/carla-ddpg

Autonomous Driving on Carla simulator using Deep Deterministic Policy Gradients. Based on Kendall, et. al. 2018.

Primary LanguageJupyter Notebook

Autonomous Driving using Deep Deterministic Policy Gradients.

Based on Kendall, et. al. 2018.

[1] Alex Kendall, Jeffrey Hawke, David Janz, Przemyslaw Mazur, Daniele Reda, John-Mark Allen, Vinh-Dieu Lam, Alex Bewley: “Learning to Drive in a Day”, 2018; [http://arxiv.org/abs/1807.00412 arXiv:1807.00412].

Installation

  1. Navigate to code/autodrive
  2. Run pip install -r requirements.txt
  3. Navigate to code/autodrive/keras-rl
  4. Run pip install .
  5. Navigate to code/autodrive/carla-client
  6. Run pip install .

Set Carla environment variable

1.$ export CARLA_ROOT=path/to/carla/directory
2. Copy mysettings.ini to CARLA_ROOT directory

Run experiment

python run_experiment.py