trackersb/deepdriving

A basic DeepDriving implementation for BeamNG

DeepDriving

A basic DeepDriving implementation for BeamNG.

Installation

This project uses many dependencies from Netzebands DeepDriving implementation. In order to ensure the proper functionality of this implementation these steps have to be followed first:

• Compile the C-Libraries
• Setup your Python 3 Environment

Download this repository and the trained model here and extract it inside the repository.

This project was tested on Python 3.5.4

Usage

Use DDController.py in order to generate driving controls:

from DDController import Controller

controller = Controller(MAX_SPEED)
controls = controller.getControl(imageData, speed)

See the example for further instructions.

Example

To run the example:

python scenario.py

Wait for the BeamNG Scenario to start up and pause, then:

python example.py

The simulation should now start and the car should follow the road.