Neural Network Car Racing Game 🏎️🧠

A 2D car racing game built with Python and Pygame, where cars are controlled by a custom-built neural network instead of hard-coded logic.
This project explores neural networks, matrix operations, and AI decision-making in a real-time game environment.

The AI learns to drive around the track using sensor-like inputs and neural network outputs to control steering and movement.

---

🚀 Features

• 🧠 Neural network–controlled cars (no ML libraries used)
• 📐 Manual implementation of:
  • Matrix multiplication
  • Weights and biases
  • Forward propagation
• 🎮 Fully playable Pygame racing game
• 🛣️ Multiple difficulty levels (Easy / Medium / Hard)

---

🧠 How the Neural Network Works

Each car is controlled by a feed-forward neural network implemented from scratch in neural_network.py.

Inputs

The network receives numerical inputs representing:

• Distance from track borders
• Car speed
• Car orientation
• Progress along the track

Outputs

The network outputs values that determine:

• Steering direction
• Acceleration and movement decisions

Architecture

• Fully connected layers
• Randomly initialised weights and biases
• Activation function applied after each layer
• Forward propagation
• Natural Selection

No external machine learning libraries were used — this was intentional to gain a deep understanding of how neural networks work internally.

---

🗂️ Project Structure

Neural-Network-Car-Racing/
├── main.py                 # Game entry point
├── neural_network.py       # Neural network implementation
├── function.py             # Math and helper functions
├── variable.py             # Game constants, buttons, objects
├── window.py               # Rendering and UI logic
│
├── easy.txt                # Easy difficulty parameters
├── medium.txt              # Medium difficulty parameters
├── hard.txt                # Hard difficulty parameters
│
├── assets/
│   ├── track.png
│   ├── cars.png
│   ├── button images
│   └── rockit.mp3
│
├── LICENSE
└── .gitignore

---

▶️ How to Run

Requirements

• Python 3.9 or later
• Pygame

Install dependencies

pip install pygame

Run the game

python main.py

---

🎯 Purpose of This Project

This project was built to:

• Learn neural networks from first principles
• Apply AI concepts in a real-time interactive system
• Strengthen understanding of:
  • Linear algebra
  • Matrix operations
  • Game loops and rendering
  • AI-driven decision making

No machine learning frameworks such as TensorFlow or PyTorch were used.

---

🧩 Possible Improvements

• Add training via genetic algorithms or reinforcement learning
• Improve physics and collision handling
• Refactor neural network for scalability

---

👤 Author

Hamzah Ibrahim

Built as a personal learning project exploring AI, neural networks, and game development from scratch.