Self-Driving Car Simulation

Overview

LaneKeeper is a self-driving car simulation built in Python.
It combines:

• Pyglet for rendering & simulation.
• Neural Networks for processing sensor data and controlling the car.
• Genetic Algorithm for training and optimizing the neural network weights.

---

Demo

Features

• Car simulation on custom tracks.
• Neural network with 5 inputs, 4 hidden layers, 2 outputs.
• Genetic algorithm with selection, crossover, and mutation.
• Training system that improves driving across generations.
• Saves the best weights in brain.json for reuse on any track.

---

Project Structure

│── images_/           # Assets (track, cars, UI, etc.)
│── brain.json         # Saved best weights (trained genes)
│── canvas.py          # Drawing / rendering helpers
│── car.py             # Car object & physics
│── evolution.py       # Genetic algorithm logic
│── hud.py             # HUD / stats display
│── network.py         # Neural network implementation
│── storage.py         # Save / load utilities
│── testdriver.py      # Run a car using saved brain.json
│── track.py           # Track & checkpoints
│── training.py        # Training loop (main entry)
│── requirements.txt   # Python dependencies
│── README.md

---

Installation

Clone the repository and install dependencies:

git clone https://github.com/yousifnimah/Self-Driving-Car-Simulation-with-Neural-Networks-Genetic-Algorithm
cd Self-Driving-Car-Simulation-with-Neural-Networks-Genetic-Algorithm
pip install -r requirements.txt

---

Usage

1. Train the cars

Run training mode (cars evolve using GA):

python training.py

2. Test a trained model

After training, the best weights are saved to brain.json.
You can test the saved model with:

python testdriver.py

Future Work

• More advanced tracks & obstacles.
• Deep Reinforcement Learning integration.
• Multi-agent driving.

---

Tags

#Python #AI #NeuralNetworks #GeneticAlgorithm #Pyglet #SelfDrivingCar