/NeuroEvolution-Vehicles

Raw code from Live Stream May 13 2019

Primary LanguageJavaScript

NeuroEvolution-Vehicles

Raw code from Live Stream May 13 2019

A neural network powered vehicle that drives around a randomly-generated path without hitting the walls.

Uses the p5.js and TensorFlow.js libraries.

https://codingtrain.github.io/NeuroEvolution-Vehicles/

To Do

  • Smooth track with OpenSimplexNoise
  • Improve the look of the track and cars
  • Add paragraph elements displaying the simulation speed and best car stats
  • Add functionality for saving/loading cars
  • Add sliders to control the features of the track before it is generated, like difficulty
    • thickness
    • jaggedness
    • height of jaggedness
  • Add button to regenerate track
  • Add fov rendering
  • Add crossover
  • create cool tracks with the supershape formula

Community Contributions