A self-driving car simulation program using a simple genetic algorithm to find and train a model to navigate independently around a track
The purpose of the project is to train a car agent to learn how to drive around a track on its own, the simulation runs indefinitely optimizing the acceleration and steering parameters to train the agent over the course of many generations. The fitness values are used to calculate and adjust the acceleration and steering values for the car, these calculations were done using the various rankings of the genomes within the genepool. Within the simulation the genenomes that are prioritized and selected first are genomes where the car traveled the furthest.