A simulation-first autonomous mobile robot built using ROS 2 (Jazzy) and Gazebo (gz-sim), focusing on core autonomy concepts such as perception, decision-making, and control through a clean ROS2 architecture and a transparent Finite State Machine (FSM).

This project was developed to:

• Gain hands-on experience with end-to-end autonomous robot architecture
• Practice ROS 2 node-based modular design
• Implement FSM-driven obstacle avoidance
• Bridge the gap between simulation and real-world robotics

The rover navigates autonomously in a simulated environment and reacts intelligently to obstacles using LiDAR-based perception.

• 🧭 Autonomous Forward Navigation
• 🚧 LiDAR-Based Obstacle Detection
• 🔄 FSM-Controlled Obstacle Avoidance
  • FORWARD → STOP → REVERSE → SCAN → TURN → FORWARD
• 🎥 Camera Integration (perception & visualization)
• 🕹️ Manual Teleoperation Support
• 🔌 ROS 2 Topic-Based Velocity Control (cmd_vel)
• 🧪 Fully Simulated in Gazebo (gz-sim)

The rover follows a modular and scalable ROS 2 architecture:

• Gazebo (gz-sim)
• Robot model using URDF / Xacro
• Sensors:
  • LiDAR
  • Camera

• Sensor data processing
• FSM-based control logic
• Velocity command publisher

• RViz2
• rqt_graph

This clear separation of concerns improves maintainability and simplifies future deployment on real hardware.

Obstacle avoidance is driven by a transparent and deterministic FSM:

1. FORWARD – Move straight
2. STOP – Pause when an obstacle is detected
3. REVERSE – Create safe distance
4. SCAN – Compare left vs right clearance using LiDAR
5. TURN_LEFT / TURN_RIGHT – Rotate toward safer direction
6. FORWARD – Resume navigation

This design avoids black-box planners and emphasizes interpretable decision-making.

• ROS 2 Jazzy
• Gazebo (gz-sim)
• Python (rclpy)
• URDF / Xacro
• RViz2
• rqt_graph

# Build the workspace
colcon build
source install/setup.bash

# Launch the simulation
ros2 launch rover_bringup rover.launch.xml

# Run the obstacle avoidance node
ros2 run rover_control obstacle_avoidance.py

This project is licensed under the MIT License. See LICENSE.md for details.