MEC-E5012 Vehicle Mechatronics: Control – TurtleBot Maze Navigation

🧠 Project Overview

This repository contains the ROS2-based implementation for the final project in MEC-E5012: Vehicle Mechatronics - Control at Aalto University.

The task is to navigate a TurtleBot through a custom 5x5 maze track in Gazebo, avoiding obstacles and finding a valid path to the goal using autonomous control strategies.

Latest exercise update: September 15, 2025

---

Project Objectives

• Implement a Python ROS2 node (navigator.py) to control the TurtleBot.
• Apply perception, control, and path planning strategies to navigate a challenging maze environment.
• Handle dynamic path planning using the A* algorithm when blocked.
• Avoid hardcoding: the solution must generalize to other mazes.
• Achieve robust and repeatable navigation performance.

---

Environment Setup

Required Files from MyCourses

1. project_track.zip – Maze track with obstacles
2. project_track_no_obs.zip – Maze track without obstacles
3. vmc2_project.zip – ROS2 package containing simulation setup and template node

---

🧭 Navigation Framework

Implementation should be structured around the following modules:

1. Perception

• Detects whether a path between two nodes is blocked or passable using sensor data.

2. Control

• Controls the robot’s motion through the maze using PID or other strategies.
• Avoids collisions with maze walls and obstacles.

3. Path Planning

• Computes the best route to the goal using A* algorithm.
• Recomputes when a path becomes blocked.

---

Robot Navigation States

The TurtleBot uses a finite state machine with four key states:

• FORWARD – Moves forward and checks passability
• BACKTRACK – Returns to the previous node if blocked
• REPLAN – Recomputes the path using A* if the current path is invalid
• DONE – Goal reached

---

Development guide

1. Start with the obstacle-free map for debugging and tuning.
2. Avoid hardcoded paths or positions.
3. Use built-in or custom PID controllers as needed.
4. Rotate the TurtleBot at nodes before moving to the next segment.
5. For testing, use teleportation commands to jump between checkpoints (see below).

---

Testing and Teleportation

Use these commands to teleport the robot (after stopping navigator.py):

Checkpoint 1

ros2 service call /set_entity_state gazebo_msgs/srv/SetEntityState "{
  state: {
    name: 'burger',
    pose: {
      position: {x: 3.560, y: -4.272, z: 0.0},
      orientation: {x: 0.0, y: 0.0, z: 0.0, w: 1.0}
    },
    reference_frame: 'world'
  }
}"

Checkpoint 2

ros2 service call /set_entity_state gazebo_msgs/srv/SetEntityState "{
  state: {
    name: 'burger',
    pose: {
      position: {x: 4.984, y: -4.272, z: 0.0},
      orientation: {x: 0.0, y: 0.0, z: 0.0, w: 1.0}
    },
    reference_frame: 'world'
  }
}"

Checkpoint 3

ros2 service call /set_entity_state gazebo_msgs/srv/SetEntityState "{
  state: {
    name: 'burger',
    pose: {
      position: {x: 7.120, y: -5.696, z: 0.0},
      orientation: {x: 0.0, y: 0.0, z: 0.0, w: 1.0}
    },
    reference_frame: 'world'
  }
}"

---

Grading Criteria (Total: 75 points)

Criteria	Points
Complete maze without collisions	60
Finish within 2.5 minutes	10
Presentation at Gala	5
Each checkpoint reached	20 each (max 60, partial credit available)

Passing Threshold: 30 points

---

📁 Repository Structure

vmc2_project/
├── launch/
│   └── project.launch.py
├── models/
│   └── (Gazebo model files)
├── src/
│   └── navigator.py
├── world/
│   └── (Custom world files)
└── README.md

---

Notes and Best Practices

• Don't hardcode the maze layout or coordinates.
• Avoid modifying or deleting the provided structure unless necessary.
• Use the PID class only if your strategy needs it.
• Be consistent across test runs to ensure reliability.
• Start simple, debug in the no-obstacle map first.

---

📄 License

This project is part of the academic course MEC-E5012 at Aalto University. Not intended for commercial or production use.

---