Unitree Go2 Gazebo Simulation (go2_gazebo_sim)

This repository contains the ROS 2 Gazebo simulation packages for the Unitree Go2 quadruped robot. It provides a complete simulation environment with Gazebo, including mapping and navigation capabilities with support for tight-space navigation testing.

Features

• Gazebo Simulation: Realistic physics simulation of the Unitree Go2.
• Navigation Stack (Nav2): Fully configured navigation stack for autonomous movement.
• SLAM: Mapping capabilities using slam_toolbox.
• LiDAR Support: Simulation of Velodyne VLP-16 and Unitree 4D LiDAR.
• Mini Maze World: Test environment with narrow corridors (0.85-0.9m), ramps, and obstacles for tight-space navigation testing.

Installation

1. Navigate to your unitree-sdk directory (assuming it is in your workspace src):

   cd ~/unitree-sdk

2. Clone this repository as a submodule:

   git clone https://github.com/OpenMind/go2_gazebo_sim.git

3. Install dependencies:

   rosdep install --from-paths . --ignore-src -r -y

4. Build the workspace:

   colcon build --symlink-install

5. Source the workspace:

   source install/setup.bash

6. Setup CycloneDDS The network interface is your WiFi adapter. Replace wlxbc071d7223b6 with your own network interface name found via ifconfig or ip a.

   <CycloneDDS>
       <Domain>
           <General>
           <Interfaces>
               <NetworkInterface name="wlxbc071d7223b6" priority="default" multicast="default" />
           </Interfaces>
           </General>
       </Domain>
   </CycloneDDS>

Usage

1. Launch SLAM (Mapping)

To launch the simulation and start mapping the environment:

ros2 launch go2_gazebo_sim go2_slam_launch.py

Default world: maze_world.sdf (mini maze with narrow corridors and slopes)

To use a different world:

ros2 launch go2_gazebo_sim go2_slam_launch.py world:=default.sdf

Control the Robot with Teleop

In a new terminal, run the teleop keyboard node to drive the robot around:

ros2 run teleop_twist_keyboard teleop_twist_keyboard

Use the keyboard controls displayed in the terminal to move the robot:

• i - Move forward
• , - Move backward
• j - Turn left
• l - Turn right
• k - Stop
• u/o/m/. - Move diagonally

Drive the robot around to cover the entire area and build a complete map.

2. Save the Map

Once you are satisfied with the map, save it using the map saver CLI:

cd ~
ros2 run nav2_map_server map_saver_cli -f my_map

This will create my_map.yaml and my_map.pgm in your home directory.

3. Launch Navigation

To launch the simulation with the navigation stack using your saved map:

ros2 launch go2_gazebo_sim go2_nav_launch.py map:=${HOME}/my_map.yaml

To use a different world:

ros2 launch go2_gazebo_sim go2_nav_launch.py world:=default.sdf map:=${HOME}/my_map.yaml

You can now set 2D Pose Estimates and Navigation Goals in RViz.

Packages

• go2_gazebo_sim: Main Gazebo simulation launch files and configurations.
• go2_description: URDF, mesh files, and world definitions for the robot.
• champ: Quadruped controller framework.

World Files

• maze_world.sdf (default): Mini maze with narrow passages (0.85-0.9m wide), upward/downward ramps (15° slopes), and step platform for comprehensive navigation testing.
• default.sdf: Simple environment with ground plane for basic testing.

Acknowledgements

Special thanks to khaledgabr77/unitree_go2_ros2 and its contributors. This repository is built upon their excellent work.