Neural Network Controlled Car in Pygame

---

Setup

1. Run main.py
2. Choose a controlling option
3. If AI controlled, choose a model
4. Confirm to begin session

In session

• If manual, use WASD or Arrow Keys to move
• Use R to reload session
• Use Q to quit session, the session will be deleted
• Use Z to save a model, this will overwrite a model saved in this session
• Use X to delete a model saved in this session

Features

• Modualar: The Car, Road, and the Network can be tweeked to a custom setup
• Good visualizations: Allows the user to see exactly how the network is functioning and iterating
• Maintainable: Sections of code are isolated

Improvements

• Self training: The script will be able to auto reload after a certain threashold save the most optimal model after serveral iterations
• Variable road: Allow more factors of the road to be controlled and altered, ie. adding lanes or curves
• Intelligent traffic: Instead of dummy cars that only goes forward, give the traffic AI brains as well