yazici/neuroevolutionary-self-driving-car

Self Driving Car equipped with Neural Network evolved using Genetic Algorithm.

NeuroEvolutionary Self Driving Car

A self driving car equipped with a Neural Network that is evolved using a Genetic Algorithm. This project is an implementation of Self Driving Simulated cars that use a Neural Network by providing the immediate sorrounding proximity as input and the output to guide them through the sorrounding.

Watch full Evolution here: Self Driving Neural Network cars evolve using a Genetic Algorithm

Requirements

To run you would require the Java Runtime Environment Or if you want to tweak the code while running you would require the Processing integrated development environment

Usage

To execute go to the application folder and run GA2.exe. The applicatiion has a gui that is easy to use.

Principle

When you first launch the application a random generation of cars will be generated. After simulating once the Genetic Algorithm (Read more: Genetic Algorithm) will kick in and apply genetic operators such as crossover, mutation etc, to give you another generation of cars derived from only the best ones in the previoius generation according to a fitness function that judges how well a car performs on the track.

Author

Osama Ashhad Azmi (Linkedin)

Inspiration

• How Artificial intelligence learns | Genetic Algorithm explained
• Evolution Simulator