Turtlesim Reinforcement Learning Environment

This project is a custom OpenAI Gymnasium environment built using ROS 2 and the Turtlesim simulator. The agent (turtle1) learns to catch randomly spawned turtles in the simulated environment.

Features

• Gymnasium-compatible environment (TurtleEnv)
• ROS 2-based communication with Turtlesim
• Reward structure:
  • +10 for catching a turtle
  • -5 for hitting the boundary
  • -0.01 time penalty per step
• Turtle is respawned after each catch
• Episode ends upon catching or failure

Requirements

• ROS 2 (e.g. Jazzy, Humble, etc.)
• Python 3.10+
• Turtlesim
• Gymnasium

Install Python dependencies:

pip install -r requirements.txt

⚠️ Ensure that ROS 2 and Turtlesim are installed and sourced before running.

How to Run

Make sure ROS 2 is sourced:

source /opt/ros/jazzy/setup.bash

#Launch Turtlesim:

ros2 run turtlesim turtlesim_node