This project is a study and implementation of neural networks' potentials appliyed on a virtual car.
With the support of a genetic algorithm based on "Darwin's natural selection law", this guy learns how to drive by itself into a set of circuits, trying over and over to find the best path.
It has been developed by Alessandro Francesconi (me) and Marco Buzzoni under the Artificial Intelligence class exam @ University of Bologna, in 2012.
If you are curious about how we did that, just download the source code and a (little) documentation here.
If you are an hacker, the GitHub project page is what you are looking for!
You can try a real time demonstration of this project in the box below (Unity Web Player is required). Here are some keyboard controls:
- Numbers from 1 to 4 to change the track
- UP-arrow to enlarge the camera view