

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

2. Move to the noVNC workspace:

   cd /workspace/docker-novnc

3. noVNC setup (First-Time):

   ./install.sh

4. Start noVNC:

   ./entrypoint.sh

5. In a browser, open http://192.168.6.2:8090/vnc.html and click Connect. A terminal window will appear.

6. Run RViz in that terminal:

   rviz2

7. In RViz, open the config via File → Open Config and choose /ros2_ws/src/ai_worker/ffw_navigation/*.rviz (use the mapping.rviz or navigation.rviz configuration as needed).

Lidar Scan Visualization

First, use RViz (via noVNC) to confirm lidar scans are visible around the robot.

Lidar scan values visualized in RViz. Use this view to confirm the sensor is publishing after connecting via noVNC.

WARNING

In RViz, set the scan topic Reliability Policy to Best Effort for this display.

Launch the Navigation Stack

All commands below run on the robot PC in a new terminal after the bringup above is running.

1. Enter the Docker container in the new terminal:

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

2. Start Nav2:

   WARNING

   Make sure the bringup terminals stay running before launching Nav2, otherwise the navigation stack will fail to connect to sensors and actuators.

   The package includes only a default Gazebo map. Create your own map with SLAM Toolbox for your environment.

   Build a map with SLAM Toolbox:

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

   Nav2 starts with SLAM enabled so you can map while driving. Pick one of the two exploration methods below:

   • Autonomous: In RViz, send a goal with 2D Goal Pose. the robot will plan, drive, and expand the map as it moves.
   • Keyboard teleop: In a new terminal, run:

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

     Drive with the keyboard to explore and build the map.

   SLAM Mapping Process

   White areas show free space discovered by SLAM, black indicates obstacles or walls, and gray is unknown space. As the robot drives, the explored (white) region expands and the unknown area shrinks while the map is refined.

   Save a generated map

   If you ran SLAM and want to reuse the map, save it with:

   cd ~/ros2_ws/src/ai_worker/ffw_navigation/maps
   ros2 run nav2_map_server map_saver_cli -f ./map

   This produces map.yaml and map.pgm in the maps directory. Change the path with -f /path/to/map_name if needed.

Navigation Using a Saved Map

• Use a generated map:

  ros2 launch ffw_navigation navigation.launch.py

  TIP

  To use a different map, place your map.yaml and image file in the maps/ folder of the ffw_navigation package.

  Set the initial pose in RViz with 2D Pose Estimate so AMCL can localize. After Nav2 is running, send a goal in RViz using 2D Goal Pose. The robot will plan and drive to the target pose. if SLAM is enabled, the map updates in real time.

RViz Goal on Saved Map

Setting 2D Goal Pose in RViz on a saved map. the plan is sent to the real robot for execution.

Real-World Execution

AI Worker following the planned curved path on the saved map and driving to the goal.

Troubleshooting

• If the /scan display is missing in RViz, set the scan topic Reliability Policy to Best Effort.
• If the map is not visible in RViz, set the map topic Durability Policy to Transient Local.

Tuning Guide

• AMCL localization: Adjust parameters in config/amcl_localization.yaml.
• Nav2 stack (costmaps, BT navigator, controller, planner): Adjust parameters in 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.