🚜 Autonomous 4WD Mobile Robot in ROS 2 & Gazebo

An autonomous 4-wheel drive (4WD) skid-steer mobile robot developed in ROS 2 and Gazebo. Featuring real-time sensor fusion across LiDAR and RGB camera streams, ROS-Gz message bridging, 2D SLAM occupancy grid mapping, and a reactive Finite State Machine (FSM) obstacle avoidance navigator.

---

🌟 Key Features

• 4WD Skid-Steer URDF: Detailed mechanical robot definition with 4-wheel drivetrain, inertia matrices, collision geometries, LiDAR tower, and front-mounted RGB camera.
• Gazebo Sim Integration: Custom physical simulation environment (world.sdf) with obstacles, wall perimeters, and realistic friction/contact dynamics.
• Bi-Directional ros_gz_bridge: Connects ROS 2 with Gazebo topics for /cmd_vel, /odom, /scan, /tf, /joint_states, and /camera/image_raw.
• FSM Obstacle Avoidance Navigator (SimpleNavigator):
  • Multi-sector LiDAR point filtering (Front 80° cone, Left, Right sectors).
  • Dynamic buffer clearance (safe_dist = 0.55m).
  • Robust state machine handling WAITING_FOR_GOAL, START_DELAY, MOVE_TO_GOAL, AVOID_OBSTACLE_TURN, CLEARING_OBSTACLE, and GOAL_REACHED_DELAY.
  • Quaternion-to-Euler yaw extraction for goal heading calculation.
• 2D SLAM & Mapping: Pre-mapped environment (map_1784564885.yaml / .pgm) ready for Nav2 waypoint navigation.
• RViz Visualization: Pre-configured displays for robot model, laser scan overlays, TF transforms, and live camera feed.

---

🏛️ System Architecture

flowchart LR
    GZ[Gazebo Sim Environment] <-->|ros_gz_bridge| ROS[ROS 2 Core]
    ROS --> SENSOR[/scan LiDAR & /odom/]
    SENSOR --> NAV[SimpleNavigator Node]
    GOAL[Goal Pose /goal_pose] --> NAV
    NAV --> CMD[/cmd_vel Velocity Commands/]
    CMD -->|ros_gz_bridge| GZ
    ROS --> RVIZ[RViz2 Visualization]

---

📁 Repository Structure

wheeled_robot_ros2/
├── config/
│   └── ros_gz_bridge.yaml          # Topic mappings between Gazebo & ROS 2
├── launch/
│   ├── display.launch.py           # RViz visualization & robot_state_publisher
│   └── gazebo.launch.py            # World spawn, bridge launch & simulation
├── map/
│   ├── map_1784564885.pgm          # Occupancy grid map bitmap
│   └── map_1784564885.yaml         # Map metadata and resolution config
├── nodes/
│   └── obstacle_avoidance.py       # FSM reactive navigator & obstacle avoider
├── rviz/
│   ├── camera.rviz                 # RViz configuration with camera stream
│   └── config.rviz                 # Base RViz configuration
├── urdf/
│   └── four_wheel.urdf             # 4WD robot kinematics & sensor URDF
├── worlds/
│   └── world.sdf                   # Gazebo simulation world with obstacles
├── CMakeLists.txt                  # Build configuration
└── package.xml                     # ROS 2 package metadata

---

⚙️ Hardware & Sensor Bridge Mapping

ROS 2 Topic	Gazebo Topic	Message Type	Direction
/clock	/clock	rosgraph_msgs/msg/Clock	Gazebo ➔ ROS 2
/cmd_vel	/cmd_vel	geometry_msgs/msg/Twist	ROS 2 ➔ Gazebo
/odom	/odom	nav_msgs/msg/Odometry	Gazebo ➔ ROS 2
/tf	/model/four_wheel/tf	tf2_msgs/msg/TFMessage	Gazebo ➔ ROS 2
/scan	/scan	sensor_msgs/msg/LaserScan	Gazebo ➔ ROS 2
/joint_states	/joint_states	sensor_msgs/msg/JointState	Gazebo ➔ ROS 2
/camera/image_raw	/camera/image_raw	sensor_msgs/msg/Image	Gazebo ➔ ROS 2

---

🚀 Quick Start Guide

1. Requirements

• ROS 2 Humble / Iron
• Gazebo Sim (Fortress / Harmonic)
• ros-humble-ros-gz or equivalent bridge package

2. Build the Package

mkdir -p ~/ros2_ws/src
cd ~/ros2_ws/src
git clone https://github.com/Arkz-Deepak/wheeled_robot_ros2.git
cd ~/ros2_ws
colcon build --symlink-install --packages-select wheeled_robot_ros2
source install/setup.bash

3. Launch the Simulation

Launch Gazebo world and bridge all sensor streams:

ros2 launch wheeled_robot_ros2 gazebo.launch.py

4. Launch RViz Visualization

ros2 launch wheeled_robot_ros2 display.launch.py

5. Run the Autonomous FSM Navigator

ros2 run wheeled_robot_ros2 obstacle_avoidance.py

Send a goal pose via RViz 2D Goal Pose or publish directly:

ros2 topic pub --once /goal_pose geometry_msgs/msg/PoseStamped "{header: {frame_id: 'map'}, pose: {position: {x: 2.0, y: 1.5, z: 0.0}}}"

---

👨‍💻 Author

• Deepak R (@Arkz-Deepak)
• Portfolio & Robotics Lab: lab.deepak-arkz.me