LICENCE: REUSE OK, ATTRIBUTION REQUIRED FOR MUSIC, CODE, VIDEOS, AND ANY OTHER FILES, DATA OR CONTENT WITHIN THIS REPO.
I designed and created a neural network I dubbed 'Pathfinder' using PHP and the FANN Library. Pathfinder is an example of a neural network that is capable of plotting an 8 direction step path (cardinal points + diagonals) from a starting position in a 5x5 grid to an ending position in that grid.
To keep the Pathfinder example simple it is not trained to deal with walls or non-traversable terrain however it would be very easy to add that by adding additional training.
Pathfinder was accepted as an example of how to use FANN on the Official PHP FANN Github.
This version of Pathfinder is implemented using FANN.js and a live version is available on GitHub Pages.
Live Preview: https://geekgirljoy.github.io/Pathfinder_Neural_Network/
This is the simplest example of Pathfinder and is the version that is included with the Official PHP FANN repo.
Note PHP FANN is required for the code contained in this repo to run.
-
Download Demonstration This video demonstrates how Pathfinder navigates around the grid. View on Youtube
-
Download Encode & Decode This video illustrates how to encode and decode grid values when working with Pathfinder. View on Youtube