Skip to content

Paseul/ml-agents-drone

Repository files navigation

Introduction

The goal of this application is to get drones to self-driving to their target points, avoiding people, trees and walls. The drone receives a negative reward when it comes into contact with an obstacle and is only rewarded when it reaches the target point.

Object explaination

Reward List

Actions

  • w - move forward.
  • s - move backward.
  • a - turn left.
  • d - turn right.
  • space - move up.
  • shift - move down

Installation

  1. Create Anaconda Environment
$conda create -n navigation python=3.6 (ml-agents requires only python 3.6)

About

self-driving drone simulation

Resources

License

Code of conduct

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published