vlepic/humanoid-robot-simulation repository overview

⁠Humanoid Robot Simulation

This repository branch contains a simulation environment for a humanoid robot ErgoCub, built using Gazebo Classic, ROS 2 Galactic, and YARP. The goal is to enable testing of sensor data, SLAM algorithms, and control in a reproducible Docker container setup.

⁠🧩 Repository Contents

• Docker image with pre-installed Gazebo, ROS 2, and YARP
• ROS 2 launch files
• YARP-to-ROS 2 bridge utilities
• Gazebo worlds and models
• .desktop shortcuts for launching inside Docker

⁠🚀 How to Launch the Simulation

⁠Option 1: Using Docker Compose (recommended)

Use the following command to launch:

docker compose up

The container will launch a KasmVNC server for GUI apps (e.g., Gazebo) avaible on https://container_ip:6901 and the following local directories are mounted:

⁠Option 2: Manual Docker Run

If you prefer not to use Compose, manually define volumes, parameters and pass usb devices:

docker run -it --rm \
  --name humanoid_robot-simulation \
  -p 6901:6901 \
  -p 11345:11345 \
  -e VNC_PW=password \
  -e TZ=Europe/Prague \
  -e PUID=1000 \
  -e PGID=1000 \
  -v /mnt/user/appdata/humanoid-robot-simulation/simulation_container/python:/opt/code \
  -v /mnt/user/appdata/humanoid-robot-simulation/simulation_container/yarp-ros2-ergocub-bridging:/opt/yarp-ros2-ergocub-bridging \
  -v /mnt/user/appdata/humanoid-robot-simulation/simulation_container/gazebo_models:/usr/share/gazebo/models \
  -v /mnt/user/appdata/humanoid-robot-simulation/simulation_container/gazebo_worlds:/opt/gazebo_worlds \
  -v /mnt/user/appdata/humanoid-robot-simulation/simulation_container/ergocub_modified:/opt/robotology-superbuild/build-base/install/share/ergoCub/robots/ergocub_modified \
  -v /mnt/user/appdata/humanoid-robot-simulation/simulation_container/ergocub_slam_bridge:/opt/ros_ws/src/ergocub_slam_bridge \
  --device=/dev/input/js0 \
  vlepic/humanoid-robot-simulation:gazeboclassic-ROS2

⁠🧠 Docker Image Content

This image vlepic/humanoid-robot-simulation:gazeboclassic-ROS2 includes:

• Ubuntu with ROS 2 Galactic
• Gazebo Classic
• YARP libraries
• Python YARP–ROS 2 bridge utilities
• SLAM and simulation tools

⁠🧭 Running the Simulation Workflow

After launching the container (either via Docker Compose or docker run), follow these steps:

1. Access the Web Interface:

   • Open your web browser and go to: http://container_ip:6901
   • Login credentials:
     • Username: kasm_user
     • Password: password

2. Launch Core Components Using Desktop Shortcuts:

   • Inside the remote desktop:
     • Start YARP Server using the .desktop launcher.
     • Start Gazebo Server using the appropriate .desktop launcher (which includes correct arguments).

3. Start Walking Modules:

   • Use the Walking Modules (Combined) launcher.
   • ⚠️ Note: The joystick must be passed to the container (see below), or the walking modules will not start.

4. Joystick Setup:

   • Ensure your joystick device is mapped when starting the container, e.g.:

     --device=/dev/input/js0
     

     Copy

5. Walking Control:

   • After walking modules load, press the "Prepare Walking" button on the joystick.
   • Wait for the console message: Robot prepared
   • Then press the "Start Walking" button on the joystick. You should see periodic telemetry output in the console, indicating the controller is active.

6. Start SLAM + RViz:

   • Use the .desktop launcher labeled SLAM with YARP.
   • This will launch SLAM Toolbox and RViz.
   • You can now walk around using the joystick and observe mapping and localization in RViz.

⁠📝 License

This project is licensed under the MIT License.

⁠📫 Contact

Author: Václav Lepic⁠