kodo-nav2-patrol

A modular ROS 2 Humble workspace for goal-based robot navigation using TurtleBot3, Gazebo, and the Nav2 stack — all running inside a reproducible Docker environment.

Package Structure

src/
├── patrol_bringup/       # Launch files & RViz config — single entry point
├── patrol_simulation/    # Gazebo world and spawn configuration
└── patrol_navigation/    # Nav2 params, map files, localization config

Prerequisites

• Docker (no ROS installation required on the host)

Setup

1. Allow X11 forwarding from Docker

xhost +local:docker

2. Build the Docker image

docker compose build

Usage

Start the container

docker compose up kodo

setup_env.bash is sourced automatically on login.

Build stack

colcon build

After first build, you have to source the environment again.

Launch the full system (single command)

ros2 launch patrol_bringup sim_nav.launch.py

Starts Gazebo, spawns the robot, brings up Nav2, and opens RViz. Gazebo runs headless by default

Sending a navigation goal

Via RViz (preferred):

1. Click the "2D Goal Pose" button in the RViz toolbar
2. Click and drag on the map to set the goal position and orientation
3. The robot will plan a path and navigate to the goal

Via CLI:

ros2 action send_goal /navigate_to_pose nav2_msgs/action/NavigateToPose \
  "{pose: {header: {frame_id: 'map'}, pose: {position: {x: 1.0, y: 0.5, z: 0.0}, orientation: {w: 1.0}}}}"

Assumptions & Limitations

• Map: Pre-built map is included in src/patrol_navigation/maps/.
• Localization: Uses AMCL. If the estimated pose drifts, use "2D Pose Estimate" in RViz.
• Display: Requires an X11 server on the host. On headless servers, use a virtual display (e.g., Xvfb).
• Robot model: Defaults to TurtleBot3 Waffle. Change TURTLEBOT3_MODEL in setup_env.bash to use burger.