COVAPSy Simulation Environment - Sorbonne Université 2025

This repository contains the simulation component of the COVAPSy course offered at Sorbonne Université in 2025.

Overview

This simulation environment is designed to assist students in building, testing, and visualizing robotic control systems, with a focus on autonomous vehicle projects. It includes:

• URDF files for RViz visualization.
• Webots simulator integration with ROS 2 using dedicated controller plugins.
• Multiple STL models of various car bodies for visual customization.
• Custom ROS 2 nodes to control vehicles and simulate realistic autonomous behavior.

Installation

To get started, first install Webots by following the official guide:
Webots Installation on Linux

File Structure

• launch/: Launch files for simulation.
• protos/: Custom Webots PROTO files, including different STL car models.
• resource/: Contains the URDF and configuration files.
• webot_simulation/: Main ROS 2 package, including drivers and controllers.
• worlds/: Contains Webots world files.
• test/: Contains optional test files or tools.

Setting up the interface between Webots and ROS2

Please refer to this tutorial Webot_ROS2_interface for more details.

Adding a New World

To add a new Webots world:

1. Save the world file in the worlds/ directory.

2. Register it in the data_files list in setup.py:

   ('share/' + package_name + '/worlds', ['worlds/your_new_world.wbt'])

3. Set it as the default in the launch file:

   DeclareLaunchArgument(
        'world',
        default_value='your_new_world.wbt',
        description='Choose one of the world files from the `/webot_simulation/world` directory'
   ),

Adding a New Controller

If you create a new plugin controller that links Webots to ROS 2, update the plugin tag in the URDF file accordingly:

1. Set the plugin path in the URDF like this:

<plugin type="your_package.your_plugin_class"/>

2. Rebuild the project:

colcon build && source install/setup.bash

Sensors and Devices

This simulation includes the following devices:

• RpLidar A2 (publishes to /TT02_jaune/RpLidarA2/point_cloud)
• IMU / Gyro
• Rear sonar sensor

ROS 2 Nodes

1. teleop.py
   Publishes AckermannDrive messages to the /cmd_ackermann topic. Ackermann messages are used because they are native to car models in Webots. This differs from the low_level_ros2 approach, which uses Twist messages.

2. obstacle_avoider.py
   Subscribes to the Lidar point cloud and performs basic obstacle avoidance. When an object is detected in front of the vehicle, it stops, reverses briefly, and turns.

Maps Available

• piste_enscopy.wbt: A replica of the ENS 2023 track.
• circular_piste.wbt: A circular loop for testing.

To create your own maps, refer to the Webots Map Creation Tutorial.

Future Directions

This simulation provides a solid foundation for developing more advanced behaviors. Future goals include:

• Reinforcement Learning-based navigation
• Autonomous driving research

Recommended Learning Resources

• TurtleBot3 DRL Navigation
• F1Tenth Reinforcement Learning

Final Notes

Simulation differs from real-world scenarios. Be patient and exercise caution when transitioning to real hardware.

Enjoy exploring and driving!

For inspiration, refer to:

• webots_ros2_tesla
• webots_ros2_turtlebot