Self-driven AI car

Self-Driven AI Car is a project that simulates an autonomous car using a custom environment built with Gymnasium. The car is trained using the PPO algorithm from the stable_baselines3 library, providing an interactive and customizable experience for AI enthusiasts.

Project written by creating custom Gymnasium environment.

Instructions

1. Clone this project.
2. Install the environment using build.bat (it fixes some path issues that I didn't have time to resolve manually).
3. Install dependencies:

   pip install pygame gymnasium stable_baselines3

4. Edit the run.py file according to your needs.
5. Navigate to the Patro_Car directory and run:

   python run.py

Customizable Environment Attributes:

• game_fps (default=300): Controls the game clock speed if render_mode="human" is set.
• rotation_amount (default=6): Defines how much the car rotates.
• max_speed (default=4): Sets the maximum speed of the car.
• acceleration_amount (default=0.1): Determines the car's acceleration rate.

Additional Resources

AI_Car.mp4

Great TUTORIAL about this topic:

Car trained by PPO algorithm from stable_baselines3.