Skip to content

Code Walkthrough – Autonomous Warehouse Patrolling Robot

This walkthrough describes the structure and functionality of our ROS2 codebase, which powers the robot's simulation, navigation, and goal-based patrolling inside a custom warehouse environment.

Packages Overview

Package Purpose
custom_world_pkg Contains the Gazebo world, models, and launch files for the simulated warehouse environment
path_planner_pkg Implements ROS2 navigation, custom goal dispatching, and maps for SLAM & AMCL

custom_world_pkg

▶ Purpose

This package defines a custom Gazebo warehouse world complete with shelves and patrol lanes. It allows realistic simulation of patrolling behavior in structured environments.

Key Structure

custom_world_pkg/
├── launch/
│   └── launch_simulation.launch.py      # Launches Gazebo with custom world
├── models/                              # Custom shelves, obstacles, etc.
├── worlds/
│   └── warehouse.world                  # Main world file used in simulation
├── package.xml / CMakeLists.txt         # Standard ROS2 package setup

Launch the Simulation

ros2 launch custom_world_pkg launch_simulation.launch.py

This opens Gazebo with a preloaded warehouse layout and the TurtleBot4 robot.

path_planner_pkg

▶ Purpose

This is the control and navigation package. It provides: - Map loading and localization - Navigation using Nav2 stack - Code to send custom navigation goals

Key Structure

path_planner_pkg/
├── launch/
│   └── nav2.launch.py                   # Launches map server, AMCL, Nav2 stack
├── maps/
│   ├── map.yaml                         # Metadata for static map
│   └── map.pgm                          # Grayscale occupancy map
├── path_planner/
│   └── custom_goal_sender.py           # Python node to send goals programmatically
├── setup.py / setup.cfg                # Python package setup
├── resource/ / test/                   # ROS2-specific folders

Core File: custom_goal_sender.py

This node: - Initializes a ROS2 node using rclpy - Publishes a PoseStamped goal to the Nav2 stack - Logs success/failure from the NavigateToPose action client

Example Usage:

ros2 run path_planner custom_goal_sender

Inside the script, you can define a goal pose like:

goal_pose.pose.position.x = 2.0
goal_pose.pose.position.y = 3.5

You can modify this to send patrol waypoints dynamically, or integrate it with a behavior tree.

How the Packages Work Together

[custom_world_pkg] → Launches Gazebo simulation with TurtleBot4
        |
        v
[path_planner_pkg] → Loads map, enables SLAM/AMCL, sends goals to Nav2
        |
        v
[Nav2 Stack] → Handles path planning and obstacle avoidance

Testing & Debugging

  • Use RViz2 with /map, /tf, and /goal_pose to visualize movement
  • Run ros2 topic echo /amcl_pose to verify localization
  • Adjust robot starting pose and goals in the launch file for better realism

🗂️ Repository Access

You can find the complete codebase here:
📦 warehouse_simulation-main