AI Cars
My first large-scale project, designed to explore neural networks and genetic algorithms through a driving simulation.
Project Overview
This project involved implementing a neural network from scratch in Python, where the "brains" of the AI cars were trained using a genetic algorithm. The cars learned to navigate a track efficiently by avoiding obstacles and finding optimal paths.
It was a fascinating challenge to simulate evolution, observe emergent behaviors, and visualize the learning process in real-time.
Technologies Used
- **Python:** Core programming language
- **NumPy:** For efficient array operations in neural network calculations
- **Pygame:** For the graphical simulation and rendering
- **Custom Neural Network Implementation:** From activation functions to forward propagation
- **Genetic Algorithm:** For optimizing neural network weights
Challenges & Learnings
- **Optimization:** Ensuring the simulation ran smoothly with many cars and complex calculations.
- **Visualizing Learning:** Designing clear graphical feedback to understand how the AI was evolving.
- **Deep Dive into AI:** Gaining a foundational understanding of neural networks and evolutionary algorithms.
- **Collision Detection:** Implementing accurate and efficient collision detection for cars and track boundaries.